trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.015521924718665115,
  "eval_steps": 500,
  "global_step": 150,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 28580.275390625,
      "learning_rate": 0.0002,
      "loss": 2.2322,
      "step": 25
    },
    {
      "epoch": 0.01,
      "grad_norm": 29493.916015625,
      "learning_rate": 0.0002,
      "loss": 1.9029,
      "step": 50
    },
    {
      "epoch": 0.01,
      "grad_norm": 50955.4375,
      "learning_rate": 0.0002,
      "loss": 1.8596,
      "step": 75
    },
    {
      "epoch": 0.01,
      "grad_norm": 26506.2109375,
      "learning_rate": 0.0002,
      "loss": 1.8312,
      "step": 100
    },
    {
      "epoch": 0.01,
      "grad_norm": 28205.296875,
      "learning_rate": 0.0002,
      "loss": 1.8088,
      "step": 125
    },
    {
      "epoch": 0.02,
      "grad_norm": 43509.62109375,
      "learning_rate": 0.0002,
      "loss": 1.8259,
      "step": 150
    }
  ],
  "logging_steps": 25,
  "max_steps": 150,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "total_flos": 1.758692546445312e+17,
  "train_batch_size": 12,
  "trial_name": null,
  "trial_params": null
}