trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9914529914529915,
  "eval_steps": 100,
  "global_step": 58,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 4.071954585641894,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -0.2491038590669632,
      "logits/rejected": -0.3451940417289734,
      "logps/chosen": -294.6504211425781,
      "logps/rejected": -404.2626647949219,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.17,
      "grad_norm": 3.905690256665182,
      "learning_rate": 4.92735454356513e-07,
      "logits/chosen": -0.2290731519460678,
      "logits/rejected": -0.42057597637176514,
      "logps/chosen": -285.1579895019531,
      "logps/rejected": -440.47698974609375,
      "loss": 0.6926,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": 0.0010758544085547328,
      "rewards/margins": 3.7795203411405964e-07,
      "rewards/rejected": 0.0010754765244200826,
      "step": 10
    },
    {
      "epoch": 0.34,
      "grad_norm": 4.061739908914029,
      "learning_rate": 4.157806645601988e-07,
      "logits/chosen": -0.24342350661754608,
      "logits/rejected": -0.45833396911621094,
      "logps/chosen": -230.465576171875,
      "logps/rejected": -421.91131591796875,
      "loss": 0.6873,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": 0.014977594837546349,
      "rewards/margins": 0.012003140524029732,
      "rewards/rejected": 0.002974454313516617,
      "step": 20
    },
    {
      "epoch": 0.51,
      "grad_norm": 3.7362242077940304,
      "learning_rate": 2.801341700638307e-07,
      "logits/chosen": -0.22939440608024597,
      "logits/rejected": -0.3433351516723633,
      "logps/chosen": -282.84619140625,
      "logps/rejected": -379.5437927246094,
      "loss": 0.6789,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": 0.03496216982603073,
      "rewards/margins": 0.027732372283935547,
      "rewards/rejected": 0.007229797542095184,
      "step": 30
    },
    {
      "epoch": 0.68,
      "grad_norm": 3.7818612304795747,
      "learning_rate": 1.3381920698905784e-07,
      "logits/chosen": -0.22702646255493164,
      "logits/rejected": -0.45289507508277893,
      "logps/chosen": -244.3043212890625,
      "logps/rejected": -453.29522705078125,
      "loss": 0.6714,
      "rewards/accuracies": 0.875,
      "rewards/chosen": 0.04880012944340706,
      "rewards/margins": 0.04337088391184807,
      "rewards/rejected": 0.005429237149655819,
      "step": 40
    },
    {
      "epoch": 0.85,
      "grad_norm": 3.8140472657170306,
      "learning_rate": 2.863599358669755e-08,
      "logits/chosen": -0.32328224182128906,
      "logits/rejected": -0.45025634765625,
      "logps/chosen": -264.197021484375,
      "logps/rejected": -470.7127380371094,
      "loss": 0.6676,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": 0.05549341440200806,
      "rewards/margins": 0.05046632140874863,
      "rewards/rejected": 0.005027088802307844,
      "step": 50
    },
    {
      "epoch": 0.99,
      "step": 58,
      "total_flos": 0.0,
      "train_loss": 0.6775914965004757,
      "train_runtime": 852.993,
      "train_samples_per_second": 4.361,
      "train_steps_per_second": 0.068
    }
  ],
  "logging_steps": 10,
  "max_steps": 58,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}