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": 7.77332439766353,
      "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": 7.492555276868743,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.7555723190307617,
      "logits/rejected": -2.746173620223999,
      "logps/chosen": -271.930419921875,
      "logps/rejected": -260.73095703125,
      "loss": 0.6914,
      "rewards/accuracies": 0.5451388955116272,
      "rewards/chosen": 0.005395061336457729,
      "rewards/margins": 0.004095336422324181,
      "rewards/rejected": 0.0012997254962101579,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 6.719853223196107,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.7869045734405518,
      "logits/rejected": -2.7681097984313965,
      "logps/chosen": -261.4552001953125,
      "logps/rejected": -250.892333984375,
      "loss": 0.6745,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 0.03226487711071968,
      "rewards/margins": 0.040763549506664276,
      "rewards/rejected": -0.008498674258589745,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 7.390169288124108,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.7878599166870117,
      "logits/rejected": -2.768826484680176,
      "logps/chosen": -293.8148498535156,
      "logps/rejected": -254.2426300048828,
      "loss": 0.6487,
      "rewards/accuracies": 0.6781250238418579,
      "rewards/chosen": -0.02358742617070675,
      "rewards/margins": 0.12725816667079926,
      "rewards/rejected": -0.15084558725357056,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 8.165501150798445,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.760540008544922,
      "logits/rejected": -2.73799467086792,
      "logps/chosen": -264.11212158203125,
      "logps/rejected": -255.53182983398438,
      "loss": 0.6355,
      "rewards/accuracies": 0.684374988079071,
      "rewards/chosen": -0.11192627251148224,
      "rewards/margins": 0.16028562188148499,
      "rewards/rejected": -0.27221187949180603,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 8.589450320852121,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.7703118324279785,
      "logits/rejected": -2.7503678798675537,
      "logps/chosen": -279.1415100097656,
      "logps/rejected": -287.29901123046875,
      "loss": 0.6202,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.22132858633995056,
      "rewards/margins": 0.13653627038002014,
      "rewards/rejected": -0.3578649163246155,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.6490624637927039,
      "train_runtime": 1626.9557,
      "train_samples_per_second": 9.394,
      "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
}