trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 5862,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.26,
      "learning_rate": 4.57352439440464e-05,
      "loss": 1.6481,
      "step": 500
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.14704878880928e-05,
      "loss": 1.5182,
      "step": 1000
    },
    {
      "epoch": 0.77,
      "learning_rate": 3.72057318321392e-05,
      "loss": 1.478,
      "step": 1500
    },
    {
      "epoch": 1.02,
      "learning_rate": 3.294097577618561e-05,
      "loss": 1.4709,
      "step": 2000
    },
    {
      "epoch": 1.28,
      "learning_rate": 2.8676219720232007e-05,
      "loss": 1.4008,
      "step": 2500
    },
    {
      "epoch": 1.54,
      "learning_rate": 2.4411463664278406e-05,
      "loss": 1.4038,
      "step": 3000
    },
    {
      "epoch": 1.79,
      "learning_rate": 2.0146707608324806e-05,
      "loss": 1.3903,
      "step": 3500
    },
    {
      "epoch": 2.05,
      "learning_rate": 1.5881951552371206e-05,
      "loss": 1.4114,
      "step": 4000
    },
    {
      "epoch": 2.3,
      "learning_rate": 1.1617195496417605e-05,
      "loss": 1.3706,
      "step": 4500
    },
    {
      "epoch": 2.56,
      "learning_rate": 7.3524394404640055e-06,
      "loss": 1.3448,
      "step": 5000
    },
    {
      "epoch": 2.81,
      "learning_rate": 3.087683384510406e-06,
      "loss": 1.3314,
      "step": 5500
    },
    {
      "epoch": 3.0,
      "step": 5862,
      "total_flos": 237909211152384.0,
      "train_loss": 1.430499308184933,
      "train_runtime": 3735.0404,
      "train_samples_per_second": 6.275,
      "train_steps_per_second": 1.569
    }
  ],
  "logging_steps": 500,
  "max_steps": 5862,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 237909211152384.0,
  "trial_name": null,
  "trial_params": null
}