trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "global_step": 4380,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.14,
      "learning_rate": 8.858447488584475e-05,
      "loss": 1.686,
      "step": 500
    },
    {
      "epoch": 2.28,
      "learning_rate": 7.716894977168951e-05,
      "loss": 0.4434,
      "step": 1000
    },
    {
      "epoch": 3.42,
      "learning_rate": 6.575342465753424e-05,
      "loss": 0.2871,
      "step": 1500
    },
    {
      "epoch": 4.57,
      "learning_rate": 5.4337899543379e-05,
      "loss": 0.2075,
      "step": 2000
    },
    {
      "epoch": 5.71,
      "learning_rate": 4.2922374429223744e-05,
      "loss": 0.1528,
      "step": 2500
    },
    {
      "epoch": 6.85,
      "learning_rate": 3.1506849315068496e-05,
      "loss": 0.1232,
      "step": 3000
    },
    {
      "epoch": 7.99,
      "learning_rate": 2.009132420091324e-05,
      "loss": 0.0981,
      "step": 3500
    },
    {
      "epoch": 9.13,
      "learning_rate": 8.675799086757991e-06,
      "loss": 0.0805,
      "step": 4000
    },
    {
      "epoch": 10.0,
      "step": 4380,
      "total_flos": 2485451858411520.0,
      "train_loss": 0.3575155772030626,
      "train_runtime": 651.5764,
      "train_samples_per_second": 107.432,
      "train_steps_per_second": 6.722
    }
  ],
  "max_steps": 4380,
  "num_train_epochs": 10,
  "total_flos": 2485451858411520.0,
  "trial_name": null,
  "trial_params": null
}