{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06,
      "learning_rate": 1.0416666666666667e-07,
      "logps/chosen": -46.291481018066406,
      "logps/rejected": -63.38591384887695,
      "loss": 0.6933,
      "losses/dpo": 0.6924217939376831,
      "losses/sft": 1.2695705890655518,
      "losses/total": 0.6924217939376831,
      "ref_logps/chosen": -46.2892951965332,
      "ref_logps/rejected": -63.38469314575195,
      "rewards/accuracies": 0.46150001883506775,
      "rewards/chosen": -0.00021846062736585736,
      "rewards/margins": -9.587412932887673e-05,
      "rewards/rejected": -0.00012258654169272631,
      "step": 25
    },
    {
      "epoch": 0.12,
      "learning_rate": 2.0833333333333333e-07,
      "logps/chosen": -44.6225471496582,
      "logps/rejected": -59.81039047241211,
      "loss": 0.6932,
      "losses/dpo": 0.6935679912567139,
      "losses/sft": 1.2453413009643555,
      "losses/total": 0.6935679912567139,
      "ref_logps/chosen": -44.617828369140625,
      "ref_logps/rejected": -59.8062744140625,
      "rewards/accuracies": 0.503000020980835,
      "rewards/chosen": -0.00047206657472997904,
      "rewards/margins": -6.079748345655389e-05,
      "rewards/rejected": -0.00041126925498247147,
      "step": 50
    },
    {
      "epoch": 0.19,
      "learning_rate": 3.1249999999999997e-07,
      "logps/chosen": -43.78951644897461,
      "logps/rejected": -59.19944381713867,
      "loss": 0.6921,
      "losses/dpo": 0.6920485496520996,
      "losses/sft": 1.2684202194213867,
      "losses/total": 0.6920485496520996,
      "ref_logps/chosen": -43.78016662597656,
      "ref_logps/rejected": -59.168540954589844,
      "rewards/accuracies": 0.5205000042915344,
      "rewards/chosen": -0.0009349272586405277,
      "rewards/margins": 0.00215632701292634,
      "rewards/rejected": -0.003091254271566868,
      "step": 75
    },
    {
      "epoch": 0.25,
      "learning_rate": 4.1666666666666667e-07,
      "logps/chosen": -44.4240837097168,
      "logps/rejected": -58.47013854980469,
      "loss": 0.6901,
      "losses/dpo": 0.6897304654121399,
      "losses/sft": 1.2445145845413208,
      "losses/total": 0.6897304654121399,
      "ref_logps/chosen": -44.39784622192383,
      "ref_logps/rejected": -58.38094711303711,
      "rewards/accuracies": 0.5600000619888306,
      "rewards/chosen": -0.0026234728284180164,
      "rewards/margins": 0.0062958355993032455,
      "rewards/rejected": -0.0089193070307374,
      "step": 100
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.976851851851852e-07,
      "logps/chosen": -44.144630432128906,
      "logps/rejected": -60.08927917480469,
      "loss": 0.6856,
      "losses/dpo": 0.6853728294372559,
      "losses/sft": 1.209934115409851,
      "losses/total": 0.6853728294372559,
      "ref_logps/chosen": -44.078399658203125,
      "ref_logps/rejected": -59.86705017089844,
      "rewards/accuracies": 0.5929999947547913,
      "rewards/chosen": -0.006622872781008482,
      "rewards/margins": 0.015600004233419895,
      "rewards/rejected": -0.02222287654876709,
      "step": 125
    },
    {
      "epoch": 0.38,
      "learning_rate": 4.861111111111111e-07,
      "logps/chosen": -45.84950256347656,
      "logps/rejected": -62.66447830200195,
      "loss": 0.6775,
      "losses/dpo": 0.6779772043228149,
      "losses/sft": 1.281442403793335,
      "losses/total": 0.6779772043228149,
      "ref_logps/chosen": -45.6851921081543,
      "ref_logps/rejected": -62.169559478759766,
      "rewards/accuracies": 0.6209999918937683,
      "rewards/chosen": -0.01643071323633194,
      "rewards/margins": 0.033061932772397995,
      "rewards/rejected": -0.049492646008729935,
      "step": 150
    },
    {
      "epoch": 0.44,
      "learning_rate": 4.74537037037037e-07,
      "logps/chosen": -44.53513717651367,
      "logps/rejected": -61.487178802490234,
      "loss": 0.6684,
      "losses/dpo": 0.6683259010314941,
      "losses/sft": 1.2735780477523804,
      "losses/total": 0.6683259010314941,
      "ref_logps/chosen": -44.296630859375,
      "ref_logps/rejected": -60.71174240112305,
      "rewards/accuracies": 0.6439999938011169,
      "rewards/chosen": -0.02384989708662033,
      "rewards/margins": 0.053693462163209915,
      "rewards/rejected": -0.07754335552453995,
      "step": 175
    },
    {
      "epoch": 0.5,
      "learning_rate": 4.6296296296296297e-07,
      "logps/chosen": -46.467018127441406,
      "logps/rejected": -63.491580963134766,
      "loss": 0.6563,
      "losses/dpo": 0.6521183252334595,
      "losses/sft": 1.2733054161071777,
      "losses/total": 0.6521183252334595,
      "ref_logps/chosen": -46.05796432495117,
      "ref_logps/rejected": -62.249488830566406,
      "rewards/accuracies": 0.6285000443458557,
      "rewards/chosen": -0.040905579924583435,
      "rewards/margins": 0.0833037793636322,
      "rewards/rejected": -0.12420935928821564,
      "step": 200
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.513888888888889e-07,
      "logps/chosen": -46.178504943847656,
      "logps/rejected": -64.19964599609375,
      "loss": 0.6429,
      "losses/dpo": 0.6435222029685974,
      "losses/sft": 1.2407046556472778,
      "losses/total": 0.6435222029685974,
      "ref_logps/chosen": -45.60074234008789,
      "ref_logps/rejected": -62.432167053222656,
      "rewards/accuracies": 0.6385000348091125,
      "rewards/chosen": -0.057775672525167465,
      "rewards/margins": 0.11897158622741699,
      "rewards/rejected": -0.17674726247787476,
      "step": 225
    },
    {
      "epoch": 0.62,
      "learning_rate": 4.398148148148148e-07,
      "logps/chosen": -46.30465316772461,
      "logps/rejected": -62.2493782043457,
      "loss": 0.6363,
      "losses/dpo": 0.634566605091095,
      "losses/sft": 1.2551387548446655,
      "losses/total": 0.634566605091095,
      "ref_logps/chosen": -45.55400085449219,
      "ref_logps/rejected": -60.08576583862305,
      "rewards/accuracies": 0.6335000395774841,
      "rewards/chosen": -0.07506560534238815,
      "rewards/margins": 0.1412954032421112,
      "rewards/rejected": -0.21636100113391876,
      "step": 250
    },
    {
      "epoch": 0.69,
      "learning_rate": 4.2824074074074075e-07,
      "logps/chosen": -45.26353073120117,
      "logps/rejected": -62.119075775146484,
      "loss": 0.6234,
      "losses/dpo": 0.6106441020965576,
      "losses/sft": 1.3094475269317627,
      "losses/total": 0.6106441020965576,
      "ref_logps/chosen": -44.300174713134766,
      "ref_logps/rejected": -59.338504791259766,
      "rewards/accuracies": 0.6434999704360962,
      "rewards/chosen": -0.09633561968803406,
      "rewards/margins": 0.1817215383052826,
      "rewards/rejected": -0.27805712819099426,
      "step": 275
    },
    {
      "epoch": 0.75,
      "learning_rate": 4.1666666666666667e-07,
      "logps/chosen": -45.76985168457031,
      "logps/rejected": -63.50913619995117,
      "loss": 0.613,
      "losses/dpo": 0.6014246940612793,
      "losses/sft": 1.2986456155776978,
      "losses/total": 0.6014246940612793,
      "ref_logps/chosen": -44.60796356201172,
      "ref_logps/rejected": -60.128543853759766,
      "rewards/accuracies": 0.6544999480247498,
      "rewards/chosen": -0.11618894338607788,
      "rewards/margins": 0.2218700349330902,
      "rewards/rejected": -0.3380589783191681,
      "step": 300
    },
    {
      "epoch": 0.81,
      "learning_rate": 4.050925925925926e-07,
      "logps/chosen": -46.350311279296875,
      "logps/rejected": -64.47545623779297,
      "loss": 0.6045,
      "losses/dpo": 0.6078605055809021,
      "losses/sft": 1.2730615139007568,
      "losses/total": 0.6078605055809021,
      "ref_logps/chosen": -44.98001480102539,
      "ref_logps/rejected": -60.557586669921875,
      "rewards/accuracies": 0.6600000262260437,
      "rewards/chosen": -0.13702912628650665,
      "rewards/margins": 0.25475841760635376,
      "rewards/rejected": -0.3917875587940216,
      "step": 325
    },
    {
      "epoch": 0.88,
      "learning_rate": 3.9351851851851854e-07,
      "logps/chosen": -46.390987396240234,
      "logps/rejected": -65.16218566894531,
      "loss": 0.5944,
      "losses/dpo": 0.5973597168922424,
      "losses/sft": 1.3051469326019287,
      "losses/total": 0.5973597168922424,
      "ref_logps/chosen": -44.8856086730957,
      "ref_logps/rejected": -60.589107513427734,
      "rewards/accuracies": 0.6424999833106995,
      "rewards/chosen": -0.15053769946098328,
      "rewards/margins": 0.30677077174186707,
      "rewards/rejected": -0.45730850100517273,
      "step": 350
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.819444444444444e-07,
      "logps/chosen": -46.03995132446289,
      "logps/rejected": -63.268062591552734,
      "loss": 0.5952,
      "losses/dpo": 0.5992311239242554,
      "losses/sft": 1.2882429361343384,
      "losses/total": 0.5992311239242554,
      "ref_logps/chosen": -44.33562469482422,
      "ref_logps/rejected": -58.439823150634766,
      "rewards/accuracies": 0.6640000343322754,
      "rewards/chosen": -0.17043215036392212,
      "rewards/margins": 0.31239163875579834,
      "rewards/rejected": -0.48282381892204285,
      "step": 375
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.703703703703703e-07,
      "logps/chosen": -45.34758377075195,
      "logps/rejected": -63.79766082763672,
      "loss": 0.5798,
      "losses/dpo": 0.573971152305603,
      "losses/sft": 1.2241010665893555,
      "losses/total": 0.573971152305603,
      "ref_logps/chosen": -43.607730865478516,
      "ref_logps/rejected": -58.310462951660156,
      "rewards/accuracies": 0.6694999933242798,
      "rewards/chosen": -0.17398566007614136,
      "rewards/margins": 0.3747338056564331,
      "rewards/rejected": -0.5487195253372192,
      "step": 400
    },
    {
      "epoch": 1.06,
      "learning_rate": 3.587962962962963e-07,
      "logps/chosen": -45.958282470703125,
      "logps/rejected": -63.922576904296875,
      "loss": 0.5783,
      "losses/dpo": 0.5742554068565369,
      "losses/sft": 1.2882492542266846,
      "losses/total": 0.5742554068565369,
      "ref_logps/chosen": -43.95150375366211,
      "ref_logps/rejected": -58.02607345581055,
      "rewards/accuracies": 0.6690000295639038,
      "rewards/chosen": -0.20067782700061798,
      "rewards/margins": 0.38897278904914856,
      "rewards/rejected": -0.5896506309509277,
      "step": 425
    },
    {
      "epoch": 1.12,
      "learning_rate": 3.472222222222222e-07,
      "logps/chosen": -47.257179260253906,
      "logps/rejected": -68.16796875,
      "loss": 0.5616,
      "losses/dpo": 0.5516221523284912,
      "losses/sft": 1.2813299894332886,
      "losses/total": 0.5516221523284912,
      "ref_logps/chosen": -45.21623611450195,
      "ref_logps/rejected": -61.444461822509766,
      "rewards/accuracies": 0.6890000104904175,
      "rewards/chosen": -0.20409366488456726,
      "rewards/margins": 0.4682568609714508,
      "rewards/rejected": -0.6723506450653076,
      "step": 450
    },
    {
      "epoch": 1.19,
      "learning_rate": 3.3564814814814815e-07,
      "logps/chosen": -46.08806610107422,
      "logps/rejected": -65.21028137207031,
      "loss": 0.5651,
      "losses/dpo": 0.554095983505249,
      "losses/sft": 1.3418306112289429,
      "losses/total": 0.554095983505249,
      "ref_logps/chosen": -43.922706604003906,
      "ref_logps/rejected": -58.401084899902344,
      "rewards/accuracies": 0.6819999814033508,
      "rewards/chosen": -0.21653667092323303,
      "rewards/margins": 0.4643844664096832,
      "rewards/rejected": -0.680921196937561,
      "step": 475
    },
    {
      "epoch": 1.25,
      "learning_rate": 3.2407407407407406e-07,
      "logps/chosen": -46.008277893066406,
      "logps/rejected": -66.00773620605469,
      "loss": 0.5556,
      "losses/dpo": 0.5685967803001404,
      "losses/sft": 1.2652785778045654,
      "losses/total": 0.5685967803001404,
      "ref_logps/chosen": -43.85118103027344,
      "ref_logps/rejected": -58.737857818603516,
      "rewards/accuracies": 0.6944999694824219,
      "rewards/chosen": -0.21570871770381927,
      "rewards/margins": 0.5112798810005188,
      "rewards/rejected": -0.7269885540008545,
      "step": 500
    },
    {
      "epoch": 1.31,
      "learning_rate": 3.1249999999999997e-07,
      "logps/chosen": -45.792293548583984,
      "logps/rejected": -68.81733703613281,
      "loss": 0.5479,
      "losses/dpo": 0.5653673410415649,
      "losses/sft": 1.2111891508102417,
      "losses/total": 0.5653673410415649,
      "ref_logps/chosen": -43.507266998291016,
      "ref_logps/rejected": -60.62786102294922,
      "rewards/accuracies": 0.6984999775886536,
      "rewards/chosen": -0.2285033017396927,
      "rewards/margins": 0.5904435515403748,
      "rewards/rejected": -0.8189470171928406,
      "step": 525
    },
    {
      "epoch": 1.38,
      "learning_rate": 3.0092592592592594e-07,
      "logps/chosen": -48.979248046875,
      "logps/rejected": -70.8954849243164,
      "loss": 0.5332,
      "losses/dpo": 0.5159104466438293,
      "losses/sft": 1.4047646522521973,
      "losses/total": 0.5159104466438293,
      "ref_logps/chosen": -46.33369064331055,
      "ref_logps/rejected": -61.91999816894531,
      "rewards/accuracies": 0.7049999833106995,
      "rewards/chosen": -0.2645554840564728,
      "rewards/margins": 0.6329929828643799,
      "rewards/rejected": -0.897548496723175,
      "step": 550
    },
    {
      "epoch": 1.44,
      "learning_rate": 2.8935185185185185e-07,
      "logps/chosen": -47.573795318603516,
      "logps/rejected": -71.3143081665039,
      "loss": 0.5276,
      "losses/dpo": 0.5273745656013489,
      "losses/sft": 1.329495906829834,
      "losses/total": 0.5273745656013489,
      "ref_logps/chosen": -44.873233795166016,
      "ref_logps/rejected": -61.62533187866211,
      "rewards/accuracies": 0.6990000009536743,
      "rewards/chosen": -0.27005600929260254,
      "rewards/margins": 0.698841392993927,
      "rewards/rejected": -0.9688975811004639,
      "step": 575
    },
    {
      "epoch": 1.5,
      "learning_rate": 2.7777777777777776e-07,
      "logps/chosen": -48.737342834472656,
      "logps/rejected": -71.09253692626953,
      "loss": 0.5234,
      "losses/dpo": 0.5140664577484131,
      "losses/sft": 1.3185381889343262,
      "losses/total": 0.5140664577484131,
      "ref_logps/chosen": -45.823787689208984,
      "ref_logps/rejected": -60.519901275634766,
      "rewards/accuracies": 0.7139999866485596,
      "rewards/chosen": -0.2913552522659302,
      "rewards/margins": 0.7659080624580383,
      "rewards/rejected": -1.0572632551193237,
      "step": 600
    },
    {
      "epoch": 1.56,
      "learning_rate": 2.662037037037037e-07,
      "logps/chosen": -46.89459991455078,
      "logps/rejected": -70.45913696289062,
      "loss": 0.5194,
      "losses/dpo": 0.5285363793373108,
      "losses/sft": 1.3135132789611816,
      "losses/total": 0.5285363793373108,
      "ref_logps/chosen": -44.250526428222656,
      "ref_logps/rejected": -59.6305046081543,
      "rewards/accuracies": 0.7170000672340393,
      "rewards/chosen": -0.2644067704677582,
      "rewards/margins": 0.8184568881988525,
      "rewards/rejected": -1.0828635692596436,
      "step": 625
    },
    {
      "epoch": 1.62,
      "learning_rate": 2.5462962962962963e-07,
      "logps/chosen": -48.4859619140625,
      "logps/rejected": -71.74458312988281,
      "loss": 0.522,
      "losses/dpo": 0.5093265771865845,
      "losses/sft": 1.342267632484436,
      "losses/total": 0.5093265771865845,
      "ref_logps/chosen": -45.3779182434082,
      "ref_logps/rejected": -60.191993713378906,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.3108051121234894,
      "rewards/margins": 0.8444538712501526,
      "rewards/rejected": -1.1552588939666748,
      "step": 650
    },
    {
      "epoch": 1.69,
      "learning_rate": 2.4305555555555555e-07,
      "logps/chosen": -47.598331451416016,
      "logps/rejected": -71.73502349853516,
      "loss": 0.5051,
      "losses/dpo": 0.5321754813194275,
      "losses/sft": 1.288547158241272,
      "losses/total": 0.5321754813194275,
      "ref_logps/chosen": -44.57024383544922,
      "ref_logps/rejected": -59.632904052734375,
      "rewards/accuracies": 0.734499990940094,
      "rewards/chosen": -0.3028090000152588,
      "rewards/margins": 0.9074033498764038,
      "rewards/rejected": -1.2102123498916626,
      "step": 675
    },
    {
      "epoch": 1.75,
      "learning_rate": 2.3148148148148148e-07,
      "logps/chosen": -48.9836540222168,
      "logps/rejected": -75.99896240234375,
      "loss": 0.5039,
      "losses/dpo": 0.4898015558719635,
      "losses/sft": 1.3348312377929688,
      "losses/total": 0.4898015558719635,
      "ref_logps/chosen": -45.258949279785156,
      "ref_logps/rejected": -62.34097671508789,
      "rewards/accuracies": 0.7209999561309814,
      "rewards/chosen": -0.37246978282928467,
      "rewards/margins": 0.993329644203186,
      "rewards/rejected": -1.3657993078231812,
      "step": 700
    },
    {
      "epoch": 1.81,
      "learning_rate": 2.199074074074074e-07,
      "logps/chosen": -49.37858581542969,
      "logps/rejected": -76.06786346435547,
      "loss": 0.5037,
      "losses/dpo": 0.4792703688144684,
      "losses/sft": 1.3326802253723145,
      "losses/total": 0.4792703688144684,
      "ref_logps/chosen": -45.24223709106445,
      "ref_logps/rejected": -61.8316764831543,
      "rewards/accuracies": 0.7134999632835388,
      "rewards/chosen": -0.4136350154876709,
      "rewards/margins": 1.0099844932556152,
      "rewards/rejected": -1.4236197471618652,
      "step": 725
    },
    {
      "epoch": 1.88,
      "learning_rate": 2.0833333333333333e-07,
      "logps/chosen": -48.57896041870117,
      "logps/rejected": -71.33642578125,
      "loss": 0.511,
      "losses/dpo": 0.5458227396011353,
      "losses/sft": 1.2892143726348877,
      "losses/total": 0.5458227396011353,
      "ref_logps/chosen": -44.61341857910156,
      "ref_logps/rejected": -57.60921859741211,
      "rewards/accuracies": 0.7119999527931213,
      "rewards/chosen": -0.39655402302742004,
      "rewards/margins": 0.9761665463447571,
      "rewards/rejected": -1.3727205991744995,
      "step": 750
    },
    {
      "epoch": 1.94,
      "learning_rate": 1.9675925925925927e-07,
      "logps/chosen": -48.83073425292969,
      "logps/rejected": -76.26995849609375,
      "loss": 0.4896,
      "losses/dpo": 0.4852687120437622,
      "losses/sft": 1.3753533363342285,
      "losses/total": 0.4852687120437622,
      "ref_logps/chosen": -44.98310089111328,
      "ref_logps/rejected": -61.261417388916016,
      "rewards/accuracies": 0.7349998950958252,
      "rewards/chosen": -0.38476306200027466,
      "rewards/margins": 1.1160913705825806,
      "rewards/rejected": -1.5008544921875,
      "step": 775
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.8518518518518516e-07,
      "logps/chosen": -49.678104400634766,
      "logps/rejected": -76.78030395507812,
      "loss": 0.5058,
      "losses/dpo": 0.4918399751186371,
      "losses/sft": 1.373268485069275,
      "losses/total": 0.4918399751186371,
      "ref_logps/chosen": -45.21029281616211,
      "ref_logps/rejected": -61.68964767456055,
      "rewards/accuracies": 0.7149999737739563,
      "rewards/chosen": -0.44678181409835815,
      "rewards/margins": 1.0622824430465698,
      "rewards/rejected": -1.5090643167495728,
      "step": 800
    },
    {
      "epoch": 2.06,
      "learning_rate": 1.736111111111111e-07,
      "logps/chosen": -49.76744079589844,
      "logps/rejected": -77.67528533935547,
      "loss": 0.4877,
      "losses/dpo": 0.47753456234931946,
      "losses/sft": 1.3492231369018555,
      "losses/total": 0.47753456234931946,
      "ref_logps/chosen": -45.20912170410156,
      "ref_logps/rejected": -61.64081954956055,
      "rewards/accuracies": 0.7359998822212219,
      "rewards/chosen": -0.45583218336105347,
      "rewards/margins": 1.1476140022277832,
      "rewards/rejected": -1.6034462451934814,
      "step": 825
    },
    {
      "epoch": 2.12,
      "learning_rate": 1.6203703703703703e-07,
      "logps/chosen": -48.476280212402344,
      "logps/rejected": -74.0830078125,
      "loss": 0.4932,
      "losses/dpo": 0.4669117331504822,
      "losses/sft": 1.3778959512710571,
      "losses/total": 0.4669117331504822,
      "ref_logps/chosen": -44.10759735107422,
      "ref_logps/rejected": -58.72689437866211,
      "rewards/accuracies": 0.7190000414848328,
      "rewards/chosen": -0.43686822056770325,
      "rewards/margins": 1.098743200302124,
      "rewards/rejected": -1.5356113910675049,
      "step": 850
    },
    {
      "epoch": 2.19,
      "learning_rate": 1.5046296296296297e-07,
      "logps/chosen": -49.38998031616211,
      "logps/rejected": -76.47419738769531,
      "loss": 0.5013,
      "losses/dpo": 0.4948745667934418,
      "losses/sft": 1.3688327074050903,
      "losses/total": 0.4948745667934418,
      "ref_logps/chosen": -44.37105941772461,
      "ref_logps/rejected": -60.47099304199219,
      "rewards/accuracies": 0.7259998917579651,
      "rewards/chosen": -0.5018922686576843,
      "rewards/margins": 1.0984277725219727,
      "rewards/rejected": -1.6003201007843018,
      "step": 875
    },
    {
      "epoch": 2.25,
      "learning_rate": 1.3888888888888888e-07,
      "logps/chosen": -50.32158279418945,
      "logps/rejected": -77.23485565185547,
      "loss": 0.4968,
      "losses/dpo": 0.5039748549461365,
      "losses/sft": 1.3766621351242065,
      "losses/total": 0.5039748549461365,
      "ref_logps/chosen": -45.48880386352539,
      "ref_logps/rejected": -60.92485809326172,
      "rewards/accuracies": 0.7320000529289246,
      "rewards/chosen": -0.48327693343162537,
      "rewards/margins": 1.1477227210998535,
      "rewards/rejected": -1.6309998035430908,
      "step": 900
    },
    {
      "epoch": 2.31,
      "learning_rate": 1.2731481481481482e-07,
      "logps/chosen": -49.106754302978516,
      "logps/rejected": -75.4212417602539,
      "loss": 0.5004,
      "losses/dpo": 0.465129017829895,
      "losses/sft": 1.396937370300293,
      "losses/total": 0.465129017829895,
      "ref_logps/chosen": -44.43598556518555,
      "ref_logps/rejected": -59.51651382446289,
      "rewards/accuracies": 0.7200000286102295,
      "rewards/chosen": -0.4670768678188324,
      "rewards/margins": 1.123395562171936,
      "rewards/rejected": -1.5904724597930908,
      "step": 925
    },
    {
      "epoch": 2.38,
      "learning_rate": 1.1574074074074074e-07,
      "logps/chosen": -49.315948486328125,
      "logps/rejected": -75.77275848388672,
      "loss": 0.4952,
      "losses/dpo": 0.48752278089523315,
      "losses/sft": 1.383622646331787,
      "losses/total": 0.48752278089523315,
      "ref_logps/chosen": -44.68310546875,
      "ref_logps/rejected": -59.506874084472656,
      "rewards/accuracies": 0.7290000915527344,
      "rewards/chosen": -0.4632847309112549,
      "rewards/margins": 1.1633038520812988,
      "rewards/rejected": -1.6265885829925537,
      "step": 950
    },
    {
      "epoch": 2.44,
      "learning_rate": 1.0416666666666667e-07,
      "logps/chosen": -49.365966796875,
      "logps/rejected": -76.18710327148438,
      "loss": 0.4945,
      "losses/dpo": 0.5000871419906616,
      "losses/sft": 1.345529317855835,
      "losses/total": 0.5000871419906616,
      "ref_logps/chosen": -44.74851989746094,
      "ref_logps/rejected": -60.123268127441406,
      "rewards/accuracies": 0.7280000448226929,
      "rewards/chosen": -0.46174487471580505,
      "rewards/margins": 1.1446377038955688,
      "rewards/rejected": -1.6063826084136963,
      "step": 975
    },
    {
      "epoch": 2.5,
      "learning_rate": 9.259259259259258e-08,
      "logps/chosen": -49.790008544921875,
      "logps/rejected": -76.8461685180664,
      "loss": 0.4891,
      "losses/dpo": 0.5052517056465149,
      "losses/sft": 1.3480595350265503,
      "losses/total": 0.5052517056465149,
      "ref_logps/chosen": -44.7381477355957,
      "ref_logps/rejected": -60.120697021484375,
      "rewards/accuracies": 0.7300000786781311,
      "rewards/chosen": -0.505186140537262,
      "rewards/margins": 1.167360782623291,
      "rewards/rejected": -1.6725467443466187,
      "step": 1000
    },
    {
      "epoch": 2.56,
      "learning_rate": 8.101851851851852e-08,
      "logps/chosen": -48.60505676269531,
      "logps/rejected": -75.21623992919922,
      "loss": 0.4956,
      "losses/dpo": 0.4796887934207916,
      "losses/sft": 1.3435637950897217,
      "losses/total": 0.4796887934207916,
      "ref_logps/chosen": -43.89369583129883,
      "ref_logps/rejected": -59.03510665893555,
      "rewards/accuracies": 0.718000054359436,
      "rewards/chosen": -0.47113633155822754,
      "rewards/margins": 1.146977186203003,
      "rewards/rejected": -1.618113398551941,
      "step": 1025
    },
    {
      "epoch": 2.62,
      "learning_rate": 6.944444444444444e-08,
      "logps/chosen": -49.14812469482422,
      "logps/rejected": -77.22814178466797,
      "loss": 0.4816,
      "losses/dpo": 0.49271050095558167,
      "losses/sft": 1.3505686521530151,
      "losses/total": 0.49271050095558167,
      "ref_logps/chosen": -44.201717376708984,
      "ref_logps/rejected": -60.2335205078125,
      "rewards/accuracies": 0.7355000376701355,
      "rewards/chosen": -0.4946403503417969,
      "rewards/margins": 1.2048207521438599,
      "rewards/rejected": -1.6994611024856567,
      "step": 1050
    },
    {
      "epoch": 2.69,
      "learning_rate": 5.787037037037037e-08,
      "logps/chosen": -50.257972717285156,
      "logps/rejected": -79.50613403320312,
      "loss": 0.4888,
      "losses/dpo": 0.4835771918296814,
      "losses/sft": 1.399458408355713,
      "losses/total": 0.4835771918296814,
      "ref_logps/chosen": -44.78007888793945,
      "ref_logps/rejected": -61.73722839355469,
      "rewards/accuracies": 0.7374999523162842,
      "rewards/chosen": -0.5477903485298157,
      "rewards/margins": 1.2291010618209839,
      "rewards/rejected": -1.7768914699554443,
      "step": 1075
    },
    {
      "epoch": 2.75,
      "learning_rate": 4.629629629629629e-08,
      "logps/chosen": -50.44590759277344,
      "logps/rejected": -80.34747314453125,
      "loss": 0.4749,
      "losses/dpo": 0.4630358815193176,
      "losses/sft": 1.3918135166168213,
      "losses/total": 0.4630358815193176,
      "ref_logps/chosen": -45.468910217285156,
      "ref_logps/rejected": -62.39448547363281,
      "rewards/accuracies": 0.7430000305175781,
      "rewards/chosen": -0.497699499130249,
      "rewards/margins": 1.297598123550415,
      "rewards/rejected": -1.7952975034713745,
      "step": 1100
    },
    {
      "epoch": 2.81,
      "learning_rate": 3.472222222222222e-08,
      "logps/chosen": -50.564212799072266,
      "logps/rejected": -77.6652603149414,
      "loss": 0.4844,
      "losses/dpo": 0.4967005252838135,
      "losses/sft": 1.3921661376953125,
      "losses/total": 0.4967005252838135,
      "ref_logps/chosen": -45.257930755615234,
      "ref_logps/rejected": -59.98822784423828,
      "rewards/accuracies": 0.7325000166893005,
      "rewards/chosen": -0.5306286811828613,
      "rewards/margins": 1.237074613571167,
      "rewards/rejected": -1.7677034139633179,
      "step": 1125
    },
    {
      "epoch": 2.88,
      "learning_rate": 2.3148148148148144e-08,
      "logps/chosen": -49.5255126953125,
      "logps/rejected": -78.27824401855469,
      "loss": 0.4747,
      "losses/dpo": 0.48569077253341675,
      "losses/sft": 1.3604434728622437,
      "losses/total": 0.48569077253341675,
      "ref_logps/chosen": -44.655574798583984,
      "ref_logps/rejected": -60.496063232421875,
      "rewards/accuracies": 0.7485000491142273,
      "rewards/chosen": -0.48699355125427246,
      "rewards/margins": 1.2912240028381348,
      "rewards/rejected": -1.7782177925109863,
      "step": 1150
    },
    {
      "epoch": 2.94,
      "learning_rate": 1.1574074074074072e-08,
      "logps/chosen": -50.011077880859375,
      "logps/rejected": -75.67598724365234,
      "loss": 0.4876,
      "losses/dpo": 0.5032810568809509,
      "losses/sft": 1.366944432258606,
      "losses/total": 0.5032810568809509,
      "ref_logps/chosen": -44.80411148071289,
      "ref_logps/rejected": -58.415794372558594,
      "rewards/accuracies": 0.7414999604225159,
      "rewards/chosen": -0.5206969976425171,
      "rewards/margins": 1.2053215503692627,
      "rewards/rejected": -1.7260186672210693,
      "step": 1175
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0,
      "logps/chosen": -51.40888595581055,
      "logps/rejected": -80.48424530029297,
      "loss": 0.469,
      "losses/dpo": 0.4639938771724701,
      "losses/sft": 1.3975021839141846,
      "losses/total": 0.4639938771724701,
      "ref_logps/chosen": -46.229427337646484,
      "ref_logps/rejected": -62.2370491027832,
      "rewards/accuracies": 0.7464999556541443,
      "rewards/chosen": -0.5179460644721985,
      "rewards/margins": 1.3067736625671387,
      "rewards/rejected": -1.8247196674346924,
      "step": 1200
    },
    {
      "epoch": 3.0,
      "step": 1200,
      "total_flos": 0.0,
      "train_loss": 0.5544587286313375,
      "train_runtime": 33858.4559,
      "train_samples_per_second": 2.835,
      "train_steps_per_second": 0.035
    }
  ],
  "logging_steps": 25,
  "max_steps": 1200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 24000,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}