last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.34223134839151265,
  "eval_steps": 500,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.11407711613050422,
      "grad_norm": 3.0874366760253906,
      "learning_rate": 9.771845767738991e-05,
      "loss": 1.2982,
      "step": 500
    },
    {
      "epoch": 0.22815423226100845,
      "grad_norm": 2.1173226833343506,
      "learning_rate": 9.543691535477983e-05,
      "loss": 0.9848,
      "step": 1000
    },
    {
      "epoch": 0.34223134839151265,
      "grad_norm": 1.83408522605896,
      "learning_rate": 9.315537303216974e-05,
      "loss": 0.9081,
      "step": 1500
    }
  ],
  "logging_steps": 500,
  "max_steps": 21915,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 2436152426496000.0,
  "train_batch_size": 12,
  "trial_name": null,
  "trial_params": null
}