BraylonDash's picture
Model save
249e3b5 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.99968,
"eval_steps": 100,
"global_step": 1562,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 3.184713375796179e-08,
"logits/chosen": -0.9295870065689087,
"logits/rejected": -0.43873703479766846,
"logps/chosen": -320.5160827636719,
"logps/rejected": -293.4969482421875,
"loss": 0.0436,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.01,
"learning_rate": 3.1847133757961787e-07,
"logits/chosen": -0.8594987988471985,
"logits/rejected": -0.762122631072998,
"logps/chosen": -363.58154296875,
"logps/rejected": -320.7466735839844,
"loss": 0.1005,
"rewards/accuracies": 0.4444444477558136,
"rewards/chosen": 0.00017958095122594386,
"rewards/margins": 0.00031936122104525566,
"rewards/rejected": -0.0001397802698193118,
"step": 10
},
{
"epoch": 0.01,
"learning_rate": 6.369426751592357e-07,
"logits/chosen": -0.8261665105819702,
"logits/rejected": -0.8099607229232788,
"logps/chosen": -329.95257568359375,
"logps/rejected": -333.58843994140625,
"loss": 0.0881,
"rewards/accuracies": 0.3812499940395355,
"rewards/chosen": 1.5792587873875163e-05,
"rewards/margins": 9.325530118076131e-05,
"rewards/rejected": -7.746272603981197e-05,
"step": 20
},
{
"epoch": 0.02,
"learning_rate": 9.554140127388537e-07,
"logits/chosen": -0.8844400644302368,
"logits/rejected": -0.8693345189094543,
"logps/chosen": -266.14617919921875,
"logps/rejected": -239.8477020263672,
"loss": 0.1076,
"rewards/accuracies": 0.33125001192092896,
"rewards/chosen": -0.00019076233729720116,
"rewards/margins": -0.00011497503146529198,
"rewards/rejected": -7.578730583190918e-05,
"step": 30
},
{
"epoch": 0.03,
"learning_rate": 1.2738853503184715e-06,
"logits/chosen": -0.9354842305183411,
"logits/rejected": -0.8044538497924805,
"logps/chosen": -337.16436767578125,
"logps/rejected": -294.0509338378906,
"loss": 0.1014,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.0002183823671657592,
"rewards/margins": 0.00015057336713653058,
"rewards/rejected": -0.00036895571975037456,
"step": 40
},
{
"epoch": 0.03,
"learning_rate": 1.5923566878980892e-06,
"logits/chosen": -0.9226251840591431,
"logits/rejected": -0.7269760966300964,
"logps/chosen": -291.8105773925781,
"logps/rejected": -249.9472198486328,
"loss": 0.0984,
"rewards/accuracies": 0.35624998807907104,
"rewards/chosen": -1.6449857866973616e-05,
"rewards/margins": -4.999707016395405e-05,
"rewards/rejected": 3.354722139192745e-05,
"step": 50
},
{
"epoch": 0.04,
"learning_rate": 1.9108280254777074e-06,
"logits/chosen": -0.7843996286392212,
"logits/rejected": -0.8503357768058777,
"logps/chosen": -314.31842041015625,
"logps/rejected": -272.9629821777344,
"loss": 0.0831,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.0004891738062724471,
"rewards/margins": 3.405969255254604e-05,
"rewards/rejected": -0.0005232334951870143,
"step": 60
},
{
"epoch": 0.04,
"learning_rate": 2.229299363057325e-06,
"logits/chosen": -0.8444937467575073,
"logits/rejected": -0.8687158823013306,
"logps/chosen": -346.148193359375,
"logps/rejected": -314.0673828125,
"loss": 0.101,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.0008814434404484928,
"rewards/margins": 0.0003942731418646872,
"rewards/rejected": -0.00127571658231318,
"step": 70
},
{
"epoch": 0.05,
"learning_rate": 2.547770700636943e-06,
"logits/chosen": -0.7672846913337708,
"logits/rejected": -0.8321496844291687,
"logps/chosen": -308.66973876953125,
"logps/rejected": -290.4983215332031,
"loss": 0.0897,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.0009476385894231498,
"rewards/margins": 0.0005286627565510571,
"rewards/rejected": -0.0014763014623895288,
"step": 80
},
{
"epoch": 0.06,
"learning_rate": 2.8662420382165605e-06,
"logits/chosen": -0.9202607870101929,
"logits/rejected": -0.6941283345222473,
"logps/chosen": -294.6612854003906,
"logps/rejected": -279.28497314453125,
"loss": 0.0967,
"rewards/accuracies": 0.4124999940395355,
"rewards/chosen": -0.0010531718144193292,
"rewards/margins": 0.0009010445210151374,
"rewards/rejected": -0.0019542162772268057,
"step": 90
},
{
"epoch": 0.06,
"learning_rate": 3.1847133757961785e-06,
"logits/chosen": -0.8830461502075195,
"logits/rejected": -0.7728676795959473,
"logps/chosen": -346.04144287109375,
"logps/rejected": -351.6465759277344,
"loss": 0.098,
"rewards/accuracies": 0.4312500059604645,
"rewards/chosen": -0.0019343973835930228,
"rewards/margins": 0.0005748984985984862,
"rewards/rejected": -0.00250929594039917,
"step": 100
},
{
"epoch": 0.06,
"eval_logits/chosen": -0.8170500993728638,
"eval_logits/rejected": -0.7225118279457092,
"eval_logps/chosen": -399.25030517578125,
"eval_logps/rejected": -370.84332275390625,
"eval_loss": 0.053326018154621124,
"eval_rewards/accuracies": 0.49799999594688416,
"eval_rewards/chosen": -0.002945071319118142,
"eval_rewards/margins": 0.0006628122646361589,
"eval_rewards/rejected": -0.0036078833509236574,
"eval_runtime": 539.8002,
"eval_samples_per_second": 3.705,
"eval_steps_per_second": 0.926,
"step": 100
},
{
"epoch": 0.07,
"learning_rate": 3.5031847133757964e-06,
"logits/chosen": -0.8748549222946167,
"logits/rejected": -0.8573201298713684,
"logps/chosen": -356.620361328125,
"logps/rejected": -334.7717590332031,
"loss": 0.0783,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.0018858186667785048,
"rewards/margins": 0.0009479810250923038,
"rewards/rejected": -0.0028337999247014523,
"step": 110
},
{
"epoch": 0.08,
"learning_rate": 3.821656050955415e-06,
"logits/chosen": -0.9678497314453125,
"logits/rejected": -0.7722519636154175,
"logps/chosen": -313.03741455078125,
"logps/rejected": -280.0140686035156,
"loss": 0.1122,
"rewards/accuracies": 0.35624998807907104,
"rewards/chosen": -0.0022745749447494745,
"rewards/margins": 0.0005548128974623978,
"rewards/rejected": -0.0028293877840042114,
"step": 120
},
{
"epoch": 0.08,
"learning_rate": 4.140127388535032e-06,
"logits/chosen": -0.8600385785102844,
"logits/rejected": -0.8846877813339233,
"logps/chosen": -355.20513916015625,
"logps/rejected": -346.33135986328125,
"loss": 0.098,
"rewards/accuracies": 0.3812499940395355,
"rewards/chosen": -0.005060167517513037,
"rewards/margins": -0.00011312137212371454,
"rewards/rejected": -0.004947046283632517,
"step": 130
},
{
"epoch": 0.09,
"learning_rate": 4.45859872611465e-06,
"logits/chosen": -0.848353385925293,
"logits/rejected": -0.8213680386543274,
"logps/chosen": -355.6943054199219,
"logps/rejected": -313.1587829589844,
"loss": 0.0975,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.005685682408511639,
"rewards/margins": 0.0019586696289479733,
"rewards/rejected": -0.007644351571798325,
"step": 140
},
{
"epoch": 0.1,
"learning_rate": 4.777070063694268e-06,
"logits/chosen": -0.9287412762641907,
"logits/rejected": -0.8769465684890747,
"logps/chosen": -329.2186584472656,
"logps/rejected": -298.8802795410156,
"loss": 0.0712,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.00595915038138628,
"rewards/margins": 0.0024761247914284468,
"rewards/rejected": -0.00843527540564537,
"step": 150
},
{
"epoch": 0.1,
"learning_rate": 4.999943753177818e-06,
"logits/chosen": -1.005480170249939,
"logits/rejected": -0.9711716771125793,
"logps/chosen": -311.71209716796875,
"logps/rejected": -299.3721923828125,
"loss": 0.0886,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.00825573317706585,
"rewards/margins": 0.0037352764047682285,
"rewards/rejected": -0.011991010047495365,
"step": 160
},
{
"epoch": 0.11,
"learning_rate": 4.998943880079481e-06,
"logits/chosen": -0.9299044609069824,
"logits/rejected": -0.9373615384101868,
"logps/chosen": -343.8020324707031,
"logps/rejected": -338.0310974121094,
"loss": 0.0924,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.01192331500351429,
"rewards/margins": 0.0040515875443816185,
"rewards/rejected": -0.015974899753928185,
"step": 170
},
{
"epoch": 0.12,
"learning_rate": 4.99669465299937e-06,
"logits/chosen": -0.9855674505233765,
"logits/rejected": -0.9634091258049011,
"logps/chosen": -309.28851318359375,
"logps/rejected": -315.94561767578125,
"loss": 0.0928,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": -0.015265477821230888,
"rewards/margins": 0.005659168586134911,
"rewards/rejected": -0.0209246464073658,
"step": 180
},
{
"epoch": 0.12,
"learning_rate": 4.993197196444851e-06,
"logits/chosen": -1.0295236110687256,
"logits/rejected": -1.1208027601242065,
"logps/chosen": -317.75494384765625,
"logps/rejected": -328.9816589355469,
"loss": 0.0805,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.020896239206194878,
"rewards/margins": 0.007452984340488911,
"rewards/rejected": -0.028349224478006363,
"step": 190
},
{
"epoch": 0.13,
"learning_rate": 4.988453258979111e-06,
"logits/chosen": -1.1082799434661865,
"logits/rejected": -1.0890588760375977,
"logps/chosen": -350.9955139160156,
"logps/rejected": -370.35955810546875,
"loss": 0.094,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -0.02722758986055851,
"rewards/margins": 0.009304001927375793,
"rewards/rejected": -0.036531589925289154,
"step": 200
},
{
"epoch": 0.13,
"eval_logits/chosen": -1.1387817859649658,
"eval_logits/rejected": -1.0754390954971313,
"eval_logps/chosen": -435.2693176269531,
"eval_logps/rejected": -419.6948547363281,
"eval_loss": 0.04912900552153587,
"eval_rewards/accuracies": 0.5525000095367432,
"eval_rewards/chosen": -0.03896407037973404,
"eval_rewards/margins": 0.013495376333594322,
"eval_rewards/rejected": -0.05245944485068321,
"eval_runtime": 539.703,
"eval_samples_per_second": 3.706,
"eval_steps_per_second": 0.926,
"step": 200
},
{
"epoch": 0.13,
"learning_rate": 4.982465212346954e-06,
"logits/chosen": -1.1744426488876343,
"logits/rejected": -1.174753189086914,
"logps/chosen": -324.98370361328125,
"logps/rejected": -325.42767333984375,
"loss": 0.1019,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": -0.031165916472673416,
"rewards/margins": 0.018221553415060043,
"rewards/rejected": -0.04938746988773346,
"step": 210
},
{
"epoch": 0.14,
"learning_rate": 4.975236050289041e-06,
"logits/chosen": -1.1808674335479736,
"logits/rejected": -1.0834966897964478,
"logps/chosen": -334.90087890625,
"logps/rejected": -318.37060546875,
"loss": 0.0697,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": -0.028991717845201492,
"rewards/margins": 0.01802077516913414,
"rewards/rejected": -0.04701249301433563,
"step": 220
},
{
"epoch": 0.15,
"learning_rate": 4.96676938704516e-06,
"logits/chosen": -1.1029061079025269,
"logits/rejected": -1.1640408039093018,
"logps/chosen": -435.73687744140625,
"logps/rejected": -423.72900390625,
"loss": 0.0749,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.040695372968912125,
"rewards/margins": 0.014833291992545128,
"rewards/rejected": -0.0555286630988121,
"step": 230
},
{
"epoch": 0.15,
"learning_rate": 4.95706945554728e-06,
"logits/chosen": -1.1478805541992188,
"logits/rejected": -1.2186776399612427,
"logps/chosen": -370.34503173828125,
"logps/rejected": -392.4799499511719,
"loss": 0.0734,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.0380186066031456,
"rewards/margins": 0.017158757895231247,
"rewards/rejected": -0.05517736077308655,
"step": 240
},
{
"epoch": 0.16,
"learning_rate": 4.9461411053032805e-06,
"logits/chosen": -1.1441445350646973,
"logits/rejected": -1.1573059558868408,
"logps/chosen": -357.66925048828125,
"logps/rejected": -347.88848876953125,
"loss": 0.1096,
"rewards/accuracies": 0.4312500059604645,
"rewards/chosen": -0.03414061293005943,
"rewards/margins": 0.009822173975408077,
"rewards/rejected": -0.043962787836790085,
"step": 250
},
{
"epoch": 0.17,
"learning_rate": 4.933989799972431e-06,
"logits/chosen": -1.1844251155853271,
"logits/rejected": -1.1023800373077393,
"logps/chosen": -362.12890625,
"logps/rejected": -344.55694580078125,
"loss": 0.087,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": -0.023841477930545807,
"rewards/margins": 0.01648232527077198,
"rewards/rejected": -0.04032380133867264,
"step": 260
},
{
"epoch": 0.17,
"learning_rate": 4.920621614633815e-06,
"logits/chosen": -1.158454418182373,
"logits/rejected": -1.1034467220306396,
"logps/chosen": -350.0291748046875,
"logps/rejected": -357.4220886230469,
"loss": 0.094,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.02827698551118374,
"rewards/margins": 0.01563875935971737,
"rewards/rejected": -0.04391574487090111,
"step": 270
},
{
"epoch": 0.18,
"learning_rate": 4.906043232749081e-06,
"logits/chosen": -1.1796420812606812,
"logits/rejected": -0.9934653043746948,
"logps/chosen": -337.7796325683594,
"logps/rejected": -324.3747863769531,
"loss": 0.0808,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": -0.02835283800959587,
"rewards/margins": 0.013374457135796547,
"rewards/rejected": -0.04172729700803757,
"step": 280
},
{
"epoch": 0.19,
"learning_rate": 4.890261942821023e-06,
"logits/chosen": -1.0947716236114502,
"logits/rejected": -1.134075403213501,
"logps/chosen": -331.79107666015625,
"logps/rejected": -309.49609375,
"loss": 0.1014,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.024125058203935623,
"rewards/margins": 0.01503480039536953,
"rewards/rejected": -0.0391598604619503,
"step": 290
},
{
"epoch": 0.19,
"learning_rate": 4.873285634749678e-06,
"logits/chosen": -1.1100094318389893,
"logits/rejected": -1.0816996097564697,
"logps/chosen": -311.70867919921875,
"logps/rejected": -315.1243591308594,
"loss": 0.0898,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": -0.017298461869359016,
"rewards/margins": 0.013444353826344013,
"rewards/rejected": -0.030742818489670753,
"step": 300
},
{
"epoch": 0.19,
"eval_logits/chosen": -1.0857516527175903,
"eval_logits/rejected": -1.0291430950164795,
"eval_logps/chosen": -414.7480163574219,
"eval_logps/rejected": -407.5087585449219,
"eval_loss": 0.045198723673820496,
"eval_rewards/accuracies": 0.578000009059906,
"eval_rewards/chosen": -0.018442772328853607,
"eval_rewards/margins": 0.021830614656209946,
"eval_rewards/rejected": -0.04027338698506355,
"eval_runtime": 539.5827,
"eval_samples_per_second": 3.707,
"eval_steps_per_second": 0.927,
"step": 300
},
{
"epoch": 0.2,
"learning_rate": 4.855122795887746e-06,
"logits/chosen": -1.1289124488830566,
"logits/rejected": -1.06770658493042,
"logps/chosen": -297.93218994140625,
"logps/rejected": -279.66107177734375,
"loss": 0.0967,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": -0.0071363793686032295,
"rewards/margins": 0.01606178656220436,
"rewards/rejected": -0.023198166862130165,
"step": 310
},
{
"epoch": 0.2,
"learning_rate": 4.83578250679731e-06,
"logits/chosen": -1.0398236513137817,
"logits/rejected": -1.0716559886932373,
"logps/chosen": -323.6569519042969,
"logps/rejected": -321.41845703125,
"loss": 0.1007,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.01684439741075039,
"rewards/margins": 0.016590449959039688,
"rewards/rejected": -0.03343484550714493,
"step": 320
},
{
"epoch": 0.21,
"learning_rate": 4.8152744367099935e-06,
"logits/chosen": -1.13577139377594,
"logits/rejected": -1.0597126483917236,
"logps/chosen": -253.326171875,
"logps/rejected": -260.37652587890625,
"loss": 0.1324,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": -0.011873602867126465,
"rewards/margins": 0.021937990561127663,
"rewards/rejected": -0.03381159156560898,
"step": 330
},
{
"epoch": 0.22,
"learning_rate": 4.793608838692792e-06,
"logits/chosen": -1.05599844455719,
"logits/rejected": -1.1052017211914062,
"logps/chosen": -330.0927734375,
"logps/rejected": -354.5337829589844,
"loss": 0.0709,
"rewards/accuracies": 0.4312500059604645,
"rewards/chosen": -0.014163943938910961,
"rewards/margins": 0.0242028646171093,
"rewards/rejected": -0.038366809487342834,
"step": 340
},
{
"epoch": 0.22,
"learning_rate": 4.770796544522026e-06,
"logits/chosen": -1.1162309646606445,
"logits/rejected": -1.0977243185043335,
"logps/chosen": -298.46929931640625,
"logps/rejected": -301.0343017578125,
"loss": 0.0828,
"rewards/accuracies": 0.4000000059604645,
"rewards/chosen": -0.014557460322976112,
"rewards/margins": 0.010886872187256813,
"rewards/rejected": -0.025444332510232925,
"step": 350
},
{
"epoch": 0.23,
"learning_rate": 4.746848959267968e-06,
"logits/chosen": -1.0554795265197754,
"logits/rejected": -0.9742960929870605,
"logps/chosen": -337.4122314453125,
"logps/rejected": -334.58367919921875,
"loss": 0.0999,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.002840764820575714,
"rewards/margins": 0.01675793156027794,
"rewards/rejected": -0.019598694518208504,
"step": 360
},
{
"epoch": 0.24,
"learning_rate": 4.721778055592841e-06,
"logits/chosen": -0.995721161365509,
"logits/rejected": -1.009927749633789,
"logps/chosen": -307.5436706542969,
"logps/rejected": -305.3434753417969,
"loss": 0.0958,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 0.0018125835340470076,
"rewards/margins": 0.01936521753668785,
"rewards/rejected": -0.017552632838487625,
"step": 370
},
{
"epoch": 0.24,
"learning_rate": 4.695596367765054e-06,
"logits/chosen": -0.9923852682113647,
"logits/rejected": -0.9759656190872192,
"logps/chosen": -369.8957824707031,
"logps/rejected": -335.42694091796875,
"loss": 0.0605,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": -0.0068010808899998665,
"rewards/margins": 0.011534233577549458,
"rewards/rejected": -0.018335314467549324,
"step": 380
},
{
"epoch": 0.25,
"learning_rate": 4.6683169853926615e-06,
"logits/chosen": -1.056213617324829,
"logits/rejected": -1.1043860912322998,
"logps/chosen": -302.78570556640625,
"logps/rejected": -315.38604736328125,
"loss": 0.0863,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.00549090001732111,
"rewards/margins": 0.01604645326733589,
"rewards/rejected": -0.021537352353334427,
"step": 390
},
{
"epoch": 0.26,
"learning_rate": 4.639953546879173e-06,
"logits/chosen": -1.011988639831543,
"logits/rejected": -0.9677278399467468,
"logps/chosen": -352.2360534667969,
"logps/rejected": -355.6038513183594,
"loss": 0.0731,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": -0.0009747882140800357,
"rewards/margins": 0.02269767038524151,
"rewards/rejected": -0.02367245778441429,
"step": 400
},
{
"epoch": 0.26,
"eval_logits/chosen": -1.0412452220916748,
"eval_logits/rejected": -0.9863678812980652,
"eval_logps/chosen": -403.1915588378906,
"eval_logps/rejected": -400.2978820800781,
"eval_loss": 0.04299690201878548,
"eval_rewards/accuracies": 0.597000002861023,
"eval_rewards/chosen": -0.006886274088174105,
"eval_rewards/margins": 0.026176199316978455,
"eval_rewards/rejected": -0.03306247293949127,
"eval_runtime": 539.7275,
"eval_samples_per_second": 3.706,
"eval_steps_per_second": 0.926,
"step": 400
},
{
"epoch": 0.26,
"learning_rate": 4.610520232605e-06,
"logits/chosen": -1.0571563243865967,
"logits/rejected": -0.9747453927993774,
"logps/chosen": -360.03936767578125,
"logps/rejected": -312.3787536621094,
"loss": 0.0706,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.007904710248112679,
"rewards/margins": 0.012304016388952732,
"rewards/rejected": -0.020208725705742836,
"step": 410
},
{
"epoch": 0.27,
"learning_rate": 4.580031757837931e-06,
"logits/chosen": -1.0687059164047241,
"logits/rejected": -1.028510570526123,
"logps/chosen": -309.14532470703125,
"logps/rejected": -316.378662109375,
"loss": 0.0713,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.0021046344190835953,
"rewards/margins": 0.0198093019425869,
"rewards/rejected": -0.021913940086960793,
"step": 420
},
{
"epoch": 0.28,
"learning_rate": 4.5485033653761936e-06,
"logits/chosen": -1.0271378755569458,
"logits/rejected": -1.0654436349868774,
"logps/chosen": -289.8176574707031,
"logps/rejected": -310.6886291503906,
"loss": 0.0907,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.001297330716624856,
"rewards/margins": 0.02493324503302574,
"rewards/rejected": -0.02363591641187668,
"step": 430
},
{
"epoch": 0.28,
"learning_rate": 4.5159508179277775e-06,
"logits/chosen": -1.0649276971817017,
"logits/rejected": -1.0275962352752686,
"logps/chosen": -287.9720458984375,
"logps/rejected": -285.8517761230469,
"loss": 0.0857,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.0025724810548126698,
"rewards/margins": 0.02007809281349182,
"rewards/rejected": -0.01750561222434044,
"step": 440
},
{
"epoch": 0.29,
"learning_rate": 4.48239039022982e-06,
"logits/chosen": -1.0931072235107422,
"logits/rejected": -0.9774026870727539,
"logps/chosen": -302.41497802734375,
"logps/rejected": -302.8403625488281,
"loss": 0.0786,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": 0.0017650052905082703,
"rewards/margins": 0.017620224505662918,
"rewards/rejected": -0.015855219215154648,
"step": 450
},
{
"epoch": 0.29,
"learning_rate": 4.447838860912011e-06,
"logits/chosen": -1.0981853008270264,
"logits/rejected": -0.9999715685844421,
"logps/chosen": -287.58599853515625,
"logps/rejected": -293.4311218261719,
"loss": 0.0808,
"rewards/accuracies": 0.38749998807907104,
"rewards/chosen": -0.007796216756105423,
"rewards/margins": 0.02398526482284069,
"rewards/rejected": -0.03178148344159126,
"step": 460
},
{
"epoch": 0.3,
"learning_rate": 4.412313504108062e-06,
"logits/chosen": -1.1234943866729736,
"logits/rejected": -1.0496309995651245,
"logps/chosen": -343.08856201171875,
"logps/rejected": -333.7749938964844,
"loss": 0.0816,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.00918793398886919,
"rewards/margins": 0.024025408551096916,
"rewards/rejected": -0.03321333974599838,
"step": 470
},
{
"epoch": 0.31,
"learning_rate": 4.375832080819465e-06,
"logits/chosen": -1.1054179668426514,
"logits/rejected": -1.063127040863037,
"logps/chosen": -352.350341796875,
"logps/rejected": -334.8752746582031,
"loss": 0.0853,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": -0.00490077119320631,
"rewards/margins": 0.02163395844399929,
"rewards/rejected": -0.026534726843237877,
"step": 480
},
{
"epoch": 0.31,
"learning_rate": 4.338412830035823e-06,
"logits/chosen": -1.0628612041473389,
"logits/rejected": -1.0398916006088257,
"logps/chosen": -347.806884765625,
"logps/rejected": -326.5227355957031,
"loss": 0.0819,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.0002868110022973269,
"rewards/margins": 0.024012237787246704,
"rewards/rejected": -0.02372542954981327,
"step": 490
},
{
"epoch": 0.32,
"learning_rate": 4.300074459616216e-06,
"logits/chosen": -1.083707332611084,
"logits/rejected": -1.1182454824447632,
"logps/chosen": -304.49078369140625,
"logps/rejected": -335.3125,
"loss": 0.0787,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.0009085664642043412,
"rewards/margins": 0.03841399401426315,
"rewards/rejected": -0.039322562515735626,
"step": 500
},
{
"epoch": 0.32,
"eval_logits/chosen": -1.0974537134170532,
"eval_logits/rejected": -1.0587471723556519,
"eval_logps/chosen": -408.4565734863281,
"eval_logps/rejected": -414.48870849609375,
"eval_loss": 0.0421723797917366,
"eval_rewards/accuracies": 0.6069999933242798,
"eval_rewards/chosen": -0.012151296250522137,
"eval_rewards/margins": 0.0351020023226738,
"eval_rewards/rejected": -0.04725329577922821,
"eval_runtime": 539.5972,
"eval_samples_per_second": 3.706,
"eval_steps_per_second": 0.927,
"step": 500
},
{
"epoch": 0.33,
"learning_rate": 4.260836136936159e-06,
"logits/chosen": -1.0836219787597656,
"logits/rejected": -1.0971637964248657,
"logps/chosen": -287.1617431640625,
"logps/rejected": -287.58148193359375,
"loss": 0.0893,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": -0.0011848447611555457,
"rewards/margins": 0.033604733645915985,
"rewards/rejected": -0.03478958457708359,
"step": 510
},
{
"epoch": 0.33,
"learning_rate": 4.220717479304816e-06,
"logits/chosen": -1.0768189430236816,
"logits/rejected": -1.0736405849456787,
"logps/chosen": -351.4105529785156,
"logps/rejected": -374.5068359375,
"loss": 0.092,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.007005206309258938,
"rewards/margins": 0.034257300198078156,
"rewards/rejected": -0.04126249998807907,
"step": 520
},
{
"epoch": 0.34,
"learning_rate": 4.179738544157272e-06,
"logits/chosen": -1.0806314945220947,
"logits/rejected": -1.0778720378875732,
"logps/chosen": -275.35101318359375,
"logps/rejected": -270.8028869628906,
"loss": 0.0745,
"rewards/accuracies": 0.4124999940395355,
"rewards/chosen": 0.0068043931387364864,
"rewards/margins": 0.013285738416016102,
"rewards/rejected": -0.006481344345957041,
"step": 530
},
{
"epoch": 0.35,
"learning_rate": 4.137919819026762e-06,
"logits/chosen": -1.0614066123962402,
"logits/rejected": -0.9721907377243042,
"logps/chosen": -313.3482971191406,
"logps/rejected": -321.7149963378906,
"loss": 0.076,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": 0.009386066347360611,
"rewards/margins": 0.01596895419061184,
"rewards/rejected": -0.006582888774573803,
"step": 540
},
{
"epoch": 0.35,
"learning_rate": 4.09528221130187e-06,
"logits/chosen": -1.0457605123519897,
"logits/rejected": -1.0901412963867188,
"logps/chosen": -314.82086181640625,
"logps/rejected": -286.0039367675781,
"loss": 0.0912,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.009857202880084515,
"rewards/margins": 0.024335484951734543,
"rewards/rejected": -0.014478283934295177,
"step": 550
},
{
"epoch": 0.36,
"learning_rate": 4.0518470377738274e-06,
"logits/chosen": -1.0764890909194946,
"logits/rejected": -1.1156491041183472,
"logps/chosen": -293.7644348144531,
"logps/rejected": -315.3866882324219,
"loss": 0.0867,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 0.011628219857811928,
"rewards/margins": 0.028168832883238792,
"rewards/rejected": -0.016540613025426865,
"step": 560
},
{
"epoch": 0.36,
"learning_rate": 4.0076360139791155e-06,
"logits/chosen": -1.0172516107559204,
"logits/rejected": -1.0550997257232666,
"logps/chosen": -340.00164794921875,
"logps/rejected": -341.64068603515625,
"loss": 0.0722,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.017406892031431198,
"rewards/margins": 0.03364910930395126,
"rewards/rejected": -0.016242217272520065,
"step": 570
},
{
"epoch": 0.37,
"learning_rate": 3.962671243342728e-06,
"logits/chosen": -1.089564323425293,
"logits/rejected": -1.038475513458252,
"logps/chosen": -314.32012939453125,
"logps/rejected": -327.6108703613281,
"loss": 0.0795,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.00651575718075037,
"rewards/margins": 0.018204618245363235,
"rewards/rejected": -0.011688861064612865,
"step": 580
},
{
"epoch": 0.38,
"learning_rate": 3.916975206127501e-06,
"logits/chosen": -1.0772429704666138,
"logits/rejected": -1.0009437799453735,
"logps/chosen": -317.83563232421875,
"logps/rejected": -302.42535400390625,
"loss": 0.0735,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.006415791809558868,
"rewards/margins": 0.03005843423306942,
"rewards/rejected": -0.0236426442861557,
"step": 590
},
{
"epoch": 0.38,
"learning_rate": 3.870570748195039e-06,
"logits/chosen": -1.1022416353225708,
"logits/rejected": -1.1078603267669678,
"logps/chosen": -338.0223083496094,
"logps/rejected": -388.9841003417969,
"loss": 0.0742,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": 0.012317690066993237,
"rewards/margins": 0.049174439162015915,
"rewards/rejected": -0.036856748163700104,
"step": 600
},
{
"epoch": 0.38,
"eval_logits/chosen": -1.0245832204818726,
"eval_logits/rejected": -0.9872007966041565,
"eval_logps/chosen": -382.8363342285156,
"eval_logps/rejected": -384.7105407714844,
"eval_loss": 0.04055745154619217,
"eval_rewards/accuracies": 0.6085000038146973,
"eval_rewards/chosen": 0.013468942604959011,
"eval_rewards/margins": 0.030944030731916428,
"eval_rewards/rejected": -0.017475087195634842,
"eval_runtime": 539.6208,
"eval_samples_per_second": 3.706,
"eval_steps_per_second": 0.927,
"step": 600
},
{
"epoch": 0.39,
"learning_rate": 3.823481069583869e-06,
"logits/chosen": -1.0397017002105713,
"logits/rejected": -0.9971572756767273,
"logps/chosen": -305.09442138671875,
"logps/rejected": -314.68621826171875,
"loss": 0.08,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.01682308129966259,
"rewards/margins": 0.026519659906625748,
"rewards/rejected": -0.009696578606963158,
"step": 610
},
{
"epoch": 0.4,
"learning_rate": 3.7757297129105087e-06,
"logits/chosen": -1.1045721769332886,
"logits/rejected": -1.0196996927261353,
"logps/chosen": -272.93206787109375,
"logps/rejected": -281.6795349121094,
"loss": 0.0878,
"rewards/accuracies": 0.4124999940395355,
"rewards/chosen": 0.009568464942276478,
"rewards/margins": 0.01820019818842411,
"rewards/rejected": -0.008631732314825058,
"step": 620
},
{
"epoch": 0.4,
"learning_rate": 3.7273405515992785e-06,
"logits/chosen": -1.0682175159454346,
"logits/rejected": -1.0213497877120972,
"logps/chosen": -284.39495849609375,
"logps/rejected": -300.61920166015625,
"loss": 0.0707,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": 0.01212714146822691,
"rewards/margins": 0.01712757535278797,
"rewards/rejected": -0.005000432953238487,
"step": 630
},
{
"epoch": 0.41,
"learning_rate": 3.678337777946706e-06,
"logits/chosen": -1.0374300479888916,
"logits/rejected": -0.9567736387252808,
"logps/chosen": -334.5086669921875,
"logps/rejected": -323.7910461425781,
"loss": 0.0941,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.02154025062918663,
"rewards/margins": 0.038109757006168365,
"rewards/rejected": -0.016569510102272034,
"step": 640
},
{
"epoch": 0.42,
"learning_rate": 3.6287458910265293e-06,
"logits/chosen": -1.0035964250564575,
"logits/rejected": -0.9258484840393066,
"logps/chosen": -309.08038330078125,
"logps/rejected": -294.2804260253906,
"loss": 0.0892,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": 0.011333728209137917,
"rewards/margins": 0.01899079605937004,
"rewards/rejected": -0.0076570697128772736,
"step": 650
},
{
"epoch": 0.42,
"learning_rate": 3.57858968444131e-06,
"logits/chosen": -1.0853677988052368,
"logits/rejected": -0.9975327253341675,
"logps/chosen": -330.30255126953125,
"logps/rejected": -312.9807434082031,
"loss": 0.0737,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": 0.014233666472136974,
"rewards/margins": 0.025601008906960487,
"rewards/rejected": -0.011367343366146088,
"step": 660
},
{
"epoch": 0.43,
"learning_rate": 3.5278942339268034e-06,
"logits/chosen": -1.0091218948364258,
"logits/rejected": -1.0755088329315186,
"logps/chosen": -276.8922424316406,
"logps/rejected": -313.3753967285156,
"loss": 0.0777,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.016201717779040337,
"rewards/margins": 0.032345883548259735,
"rewards/rejected": -0.016144167631864548,
"step": 670
},
{
"epoch": 0.44,
"learning_rate": 3.476684884815279e-06,
"logits/chosen": -1.0601623058319092,
"logits/rejected": -0.9890631437301636,
"logps/chosen": -280.58538818359375,
"logps/rejected": -280.070068359375,
"loss": 0.1082,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.00014938972890377045,
"rewards/margins": 0.00774354999884963,
"rewards/rejected": -0.007594159804284573,
"step": 680
},
{
"epoch": 0.44,
"learning_rate": 3.424987239364044e-06,
"logits/chosen": -1.0564385652542114,
"logits/rejected": -1.0079165697097778,
"logps/chosen": -290.0457458496094,
"logps/rejected": -317.35809326171875,
"loss": 0.0844,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.01850438304245472,
"rewards/margins": 0.03815234825015068,
"rewards/rejected": -0.01964796707034111,
"step": 690
},
{
"epoch": 0.45,
"learning_rate": 3.3728271439555277e-06,
"logits/chosen": -1.061989426612854,
"logits/rejected": -1.0436265468597412,
"logps/chosen": -300.4573669433594,
"logps/rejected": -291.84320068359375,
"loss": 0.0635,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": 0.012897541746497154,
"rewards/margins": 0.022769348695874214,
"rewards/rejected": -0.009871806018054485,
"step": 700
},
{
"epoch": 0.45,
"eval_logits/chosen": -1.022524118423462,
"eval_logits/rejected": -0.9903200268745422,
"eval_logps/chosen": -379.6695556640625,
"eval_logps/rejected": -386.0258483886719,
"eval_loss": 0.04010883718729019,
"eval_rewards/accuracies": 0.609499990940094,
"eval_rewards/chosen": 0.01663573645055294,
"eval_rewards/margins": 0.035426173359155655,
"eval_rewards/rejected": -0.018790436908602715,
"eval_runtime": 539.7506,
"eval_samples_per_second": 3.705,
"eval_steps_per_second": 0.926,
"step": 700
},
{
"epoch": 0.45,
"learning_rate": 3.3202306761753078e-06,
"logits/chosen": -1.056015133857727,
"logits/rejected": -1.044135570526123,
"logps/chosen": -291.650634765625,
"logps/rejected": -287.6162414550781,
"loss": 0.0908,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.012986931018531322,
"rewards/margins": 0.0346546433866024,
"rewards/rejected": -0.021667715162038803,
"step": 710
},
{
"epoch": 0.46,
"learning_rate": 3.2672241317745513e-06,
"logits/chosen": -0.9736151695251465,
"logits/rejected": -0.9646120071411133,
"logps/chosen": -311.49346923828125,
"logps/rejected": -341.27496337890625,
"loss": 0.0756,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.017372317612171173,
"rewards/margins": 0.028918754309415817,
"rewards/rejected": -0.011546434834599495,
"step": 720
},
{
"epoch": 0.47,
"learning_rate": 3.213834011523378e-06,
"logits/chosen": -1.1018531322479248,
"logits/rejected": -1.025577187538147,
"logps/chosen": -319.1675720214844,
"logps/rejected": -316.7040100097656,
"loss": 0.0765,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.014244127087295055,
"rewards/margins": 0.03847852349281311,
"rewards/rejected": -0.02423439547419548,
"step": 730
},
{
"epoch": 0.47,
"learning_rate": 3.160087007961724e-06,
"logits/chosen": -1.048353910446167,
"logits/rejected": -0.990998387336731,
"logps/chosen": -310.2376403808594,
"logps/rejected": -319.0958557128906,
"loss": 0.0929,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": 0.01772814430296421,
"rewards/margins": 0.029912665486335754,
"rewards/rejected": -0.012184521183371544,
"step": 740
},
{
"epoch": 0.48,
"learning_rate": 3.1060099920543404e-06,
"logits/chosen": -0.9734305143356323,
"logits/rejected": -1.053117275238037,
"logps/chosen": -253.1719207763672,
"logps/rejected": -268.72723388671875,
"loss": 0.1126,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 0.019246799871325493,
"rewards/margins": 0.028338003903627396,
"rewards/rejected": -0.009091204032301903,
"step": 750
},
{
"epoch": 0.49,
"learning_rate": 3.0516299997565675e-06,
"logits/chosen": -0.9801149368286133,
"logits/rejected": -1.053255558013916,
"logps/chosen": -300.00152587890625,
"logps/rejected": -316.54656982421875,
"loss": 0.0736,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": 0.013104112818837166,
"rewards/margins": 0.0340370312333107,
"rewards/rejected": -0.020932912826538086,
"step": 760
},
{
"epoch": 0.49,
"learning_rate": 2.996974218497643e-06,
"logits/chosen": -0.9665297269821167,
"logits/rejected": -1.033067226409912,
"logps/chosen": -327.8155822753906,
"logps/rejected": -291.17840576171875,
"loss": 0.0969,
"rewards/accuracies": 0.4312500059604645,
"rewards/chosen": 0.013064196333289146,
"rewards/margins": 0.014280739240348339,
"rewards/rejected": -0.0012165403459221125,
"step": 770
},
{
"epoch": 0.5,
"learning_rate": 2.9420699735882673e-06,
"logits/chosen": -1.103208065032959,
"logits/rejected": -1.0667420625686646,
"logps/chosen": -284.13043212890625,
"logps/rejected": -288.5975646972656,
"loss": 0.0726,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": 0.021614065393805504,
"rewards/margins": 0.022522414103150368,
"rewards/rejected": -0.000908347952645272,
"step": 780
},
{
"epoch": 0.51,
"learning_rate": 2.8869447145592345e-06,
"logits/chosen": -0.9419771432876587,
"logits/rejected": -0.9063804745674133,
"logps/chosen": -275.3219299316406,
"logps/rejected": -281.0148620605469,
"loss": 0.0885,
"rewards/accuracies": 0.4124999940395355,
"rewards/chosen": 0.021688418462872505,
"rewards/margins": 0.02460251934826374,
"rewards/rejected": -0.0029140994884073734,
"step": 790
},
{
"epoch": 0.51,
"learning_rate": 2.831626001437969e-06,
"logits/chosen": -0.9940840601921082,
"logits/rejected": -0.9761594533920288,
"logps/chosen": -284.7848205566406,
"logps/rejected": -305.55279541015625,
"loss": 0.0881,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.02503793314099312,
"rewards/margins": 0.03414962440729141,
"rewards/rejected": -0.009111693128943443,
"step": 800
},
{
"epoch": 0.51,
"eval_logits/chosen": -0.9974508881568909,
"eval_logits/rejected": -0.9657922387123108,
"eval_logps/chosen": -371.2671813964844,
"eval_logps/rejected": -377.4323425292969,
"eval_loss": 0.03950659930706024,
"eval_rewards/accuracies": 0.6085000038146973,
"eval_rewards/chosen": 0.025038093328475952,
"eval_rewards/margins": 0.03523498401045799,
"eval_rewards/rejected": -0.010196887888014317,
"eval_runtime": 539.4827,
"eval_samples_per_second": 3.707,
"eval_steps_per_second": 0.927,
"step": 800
},
{
"epoch": 0.52,
"learning_rate": 2.776141490969808e-06,
"logits/chosen": -0.9662303924560547,
"logits/rejected": -1.007033109664917,
"logps/chosen": -273.5274963378906,
"logps/rejected": -297.3504943847656,
"loss": 0.0811,
"rewards/accuracies": 0.4124999940395355,
"rewards/chosen": 0.017237264662981033,
"rewards/margins": 0.020813334733247757,
"rewards/rejected": -0.0035760707687586546,
"step": 810
},
{
"epoch": 0.52,
"learning_rate": 2.720518922790937e-06,
"logits/chosen": -1.0530205965042114,
"logits/rejected": -0.9808292388916016,
"logps/chosen": -260.7068176269531,
"logps/rejected": -254.8417205810547,
"loss": 0.1111,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": 0.015834391117095947,
"rewards/margins": 0.030566086992621422,
"rewards/rejected": -0.014731695875525475,
"step": 820
},
{
"epoch": 0.53,
"learning_rate": 2.66478610555988e-06,
"logits/chosen": -0.973135769367218,
"logits/rejected": -1.0151408910751343,
"logps/chosen": -315.51763916015625,
"logps/rejected": -324.4259338378906,
"loss": 0.0997,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.022444238886237144,
"rewards/margins": 0.034130193293094635,
"rewards/rejected": -0.011685955338180065,
"step": 830
},
{
"epoch": 0.54,
"learning_rate": 2.608970903054482e-06,
"logits/chosen": -0.9921044111251831,
"logits/rejected": -0.996437668800354,
"logps/chosen": -298.2916564941406,
"logps/rejected": -307.67156982421875,
"loss": 0.0765,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": 0.020961161702871323,
"rewards/margins": 0.020338475704193115,
"rewards/rejected": 0.0006226839614100754,
"step": 840
},
{
"epoch": 0.54,
"learning_rate": 2.553101220241337e-06,
"logits/chosen": -0.955792248249054,
"logits/rejected": -0.9853283166885376,
"logps/chosen": -322.03619384765625,
"logps/rejected": -341.2371520996094,
"loss": 0.0739,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.02565811201930046,
"rewards/margins": 0.01996433734893799,
"rewards/rejected": 0.005693775601685047,
"step": 850
},
{
"epoch": 0.55,
"learning_rate": 2.4972049893246218e-06,
"logits/chosen": -0.9531941413879395,
"logits/rejected": -0.9866918325424194,
"logps/chosen": -299.243896484375,
"logps/rejected": -302.88336181640625,
"loss": 0.0889,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.02744477614760399,
"rewards/margins": 0.03810085728764534,
"rewards/rejected": -0.0106560830026865,
"step": 860
},
{
"epoch": 0.56,
"learning_rate": 2.4413101557813095e-06,
"logits/chosen": -0.9942695498466492,
"logits/rejected": -1.0122566223144531,
"logps/chosen": -291.08892822265625,
"logps/rejected": -301.3489074707031,
"loss": 0.076,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 0.027957703918218613,
"rewards/margins": 0.030374949797987938,
"rewards/rejected": -0.0024172496050596237,
"step": 870
},
{
"epoch": 0.56,
"learning_rate": 2.3854446643897566e-06,
"logits/chosen": -0.935819149017334,
"logits/rejected": -0.8908926248550415,
"logps/chosen": -274.655517578125,
"logps/rejected": -301.8233642578125,
"loss": 0.093,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.02359806001186371,
"rewards/margins": 0.040579214692115784,
"rewards/rejected": -0.016981154680252075,
"step": 880
},
{
"epoch": 0.57,
"learning_rate": 2.3296364452586246e-06,
"logits/chosen": -0.9394947290420532,
"logits/rejected": -1.015801191329956,
"logps/chosen": -244.6657257080078,
"logps/rejected": -255.62155151367188,
"loss": 0.0973,
"rewards/accuracies": 0.40625,
"rewards/chosen": 0.02046729251742363,
"rewards/margins": 0.011587701737880707,
"rewards/rejected": 0.008879591710865498,
"step": 890
},
{
"epoch": 0.58,
"learning_rate": 2.273913399863151e-06,
"logits/chosen": -0.9255924224853516,
"logits/rejected": -0.9949308633804321,
"logps/chosen": -319.0769958496094,
"logps/rejected": -335.22125244140625,
"loss": 0.0753,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0340295247733593,
"rewards/margins": 0.03226994723081589,
"rewards/rejected": 0.0017595753306522965,
"step": 900
},
{
"epoch": 0.58,
"eval_logits/chosen": -0.9455928802490234,
"eval_logits/rejected": -0.9026016592979431,
"eval_logps/chosen": -365.869873046875,
"eval_logps/rejected": -371.78717041015625,
"eval_loss": 0.03934764117002487,
"eval_rewards/accuracies": 0.5989999771118164,
"eval_rewards/chosen": 0.0304353516548872,
"eval_rewards/margins": 0.03498707711696625,
"eval_rewards/rejected": -0.00455172173678875,
"eval_runtime": 539.932,
"eval_samples_per_second": 3.704,
"eval_steps_per_second": 0.926,
"step": 900
},
{
"epoch": 0.58,
"learning_rate": 2.2183033870957237e-06,
"logits/chosen": -0.9681524038314819,
"logits/rejected": -0.9325584173202515,
"logps/chosen": -338.40325927734375,
"logps/rejected": -328.984619140625,
"loss": 0.0758,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.017296601086854935,
"rewards/margins": 0.03395865857601166,
"rewards/rejected": -0.016662055626511574,
"step": 910
},
{
"epoch": 0.59,
"learning_rate": 2.1628342093377533e-06,
"logits/chosen": -0.9232437014579773,
"logits/rejected": -0.9427019357681274,
"logps/chosen": -301.0130615234375,
"logps/rejected": -300.59088134765625,
"loss": 0.0675,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.03031068481504917,
"rewards/margins": 0.030671527609229088,
"rewards/rejected": -0.00036084355087950826,
"step": 920
},
{
"epoch": 0.6,
"learning_rate": 2.1075335985597954e-06,
"logits/chosen": -0.9461824297904968,
"logits/rejected": -0.9701375961303711,
"logps/chosen": -309.2986755371094,
"logps/rejected": -296.10052490234375,
"loss": 0.0781,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": 0.02147866040468216,
"rewards/margins": 0.023677725344896317,
"rewards/rejected": -0.002199061680585146,
"step": 930
},
{
"epoch": 0.6,
"learning_rate": 2.0524292024568687e-06,
"logits/chosen": -0.8931276202201843,
"logits/rejected": -0.9411805272102356,
"logps/chosen": -305.31793212890625,
"logps/rejected": -297.4659729003906,
"loss": 0.076,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": 0.024056371301412582,
"rewards/margins": 0.024243740364909172,
"rewards/rejected": -0.00018737166828941554,
"step": 940
},
{
"epoch": 0.61,
"learning_rate": 1.9975485706259194e-06,
"logits/chosen": -0.9346133470535278,
"logits/rejected": -0.9174652099609375,
"logps/chosen": -243.5562744140625,
"logps/rejected": -260.65185546875,
"loss": 0.0812,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.036083243787288666,
"rewards/margins": 0.028012529015541077,
"rewards/rejected": 0.00807071290910244,
"step": 950
},
{
"epoch": 0.61,
"learning_rate": 1.942919140792319e-06,
"logits/chosen": -0.9425550699234009,
"logits/rejected": -0.8944212198257446,
"logps/chosen": -263.16070556640625,
"logps/rejected": -296.932861328125,
"loss": 0.0774,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.03240296244621277,
"rewards/margins": 0.03853650763630867,
"rewards/rejected": -0.006133544258773327,
"step": 960
},
{
"epoch": 0.62,
"learning_rate": 1.888568225092296e-06,
"logits/chosen": -0.9160875082015991,
"logits/rejected": -0.9029878377914429,
"logps/chosen": -259.0009460449219,
"logps/rejected": -284.0046081542969,
"loss": 0.071,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": 0.031025772914290428,
"rewards/margins": 0.026529574766755104,
"rewards/rejected": 0.004496193490922451,
"step": 970
},
{
"epoch": 0.63,
"learning_rate": 1.8345229964181628e-06,
"logits/chosen": -1.0157150030136108,
"logits/rejected": -0.9280340075492859,
"logps/chosen": -280.16455078125,
"logps/rejected": -297.15655517578125,
"loss": 0.0771,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.02662072703242302,
"rewards/margins": 0.036351434886455536,
"rewards/rejected": -0.009730703197419643,
"step": 980
},
{
"epoch": 0.63,
"learning_rate": 1.7808104748331459e-06,
"logits/chosen": -0.9648904800415039,
"logits/rejected": -0.9164671897888184,
"logps/chosen": -303.514892578125,
"logps/rejected": -292.10748291015625,
"loss": 0.0935,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": 0.025951769202947617,
"rewards/margins": 0.030179208144545555,
"rewards/rejected": -0.004227438475936651,
"step": 990
},
{
"epoch": 0.64,
"learning_rate": 1.7274575140626318e-06,
"logits/chosen": -1.048002004623413,
"logits/rejected": -0.9596036672592163,
"logps/chosen": -297.2816467285156,
"logps/rejected": -316.36981201171875,
"loss": 0.0922,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 0.01601681485772133,
"rewards/margins": 0.02387891337275505,
"rewards/rejected": -0.007862097583711147,
"step": 1000
},
{
"epoch": 0.64,
"eval_logits/chosen": -0.9184301495552063,
"eval_logits/rejected": -0.8800927996635437,
"eval_logps/chosen": -367.73187255859375,
"eval_logps/rejected": -374.7669372558594,
"eval_loss": 0.03902236372232437,
"eval_rewards/accuracies": 0.5989999771118164,
"eval_rewards/chosen": 0.028573375195264816,
"eval_rewards/margins": 0.03610490262508392,
"eval_rewards/rejected": -0.007531528826802969,
"eval_runtime": 539.7163,
"eval_samples_per_second": 3.706,
"eval_steps_per_second": 0.926,
"step": 1000
},
{
"epoch": 0.65,
"learning_rate": 1.6744907880685735e-06,
"logits/chosen": -0.9194186925888062,
"logits/rejected": -0.8924249410629272,
"logps/chosen": -283.1637878417969,
"logps/rejected": -281.9581604003906,
"loss": 0.0699,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": 0.02727659046649933,
"rewards/margins": 0.020740380510687828,
"rewards/rejected": 0.006536208093166351,
"step": 1010
},
{
"epoch": 0.65,
"learning_rate": 1.6219367777137652e-06,
"logits/chosen": -0.9643144607543945,
"logits/rejected": -0.9084329605102539,
"logps/chosen": -296.56732177734375,
"logps/rejected": -299.80718994140625,
"loss": 0.0756,
"rewards/accuracies": 0.4437499940395355,
"rewards/chosen": 0.02761662006378174,
"rewards/margins": 0.022495564073324203,
"rewards/rejected": 0.00512105505913496,
"step": 1020
},
{
"epoch": 0.66,
"learning_rate": 1.569821757522666e-06,
"logits/chosen": -0.9250129461288452,
"logits/rejected": -0.8530179262161255,
"logps/chosen": -305.6438903808594,
"logps/rejected": -310.8564758300781,
"loss": 0.0796,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": 0.023295477032661438,
"rewards/margins": 0.03615058213472366,
"rewards/rejected": -0.012855103239417076,
"step": 1030
},
{
"epoch": 0.67,
"learning_rate": 1.5181717825453732e-06,
"logits/chosen": -0.9920689463615417,
"logits/rejected": -0.9287697076797485,
"logps/chosen": -268.88922119140625,
"logps/rejected": -297.29791259765625,
"loss": 0.0907,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 0.025286570191383362,
"rewards/margins": 0.03552253916859627,
"rewards/rejected": -0.010235967114567757,
"step": 1040
},
{
"epoch": 0.67,
"learning_rate": 1.4670126753313286e-06,
"logits/chosen": -0.9325569272041321,
"logits/rejected": -0.9226212501525879,
"logps/chosen": -273.0399169921875,
"logps/rejected": -300.64337158203125,
"loss": 0.0829,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.02183723822236061,
"rewards/margins": 0.02524760365486145,
"rewards/rejected": -0.0034103658981621265,
"step": 1050
},
{
"epoch": 0.68,
"learning_rate": 1.4163700130192627e-06,
"logits/chosen": -0.9867246747016907,
"logits/rejected": -0.9270626306533813,
"logps/chosen": -240.3236846923828,
"logps/rejected": -277.5113525390625,
"loss": 0.1007,
"rewards/accuracies": 0.4000000059604645,
"rewards/chosen": 0.0198238343000412,
"rewards/margins": 0.027476917952299118,
"rewards/rejected": -0.007653082255274057,
"step": 1060
},
{
"epoch": 0.68,
"learning_rate": 1.366269114549833e-06,
"logits/chosen": -0.9571270942687988,
"logits/rejected": -0.9630721807479858,
"logps/chosen": -282.31561279296875,
"logps/rejected": -273.7547607421875,
"loss": 0.0832,
"rewards/accuracies": 0.4312500059604645,
"rewards/chosen": 0.014912809245288372,
"rewards/margins": 0.02561432123184204,
"rewards/rejected": -0.01070151012390852,
"step": 1070
},
{
"epoch": 0.69,
"learning_rate": 1.3167350280073514e-06,
"logits/chosen": -0.9449627995491028,
"logits/rejected": -0.9669274091720581,
"logps/chosen": -290.40899658203125,
"logps/rejected": -305.1846923828125,
"loss": 0.0881,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": 0.020313743501901627,
"rewards/margins": 0.03765721619129181,
"rewards/rejected": -0.017343472689390182,
"step": 1080
},
{
"epoch": 0.7,
"learning_rate": 1.267792518096918e-06,
"logits/chosen": -0.954971194267273,
"logits/rejected": -1.0019423961639404,
"logps/chosen": -314.25335693359375,
"logps/rejected": -375.009521484375,
"loss": 0.0652,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": 0.023724760860204697,
"rewards/margins": 0.04069076478481293,
"rewards/rejected": -0.01696600392460823,
"step": 1090
},
{
"epoch": 0.7,
"learning_rate": 1.2194660537632423e-06,
"logits/chosen": -0.9429014325141907,
"logits/rejected": -0.9538652300834656,
"logps/chosen": -279.4613342285156,
"logps/rejected": -323.7087097167969,
"loss": 0.0703,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": 0.02060026489198208,
"rewards/margins": 0.03607643395662308,
"rewards/rejected": -0.015476171858608723,
"step": 1100
},
{
"epoch": 0.7,
"eval_logits/chosen": -0.9601659178733826,
"eval_logits/rejected": -0.9299748539924622,
"eval_logps/chosen": -373.62261962890625,
"eval_logps/rejected": -383.30255126953125,
"eval_loss": 0.03885647654533386,
"eval_rewards/accuracies": 0.6000000238418579,
"eval_rewards/chosen": 0.022682595998048782,
"eval_rewards/margins": 0.03874973580241203,
"eval_rewards/rejected": -0.016067136079072952,
"eval_runtime": 539.6341,
"eval_samples_per_second": 3.706,
"eval_steps_per_second": 0.927,
"step": 1100
},
{
"epoch": 0.71,
"learning_rate": 1.1717797959573262e-06,
"logits/chosen": -1.003103494644165,
"logits/rejected": -0.9326059222221375,
"logps/chosen": -297.8578186035156,
"logps/rejected": -297.1626892089844,
"loss": 0.0771,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.017790749669075012,
"rewards/margins": 0.035782478749752045,
"rewards/rejected": -0.017991727218031883,
"step": 1110
},
{
"epoch": 0.72,
"learning_rate": 1.1247575855571251e-06,
"logits/chosen": -0.9976641535758972,
"logits/rejected": -0.9637433886528015,
"logps/chosen": -306.8631896972656,
"logps/rejected": -350.14324951171875,
"loss": 0.081,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.029013922438025475,
"rewards/margins": 0.05382751673460007,
"rewards/rejected": -0.02481359988451004,
"step": 1120
},
{
"epoch": 0.72,
"learning_rate": 1.07842293144824e-06,
"logits/chosen": -0.9757798314094543,
"logits/rejected": -0.9444389343261719,
"logps/chosen": -312.77191162109375,
"logps/rejected": -317.60406494140625,
"loss": 0.0816,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.015693334862589836,
"rewards/margins": 0.029305079951882362,
"rewards/rejected": -0.013611746951937675,
"step": 1130
},
{
"epoch": 0.73,
"learning_rate": 1.0327989987705781e-06,
"logits/chosen": -1.0041489601135254,
"logits/rejected": -1.006225824356079,
"logps/chosen": -291.49542236328125,
"logps/rejected": -287.52978515625,
"loss": 0.0888,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": 0.01858539506793022,
"rewards/margins": 0.029126202687621117,
"rewards/rejected": -0.01054080855101347,
"step": 1140
},
{
"epoch": 0.74,
"learning_rate": 9.879085973368805e-07,
"logits/chosen": -1.0146872997283936,
"logits/rejected": -0.9577957987785339,
"logps/chosen": -322.76556396484375,
"logps/rejected": -336.08636474609375,
"loss": 0.0665,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.023015085607767105,
"rewards/margins": 0.05666361376643181,
"rewards/rejected": -0.0336485281586647,
"step": 1150
},
{
"epoch": 0.74,
"learning_rate": 9.437741702288908e-07,
"logits/chosen": -1.0026918649673462,
"logits/rejected": -0.9933171272277832,
"logps/chosen": -299.72308349609375,
"logps/rejected": -311.6734924316406,
"loss": 0.0794,
"rewards/accuracies": 0.4437499940395355,
"rewards/chosen": 0.01335589587688446,
"rewards/margins": 0.025558674708008766,
"rewards/rejected": -0.012202778831124306,
"step": 1160
},
{
"epoch": 0.75,
"learning_rate": 9.004177825768751e-07,
"logits/chosen": -1.0369510650634766,
"logits/rejected": -0.9392277002334595,
"logps/chosen": -267.6405334472656,
"logps/rejected": -298.1479797363281,
"loss": 0.0757,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": 0.012007731013000011,
"rewards/margins": 0.024681296199560165,
"rewards/rejected": -0.012673566117882729,
"step": 1170
},
{
"epoch": 0.76,
"learning_rate": 8.578611105280987e-07,
"logits/chosen": -1.0083134174346924,
"logits/rejected": -0.992781937122345,
"logps/chosen": -316.5964660644531,
"logps/rejected": -365.96661376953125,
"loss": 0.0643,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.02335866168141365,
"rewards/margins": 0.050595641136169434,
"rewards/rejected": -0.027236973866820335,
"step": 1180
},
{
"epoch": 0.76,
"learning_rate": 8.161254304097715e-07,
"logits/chosen": -1.026871919631958,
"logits/rejected": -0.9791032075881958,
"logps/chosen": -277.67547607421875,
"logps/rejected": -294.22369384765625,
"loss": 0.0853,
"rewards/accuracies": 0.42500001192092896,
"rewards/chosen": 0.02239377610385418,
"rewards/margins": 0.026370327919721603,
"rewards/rejected": -0.003976552281528711,
"step": 1190
},
{
"epoch": 0.77,
"learning_rate": 7.752316080918934e-07,
"logits/chosen": -0.942740797996521,
"logits/rejected": -0.9156352877616882,
"logps/chosen": -290.0126953125,
"logps/rejected": -303.8479919433594,
"loss": 0.0746,
"rewards/accuracies": 0.4375,
"rewards/chosen": 0.021792907267808914,
"rewards/margins": 0.02630285918712616,
"rewards/rejected": -0.004509954713284969,
"step": 1200
},
{
"epoch": 0.77,
"eval_logits/chosen": -0.930611789226532,
"eval_logits/rejected": -0.8943980932235718,
"eval_logps/chosen": -373.7153015136719,
"eval_logps/rejected": -385.1601257324219,
"eval_loss": 0.03884938731789589,
"eval_rewards/accuracies": 0.6050000190734863,
"eval_rewards/chosen": 0.022589918226003647,
"eval_rewards/margins": 0.04051463305950165,
"eval_rewards/rejected": -0.017924712970852852,
"eval_runtime": 539.5465,
"eval_samples_per_second": 3.707,
"eval_steps_per_second": 0.927,
"step": 1200
},
{
"epoch": 0.77,
"learning_rate": 7.352000885553012e-07,
"logits/chosen": -0.9893749356269836,
"logits/rejected": -0.9994084239006042,
"logps/chosen": -307.1074523925781,
"logps/rejected": -315.70770263671875,
"loss": 0.0767,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": 0.016450155526399612,
"rewards/margins": 0.02466857247054577,
"rewards/rejected": -0.008218420669436455,
"step": 1210
},
{
"epoch": 0.78,
"learning_rate": 6.960508856701464e-07,
"logits/chosen": -1.0321322679519653,
"logits/rejected": -0.9207181930541992,
"logps/chosen": -291.80029296875,
"logps/rejected": -298.3025817871094,
"loss": 0.0945,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.015555900521576405,
"rewards/margins": 0.02910393476486206,
"rewards/rejected": -0.013548034243285656,
"step": 1220
},
{
"epoch": 0.79,
"learning_rate": 6.578035721899029e-07,
"logits/chosen": -0.9535512924194336,
"logits/rejected": -1.008029580116272,
"logps/chosen": -324.25152587890625,
"logps/rejected": -346.2152099609375,
"loss": 0.0693,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.012142010033130646,
"rewards/margins": 0.03695748746395111,
"rewards/rejected": -0.024815475568175316,
"step": 1230
},
{
"epoch": 0.79,
"learning_rate": 6.204772699659126e-07,
"logits/chosen": -1.0189663171768188,
"logits/rejected": -1.051673173904419,
"logps/chosen": -268.8255920410156,
"logps/rejected": -264.9363708496094,
"loss": 0.0811,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 0.022226419299840927,
"rewards/margins": 0.02963913045823574,
"rewards/rejected": -0.007412709295749664,
"step": 1240
},
{
"epoch": 0.8,
"learning_rate": 5.840906403873648e-07,
"logits/chosen": -0.9550544023513794,
"logits/rejected": -0.9686701893806458,
"logps/chosen": -311.6850280761719,
"logps/rejected": -331.0824279785156,
"loss": 0.0602,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": 0.01663174293935299,
"rewards/margins": 0.03210796043276787,
"rewards/rejected": -0.015476214699447155,
"step": 1250
},
{
"epoch": 0.81,
"learning_rate": 5.486618750514813e-07,
"logits/chosen": -0.9893242716789246,
"logits/rejected": -0.9823307991027832,
"logps/chosen": -319.13421630859375,
"logps/rejected": -335.4979553222656,
"loss": 0.0904,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 0.019062984734773636,
"rewards/margins": 0.03206023946404457,
"rewards/rejected": -0.012997254729270935,
"step": 1260
},
{
"epoch": 0.81,
"learning_rate": 5.142086866685783e-07,
"logits/chosen": -0.9724270701408386,
"logits/rejected": -0.9491437673568726,
"logps/chosen": -279.63824462890625,
"logps/rejected": -299.930908203125,
"loss": 0.0748,
"rewards/accuracies": 0.4437499940395355,
"rewards/chosen": 0.01930154860019684,
"rewards/margins": 0.022973302751779556,
"rewards/rejected": -0.0036717529874294996,
"step": 1270
},
{
"epoch": 0.82,
"learning_rate": 4.807483002065439e-07,
"logits/chosen": -0.9794095754623413,
"logits/rejected": -0.9791723489761353,
"logps/chosen": -263.2690124511719,
"logps/rejected": -264.7826232910156,
"loss": 0.0906,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 0.020124919712543488,
"rewards/margins": 0.028638970106840134,
"rewards/rejected": -0.008514048531651497,
"step": 1280
},
{
"epoch": 0.83,
"learning_rate": 4.4829744427917153e-07,
"logits/chosen": -1.0250556468963623,
"logits/rejected": -0.9449254870414734,
"logps/chosen": -262.35614013671875,
"logps/rejected": -292.6833190917969,
"loss": 0.0725,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": 0.029375359416007996,
"rewards/margins": 0.04200378805398941,
"rewards/rejected": -0.012628423981368542,
"step": 1290
},
{
"epoch": 0.83,
"learning_rate": 4.168723427826382e-07,
"logits/chosen": -0.964804470539093,
"logits/rejected": -0.9295471906661987,
"logps/chosen": -272.17938232421875,
"logps/rejected": -284.2236022949219,
"loss": 0.0925,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.01671653613448143,
"rewards/margins": 0.0240671094506979,
"rewards/rejected": -0.007350574247539043,
"step": 1300
},
{
"epoch": 0.83,
"eval_logits/chosen": -0.949418306350708,
"eval_logits/rejected": -0.9171380400657654,
"eval_logps/chosen": -370.0339660644531,
"eval_logps/rejected": -380.30718994140625,
"eval_loss": 0.03874335065484047,
"eval_rewards/accuracies": 0.6029999852180481,
"eval_rewards/chosen": 0.02627129666507244,
"eval_rewards/margins": 0.039343029260635376,
"eval_rewards/rejected": -0.013071730732917786,
"eval_runtime": 539.7909,
"eval_samples_per_second": 3.705,
"eval_steps_per_second": 0.926,
"step": 1300
},
{
"epoch": 0.84,
"learning_rate": 3.864887067843251e-07,
"logits/chosen": -0.9890697598457336,
"logits/rejected": -0.8923286199569702,
"logps/chosen": -310.97125244140625,
"logps/rejected": -309.845703125,
"loss": 0.0789,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 0.019753634929656982,
"rewards/margins": 0.031969424337148666,
"rewards/rejected": -0.01221578847616911,
"step": 1310
},
{
"epoch": 0.84,
"learning_rate": 3.5716172666802637e-07,
"logits/chosen": -1.0090614557266235,
"logits/rejected": -0.9583064317703247,
"logps/chosen": -266.8111267089844,
"logps/rejected": -302.0957946777344,
"loss": 0.0764,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": 0.023147406056523323,
"rewards/margins": 0.04169207811355591,
"rewards/rejected": -0.018544670194387436,
"step": 1320
},
{
"epoch": 0.85,
"learning_rate": 3.289060645394704e-07,
"logits/chosen": -0.9466385841369629,
"logits/rejected": -0.9388877749443054,
"logps/chosen": -291.6334533691406,
"logps/rejected": -310.3324279785156,
"loss": 0.094,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": 0.021957775577902794,
"rewards/margins": 0.026331758126616478,
"rewards/rejected": -0.0043739816173911095,
"step": 1330
},
{
"epoch": 0.86,
"learning_rate": 3.0173584689596246e-07,
"logits/chosen": -0.9820948839187622,
"logits/rejected": -0.9631234407424927,
"logps/chosen": -290.67987060546875,
"logps/rejected": -281.71405029296875,
"loss": 0.0816,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.026712050661444664,
"rewards/margins": 0.026149820536375046,
"rewards/rejected": 0.0005622319877147675,
"step": 1340
},
{
"epoch": 0.86,
"learning_rate": 2.756646575638025e-07,
"logits/chosen": -0.9123773574829102,
"logits/rejected": -0.9757976531982422,
"logps/chosen": -270.1868896484375,
"logps/rejected": -303.7170715332031,
"loss": 0.0753,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.025742124766111374,
"rewards/margins": 0.034964997321367264,
"rewards/rejected": -0.009222874417901039,
"step": 1350
},
{
"epoch": 0.87,
"learning_rate": 2.507055309070111e-07,
"logits/chosen": -0.9779159426689148,
"logits/rejected": -0.9737545847892761,
"logps/chosen": -285.8861083984375,
"logps/rejected": -280.3493347167969,
"loss": 0.0912,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": 0.021084221079945564,
"rewards/margins": 0.029098382219672203,
"rewards/rejected": -0.008014162071049213,
"step": 1360
},
{
"epoch": 0.88,
"learning_rate": 2.2687094531076565e-07,
"logits/chosen": -1.0015113353729248,
"logits/rejected": -0.9118536114692688,
"logps/chosen": -292.0755920410156,
"logps/rejected": -336.7613830566406,
"loss": 0.0668,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": 0.023335173726081848,
"rewards/margins": 0.042082760483026505,
"rewards/rejected": -0.018747588619589806,
"step": 1370
},
{
"epoch": 0.88,
"learning_rate": 2.0417281694279424e-07,
"logits/chosen": -0.9744800329208374,
"logits/rejected": -0.8867910504341125,
"logps/chosen": -322.03643798828125,
"logps/rejected": -347.53192138671875,
"loss": 0.0651,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.016095371916890144,
"rewards/margins": 0.03282975032925606,
"rewards/rejected": -0.016734374687075615,
"step": 1380
},
{
"epoch": 0.89,
"learning_rate": 1.8262249379585484e-07,
"logits/chosen": -0.8901314735412598,
"logits/rejected": -0.9079896807670593,
"logps/chosen": -318.6395568847656,
"logps/rejected": -306.5535583496094,
"loss": 0.0696,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": 0.02375376783311367,
"rewards/margins": 0.02338986098766327,
"rewards/rejected": 0.00036390620516613126,
"step": 1390
},
{
"epoch": 0.9,
"learning_rate": 1.6223075001427667e-07,
"logits/chosen": -0.9552518725395203,
"logits/rejected": -0.9494439959526062,
"logps/chosen": -281.0766906738281,
"logps/rejected": -289.63055419921875,
"loss": 0.0863,
"rewards/accuracies": 0.40625,
"rewards/chosen": 0.010858943685889244,
"rewards/margins": 0.015300577506422997,
"rewards/rejected": -0.004441632889211178,
"step": 1400
},
{
"epoch": 0.9,
"eval_logits/chosen": -0.9446871280670166,
"eval_logits/rejected": -0.9120718240737915,
"eval_logps/chosen": -369.4450378417969,
"eval_logps/rejected": -379.5608215332031,
"eval_loss": 0.038691576570272446,
"eval_rewards/accuracies": 0.6054999828338623,
"eval_rewards/chosen": 0.02686023712158203,
"eval_rewards/margins": 0.03918563574552536,
"eval_rewards/rejected": -0.01232539676129818,
"eval_runtime": 539.9174,
"eval_samples_per_second": 3.704,
"eval_steps_per_second": 0.926,
"step": 1400
},
{
"epoch": 0.9,
"learning_rate": 1.4300778050739317e-07,
"logits/chosen": -0.9774508476257324,
"logits/rejected": -0.9408830404281616,
"logps/chosen": -298.72247314453125,
"logps/rejected": -294.75408935546875,
"loss": 0.0665,
"rewards/accuracies": 0.4437499940395355,
"rewards/chosen": 0.01832416281104088,
"rewards/margins": 0.014284146018326283,
"rewards/rejected": 0.004040017258375883,
"step": 1410
},
{
"epoch": 0.91,
"learning_rate": 1.2496319585257183e-07,
"logits/chosen": -0.9892775416374207,
"logits/rejected": -0.9716461300849915,
"logps/chosen": -323.1170349121094,
"logps/rejected": -303.7762451171875,
"loss": 0.0854,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": 0.02715294435620308,
"rewards/margins": 0.03064989112317562,
"rewards/rejected": -0.0034969463013112545,
"step": 1420
},
{
"epoch": 0.92,
"learning_rate": 1.0810601749037669e-07,
"logits/chosen": -0.9020091891288757,
"logits/rejected": -0.9708755612373352,
"logps/chosen": -289.08258056640625,
"logps/rejected": -339.64703369140625,
"loss": 0.0713,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": 0.023048024624586105,
"rewards/margins": 0.04538671672344208,
"rewards/rejected": -0.022338688373565674,
"step": 1430
},
{
"epoch": 0.92,
"learning_rate": 9.244467321427585e-08,
"logits/chosen": -1.0033928155899048,
"logits/rejected": -1.0245609283447266,
"logps/chosen": -276.730712890625,
"logps/rejected": -303.7331237792969,
"loss": 0.0785,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.033762019127607346,
"rewards/margins": 0.05797597020864487,
"rewards/rejected": -0.024213949218392372,
"step": 1440
},
{
"epoch": 0.93,
"learning_rate": 7.798699295714002e-08,
"logits/chosen": -0.9668358564376831,
"logits/rejected": -0.9086298942565918,
"logps/chosen": -320.6669006347656,
"logps/rejected": -296.5704345703125,
"loss": 0.0777,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.023420458659529686,
"rewards/margins": 0.03309481590986252,
"rewards/rejected": -0.009674356319010258,
"step": 1450
},
{
"epoch": 0.93,
"learning_rate": 6.474020487664934e-08,
"logits/chosen": -0.9561643600463867,
"logits/rejected": -0.9422094225883484,
"logps/chosen": -276.42156982421875,
"logps/rejected": -291.9980163574219,
"loss": 0.076,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.023049993440508842,
"rewards/margins": 0.038113731890916824,
"rewards/rejected": -0.01506374217569828,
"step": 1460
},
{
"epoch": 0.94,
"learning_rate": 5.271093174155223e-08,
"logits/chosen": -1.0463026762008667,
"logits/rejected": -0.9269806742668152,
"logps/chosen": -271.7001037597656,
"logps/rejected": -273.11419677734375,
"loss": 0.0773,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": 0.02771763503551483,
"rewards/margins": 0.02654975652694702,
"rewards/rejected": 0.0011678790906444192,
"step": 1470
},
{
"epoch": 0.95,
"learning_rate": 4.190518762059587e-08,
"logits/chosen": -0.9642139673233032,
"logits/rejected": -0.9358107447624207,
"logps/chosen": -301.32647705078125,
"logps/rejected": -310.1565856933594,
"loss": 0.0955,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": 0.019302543252706528,
"rewards/margins": 0.034524787217378616,
"rewards/rejected": -0.015222239308059216,
"step": 1480
},
{
"epoch": 0.95,
"learning_rate": 3.232837487577384e-08,
"logits/chosen": -0.9852960705757141,
"logits/rejected": -0.9471799731254578,
"logps/chosen": -306.68548583984375,
"logps/rejected": -294.8326416015625,
"loss": 0.0895,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": 0.018351960927248,
"rewards/margins": 0.04016681760549545,
"rewards/rejected": -0.0218148622661829,
"step": 1490
},
{
"epoch": 0.96,
"learning_rate": 2.39852814614025e-08,
"logits/chosen": -0.9789448976516724,
"logits/rejected": -0.9438329935073853,
"logps/chosen": -281.9150085449219,
"logps/rejected": -267.4187316894531,
"loss": 0.0904,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": 0.019616561010479927,
"rewards/margins": 0.023770466446876526,
"rewards/rejected": -0.004153906367719173,
"step": 1500
},
{
"epoch": 0.96,
"eval_logits/chosen": -0.9535864591598511,
"eval_logits/rejected": -0.9203009009361267,
"eval_logps/chosen": -369.494384765625,
"eval_logps/rejected": -379.5999755859375,
"eval_loss": 0.03863796219229698,
"eval_rewards/accuracies": 0.6044999957084656,
"eval_rewards/chosen": 0.026810916140675545,
"eval_rewards/margins": 0.03917544335126877,
"eval_rewards/rejected": -0.012364527210593224,
"eval_runtime": 539.9867,
"eval_samples_per_second": 3.704,
"eval_steps_per_second": 0.926,
"step": 1500
},
{
"epoch": 0.97,
"learning_rate": 1.6880078530367716e-08,
"logits/chosen": -0.9464886784553528,
"logits/rejected": -0.9203370213508606,
"logps/chosen": -291.42730712890625,
"logps/rejected": -302.47357177734375,
"loss": 0.0968,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": 0.018558986485004425,
"rewards/margins": 0.03005158342421055,
"rewards/rejected": -0.011492597870528698,
"step": 1510
},
{
"epoch": 0.97,
"learning_rate": 1.1016318348746058e-08,
"logits/chosen": -0.9404022097587585,
"logits/rejected": -0.9021556973457336,
"logps/chosen": -266.0870666503906,
"logps/rejected": -277.4389343261719,
"loss": 0.1015,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 0.021859928965568542,
"rewards/margins": 0.028949573636054993,
"rewards/rejected": -0.007089647464454174,
"step": 1520
},
{
"epoch": 0.98,
"learning_rate": 6.396932519840693e-09,
"logits/chosen": -0.9894587397575378,
"logits/rejected": -1.043691873550415,
"logps/chosen": -284.725830078125,
"logps/rejected": -302.0658874511719,
"loss": 0.0885,
"rewards/accuracies": 0.4375,
"rewards/chosen": 0.01785745844244957,
"rewards/margins": 0.0238940566778183,
"rewards/rejected": -0.0060366010293364525,
"step": 1530
},
{
"epoch": 0.99,
"learning_rate": 3.024230518515192e-09,
"logits/chosen": -1.0197975635528564,
"logits/rejected": -0.9348329305648804,
"logps/chosen": -242.8181915283203,
"logps/rejected": -304.18841552734375,
"loss": 0.115,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 0.027777481824159622,
"rewards/margins": 0.04096692427992821,
"rewards/rejected": -0.013189440593123436,
"step": 1540
},
{
"epoch": 0.99,
"learning_rate": 8.998985365679669e-10,
"logits/chosen": -0.918280303478241,
"logits/rejected": -0.9589518308639526,
"logps/chosen": -293.98236083984375,
"logps/rejected": -301.8656921386719,
"loss": 0.0665,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": 0.025340059772133827,
"rewards/margins": 0.033479273319244385,
"rewards/rejected": -0.008139212615787983,
"step": 1550
},
{
"epoch": 1.0,
"learning_rate": 2.499863971494598e-11,
"logits/chosen": -0.9552096128463745,
"logits/rejected": -0.8985775113105774,
"logps/chosen": -317.04779052734375,
"logps/rejected": -310.51031494140625,
"loss": 0.0696,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.025858771055936813,
"rewards/margins": 0.03246783837676048,
"rewards/rejected": -0.006609070114791393,
"step": 1560
},
{
"epoch": 1.0,
"step": 1562,
"total_flos": 0.0,
"train_loss": 0.08419492204701015,
"train_runtime": 22113.8812,
"train_samples_per_second": 1.131,
"train_steps_per_second": 0.071
}
],
"logging_steps": 10,
"max_steps": 1562,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 100,
"total_flos": 0.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}