{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.986666666666667,
  "eval_steps": 500,
  "global_step": 168,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.17777777777777778,
      "grad_norm": 3.8474488258361816,
      "learning_rate": 2.9411764705882355e-06,
      "logits/chosen": -2.332357883453369,
      "logits/rejected": -2.368460178375244,
      "logps/chosen": -1.2429828643798828,
      "logps/rejected": -1.659155249595642,
      "loss": 1.2935,
      "odds_ratio_loss": 11.457437515258789,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.12429828941822052,
      "rewards/margins": 0.04161724075675011,
      "rewards/rejected": -0.16591551899909973,
      "sft_loss": 0.14774402976036072,
      "step": 10
    },
    {
      "epoch": 0.35555555555555557,
      "grad_norm": 5.139791965484619,
      "learning_rate": 4.995131923687488e-06,
      "logits/chosen": -2.292804002761841,
      "logits/rejected": -2.327223300933838,
      "logps/chosen": -1.2883718013763428,
      "logps/rejected": -1.7239799499511719,
      "loss": 1.337,
      "odds_ratio_loss": 11.885274887084961,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.12883718311786652,
      "rewards/margins": 0.043560806661844254,
      "rewards/rejected": -0.17239800095558167,
      "sft_loss": 0.14851602911949158,
      "step": 20
    },
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 5.0462751388549805,
      "learning_rate": 4.90911473983908e-06,
      "logits/chosen": -2.3472771644592285,
      "logits/rejected": -2.3845486640930176,
      "logps/chosen": -1.186187982559204,
      "logps/rejected": -1.59175705909729,
      "loss": 1.2375,
      "odds_ratio_loss": 10.995233535766602,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -0.11861880123615265,
      "rewards/margins": 0.04055692255496979,
      "rewards/rejected": -0.15917572379112244,
      "sft_loss": 0.13796505331993103,
      "step": 30
    },
    {
      "epoch": 0.7111111111111111,
      "grad_norm": 3.613401412963867,
      "learning_rate": 4.71919261421297e-06,
      "logits/chosen": -2.370697021484375,
      "logits/rejected": -2.4233319759368896,
      "logps/chosen": -0.9008461236953735,
      "logps/rejected": -1.4053981304168701,
      "loss": 0.9426,
      "odds_ratio_loss": 8.488945007324219,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": -0.09008461236953735,
      "rewards/margins": 0.05045522004365921,
      "rewards/rejected": -0.14053983986377716,
      "sft_loss": 0.09369887411594391,
      "step": 40
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 3.07717227935791,
      "learning_rate": 4.43355687413747e-06,
      "logits/chosen": -2.4286742210388184,
      "logits/rejected": -2.4549202919006348,
      "logps/chosen": -0.6461768746376038,
      "logps/rejected": -1.0805357694625854,
      "loss": 0.6945,
      "odds_ratio_loss": 6.29970645904541,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.06461768597364426,
      "rewards/margins": 0.04343589395284653,
      "rewards/rejected": -0.10805357992649078,
      "sft_loss": 0.06456876546144485,
      "step": 50
    },
    {
      "epoch": 1.0666666666666667,
      "grad_norm": 2.27549409866333,
      "learning_rate": 4.064526968101844e-06,
      "logits/chosen": -2.317702531814575,
      "logits/rejected": -2.337463855743408,
      "logps/chosen": -0.5583322048187256,
      "logps/rejected": -1.0102574825286865,
      "loss": 0.6081,
      "odds_ratio_loss": 5.508663177490234,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.05583322048187256,
      "rewards/margins": 0.04519252851605415,
      "rewards/rejected": -0.10102574527263641,
      "sft_loss": 0.05719046667218208,
      "step": 60
    },
    {
      "epoch": 1.2444444444444445,
      "grad_norm": 2.98760724067688,
      "learning_rate": 3.6280191288478437e-06,
      "logits/chosen": -2.4034271240234375,
      "logits/rejected": -2.4294583797454834,
      "logps/chosen": -0.35430365800857544,
      "logps/rejected": -0.7878178358078003,
      "loss": 0.3968,
      "odds_ratio_loss": 3.5909438133239746,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.035430364310741425,
      "rewards/margins": 0.0433514229953289,
      "rewards/rejected": -0.07878179103136063,
      "sft_loss": 0.037743426859378815,
      "step": 70
    },
    {
      "epoch": 1.4222222222222223,
      "grad_norm": 2.200446605682373,
      "learning_rate": 3.142859907420615e-06,
      "logits/chosen": -2.336760997772217,
      "logits/rejected": -2.365061044692993,
      "logps/chosen": -0.2458254098892212,
      "logps/rejected": -0.6315831542015076,
      "loss": 0.2882,
      "odds_ratio_loss": 2.5942039489746094,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.02458254061639309,
      "rewards/margins": 0.03857577592134476,
      "rewards/rejected": -0.063158318400383,
      "sft_loss": 0.028748363256454468,
      "step": 80
    },
    {
      "epoch": 1.6,
      "grad_norm": 1.6536734104156494,
      "learning_rate": 2.629974185404951e-06,
      "logits/chosen": -2.313284397125244,
      "logits/rejected": -2.3488173484802246,
      "logps/chosen": -0.1942504495382309,
      "logps/rejected": -0.6369145512580872,
      "loss": 0.2406,
      "odds_ratio_loss": 2.1658334732055664,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.01942504197359085,
      "rewards/margins": 0.04426640644669533,
      "rewards/rejected": -0.06369145214557648,
      "sft_loss": 0.023997236043214798,
      "step": 90
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 1.2765487432479858,
      "learning_rate": 2.1114826863194882e-06,
      "logits/chosen": -2.2804689407348633,
      "logits/rejected": -2.3144404888153076,
      "logps/chosen": -0.13879844546318054,
      "logps/rejected": -0.6243221163749695,
      "loss": 0.1759,
      "odds_ratio_loss": 1.5965832471847534,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": -0.013879844918847084,
      "rewards/margins": 0.048552367836236954,
      "rewards/rejected": -0.06243220716714859,
      "sft_loss": 0.01626196689903736,
      "step": 100
    },
    {
      "epoch": 1.9555555555555557,
      "grad_norm": 1.2211233377456665,
      "learning_rate": 1.6097479104361328e-06,
      "logits/chosen": -2.317593812942505,
      "logits/rejected": -2.336158037185669,
      "logps/chosen": -0.1410079300403595,
      "logps/rejected": -0.5155030488967896,
      "loss": 0.1836,
      "odds_ratio_loss": 1.6868371963500977,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.014100792817771435,
      "rewards/margins": 0.037449512630701065,
      "rewards/rejected": -0.051550306379795074,
      "sft_loss": 0.01492035947740078,
      "step": 110
    },
    {
      "epoch": 2.1333333333333333,
      "grad_norm": 1.3183128833770752,
      "learning_rate": 1.1464096417858821e-06,
      "logits/chosen": -2.290233850479126,
      "logits/rejected": -2.3088765144348145,
      "logps/chosen": -0.10447756201028824,
      "logps/rejected": -0.6317979097366333,
      "loss": 0.1326,
      "odds_ratio_loss": 1.1886184215545654,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": -0.010447757318615913,
      "rewards/margins": 0.052732039242982864,
      "rewards/rejected": -0.06317979097366333,
      "sft_loss": 0.013712344691157341,
      "step": 120
    },
    {
      "epoch": 2.311111111111111,
      "grad_norm": 1.7163466215133667,
      "learning_rate": 7.414516258630245e-07,
      "logits/chosen": -2.262594223022461,
      "logits/rejected": -2.284545421600342,
      "logps/chosen": -0.12122899293899536,
      "logps/rejected": -0.5315740704536438,
      "loss": 0.1618,
      "odds_ratio_loss": 1.4604320526123047,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.012122899293899536,
      "rewards/margins": 0.0410345084965229,
      "rewards/rejected": -0.05315741151571274,
      "sft_loss": 0.015804503113031387,
      "step": 130
    },
    {
      "epoch": 2.488888888888889,
      "grad_norm": 3.0588433742523193,
      "learning_rate": 4.123396721497977e-07,
      "logits/chosen": -2.3231940269470215,
      "logits/rejected": -2.34096622467041,
      "logps/chosen": -0.12840591371059418,
      "logps/rejected": -0.5461179614067078,
      "loss": 0.1688,
      "odds_ratio_loss": 1.521639108657837,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.012840591371059418,
      "rewards/margins": 0.04177120327949524,
      "rewards/rejected": -0.05461179465055466,
      "sft_loss": 0.016592377796769142,
      "step": 140
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 2.731790065765381,
      "learning_rate": 1.7326835503629542e-07,
      "logits/chosen": -2.2994039058685303,
      "logits/rejected": -2.324022054672241,
      "logps/chosen": -0.12199757248163223,
      "logps/rejected": -0.5101950168609619,
      "loss": 0.1648,
      "odds_ratio_loss": 1.4926129579544067,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.012199757620692253,
      "rewards/margins": 0.03881974145770073,
      "rewards/rejected": -0.05101950094103813,
      "sft_loss": 0.015529977157711983,
      "step": 150
    },
    {
      "epoch": 2.8444444444444446,
      "grad_norm": 1.6209518909454346,
      "learning_rate": 3.4548802869627806e-08,
      "logits/chosen": -2.19964599609375,
      "logits/rejected": -2.222959041595459,
      "logps/chosen": -0.11084076017141342,
      "logps/rejected": -0.5317128300666809,
      "loss": 0.1456,
      "odds_ratio_loss": 1.3233401775360107,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.011084076017141342,
      "rewards/margins": 0.042087212204933167,
      "rewards/rejected": -0.05317128449678421,
      "sft_loss": 0.013311339542269707,
      "step": 160
    },
    {
      "epoch": 2.986666666666667,
      "step": 168,
      "total_flos": 4.410496642646016e+16,
      "train_loss": 0.4940077399923688,
      "train_runtime": 519.2548,
      "train_samples_per_second": 5.2,
      "train_steps_per_second": 0.324
    }
  ],
  "logging_steps": 10,
  "max_steps": 168,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4.410496642646016e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}