trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 52,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 6.388934227309893,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -1.2247042655944824,
      "logits/rejected": -1.0684211254119873,
      "logps/chosen": -569.8499145507812,
      "logps/rejected": -1057.6484375,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.19,
      "grad_norm": 6.575619838412587,
      "learning_rate": 4.907293218369498e-07,
      "logits/chosen": -1.1266288757324219,
      "logits/rejected": -0.9588133692741394,
      "logps/chosen": -644.7501831054688,
      "logps/rejected": -896.37548828125,
      "loss": 0.6907,
      "rewards/accuracies": 0.5347222089767456,
      "rewards/chosen": -0.0003651145671028644,
      "rewards/margins": 0.003978920169174671,
      "rewards/rejected": -0.004344034940004349,
      "step": 10
    },
    {
      "epoch": 0.38,
      "grad_norm": 6.33044990257594,
      "learning_rate": 3.941700805287168e-07,
      "logits/chosen": -1.1694377660751343,
      "logits/rejected": -0.9742730855941772,
      "logps/chosen": -553.2115478515625,
      "logps/rejected": -933.1784057617188,
      "loss": 0.66,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": -0.0070250085555016994,
      "rewards/margins": 0.06466411054134369,
      "rewards/rejected": -0.07168911397457123,
      "step": 20
    },
    {
      "epoch": 0.58,
      "grad_norm": 7.898094155565296,
      "learning_rate": 2.3293939665883228e-07,
      "logits/chosen": -1.121983528137207,
      "logits/rejected": -0.9881827235221863,
      "logps/chosen": -594.9716186523438,
      "logps/rejected": -1039.31982421875,
      "loss": 0.5638,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -0.06966790556907654,
      "rewards/margins": 0.3061855733394623,
      "rewards/rejected": -0.3758534789085388,
      "step": 30
    },
    {
      "epoch": 0.77,
      "grad_norm": 9.30835250885683,
      "learning_rate": 7.936171419533652e-08,
      "logits/chosen": -1.0671305656433105,
      "logits/rejected": -0.9684460759162903,
      "logps/chosen": -680.4083251953125,
      "logps/rejected": -1006.2408447265625,
      "loss": 0.4643,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.20575208961963654,
      "rewards/margins": 0.7841703295707703,
      "rewards/rejected": -0.9899223446846008,
      "step": 40
    },
    {
      "epoch": 0.96,
      "grad_norm": 8.502404185464833,
      "learning_rate": 2.328513490917311e-09,
      "logits/chosen": -1.0291626453399658,
      "logits/rejected": -0.9718970060348511,
      "logps/chosen": -671.4656372070312,
      "logps/rejected": -1046.7633056640625,
      "loss": 0.4299,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": -0.2569717466831207,
      "rewards/margins": 0.8754018545150757,
      "rewards/rejected": -1.132373571395874,
      "step": 50
    },
    {
      "epoch": 1.0,
      "step": 52,
      "total_flos": 0.0,
      "train_loss": 0.5552032154340011,
      "train_runtime": 696.2336,
      "train_samples_per_second": 4.777,
      "train_steps_per_second": 0.075
    }
  ],
  "logging_steps": 10,
  "max_steps": 52,
  "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
}