trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "global_step": 6650,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.75,
      "learning_rate": 4.62406015037594e-05,
      "loss": 2.8619,
      "step": 500
    },
    {
      "epoch": 1.5,
      "learning_rate": 4.24812030075188e-05,
      "loss": 2.1471,
      "step": 1000
    },
    {
      "epoch": 2.26,
      "learning_rate": 3.87218045112782e-05,
      "loss": 1.8123,
      "step": 1500
    },
    {
      "epoch": 3.01,
      "learning_rate": 3.49624060150376e-05,
      "loss": 1.5294,
      "step": 2000
    },
    {
      "epoch": 3.76,
      "learning_rate": 3.120300751879699e-05,
      "loss": 1.1856,
      "step": 2500
    },
    {
      "epoch": 4.51,
      "learning_rate": 2.7443609022556393e-05,
      "loss": 0.9924,
      "step": 3000
    },
    {
      "epoch": 5.26,
      "learning_rate": 2.368421052631579e-05,
      "loss": 0.8382,
      "step": 3500
    },
    {
      "epoch": 6.02,
      "learning_rate": 1.9924812030075188e-05,
      "loss": 0.6972,
      "step": 4000
    },
    {
      "epoch": 6.77,
      "learning_rate": 1.6165413533834585e-05,
      "loss": 0.5187,
      "step": 4500
    },
    {
      "epoch": 7.52,
      "learning_rate": 1.2406015037593984e-05,
      "loss": 0.4422,
      "step": 5000
    },
    {
      "epoch": 8.27,
      "learning_rate": 8.646616541353383e-06,
      "loss": 0.3649,
      "step": 5500
    },
    {
      "epoch": 9.02,
      "learning_rate": 4.887218045112782e-06,
      "loss": 0.3106,
      "step": 6000
    },
    {
      "epoch": 9.77,
      "learning_rate": 1.1278195488721805e-06,
      "loss": 0.2562,
      "step": 6500
    },
    {
      "epoch": 10.0,
      "step": 6650,
      "total_flos": 1.7426815862833152e+16,
      "train_loss": 1.0550857498054218,
      "train_runtime": 2499.7825,
      "train_samples_per_second": 26.578,
      "train_steps_per_second": 2.66
    }
  ],
  "max_steps": 6650,
  "num_train_epochs": 10,
  "total_flos": 1.7426815862833152e+16,
  "trial_name": null,
  "trial_params": null
}