{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.981366459627329,
  "eval_steps": 50,
  "global_step": 120,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12422360248447205,
      "grad_norm": 62.403607830645235,
      "learning_rate": 5e-07,
      "logits/chosen": -2.756077289581299,
      "logits/rejected": -2.75536847114563,
      "logps/chosen": -266.15899658203125,
      "logps/rejected": -237.189697265625,
      "loss": 0.6905,
      "rewards/accuracies": 0.36250001192092896,
      "rewards/chosen": 0.020162902772426605,
      "rewards/margins": 0.009148921817541122,
      "rewards/rejected": 0.011013981886208057,
      "step": 5
    },
    {
      "epoch": 0.2484472049689441,
      "grad_norm": 80.40658980559981,
      "learning_rate": 1e-06,
      "logits/chosen": -2.689826250076294,
      "logits/rejected": -2.672236204147339,
      "logps/chosen": -256.6275939941406,
      "logps/rejected": -209.4879608154297,
      "loss": 0.6412,
      "rewards/accuracies": 0.6187499761581421,
      "rewards/chosen": 0.5279896259307861,
      "rewards/margins": 0.1352878212928772,
      "rewards/rejected": 0.3927018344402313,
      "step": 10
    },
    {
      "epoch": 0.37267080745341613,
      "grad_norm": 42.882430184568136,
      "learning_rate": 9.949107209404663e-07,
      "logits/chosen": -2.516449213027954,
      "logits/rejected": -2.5170836448669434,
      "logps/chosen": -253.22061157226562,
      "logps/rejected": -214.71920776367188,
      "loss": 0.6132,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 1.5385050773620605,
      "rewards/margins": 0.7560700178146362,
      "rewards/rejected": 0.7824350595474243,
      "step": 15
    },
    {
      "epoch": 0.4968944099378882,
      "grad_norm": 45.56060438591064,
      "learning_rate": 9.797464868072486e-07,
      "logits/chosen": -2.4393157958984375,
      "logits/rejected": -2.4060797691345215,
      "logps/chosen": -257.09130859375,
      "logps/rejected": -198.79998779296875,
      "loss": 0.644,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 1.761357307434082,
      "rewards/margins": 1.1699168682098389,
      "rewards/rejected": 0.5914403796195984,
      "step": 20
    },
    {
      "epoch": 0.6211180124223602,
      "grad_norm": 41.36750304040445,
      "learning_rate": 9.548159976772592e-07,
      "logits/chosen": -2.3880293369293213,
      "logits/rejected": -2.402343273162842,
      "logps/chosen": -246.7839813232422,
      "logps/rejected": -230.6208953857422,
      "loss": 0.5949,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": 1.8271690607070923,
      "rewards/margins": 1.592099905014038,
      "rewards/rejected": 0.23506923019886017,
      "step": 25
    },
    {
      "epoch": 0.7453416149068323,
      "grad_norm": 45.4963965931298,
      "learning_rate": 9.206267664155906e-07,
      "logits/chosen": -2.424156427383423,
      "logits/rejected": -2.3829989433288574,
      "logps/chosen": -245.9100341796875,
      "logps/rejected": -231.0552978515625,
      "loss": 0.5972,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 1.6904218196868896,
      "rewards/margins": 1.704167366027832,
      "rewards/rejected": -0.013745462521910667,
      "step": 30
    },
    {
      "epoch": 0.8695652173913043,
      "grad_norm": 34.827462800688885,
      "learning_rate": 8.778747871771291e-07,
      "logits/chosen": -2.452261447906494,
      "logits/rejected": -2.428560972213745,
      "logps/chosen": -258.96221923828125,
      "logps/rejected": -219.5812530517578,
      "loss": 0.5609,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 1.5247328281402588,
      "rewards/margins": 1.6348918676376343,
      "rewards/rejected": -0.11015894263982773,
      "step": 35
    },
    {
      "epoch": 0.9937888198757764,
      "grad_norm": 41.54362761779536,
      "learning_rate": 8.274303669726426e-07,
      "logits/chosen": -2.435807228088379,
      "logits/rejected": -2.425825834274292,
      "logps/chosen": -246.02548217773438,
      "logps/rejected": -225.33700561523438,
      "loss": 0.6521,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": 1.2175686359405518,
      "rewards/margins": 0.9398768544197083,
      "rewards/rejected": 0.2776917815208435,
      "step": 40
    },
    {
      "epoch": 1.1180124223602483,
      "grad_norm": 18.152319532594465,
      "learning_rate": 7.703204087277988e-07,
      "logits/chosen": -2.3833680152893066,
      "logits/rejected": -2.36021089553833,
      "logps/chosen": -239.3954315185547,
      "logps/rejected": -245.8417205810547,
      "loss": 0.244,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 2.0914113521575928,
      "rewards/margins": 3.1723241806030273,
      "rewards/rejected": -1.0809123516082764,
      "step": 45
    },
    {
      "epoch": 1.2422360248447206,
      "grad_norm": 21.9049414703945,
      "learning_rate": 7.077075065009433e-07,
      "logits/chosen": -2.3475632667541504,
      "logits/rejected": -2.3293070793151855,
      "logps/chosen": -239.13436889648438,
      "logps/rejected": -224.94949340820312,
      "loss": 0.258,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": 2.809220552444458,
      "rewards/margins": 3.1373844146728516,
      "rewards/rejected": -0.32816413044929504,
      "step": 50
    },
    {
      "epoch": 1.2422360248447206,
      "eval_logits/chosen": -2.3354883193969727,
      "eval_logits/rejected": -2.3212666511535645,
      "eval_logps/chosen": -266.6775207519531,
      "eval_logps/rejected": -253.1901092529297,
      "eval_loss": 0.6212884783744812,
      "eval_rewards/accuracies": 0.7847222089767456,
      "eval_rewards/chosen": 2.382004737854004,
      "eval_rewards/margins": 1.986113429069519,
      "eval_rewards/rejected": 0.39589133858680725,
      "eval_runtime": 75.1943,
      "eval_samples_per_second": 15.161,
      "eval_steps_per_second": 0.239,
      "step": 50
    },
    {
      "epoch": 1.3664596273291925,
      "grad_norm": 19.456900162913456,
      "learning_rate": 6.408662784207149e-07,
      "logits/chosen": -2.3208935260772705,
      "logits/rejected": -2.2951972484588623,
      "logps/chosen": -247.5399932861328,
      "logps/rejected": -237.8445281982422,
      "loss": 0.2857,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": 2.739036798477173,
      "rewards/margins": 3.3734939098358154,
      "rewards/rejected": -0.6344569325447083,
      "step": 55
    },
    {
      "epoch": 1.4906832298136645,
      "grad_norm": 26.00493649217999,
      "learning_rate": 5.711574191366427e-07,
      "logits/chosen": -2.2791221141815186,
      "logits/rejected": -2.2677152156829834,
      "logps/chosen": -236.4568634033203,
      "logps/rejected": -235.218505859375,
      "loss": 0.2777,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": 2.477545976638794,
      "rewards/margins": 3.141633987426758,
      "rewards/rejected": -0.6640880703926086,
      "step": 60
    },
    {
      "epoch": 1.6149068322981366,
      "grad_norm": 22.738140845846544,
      "learning_rate": 5e-07,
      "logits/chosen": -2.2787652015686035,
      "logits/rejected": -2.227478265762329,
      "logps/chosen": -228.0373992919922,
      "logps/rejected": -255.89987182617188,
      "loss": 0.289,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 2.602055072784424,
      "rewards/margins": 3.5908477306365967,
      "rewards/rejected": -0.9887924194335938,
      "step": 65
    },
    {
      "epoch": 1.7391304347826086,
      "grad_norm": 23.007905633616,
      "learning_rate": 4.2884258086335745e-07,
      "logits/chosen": -2.259939670562744,
      "logits/rejected": -2.263455867767334,
      "logps/chosen": -245.46206665039062,
      "logps/rejected": -223.74270629882812,
      "loss": 0.2974,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 2.2899632453918457,
      "rewards/margins": 3.3291678428649902,
      "rewards/rejected": -1.0392045974731445,
      "step": 70
    },
    {
      "epoch": 1.8633540372670807,
      "grad_norm": 22.7179826338975,
      "learning_rate": 3.591337215792851e-07,
      "logits/chosen": -2.2751305103302,
      "logits/rejected": -2.2498066425323486,
      "logps/chosen": -237.3503875732422,
      "logps/rejected": -236.15951538085938,
      "loss": 0.2561,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": 2.172248363494873,
      "rewards/margins": 3.147101879119873,
      "rewards/rejected": -0.9748538732528687,
      "step": 75
    },
    {
      "epoch": 1.9875776397515528,
      "grad_norm": 26.637593255483548,
      "learning_rate": 2.922924934990568e-07,
      "logits/chosen": -2.273965358734131,
      "logits/rejected": -2.2656917572021484,
      "logps/chosen": -257.8080139160156,
      "logps/rejected": -217.7511444091797,
      "loss": 0.3499,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 2.5381007194519043,
      "rewards/margins": 3.5010929107666016,
      "rewards/rejected": -0.9629920721054077,
      "step": 80
    },
    {
      "epoch": 2.111801242236025,
      "grad_norm": 13.915541615199986,
      "learning_rate": 2.2967959127220137e-07,
      "logits/chosen": -2.2352209091186523,
      "logits/rejected": -2.197958469390869,
      "logps/chosen": -236.0648193359375,
      "logps/rejected": -232.01242065429688,
      "loss": 0.1697,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 2.252695083618164,
      "rewards/margins": 3.7181904315948486,
      "rewards/rejected": -1.4654955863952637,
      "step": 85
    },
    {
      "epoch": 2.2360248447204967,
      "grad_norm": 15.499960854642854,
      "learning_rate": 1.725696330273575e-07,
      "logits/chosen": -2.197455883026123,
      "logits/rejected": -2.1763360500335693,
      "logps/chosen": -239.3179473876953,
      "logps/rejected": -238.17578125,
      "loss": 0.1491,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 2.6867127418518066,
      "rewards/margins": 4.167205810546875,
      "rewards/rejected": -1.480492115020752,
      "step": 90
    },
    {
      "epoch": 2.360248447204969,
      "grad_norm": 16.718898555042905,
      "learning_rate": 1.2212521282287093e-07,
      "logits/chosen": -2.1886236667633057,
      "logits/rejected": -2.1752870082855225,
      "logps/chosen": -246.4811553955078,
      "logps/rejected": -212.0710906982422,
      "loss": 0.158,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 2.7335801124572754,
      "rewards/margins": 3.6828677654266357,
      "rewards/rejected": -0.9492877125740051,
      "step": 95
    },
    {
      "epoch": 2.4844720496894412,
      "grad_norm": 19.465292915783245,
      "learning_rate": 7.937323358440934e-08,
      "logits/chosen": -2.1609156131744385,
      "logits/rejected": -2.1472043991088867,
      "logps/chosen": -224.5920867919922,
      "logps/rejected": -214.02554321289062,
      "loss": 0.1741,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 2.6809496879577637,
      "rewards/margins": 3.8545315265655518,
      "rewards/rejected": -1.1735817193984985,
      "step": 100
    },
    {
      "epoch": 2.4844720496894412,
      "eval_logits/chosen": -2.201831340789795,
      "eval_logits/rejected": -2.172940254211426,
      "eval_logps/chosen": -268.1634216308594,
      "eval_logps/rejected": -258.27825927734375,
      "eval_loss": 0.6181273460388184,
      "eval_rewards/accuracies": 0.7777777910232544,
      "eval_rewards/chosen": 2.2334184646606445,
      "eval_rewards/margins": 2.346345901489258,
      "eval_rewards/rejected": -0.11292734742164612,
      "eval_runtime": 74.9234,
      "eval_samples_per_second": 15.216,
      "eval_steps_per_second": 0.24,
      "step": 100
    },
    {
      "epoch": 2.608695652173913,
      "grad_norm": 19.995154150155244,
      "learning_rate": 4.518400232274078e-08,
      "logits/chosen": -2.207038402557373,
      "logits/rejected": -2.1583220958709717,
      "logps/chosen": -251.38162231445312,
      "logps/rejected": -263.5743408203125,
      "loss": 0.1583,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 2.9800047874450684,
      "rewards/margins": 4.533116817474365,
      "rewards/rejected": -1.5531116724014282,
      "step": 105
    },
    {
      "epoch": 2.732919254658385,
      "grad_norm": 16.500504970360172,
      "learning_rate": 2.025351319275137e-08,
      "logits/chosen": -2.172910690307617,
      "logits/rejected": -2.1606621742248535,
      "logps/chosen": -211.9971160888672,
      "logps/rejected": -253.26974487304688,
      "loss": 0.1399,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 2.546518325805664,
      "rewards/margins": 4.063871383666992,
      "rewards/rejected": -1.5173530578613281,
      "step": 110
    },
    {
      "epoch": 2.857142857142857,
      "grad_norm": 19.536794540135528,
      "learning_rate": 5.0892790595336575e-09,
      "logits/chosen": -2.167283296585083,
      "logits/rejected": -2.1639857292175293,
      "logps/chosen": -246.7537841796875,
      "logps/rejected": -237.97262573242188,
      "loss": 0.1617,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 2.768967866897583,
      "rewards/margins": 4.068437576293945,
      "rewards/rejected": -1.2994694709777832,
      "step": 115
    },
    {
      "epoch": 2.981366459627329,
      "grad_norm": 13.712198521871002,
      "learning_rate": 0.0,
      "logits/chosen": -2.1904730796813965,
      "logits/rejected": -2.1649465560913086,
      "logps/chosen": -250.57174682617188,
      "logps/rejected": -253.6211395263672,
      "loss": 0.1699,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 2.733159065246582,
      "rewards/margins": 3.944279193878174,
      "rewards/rejected": -1.2111196517944336,
      "step": 120
    },
    {
      "epoch": 2.981366459627329,
      "step": 120,
      "total_flos": 1414680891359232.0,
      "train_loss": 0.3555192892750104,
      "train_runtime": 4293.0666,
      "train_samples_per_second": 7.168,
      "train_steps_per_second": 0.028
    }
  ],
  "logging_steps": 5,
  "max_steps": 120,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1414680891359232.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}