trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 16.0,
  "eval_steps": 500,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.0,
      "grad_norm": 0.12720850110054016,
      "learning_rate": 0.0001953038210948861,
      "loss": 1.3526,
      "step": 25
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.1418006867170334,
      "learning_rate": 0.00017567128158176953,
      "loss": 1.1521,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.19072043895721436,
      "learning_rate": 0.00014380476768566824,
      "loss": 1.0911,
      "step": 75
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.26865869760513306,
      "learning_rate": 0.00010485622221144484,
      "loss": 1.0212,
      "step": 100
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.4043382704257965,
      "learning_rate": 6.512255856701177e-05,
      "loss": 0.9493,
      "step": 125
    },
    {
      "epoch": 12.0,
      "grad_norm": 0.3770226836204529,
      "learning_rate": 3.102762227218957e-05,
      "loss": 0.8919,
      "step": 150
    },
    {
      "epoch": 14.0,
      "grad_norm": 0.4045465886592865,
      "learning_rate": 8.083631020418791e-06,
      "loss": 0.8576,
      "step": 175
    },
    {
      "epoch": 16.0,
      "grad_norm": 0.36538949608802795,
      "learning_rate": 0.0,
      "loss": 0.8455,
      "step": 200
    }
  ],
  "logging_steps": 25,
  "max_steps": 200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 17,
  "save_steps": 50,
  "total_flos": 1.408501775597568e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}