trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 5268,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.28,
      "learning_rate": 4.525436598329537e-05,
      "loss": 0.1662,
      "step": 500
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.050873196659074e-05,
      "loss": 0.0736,
      "step": 1000
    },
    {
      "epoch": 0.85,
      "learning_rate": 3.5763097949886106e-05,
      "loss": 0.0629,
      "step": 1500
    },
    {
      "epoch": 1.14,
      "learning_rate": 3.1017463933181475e-05,
      "loss": 0.042,
      "step": 2000
    },
    {
      "epoch": 1.42,
      "learning_rate": 2.6271829916476843e-05,
      "loss": 0.028,
      "step": 2500
    },
    {
      "epoch": 1.71,
      "learning_rate": 2.152619589977221e-05,
      "loss": 0.0283,
      "step": 3000
    },
    {
      "epoch": 1.99,
      "learning_rate": 1.678056188306758e-05,
      "loss": 0.0244,
      "step": 3500
    },
    {
      "epoch": 2.28,
      "learning_rate": 1.2034927866362947e-05,
      "loss": 0.012,
      "step": 4000
    },
    {
      "epoch": 2.56,
      "learning_rate": 7.289293849658315e-06,
      "loss": 0.0112,
      "step": 4500
    },
    {
      "epoch": 2.85,
      "learning_rate": 2.5436598329536827e-06,
      "loss": 0.0135,
      "step": 5000
    },
    {
      "epoch": 3.0,
      "step": 5268,
      "total_flos": 445788652997016.0,
      "train_loss": 0.044327347904964204,
      "train_runtime": 311.2643,
      "train_samples_per_second": 135.329,
      "train_steps_per_second": 16.925
    }
  ],
  "logging_steps": 500,
  "max_steps": 5268,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 445788652997016.0,
  "trial_name": null,
  "trial_params": null
}