trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "global_step": 1840640,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 4.5e-05,
      "loss": 6.4348,
      "step": 184064
    },
    {
      "epoch": 2.0,
      "learning_rate": 4e-05,
      "loss": 5.3925,
      "step": 368128
    },
    {
      "epoch": 3.0,
      "learning_rate": 3.5e-05,
      "loss": 5.0967,
      "step": 552192
    },
    {
      "epoch": 4.0,
      "learning_rate": 3e-05,
      "loss": 4.9327,
      "step": 736256
    },
    {
      "epoch": 5.0,
      "learning_rate": 2.5e-05,
      "loss": 4.8269,
      "step": 920320
    },
    {
      "epoch": 6.0,
      "learning_rate": 2e-05,
      "loss": 4.7486,
      "step": 1104384
    },
    {
      "epoch": 7.0,
      "learning_rate": 1.5e-05,
      "loss": 4.6854,
      "step": 1288448
    },
    {
      "epoch": 8.0,
      "learning_rate": 1e-05,
      "loss": 4.6333,
      "step": 1472512
    },
    {
      "epoch": 9.0,
      "learning_rate": 5e-06,
      "loss": 4.5918,
      "step": 1656576
    },
    {
      "epoch": 10.0,
      "learning_rate": 0.0,
      "loss": 4.5572,
      "step": 1840640
    }
  ],
  "max_steps": 1840640,
  "num_train_epochs": 10,
  "total_flos": 3.184855687291592e+17,
  "trial_name": null,
  "trial_params": null
}