trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9874476987447699,
  "eval_steps": 500,
  "global_step": 59,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.016736401673640166,
      "grad_norm": 7.772482099050406,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.807276487350464,
      "logits/rejected": -2.7759768962860107,
      "logps/chosen": -315.42626953125,
      "logps/rejected": -227.5915985107422,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.16736401673640167,
      "grad_norm": 7.4630481490893485,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.7555835247039795,
      "logits/rejected": -2.746177911758423,
      "logps/chosen": -271.9296569824219,
      "logps/rejected": -260.67138671875,
      "loss": 0.6914,
      "rewards/accuracies": 0.5347222089767456,
      "rewards/chosen": 0.0054025910794734955,
      "rewards/margins": 0.0035073086619377136,
      "rewards/rejected": 0.0018952824175357819,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 6.708300371266294,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.787097930908203,
      "logits/rejected": -2.7682840824127197,
      "logps/chosen": -261.45843505859375,
      "logps/rejected": -250.8958740234375,
      "loss": 0.6745,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.03223257511854172,
      "rewards/margins": 0.040766604244709015,
      "rewards/rejected": -0.008534022606909275,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 7.409956341692702,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.7880468368530273,
      "logits/rejected": -2.7689902782440186,
      "logps/chosen": -293.8247985839844,
      "logps/rejected": -254.217529296875,
      "loss": 0.6487,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -0.023687291890382767,
      "rewards/margins": 0.12690749764442444,
      "rewards/rejected": -0.1505947858095169,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 8.167332409118467,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.760864734649658,
      "logits/rejected": -2.7383790016174316,
      "logps/chosen": -264.1250915527344,
      "logps/rejected": -255.5350799560547,
      "loss": 0.6354,
      "rewards/accuracies": 0.684374988079071,
      "rewards/chosen": -0.11205615103244781,
      "rewards/margins": 0.16018818318843842,
      "rewards/rejected": -0.27224433422088623,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 8.545155858110496,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.770779848098755,
      "logits/rejected": -2.750847339630127,
      "logps/chosen": -279.1131286621094,
      "logps/rejected": -287.27044677734375,
      "loss": 0.6202,
      "rewards/accuracies": 0.628125011920929,
      "rewards/chosen": -0.22104480862617493,
      "rewards/margins": 0.13653476536273956,
      "rewards/rejected": -0.3575795590877533,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.6490178754774191,
      "train_runtime": 1630.8514,
      "train_samples_per_second": 9.371,
      "train_steps_per_second": 0.036
    }
  ],
  "logging_steps": 10,
  "max_steps": 59,
  "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
}