{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.97196261682243,
  "eval_steps": 50,
  "global_step": 159,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09345794392523364,
      "grad_norm": 66.9687943936917,
      "learning_rate": 5e-07,
      "logits/chosen": -2.720803737640381,
      "logits/rejected": -2.7183666229248047,
      "logps/chosen": -237.3436279296875,
      "logps/rejected": -190.54464721679688,
      "loss": 0.6913,
      "rewards/accuracies": 0.32499998807907104,
      "rewards/chosen": 0.013445606455206871,
      "rewards/margins": 0.008647488430142403,
      "rewards/rejected": 0.00479811942204833,
      "step": 5
    },
    {
      "epoch": 0.18691588785046728,
      "grad_norm": 48.6419981873445,
      "learning_rate": 1e-06,
      "logits/chosen": -2.679405689239502,
      "logits/rejected": -2.670754909515381,
      "logps/chosen": -279.81866455078125,
      "logps/rejected": -226.22573852539062,
      "loss": 0.6556,
      "rewards/accuracies": 0.6187499761581421,
      "rewards/chosen": 0.4765666127204895,
      "rewards/margins": 0.12125828117132187,
      "rewards/rejected": 0.35530832409858704,
      "step": 10
    },
    {
      "epoch": 0.2803738317757009,
      "grad_norm": 44.827387730520904,
      "learning_rate": 9.972240926774166e-07,
      "logits/chosen": -2.5193655490875244,
      "logits/rejected": -2.510051965713501,
      "logps/chosen": -236.1126708984375,
      "logps/rejected": -201.1164093017578,
      "loss": 0.6375,
      "rewards/accuracies": 0.643750011920929,
      "rewards/chosen": 1.2661212682724,
      "rewards/margins": 0.5709505677223206,
      "rewards/rejected": 0.6951709985733032,
      "step": 15
    },
    {
      "epoch": 0.37383177570093457,
      "grad_norm": 51.33633927747967,
      "learning_rate": 9.889271933555212e-07,
      "logits/chosen": -2.4093480110168457,
      "logits/rejected": -2.381843090057373,
      "logps/chosen": -258.9214782714844,
      "logps/rejected": -220.68408203125,
      "loss": 0.6632,
      "rewards/accuracies": 0.6187499761581421,
      "rewards/chosen": 1.2196061611175537,
      "rewards/margins": 0.9229635000228882,
      "rewards/rejected": 0.2966426610946655,
      "step": 20
    },
    {
      "epoch": 0.4672897196261682,
      "grad_norm": 58.55983526769123,
      "learning_rate": 9.752014277286431e-07,
      "logits/chosen": -2.319462299346924,
      "logits/rejected": -2.303922653198242,
      "logps/chosen": -254.1189422607422,
      "logps/rejected": -196.4254608154297,
      "loss": 0.6806,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 1.0977602005004883,
      "rewards/margins": 0.803708553314209,
      "rewards/rejected": 0.2940516173839569,
      "step": 25
    },
    {
      "epoch": 0.5607476635514018,
      "grad_norm": 38.96884797085844,
      "learning_rate": 9.561992016100291e-07,
      "logits/chosen": -2.380964994430542,
      "logits/rejected": -2.357675075531006,
      "logps/chosen": -242.76406860351562,
      "logps/rejected": -221.53903198242188,
      "loss": 0.6103,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 0.37621065974235535,
      "rewards/margins": 0.7926680445671082,
      "rewards/rejected": -0.4164574146270752,
      "step": 30
    },
    {
      "epoch": 0.6542056074766355,
      "grad_norm": 49.029442328071866,
      "learning_rate": 9.321315086741915e-07,
      "logits/chosen": -2.3943734169006348,
      "logits/rejected": -2.386823892593384,
      "logps/chosen": -251.5531463623047,
      "logps/rejected": -226.2257537841797,
      "loss": 0.61,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.37054210901260376,
      "rewards/margins": 0.955781102180481,
      "rewards/rejected": -0.5852389931678772,
      "step": 35
    },
    {
      "epoch": 0.7476635514018691,
      "grad_norm": 47.16621948348876,
      "learning_rate": 9.032655876613635e-07,
      "logits/chosen": -2.323800802230835,
      "logits/rejected": -2.296937942504883,
      "logps/chosen": -255.12149047851562,
      "logps/rejected": -216.43679809570312,
      "loss": 0.5697,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.7852829694747925,
      "rewards/margins": 1.0229356288909912,
      "rewards/rejected": -0.2376527041196823,
      "step": 40
    },
    {
      "epoch": 0.8411214953271028,
      "grad_norm": 45.06757129062354,
      "learning_rate": 8.699219550575952e-07,
      "logits/chosen": -2.2773144245147705,
      "logits/rejected": -2.2689876556396484,
      "logps/chosen": -242.54403686523438,
      "logps/rejected": -198.2088165283203,
      "loss": 0.5913,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": 1.0134313106536865,
      "rewards/margins": 0.8711115121841431,
      "rewards/rejected": 0.1423199325799942,
      "step": 45
    },
    {
      "epoch": 0.9345794392523364,
      "grad_norm": 37.25379822397818,
      "learning_rate": 8.324708461985124e-07,
      "logits/chosen": -2.3120808601379395,
      "logits/rejected": -2.325521945953369,
      "logps/chosen": -244.45162963867188,
      "logps/rejected": -233.9671173095703,
      "loss": 0.5323,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.8873047828674316,
      "rewards/margins": 1.0542502403259277,
      "rewards/rejected": -0.16694557666778564,
      "step": 50
    },
    {
      "epoch": 0.9345794392523364,
      "eval_logits/chosen": -2.3802032470703125,
      "eval_logits/rejected": -2.3511736392974854,
      "eval_logps/chosen": -257.5635070800781,
      "eval_logps/rejected": -242.5312957763672,
      "eval_loss": 0.5823682546615601,
      "eval_rewards/accuracies": 0.7395833134651184,
      "eval_rewards/chosen": 0.9205262660980225,
      "eval_rewards/margins": 1.2713056802749634,
      "eval_rewards/rejected": -0.35077938437461853,
      "eval_runtime": 101.9106,
      "eval_samples_per_second": 14.915,
      "eval_steps_per_second": 0.236,
      "step": 50
    },
    {
      "epoch": 1.02803738317757,
      "grad_norm": 17.588268217174,
      "learning_rate": 7.913281043133977e-07,
      "logits/chosen": -2.377732038497925,
      "logits/rejected": -2.4019691944122314,
      "logps/chosen": -251.0818328857422,
      "logps/rejected": -231.4687957763672,
      "loss": 0.4699,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 1.248857021331787,
      "rewards/margins": 1.5204874277114868,
      "rewards/rejected": -0.2716304361820221,
      "step": 55
    },
    {
      "epoch": 1.1214953271028036,
      "grad_norm": 17.6740148393279,
      "learning_rate": 7.469505631561317e-07,
      "logits/chosen": -2.4365363121032715,
      "logits/rejected": -2.3999149799346924,
      "logps/chosen": -235.5603790283203,
      "logps/rejected": -217.8879852294922,
      "loss": 0.2443,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 1.6067253351211548,
      "rewards/margins": 2.39015531539917,
      "rewards/rejected": -0.7834302186965942,
      "step": 60
    },
    {
      "epoch": 1.2149532710280373,
      "grad_norm": 19.25307975460043,
      "learning_rate": 6.998309744925411e-07,
      "logits/chosen": -2.459545612335205,
      "logits/rejected": -2.443091869354248,
      "logps/chosen": -234.2629852294922,
      "logps/rejected": -229.77243041992188,
      "loss": 0.2014,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 2.0133259296417236,
      "rewards/margins": 3.0357606410980225,
      "rewards/rejected": -1.0224347114562988,
      "step": 65
    },
    {
      "epoch": 1.308411214953271,
      "grad_norm": 18.33160467378068,
      "learning_rate": 6.504925367674594e-07,
      "logits/chosen": -2.5056710243225098,
      "logits/rejected": -2.481133222579956,
      "logps/chosen": -239.59219360351562,
      "logps/rejected": -222.05429077148438,
      "loss": 0.2154,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.9632623195648193,
      "rewards/margins": 3.0751612186431885,
      "rewards/rejected": -1.1118988990783691,
      "step": 70
    },
    {
      "epoch": 1.4018691588785046,
      "grad_norm": 26.100443454167603,
      "learning_rate": 5.994830857031499e-07,
      "logits/chosen": -2.4787604808807373,
      "logits/rejected": -2.477220058441162,
      "logps/chosen": -242.30496215820312,
      "logps/rejected": -246.5374298095703,
      "loss": 0.2167,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 2.2068536281585693,
      "rewards/margins": 3.7736289501190186,
      "rewards/rejected": -1.5667749643325806,
      "step": 75
    },
    {
      "epoch": 1.4953271028037383,
      "grad_norm": 23.947057848275417,
      "learning_rate": 5.473690113345342e-07,
      "logits/chosen": -2.4580140113830566,
      "logits/rejected": -2.4340569972991943,
      "logps/chosen": -232.0405731201172,
      "logps/rejected": -232.50894165039062,
      "loss": 0.2133,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.6286522150039673,
      "rewards/margins": 3.344123363494873,
      "rewards/rejected": -1.7154712677001953,
      "step": 80
    },
    {
      "epoch": 1.588785046728972,
      "grad_norm": 25.173894130852474,
      "learning_rate": 4.947289690242102e-07,
      "logits/chosen": -2.3979992866516113,
      "logits/rejected": -2.372950553894043,
      "logps/chosen": -234.0653533935547,
      "logps/rejected": -226.2272491455078,
      "loss": 0.2526,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 2.000070333480835,
      "rewards/margins": 3.399747371673584,
      "rewards/rejected": -1.3996769189834595,
      "step": 85
    },
    {
      "epoch": 1.6822429906542056,
      "grad_norm": 25.381019801326666,
      "learning_rate": 4.421474542878194e-07,
      "logits/chosen": -2.402013063430786,
      "logits/rejected": -2.358625888824463,
      "logps/chosen": -240.19235229492188,
      "logps/rejected": -235.06942749023438,
      "loss": 0.2381,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 2.031113624572754,
      "rewards/margins": 3.7032268047332764,
      "rewards/rejected": -1.6721128225326538,
      "step": 90
    },
    {
      "epoch": 1.7757009345794392,
      "grad_norm": 22.605364041509517,
      "learning_rate": 3.902083127725186e-07,
      "logits/chosen": -2.3787314891815186,
      "logits/rejected": -2.382676601409912,
      "logps/chosen": -231.67434692382812,
      "logps/rejected": -207.5229949951172,
      "loss": 0.2348,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 2.045062780380249,
      "rewards/margins": 3.560044765472412,
      "rewards/rejected": -1.5149818658828735,
      "step": 95
    },
    {
      "epoch": 1.8691588785046729,
      "grad_norm": 26.859387828548858,
      "learning_rate": 3.394882574513519e-07,
      "logits/chosen": -2.3952929973602295,
      "logits/rejected": -2.3831348419189453,
      "logps/chosen": -234.7982940673828,
      "logps/rejected": -259.3247375488281,
      "loss": 0.2441,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.9878575801849365,
      "rewards/margins": 3.512976884841919,
      "rewards/rejected": -1.525119423866272,
      "step": 100
    },
    {
      "epoch": 1.8691588785046729,
      "eval_logits/chosen": -2.3957111835479736,
      "eval_logits/rejected": -2.363419532775879,
      "eval_logps/chosen": -256.04901123046875,
      "eval_logps/rejected": -246.68414306640625,
      "eval_loss": 0.5841386318206787,
      "eval_rewards/accuracies": 0.7708333134651184,
      "eval_rewards/chosen": 1.071976661682129,
      "eval_rewards/margins": 1.8380416631698608,
      "eval_rewards/rejected": -0.7660649418830872,
      "eval_runtime": 100.9496,
      "eval_samples_per_second": 15.057,
      "eval_steps_per_second": 0.238,
      "step": 100
    },
    {
      "epoch": 1.9626168224299065,
      "grad_norm": 30.452454043277406,
      "learning_rate": 2.9055046501619083e-07,
      "logits/chosen": -2.3725104331970215,
      "logits/rejected": -2.3521809577941895,
      "logps/chosen": -251.14065551757812,
      "logps/rejected": -223.05380249023438,
      "loss": 0.2356,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 2.067333459854126,
      "rewards/margins": 3.433384656906128,
      "rewards/rejected": -1.366051435470581,
      "step": 105
    },
    {
      "epoch": 2.05607476635514,
      "grad_norm": 14.133552406481272,
      "learning_rate": 2.439383225725225e-07,
      "logits/chosen": -2.374899387359619,
      "logits/rejected": -2.3513777256011963,
      "logps/chosen": -229.90478515625,
      "logps/rejected": -230.5806884765625,
      "loss": 0.1857,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.9912045001983643,
      "rewards/margins": 3.692650318145752,
      "rewards/rejected": -1.7014458179473877,
      "step": 110
    },
    {
      "epoch": 2.149532710280374,
      "grad_norm": 11.316511114391215,
      "learning_rate": 2.0016939407046986e-07,
      "logits/chosen": -2.380131959915161,
      "logits/rejected": -2.3474698066711426,
      "logps/chosen": -223.51254272460938,
      "logps/rejected": -238.4824676513672,
      "loss": 0.1132,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 1.8329362869262695,
      "rewards/margins": 3.968013286590576,
      "rewards/rejected": -2.135077476501465,
      "step": 115
    },
    {
      "epoch": 2.2429906542056073,
      "grad_norm": 19.797163444062072,
      "learning_rate": 1.5972967346655448e-07,
      "logits/chosen": -2.371460437774658,
      "logits/rejected": -2.36273193359375,
      "logps/chosen": -231.41104125976562,
      "logps/rejected": -190.1511993408203,
      "loss": 0.1337,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 2.0867326259613037,
      "rewards/margins": 3.7511298656463623,
      "rewards/rejected": -1.6643966436386108,
      "step": 120
    },
    {
      "epoch": 2.336448598130841,
      "grad_norm": 13.870872057083783,
      "learning_rate": 1.2306818842696715e-07,
      "logits/chosen": -2.391444683074951,
      "logits/rejected": -2.373061180114746,
      "logps/chosen": -235.8319854736328,
      "logps/rejected": -220.6596221923828,
      "loss": 0.1358,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 2.6346383094787598,
      "rewards/margins": 4.137408256530762,
      "rewards/rejected": -1.502769112586975,
      "step": 125
    },
    {
      "epoch": 2.4299065420560746,
      "grad_norm": 12.23774644882284,
      "learning_rate": 9.059201449082043e-08,
      "logits/chosen": -2.3761117458343506,
      "logits/rejected": -2.363049268722534,
      "logps/chosen": -219.3924560546875,
      "logps/rejected": -233.48605346679688,
      "loss": 0.1281,
      "rewards/accuracies": 0.981249988079071,
      "rewards/chosen": 2.4915733337402344,
      "rewards/margins": 4.301194190979004,
      "rewards/rejected": -1.8096210956573486,
      "step": 130
    },
    {
      "epoch": 2.5233644859813085,
      "grad_norm": 18.769111889041984,
      "learning_rate": 6.266175505426957e-08,
      "logits/chosen": -2.3567309379577637,
      "logits/rejected": -2.3618130683898926,
      "logps/chosen": -228.6204071044922,
      "logps/rejected": -231.1812286376953,
      "loss": 0.1246,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 2.454102039337158,
      "rewards/margins": 4.2184247970581055,
      "rewards/rejected": -1.7643229961395264,
      "step": 135
    },
    {
      "epoch": 2.616822429906542,
      "grad_norm": 16.78135239414831,
      "learning_rate": 3.958753736408105e-08,
      "logits/chosen": -2.389759063720703,
      "logits/rejected": -2.372283458709717,
      "logps/chosen": -221.2369842529297,
      "logps/rejected": -223.2550811767578,
      "loss": 0.1219,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 2.3341567516326904,
      "rewards/margins": 4.09852409362793,
      "rewards/rejected": -1.7643673419952393,
      "step": 140
    },
    {
      "epoch": 2.710280373831776,
      "grad_norm": 15.32254073396557,
      "learning_rate": 2.162556897965101e-08,
      "logits/chosen": -2.3783352375030518,
      "logits/rejected": -2.360616445541382,
      "logps/chosen": -234.25350952148438,
      "logps/rejected": -253.4725799560547,
      "loss": 0.1048,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 2.511007308959961,
      "rewards/margins": 4.834166049957275,
      "rewards/rejected": -2.3231587409973145,
      "step": 145
    },
    {
      "epoch": 2.803738317757009,
      "grad_norm": 14.732641015320004,
      "learning_rate": 8.975292939244927e-09,
      "logits/chosen": -2.3658361434936523,
      "logits/rejected": -2.349327564239502,
      "logps/chosen": -226.5908660888672,
      "logps/rejected": -223.129638671875,
      "loss": 0.1203,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 2.179231643676758,
      "rewards/margins": 4.109222412109375,
      "rewards/rejected": -1.9299901723861694,
      "step": 150
    },
    {
      "epoch": 2.803738317757009,
      "eval_logits/chosen": -2.393800735473633,
      "eval_logits/rejected": -2.363922357559204,
      "eval_logps/chosen": -255.3956756591797,
      "eval_logps/rejected": -247.4010467529297,
      "eval_loss": 0.5898510813713074,
      "eval_rewards/accuracies": 0.7760416865348816,
      "eval_rewards/chosen": 1.1373103857040405,
      "eval_rewards/margins": 1.9750633239746094,
      "eval_rewards/rejected": -0.8377528190612793,
      "eval_runtime": 100.7549,
      "eval_samples_per_second": 15.086,
      "eval_steps_per_second": 0.238,
      "step": 150
    },
    {
      "epoch": 2.897196261682243,
      "grad_norm": 12.576875515019212,
      "learning_rate": 1.7771732184357901e-09,
      "logits/chosen": -2.388183355331421,
      "logits/rejected": -2.389042615890503,
      "logps/chosen": -246.50845336914062,
      "logps/rejected": -217.6068878173828,
      "loss": 0.1134,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 2.572183132171631,
      "rewards/margins": 4.353964805603027,
      "rewards/rejected": -1.781781554222107,
      "step": 155
    },
    {
      "epoch": 2.97196261682243,
      "step": 159,
      "total_flos": 1874604243025920.0,
      "train_loss": 0.33037466410570926,
      "train_runtime": 5730.1243,
      "train_samples_per_second": 7.16,
      "train_steps_per_second": 0.028
    }
  ],
  "logging_steps": 5,
  "max_steps": 159,
  "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": 1874604243025920.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}