checkpoint-3000/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.1779661016949152,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.53,
      "learning_rate": 1.978813559322034e-05,
      "loss": 2.0034,
      "step": 500
    },
    {
      "epoch": 1.06,
      "learning_rate": 1.9576271186440678e-05,
      "loss": 1.3659,
      "step": 1000
    },
    {
      "epoch": 1.59,
      "learning_rate": 1.9364406779661017e-05,
      "loss": 1.2089,
      "step": 1500
    },
    {
      "epoch": 2.12,
      "learning_rate": 1.9152542372881357e-05,
      "loss": 1.1463,
      "step": 2000
    },
    {
      "epoch": 2.65,
      "learning_rate": 1.8940677966101697e-05,
      "loss": 1.0185,
      "step": 2500
    },
    {
      "epoch": 3.18,
      "learning_rate": 1.8728813559322033e-05,
      "loss": 0.9664,
      "step": 3000
    }
  ],
  "max_steps": 47200,
  "num_train_epochs": 50,
  "total_flos": 9490525857741888.0,
  "trial_name": null,
  "trial_params": null
}