zephyr-7b-gpo-v7-i3 / trainer_state.json
lole25's picture
Model save
e582bc1 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 500,
"global_step": 1875,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 2.6595744680851065e-08,
"logits/chosen": -1.4655269384384155,
"logits/rejected": -1.033621907234192,
"logps/chosen": -459.0647277832031,
"logps/rejected": -1330.877197265625,
"loss": 0.3828,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.01,
"learning_rate": 2.6595744680851066e-07,
"logits/chosen": -1.5051629543304443,
"logits/rejected": -0.712297260761261,
"logps/chosen": -705.1040649414062,
"logps/rejected": -1360.2410888671875,
"loss": 0.3114,
"rewards/accuracies": 0.3888888955116272,
"rewards/chosen": 0.0001662555441725999,
"rewards/margins": 2.4516626581316814e-05,
"rewards/rejected": 0.00014173888484947383,
"step": 10
},
{
"epoch": 0.01,
"learning_rate": 5.319148936170213e-07,
"logits/chosen": -1.3879754543304443,
"logits/rejected": -0.8259027600288391,
"logps/chosen": -617.7056884765625,
"logps/rejected": -1266.44580078125,
"loss": 0.3413,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.00024980775197036564,
"rewards/margins": 0.0005200988380238414,
"rewards/rejected": -0.0007699066773056984,
"step": 20
},
{
"epoch": 0.02,
"learning_rate": 7.97872340425532e-07,
"logits/chosen": -1.4345638751983643,
"logits/rejected": -0.8269219398498535,
"logps/chosen": -646.2857055664062,
"logps/rejected": -1389.702392578125,
"loss": 0.3747,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 0.0002903448767028749,
"rewards/margins": 0.00816894881427288,
"rewards/rejected": -0.007878603413701057,
"step": 30
},
{
"epoch": 0.02,
"learning_rate": 1.0638297872340427e-06,
"logits/chosen": -1.4127135276794434,
"logits/rejected": -0.9716196060180664,
"logps/chosen": -517.2140502929688,
"logps/rejected": -1226.396728515625,
"loss": 0.3177,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.001012824010103941,
"rewards/margins": 0.01804637908935547,
"rewards/rejected": -0.019059203565120697,
"step": 40
},
{
"epoch": 0.03,
"learning_rate": 1.3297872340425533e-06,
"logits/chosen": -1.5345873832702637,
"logits/rejected": -0.6556589603424072,
"logps/chosen": -707.1253051757812,
"logps/rejected": -1423.584228515625,
"loss": 0.2566,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.006385817192494869,
"rewards/margins": 0.05405501276254654,
"rewards/rejected": -0.06044083088636398,
"step": 50
},
{
"epoch": 0.03,
"learning_rate": 1.595744680851064e-06,
"logits/chosen": -1.244545340538025,
"logits/rejected": -0.7255198359489441,
"logps/chosen": -639.6641845703125,
"logps/rejected": -1549.7298583984375,
"loss": 0.2391,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.07489156723022461,
"rewards/margins": 0.12882688641548157,
"rewards/rejected": -0.20371845364570618,
"step": 60
},
{
"epoch": 0.04,
"learning_rate": 1.8617021276595745e-06,
"logits/chosen": -1.3547617197036743,
"logits/rejected": -0.35274913907051086,
"logps/chosen": -767.3015747070312,
"logps/rejected": -1636.340087890625,
"loss": 0.2003,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.19739916920661926,
"rewards/margins": 0.2098737508058548,
"rewards/rejected": -0.40727296471595764,
"step": 70
},
{
"epoch": 0.04,
"learning_rate": 2.1276595744680853e-06,
"logits/chosen": -1.2055673599243164,
"logits/rejected": -0.5075792074203491,
"logps/chosen": -786.0068969726562,
"logps/rejected": -1766.186767578125,
"loss": 0.1685,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.15041813254356384,
"rewards/margins": 0.2894287705421448,
"rewards/rejected": -0.4398469030857086,
"step": 80
},
{
"epoch": 0.05,
"learning_rate": 2.393617021276596e-06,
"logits/chosen": -1.478634238243103,
"logits/rejected": -0.5144511461257935,
"logps/chosen": -730.87890625,
"logps/rejected": -1668.932861328125,
"loss": 0.1939,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.16476184129714966,
"rewards/margins": 0.24078884720802307,
"rewards/rejected": -0.4055507183074951,
"step": 90
},
{
"epoch": 0.05,
"learning_rate": 2.6595744680851065e-06,
"logits/chosen": -1.4973030090332031,
"logits/rejected": -0.6912057399749756,
"logps/chosen": -791.8748779296875,
"logps/rejected": -1806.0367431640625,
"loss": 0.2125,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.21075236797332764,
"rewards/margins": 0.25144320726394653,
"rewards/rejected": -0.46219557523727417,
"step": 100
},
{
"epoch": 0.06,
"learning_rate": 2.9255319148936174e-06,
"logits/chosen": -1.3953739404678345,
"logits/rejected": -0.6119239926338196,
"logps/chosen": -722.7054443359375,
"logps/rejected": -1811.979736328125,
"loss": 0.1752,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.16878579556941986,
"rewards/margins": 0.3092211186885834,
"rewards/rejected": -0.47800692915916443,
"step": 110
},
{
"epoch": 0.06,
"learning_rate": 3.191489361702128e-06,
"logits/chosen": -1.4477260112762451,
"logits/rejected": -1.145725965499878,
"logps/chosen": -636.2543334960938,
"logps/rejected": -1560.247802734375,
"loss": 0.2261,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.21692097187042236,
"rewards/margins": 0.2692717909812927,
"rewards/rejected": -0.4861927628517151,
"step": 120
},
{
"epoch": 0.07,
"learning_rate": 3.457446808510639e-06,
"logits/chosen": -1.4003589153289795,
"logits/rejected": -0.4936106204986572,
"logps/chosen": -758.2838134765625,
"logps/rejected": -1723.779052734375,
"loss": 0.1853,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.25890326499938965,
"rewards/margins": 0.23604793846607208,
"rewards/rejected": -0.4949512481689453,
"step": 130
},
{
"epoch": 0.07,
"learning_rate": 3.723404255319149e-06,
"logits/chosen": -1.0825198888778687,
"logits/rejected": 0.35903090238571167,
"logps/chosen": -885.7689208984375,
"logps/rejected": -1848.095947265625,
"loss": 0.2007,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.23516646027565002,
"rewards/margins": 0.2823827266693115,
"rewards/rejected": -0.5175491571426392,
"step": 140
},
{
"epoch": 0.08,
"learning_rate": 3.98936170212766e-06,
"logits/chosen": -1.4162981510162354,
"logits/rejected": -0.7130796313285828,
"logps/chosen": -671.996337890625,
"logps/rejected": -1514.2220458984375,
"loss": 0.2171,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.09221184998750687,
"rewards/margins": 0.2251388132572174,
"rewards/rejected": -0.31735068559646606,
"step": 150
},
{
"epoch": 0.09,
"learning_rate": 4.255319148936171e-06,
"logits/chosen": -1.5083972215652466,
"logits/rejected": -0.1440640687942505,
"logps/chosen": -748.1359252929688,
"logps/rejected": -1689.578369140625,
"loss": 0.2096,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.1878463327884674,
"rewards/margins": 0.28401127457618713,
"rewards/rejected": -0.47185754776000977,
"step": 160
},
{
"epoch": 0.09,
"learning_rate": 4.521276595744681e-06,
"logits/chosen": -1.432492733001709,
"logits/rejected": -0.35446396470069885,
"logps/chosen": -667.8497314453125,
"logps/rejected": -1559.5621337890625,
"loss": 0.2264,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.10573254525661469,
"rewards/margins": 0.13765206933021545,
"rewards/rejected": -0.24338459968566895,
"step": 170
},
{
"epoch": 0.1,
"learning_rate": 4.787234042553192e-06,
"logits/chosen": -1.4102504253387451,
"logits/rejected": -0.08080820739269257,
"logps/chosen": -640.1746215820312,
"logps/rejected": -1653.6851806640625,
"loss": 0.4574,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.04140184074640274,
"rewards/margins": 0.3433176875114441,
"rewards/rejected": -0.3847195506095886,
"step": 180
},
{
"epoch": 0.1,
"learning_rate": 4.999982660399688e-06,
"logits/chosen": -1.5038793087005615,
"logits/rejected": -0.5108478665351868,
"logps/chosen": -650.073974609375,
"logps/rejected": -1593.111083984375,
"loss": 0.2165,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.015623539686203003,
"rewards/margins": 0.13894978165626526,
"rewards/rejected": -0.15457329154014587,
"step": 190
},
{
"epoch": 0.11,
"learning_rate": 4.99937579964398e-06,
"logits/chosen": -1.4442822933197021,
"logits/rejected": -0.47450265288352966,
"logps/chosen": -636.0139770507812,
"logps/rejected": -1700.844482421875,
"loss": 0.2143,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.03425021469593048,
"rewards/margins": 0.2377198487520218,
"rewards/rejected": -0.27197009325027466,
"step": 200
},
{
"epoch": 0.11,
"learning_rate": 4.9979021993870645e-06,
"logits/chosen": -1.405047059059143,
"logits/rejected": -0.5784617066383362,
"logps/chosen": -678.7222290039062,
"logps/rejected": -1632.005615234375,
"loss": 0.2189,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.10118959844112396,
"rewards/margins": 0.24721214175224304,
"rewards/rejected": -0.3484017848968506,
"step": 210
},
{
"epoch": 0.12,
"learning_rate": 4.995562370647553e-06,
"logits/chosen": -1.4955471754074097,
"logits/rejected": -0.4132479727268219,
"logps/chosen": -838.4027099609375,
"logps/rejected": -1668.740966796875,
"loss": 0.2038,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.18389900028705597,
"rewards/margins": 0.2373250424861908,
"rewards/rejected": -0.42122405767440796,
"step": 220
},
{
"epoch": 0.12,
"learning_rate": 4.992357124836838e-06,
"logits/chosen": -1.2583847045898438,
"logits/rejected": -0.5125802755355835,
"logps/chosen": -804.1488037109375,
"logps/rejected": -1768.70703125,
"loss": 0.2087,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.16385242342948914,
"rewards/margins": 0.26760467886924744,
"rewards/rejected": -0.4314571022987366,
"step": 230
},
{
"epoch": 0.13,
"learning_rate": 4.9882875734777044e-06,
"logits/chosen": -1.3879817724227905,
"logits/rejected": -0.7601082921028137,
"logps/chosen": -643.2053833007812,
"logps/rejected": -1582.4229736328125,
"loss": 0.2098,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.08198627084493637,
"rewards/margins": 0.24419136345386505,
"rewards/rejected": -0.3261776566505432,
"step": 240
},
{
"epoch": 0.13,
"learning_rate": 4.983355127818882e-06,
"logits/chosen": -1.2820708751678467,
"logits/rejected": -0.7363845109939575,
"logps/chosen": -591.2818603515625,
"logps/rejected": -1653.382080078125,
"loss": 0.1716,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.055700987577438354,
"rewards/margins": 0.2979794144630432,
"rewards/rejected": -0.35368040204048157,
"step": 250
},
{
"epoch": 0.14,
"learning_rate": 4.977561498345639e-06,
"logits/chosen": -1.5291858911514282,
"logits/rejected": -0.19423088431358337,
"logps/chosen": -732.954833984375,
"logps/rejected": -1630.454833984375,
"loss": 0.227,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.09654470533132553,
"rewards/margins": 0.2725062370300293,
"rewards/rejected": -0.36905092000961304,
"step": 260
},
{
"epoch": 0.14,
"learning_rate": 4.970908694186624e-06,
"logits/chosen": -1.2763826847076416,
"logits/rejected": -0.358491986989975,
"logps/chosen": -659.6267700195312,
"logps/rejected": -1627.0989990234375,
"loss": 0.1989,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.10589329153299332,
"rewards/margins": 0.22936058044433594,
"rewards/rejected": -0.33525389432907104,
"step": 270
},
{
"epoch": 0.15,
"learning_rate": 4.9633990224171305e-06,
"logits/chosen": -1.4150131940841675,
"logits/rejected": -0.5769174695014954,
"logps/chosen": -790.4682006835938,
"logps/rejected": -1594.6312255859375,
"loss": 0.1994,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.1530950367450714,
"rewards/margins": 0.20400938391685486,
"rewards/rejected": -0.3571043908596039,
"step": 280
},
{
"epoch": 0.15,
"learning_rate": 4.955035087259046e-06,
"logits/chosen": -1.5317466259002686,
"logits/rejected": -0.8124315142631531,
"logps/chosen": -732.0606079101562,
"logps/rejected": -1573.697021484375,
"loss": 0.192,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.16247740387916565,
"rewards/margins": 0.23650550842285156,
"rewards/rejected": -0.3989828824996948,
"step": 290
},
{
"epoch": 0.16,
"learning_rate": 4.945819789177756e-06,
"logits/chosen": -1.3253892660140991,
"logits/rejected": -0.6317247152328491,
"logps/chosen": -800.7886962890625,
"logps/rejected": -1812.166259765625,
"loss": 0.1507,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.15501004457473755,
"rewards/margins": 0.2739163339138031,
"rewards/rejected": -0.42892640829086304,
"step": 300
},
{
"epoch": 0.17,
"learning_rate": 4.935756323876306e-06,
"logits/chosen": -1.4076381921768188,
"logits/rejected": -0.5177509188652039,
"logps/chosen": -560.181396484375,
"logps/rejected": -1555.759521484375,
"loss": 0.1571,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.08090414851903915,
"rewards/margins": 0.30350688099861145,
"rewards/rejected": -0.3844110071659088,
"step": 310
},
{
"epoch": 0.17,
"learning_rate": 4.924848181187199e-06,
"logits/chosen": -1.5608190298080444,
"logits/rejected": -0.4406317174434662,
"logps/chosen": -630.440673828125,
"logps/rejected": -1558.788818359375,
"loss": 0.2186,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.11196412891149521,
"rewards/margins": 0.19183890521526337,
"rewards/rejected": -0.303803026676178,
"step": 320
},
{
"epoch": 0.18,
"learning_rate": 4.913099143862173e-06,
"logits/chosen": -1.414440631866455,
"logits/rejected": -0.49061527848243713,
"logps/chosen": -724.7615966796875,
"logps/rejected": -1786.733642578125,
"loss": 0.172,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.12826617062091827,
"rewards/margins": 0.3003011643886566,
"rewards/rejected": -0.4285673499107361,
"step": 330
},
{
"epoch": 0.18,
"learning_rate": 4.900513286260416e-06,
"logits/chosen": -1.4440535306930542,
"logits/rejected": -0.4736519753932953,
"logps/chosen": -700.4628295898438,
"logps/rejected": -1750.1715087890625,
"loss": 0.1702,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.10472197830677032,
"rewards/margins": 0.2777271866798401,
"rewards/rejected": -0.38244912028312683,
"step": 340
},
{
"epoch": 0.19,
"learning_rate": 4.887094972935645e-06,
"logits/chosen": -1.5601012706756592,
"logits/rejected": -0.19679054617881775,
"logps/chosen": -773.9390869140625,
"logps/rejected": -1652.917724609375,
"loss": 0.1594,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.13721729815006256,
"rewards/margins": 0.26322948932647705,
"rewards/rejected": -0.40044674277305603,
"step": 350
},
{
"epoch": 0.19,
"learning_rate": 4.87284885712256e-06,
"logits/chosen": -1.5175647735595703,
"logits/rejected": -0.5158883333206177,
"logps/chosen": -730.0354614257812,
"logps/rejected": -1762.952880859375,
"loss": 0.226,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.15878579020500183,
"rewards/margins": 0.2685873210430145,
"rewards/rejected": -0.42737311124801636,
"step": 360
},
{
"epoch": 0.2,
"learning_rate": 4.857779879123181e-06,
"logits/chosen": -1.2855349779129028,
"logits/rejected": -0.4378298819065094,
"logps/chosen": -668.7643432617188,
"logps/rejected": -1692.6859130859375,
"loss": 0.1967,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.16277587413787842,
"rewards/margins": 0.2703282833099365,
"rewards/rejected": -0.43310412764549255,
"step": 370
},
{
"epoch": 0.2,
"learning_rate": 4.841893264593643e-06,
"logits/chosen": -1.5854368209838867,
"logits/rejected": -0.7641209959983826,
"logps/chosen": -760.0360107421875,
"logps/rejected": -1679.648681640625,
"loss": 0.2054,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.1432400643825531,
"rewards/margins": 0.27862492203712463,
"rewards/rejected": -0.42186498641967773,
"step": 380
},
{
"epoch": 0.21,
"learning_rate": 4.825194522732023e-06,
"logits/chosen": -1.2518047094345093,
"logits/rejected": 0.1541678011417389,
"logps/chosen": -620.4710693359375,
"logps/rejected": -1564.021240234375,
"loss": 0.163,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.08593844622373581,
"rewards/margins": 0.3075733780860901,
"rewards/rejected": -0.3935117721557617,
"step": 390
},
{
"epoch": 0.21,
"learning_rate": 4.807689444367853e-06,
"logits/chosen": -1.3414735794067383,
"logits/rejected": 0.27303165197372437,
"logps/chosen": -734.915771484375,
"logps/rejected": -1726.513671875,
"loss": 0.1809,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.1514008492231369,
"rewards/margins": 0.3023954927921295,
"rewards/rejected": -0.45379629731178284,
"step": 400
},
{
"epoch": 0.22,
"learning_rate": 4.78938409995396e-06,
"logits/chosen": -1.6200311183929443,
"logits/rejected": -0.9766647219657898,
"logps/chosen": -630.611328125,
"logps/rejected": -1720.3140869140625,
"loss": 0.1582,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.14615099132061005,
"rewards/margins": 0.30823230743408203,
"rewards/rejected": -0.45438331365585327,
"step": 410
},
{
"epoch": 0.22,
"learning_rate": 4.770284837461342e-06,
"logits/chosen": -1.6855354309082031,
"logits/rejected": -0.5985993146896362,
"logps/chosen": -717.0406494140625,
"logps/rejected": -1674.328369140625,
"loss": 0.157,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.14980314671993256,
"rewards/margins": 0.27969104051589966,
"rewards/rejected": -0.42949414253234863,
"step": 420
},
{
"epoch": 0.23,
"learning_rate": 4.7503982801778015e-06,
"logits/chosen": -1.5795036554336548,
"logits/rejected": -0.012340852990746498,
"logps/chosen": -767.9241943359375,
"logps/rejected": -1746.126220703125,
"loss": 0.1509,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.19612105190753937,
"rewards/margins": 0.31204092502593994,
"rewards/rejected": -0.5081619024276733,
"step": 430
},
{
"epoch": 0.23,
"learning_rate": 4.729731324411104e-06,
"logits/chosen": -1.335930585861206,
"logits/rejected": 0.10618801414966583,
"logps/chosen": -792.5811767578125,
"logps/rejected": -1725.929443359375,
"loss": 0.1641,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.20740894973278046,
"rewards/margins": 0.2866581678390503,
"rewards/rejected": -0.49406710267066956,
"step": 440
},
{
"epoch": 0.24,
"learning_rate": 4.7082911370974645e-06,
"logits/chosen": -1.4894254207611084,
"logits/rejected": -0.4472936987876892,
"logps/chosen": -812.9591674804688,
"logps/rejected": -1778.461181640625,
"loss": 0.1515,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.16330209374427795,
"rewards/margins": 0.321885883808136,
"rewards/rejected": -0.48518800735473633,
"step": 450
},
{
"epoch": 0.25,
"learning_rate": 4.68608515331618e-06,
"logits/chosen": -1.3552906513214111,
"logits/rejected": 0.43209949135780334,
"logps/chosen": -839.5066528320312,
"logps/rejected": -1581.7581787109375,
"loss": 0.1617,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.18863311409950256,
"rewards/margins": 0.24536097049713135,
"rewards/rejected": -0.4339941143989563,
"step": 460
},
{
"epoch": 0.25,
"learning_rate": 4.663121073711269e-06,
"logits/chosen": -1.4539438486099243,
"logits/rejected": -0.8116742968559265,
"logps/chosen": -729.900634765625,
"logps/rejected": -1769.510986328125,
"loss": 0.1622,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.18170835077762604,
"rewards/margins": 0.3435843884944916,
"rewards/rejected": -0.5252927541732788,
"step": 470
},
{
"epoch": 0.26,
"learning_rate": 4.63940686182103e-06,
"logits/chosen": -1.2978649139404297,
"logits/rejected": -0.1232295036315918,
"logps/chosen": -824.3473510742188,
"logps/rejected": -1865.825439453125,
"loss": 0.1567,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.14898522198200226,
"rewards/margins": 0.2702946960926056,
"rewards/rejected": -0.41927990317344666,
"step": 480
},
{
"epoch": 0.26,
"learning_rate": 4.614950741316425e-06,
"logits/chosen": -1.4946939945220947,
"logits/rejected": 0.050956688821315765,
"logps/chosen": -708.5657958984375,
"logps/rejected": -1557.0811767578125,
"loss": 0.2028,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.09371930360794067,
"rewards/margins": 0.19662004709243774,
"rewards/rejected": -0.2903393507003784,
"step": 490
},
{
"epoch": 0.27,
"learning_rate": 4.589761193149254e-06,
"logits/chosen": -1.3152101039886475,
"logits/rejected": 0.32155555486679077,
"logps/chosen": -701.7420654296875,
"logps/rejected": -1673.237548828125,
"loss": 0.1656,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.127631276845932,
"rewards/margins": 0.2853541970252991,
"rewards/rejected": -0.4129854738712311,
"step": 500
},
{
"epoch": 0.27,
"learning_rate": 4.563846952611112e-06,
"logits/chosen": -1.159698247909546,
"logits/rejected": 1.021859884262085,
"logps/chosen": -789.81884765625,
"logps/rejected": -1585.963623046875,
"loss": 0.1666,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.21442024409770966,
"rewards/margins": 0.22131362557411194,
"rewards/rejected": -0.4357338845729828,
"step": 510
},
{
"epoch": 0.28,
"learning_rate": 4.537217006304141e-06,
"logits/chosen": -1.3945280313491821,
"logits/rejected": -0.13302992284297943,
"logps/chosen": -911.0645751953125,
"logps/rejected": -1837.2916259765625,
"loss": 0.1831,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.24346116185188293,
"rewards/margins": 0.2620731592178345,
"rewards/rejected": -0.505534291267395,
"step": 520
},
{
"epoch": 0.28,
"learning_rate": 4.50988058902464e-06,
"logits/chosen": -1.38224458694458,
"logits/rejected": -0.5236114859580994,
"logps/chosen": -886.48583984375,
"logps/rejected": -1802.9521484375,
"loss": 0.2098,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.29623985290527344,
"rewards/margins": 0.2730676829814911,
"rewards/rejected": -0.5693076252937317,
"step": 530
},
{
"epoch": 0.29,
"learning_rate": 4.481847180560593e-06,
"logits/chosen": -1.3253287076950073,
"logits/rejected": 0.04275239259004593,
"logps/chosen": -765.50048828125,
"logps/rejected": -1691.4166259765625,
"loss": 0.1454,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.12114331871271133,
"rewards/margins": 0.28145861625671387,
"rewards/rejected": -0.4026018977165222,
"step": 540
},
{
"epoch": 0.29,
"learning_rate": 4.453126502404253e-06,
"logits/chosen": -1.505723237991333,
"logits/rejected": -0.5302447080612183,
"logps/chosen": -686.806396484375,
"logps/rejected": -1857.9820556640625,
"loss": 0.1407,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.05712117999792099,
"rewards/margins": 0.3960820138454437,
"rewards/rejected": -0.4532032012939453,
"step": 550
},
{
"epoch": 0.3,
"learning_rate": 4.423728514380892e-06,
"logits/chosen": -1.3196251392364502,
"logits/rejected": -0.7369681596755981,
"logps/chosen": -761.491943359375,
"logps/rejected": -1692.483642578125,
"loss": 0.1989,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.09536974877119064,
"rewards/margins": 0.2588408887386322,
"rewards/rejected": -0.35421064496040344,
"step": 560
},
{
"epoch": 0.3,
"learning_rate": 4.393663411194918e-06,
"logits/chosen": -1.2871177196502686,
"logits/rejected": -0.42333096265792847,
"logps/chosen": -770.8462524414062,
"logps/rejected": -1845.9010009765625,
"loss": 0.1735,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.19140848517417908,
"rewards/margins": 0.343218058347702,
"rewards/rejected": -0.5346266031265259,
"step": 570
},
{
"epoch": 0.31,
"learning_rate": 4.362941618894523e-06,
"logits/chosen": -1.2731715440750122,
"logits/rejected": -0.4260900616645813,
"logps/chosen": -621.0642700195312,
"logps/rejected": -1537.65966796875,
"loss": 0.1433,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.14984165132045746,
"rewards/margins": 0.2771284878253937,
"rewards/rejected": -0.42697015404701233,
"step": 580
},
{
"epoch": 0.31,
"learning_rate": 4.331573791256116e-06,
"logits/chosen": -1.1939516067504883,
"logits/rejected": -0.4788111746311188,
"logps/chosen": -734.4415893554688,
"logps/rejected": -1785.6878662109375,
"loss": 0.1775,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.16298024356365204,
"rewards/margins": 0.288863867521286,
"rewards/rejected": -0.45184412598609924,
"step": 590
},
{
"epoch": 0.32,
"learning_rate": 4.299570806089786e-06,
"logits/chosen": -1.243289828300476,
"logits/rejected": -0.7459559440612793,
"logps/chosen": -627.3175048828125,
"logps/rejected": -1805.260498046875,
"loss": 0.1346,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.09667791426181793,
"rewards/margins": 0.37870708107948303,
"rewards/rejected": -0.47538501024246216,
"step": 600
},
{
"epoch": 0.33,
"learning_rate": 4.266943761467057e-06,
"logits/chosen": -1.5651763677597046,
"logits/rejected": -0.5712302327156067,
"logps/chosen": -705.9287719726562,
"logps/rejected": -1588.61279296875,
"loss": 0.1316,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.10819022357463837,
"rewards/margins": 0.26426267623901367,
"rewards/rejected": -0.37245288491249084,
"step": 610
},
{
"epoch": 0.33,
"learning_rate": 4.233703971872287e-06,
"logits/chosen": -1.372422218322754,
"logits/rejected": -0.3783385157585144,
"logps/chosen": -720.2787475585938,
"logps/rejected": -1659.054931640625,
"loss": 0.1541,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.09280645102262497,
"rewards/margins": 0.30042776465415955,
"rewards/rejected": -0.3932341933250427,
"step": 620
},
{
"epoch": 0.34,
"learning_rate": 4.1998629642789925e-06,
"logits/chosen": -1.4179317951202393,
"logits/rejected": -0.7052455544471741,
"logps/chosen": -801.2463989257812,
"logps/rejected": -1756.431884765625,
"loss": 0.169,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.1531040519475937,
"rewards/margins": 0.31313425302505493,
"rewards/rejected": -0.4662383496761322,
"step": 630
},
{
"epoch": 0.34,
"learning_rate": 4.165432474152505e-06,
"logits/chosen": -1.1481022834777832,
"logits/rejected": 0.5914746522903442,
"logps/chosen": -755.81103515625,
"logps/rejected": -1772.6920166015625,
"loss": 0.2111,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.15703272819519043,
"rewards/margins": 0.31566134095191956,
"rewards/rejected": -0.47269406914711,
"step": 640
},
{
"epoch": 0.35,
"learning_rate": 4.130424441380308e-06,
"logits/chosen": -1.2507625818252563,
"logits/rejected": -0.3694377839565277,
"logps/chosen": -659.0304565429688,
"logps/rejected": -1531.488525390625,
"loss": 0.1981,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.08185335248708725,
"rewards/margins": 0.20337195694446564,
"rewards/rejected": -0.2852253317832947,
"step": 650
},
{
"epoch": 0.35,
"learning_rate": 4.09485100613151e-06,
"logits/chosen": -1.4763983488082886,
"logits/rejected": -0.03595398738980293,
"logps/chosen": -728.7596435546875,
"logps/rejected": -1576.779052734375,
"loss": 0.1871,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.1268945336341858,
"rewards/margins": 0.21409745514392853,
"rewards/rejected": -0.3409919738769531,
"step": 660
},
{
"epoch": 0.36,
"learning_rate": 4.058724504646834e-06,
"logits/chosen": -1.0833264589309692,
"logits/rejected": -0.46659836173057556,
"logps/chosen": -780.9652099609375,
"logps/rejected": -1779.447998046875,
"loss": 0.1612,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.17059431970119476,
"rewards/margins": 0.328458696603775,
"rewards/rejected": -0.4990530014038086,
"step": 670
},
{
"epoch": 0.36,
"learning_rate": 4.022057464960632e-06,
"logits/chosen": -1.399720549583435,
"logits/rejected": -0.561631977558136,
"logps/chosen": -789.7105102539062,
"logps/rejected": -1804.3443603515625,
"loss": 0.1701,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.12060783058404922,
"rewards/margins": 0.3465253710746765,
"rewards/rejected": -0.46713313460350037,
"step": 680
},
{
"epoch": 0.37,
"learning_rate": 3.984862602556383e-06,
"logits/chosen": -1.2853963375091553,
"logits/rejected": -0.2969803214073181,
"logps/chosen": -711.6168823242188,
"logps/rejected": -1566.543701171875,
"loss": 0.1855,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.159945547580719,
"rewards/margins": 0.22837324440479279,
"rewards/rejected": -0.3883187770843506,
"step": 690
},
{
"epoch": 0.37,
"learning_rate": 3.947152815957187e-06,
"logits/chosen": -1.2532896995544434,
"logits/rejected": -0.5551369786262512,
"logps/chosen": -636.3924560546875,
"logps/rejected": -1902.0989990234375,
"loss": 0.17,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.1209067553281784,
"rewards/margins": 0.3948102295398712,
"rewards/rejected": -0.5157170295715332,
"step": 700
},
{
"epoch": 0.38,
"learning_rate": 3.908941182252785e-06,
"logits/chosen": -1.4191337823867798,
"logits/rejected": -0.11001946777105331,
"logps/chosen": -713.93212890625,
"logps/rejected": -1656.491943359375,
"loss": 0.1692,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.07871367037296295,
"rewards/margins": 0.2747347354888916,
"rewards/rejected": -0.35344842076301575,
"step": 710
},
{
"epoch": 0.38,
"learning_rate": 3.8702409525646535e-06,
"logits/chosen": -1.3239977359771729,
"logits/rejected": -0.34808486700057983,
"logps/chosen": -781.5916137695312,
"logps/rejected": -1643.9801025390625,
"loss": 0.1728,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.1793273687362671,
"rewards/margins": 0.2667609453201294,
"rewards/rejected": -0.4460883140563965,
"step": 720
},
{
"epoch": 0.39,
"learning_rate": 3.8310655474507495e-06,
"logits/chosen": -1.5411983728408813,
"logits/rejected": -0.2094588726758957,
"logps/chosen": -853.9683837890625,
"logps/rejected": -1529.486572265625,
"loss": 0.8403,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.16295142471790314,
"rewards/margins": 0.15665379166603088,
"rewards/rejected": -0.3196052312850952,
"step": 730
},
{
"epoch": 0.39,
"learning_rate": 3.7914285522515002e-06,
"logits/chosen": -1.6163041591644287,
"logits/rejected": -0.5802554488182068,
"logps/chosen": -680.8773193359375,
"logps/rejected": -1595.187744140625,
"loss": 0.247,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.054552387446165085,
"rewards/margins": 0.15820157527923584,
"rewards/rejected": -0.21275396645069122,
"step": 740
},
{
"epoch": 0.4,
"learning_rate": 3.751343712378639e-06,
"logits/chosen": -1.4396401643753052,
"logits/rejected": -0.4410804212093353,
"logps/chosen": -692.0823974609375,
"logps/rejected": -1640.334228515625,
"loss": 0.2633,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.04884466528892517,
"rewards/margins": 0.16748689115047455,
"rewards/rejected": -0.2163315713405609,
"step": 750
},
{
"epoch": 0.41,
"learning_rate": 3.710824928548546e-06,
"logits/chosen": -1.5297174453735352,
"logits/rejected": -0.6668530106544495,
"logps/chosen": -611.5433349609375,
"logps/rejected": -1556.078125,
"loss": 0.2196,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.033245183527469635,
"rewards/margins": 0.21886444091796875,
"rewards/rejected": -0.2521096169948578,
"step": 760
},
{
"epoch": 0.41,
"learning_rate": 3.6698862519617225e-06,
"logits/chosen": -1.5771548748016357,
"logits/rejected": -0.4002736508846283,
"logps/chosen": -641.4221801757812,
"logps/rejected": -1400.081298828125,
"loss": 0.2114,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.03950526565313339,
"rewards/margins": 0.1628362238407135,
"rewards/rejected": -0.2023414820432663,
"step": 770
},
{
"epoch": 0.42,
"learning_rate": 3.6285418794300793e-06,
"logits/chosen": -1.587998628616333,
"logits/rejected": -0.8064206838607788,
"logps/chosen": -660.1177978515625,
"logps/rejected": -1664.508056640625,
"loss": 0.2068,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.0692194476723671,
"rewards/margins": 0.28254234790802,
"rewards/rejected": -0.3517618179321289,
"step": 780
},
{
"epoch": 0.42,
"learning_rate": 3.5868061484537365e-06,
"logits/chosen": -1.5779733657836914,
"logits/rejected": -0.9056744575500488,
"logps/chosen": -644.6177978515625,
"logps/rejected": -1487.890869140625,
"loss": 0.1864,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.09771857410669327,
"rewards/margins": 0.18923287093639374,
"rewards/rejected": -0.2869514524936676,
"step": 790
},
{
"epoch": 0.43,
"learning_rate": 3.5446935322490285e-06,
"logits/chosen": -1.4548509120941162,
"logits/rejected": -0.490948349237442,
"logps/chosen": -806.4798583984375,
"logps/rejected": -1708.6116943359375,
"loss": 0.1662,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.1382991373538971,
"rewards/margins": 0.2544620931148529,
"rewards/rejected": -0.3927612006664276,
"step": 800
},
{
"epoch": 0.43,
"learning_rate": 3.502218634729447e-06,
"logits/chosen": -1.2813379764556885,
"logits/rejected": -0.6146308183670044,
"logps/chosen": -652.1132202148438,
"logps/rejected": -1596.59619140625,
"loss": 0.2302,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.10586214065551758,
"rewards/margins": 0.2745637595653534,
"rewards/rejected": -0.38042590022087097,
"step": 810
},
{
"epoch": 0.44,
"learning_rate": 3.459396185441265e-06,
"logits/chosen": -1.5315301418304443,
"logits/rejected": -0.7480667233467102,
"logps/chosen": -580.1861572265625,
"logps/rejected": -1714.6907958984375,
"loss": 0.1641,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.04294740408658981,
"rewards/margins": 0.34141218662261963,
"rewards/rejected": -0.38435959815979004,
"step": 820
},
{
"epoch": 0.44,
"learning_rate": 3.4162410344555834e-06,
"logits/chosen": -1.462127447128296,
"logits/rejected": -0.8618795275688171,
"logps/chosen": -741.3155517578125,
"logps/rejected": -1888.650146484375,
"loss": 0.1266,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.09163922816514969,
"rewards/margins": 0.3773654103279114,
"rewards/rejected": -0.46900463104248047,
"step": 830
},
{
"epoch": 0.45,
"learning_rate": 3.3727681472185937e-06,
"logits/chosen": -1.6406835317611694,
"logits/rejected": -0.5775340795516968,
"logps/chosen": -797.4598999023438,
"logps/rejected": -1779.310302734375,
"loss": 0.1442,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.11959085613489151,
"rewards/margins": 0.32207822799682617,
"rewards/rejected": -0.4416690468788147,
"step": 840
},
{
"epoch": 0.45,
"learning_rate": 3.3289925993618217e-06,
"logits/chosen": -1.3674832582473755,
"logits/rejected": -0.03007361851632595,
"logps/chosen": -736.8541870117188,
"logps/rejected": -1707.4468994140625,
"loss": 0.1866,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.08552592992782593,
"rewards/margins": 0.30176034569740295,
"rewards/rejected": -0.3872862458229065,
"step": 850
},
{
"epoch": 0.46,
"learning_rate": 3.2849295714741643e-06,
"logits/chosen": -1.206857442855835,
"logits/rejected": -0.4066374897956848,
"logps/chosen": -732.4100341796875,
"logps/rejected": -1567.50048828125,
"loss": 0.193,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.09428885579109192,
"rewards/margins": 0.255453884601593,
"rewards/rejected": -0.34974271059036255,
"step": 860
},
{
"epoch": 0.46,
"learning_rate": 3.2405943438375287e-06,
"logits/chosen": -1.2475948333740234,
"logits/rejected": -0.3390772342681885,
"logps/chosen": -550.7503662109375,
"logps/rejected": -1530.1998291015625,
"loss": 0.1633,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.018922504037618637,
"rewards/margins": 0.2624686658382416,
"rewards/rejected": -0.2813911437988281,
"step": 870
},
{
"epoch": 0.47,
"learning_rate": 3.1960022911279036e-06,
"logits/chosen": -1.1934720277786255,
"logits/rejected": -0.11626192182302475,
"logps/chosen": -602.2255859375,
"logps/rejected": -1660.957763671875,
"loss": 0.1537,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.03315038979053497,
"rewards/margins": 0.32097241282463074,
"rewards/rejected": -0.3541228175163269,
"step": 880
},
{
"epoch": 0.47,
"learning_rate": 3.1511688770836844e-06,
"logits/chosen": -1.501976728439331,
"logits/rejected": -0.6784012913703918,
"logps/chosen": -589.7467041015625,
"logps/rejected": -1566.115478515625,
"loss": 0.1589,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.06981668621301651,
"rewards/margins": 0.30536967515945435,
"rewards/rejected": -0.3751864433288574,
"step": 890
},
{
"epoch": 0.48,
"learning_rate": 3.1061096491431307e-06,
"logits/chosen": -1.683624267578125,
"logits/rejected": -0.6069055795669556,
"logps/chosen": -765.5227661132812,
"logps/rejected": -1787.843017578125,
"loss": 0.1679,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.0950092300772667,
"rewards/margins": 0.34752824902534485,
"rewards/rejected": -0.44253745675086975,
"step": 900
},
{
"epoch": 0.49,
"learning_rate": 3.0608402330527796e-06,
"logits/chosen": -1.5381743907928467,
"logits/rejected": -0.275674045085907,
"logps/chosen": -740.3082885742188,
"logps/rejected": -1551.924560546875,
"loss": 0.169,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.12154094874858856,
"rewards/margins": 0.24538445472717285,
"rewards/rejected": -0.3669254183769226,
"step": 910
},
{
"epoch": 0.49,
"learning_rate": 3.0153763274487176e-06,
"logits/chosen": -1.3017113208770752,
"logits/rejected": -0.47879767417907715,
"logps/chosen": -677.2306518554688,
"logps/rejected": -1834.7994384765625,
"loss": 0.1478,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.09329520165920258,
"rewards/margins": 0.37344565987586975,
"rewards/rejected": -0.4667409062385559,
"step": 920
},
{
"epoch": 0.5,
"learning_rate": 2.9697336984125683e-06,
"logits/chosen": -1.5965297222137451,
"logits/rejected": -0.08967357873916626,
"logps/chosen": -750.0841064453125,
"logps/rejected": -1768.0572509765625,
"loss": 0.1495,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.0852770283818245,
"rewards/margins": 0.3263811767101288,
"rewards/rejected": -0.4116581380367279,
"step": 930
},
{
"epoch": 0.5,
"learning_rate": 2.923928174004094e-06,
"logits/chosen": -1.5777353048324585,
"logits/rejected": -0.23106975853443146,
"logps/chosen": -790.501953125,
"logps/rejected": -1668.7314453125,
"loss": 0.2048,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.14181482791900635,
"rewards/margins": 0.2708645462989807,
"rewards/rejected": -0.41267937421798706,
"step": 940
},
{
"epoch": 0.51,
"learning_rate": 2.8779756387723036e-06,
"logits/chosen": -1.1467679738998413,
"logits/rejected": 0.07232928276062012,
"logps/chosen": -592.8287963867188,
"logps/rejected": -1479.1529541015625,
"loss": 0.1978,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.032741207629442215,
"rewards/margins": 0.26754575967788696,
"rewards/rejected": -0.3002869486808777,
"step": 950
},
{
"epoch": 0.51,
"learning_rate": 2.831892028246968e-06,
"logits/chosen": -1.6411889791488647,
"logits/rejected": -0.334705650806427,
"logps/chosen": -637.6863403320312,
"logps/rejected": -1307.0592041015625,
"loss": 0.1838,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.059777967631816864,
"rewards/margins": 0.1975717544555664,
"rewards/rejected": -0.25734972953796387,
"step": 960
},
{
"epoch": 0.52,
"learning_rate": 2.7856933234124617e-06,
"logits/chosen": -1.4477770328521729,
"logits/rejected": -0.447587788105011,
"logps/chosen": -725.3448486328125,
"logps/rejected": -1688.343994140625,
"loss": 0.1638,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.07935594022274017,
"rewards/margins": 0.2956528663635254,
"rewards/rejected": -0.37500879168510437,
"step": 970
},
{
"epoch": 0.52,
"learning_rate": 2.7393955451658387e-06,
"logits/chosen": -1.441928505897522,
"logits/rejected": -0.4485958516597748,
"logps/chosen": -674.2760620117188,
"logps/rejected": -1747.831787109375,
"loss": 0.1706,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.04113630950450897,
"rewards/margins": 0.3252195417881012,
"rewards/rejected": -0.366355836391449,
"step": 980
},
{
"epoch": 0.53,
"learning_rate": 2.6930147487610667e-06,
"logits/chosen": -1.560974359512329,
"logits/rejected": -0.35725072026252747,
"logps/chosen": -719.2286376953125,
"logps/rejected": -1786.9176025390625,
"loss": 0.1374,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.06295976042747498,
"rewards/margins": 0.3578505218029022,
"rewards/rejected": -0.4208102822303772,
"step": 990
},
{
"epoch": 0.53,
"learning_rate": 2.6465670182413487e-06,
"logits/chosen": -1.5361359119415283,
"logits/rejected": -0.5901980996131897,
"logps/chosen": -779.523193359375,
"logps/rejected": -1766.1044921875,
"loss": 0.1344,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.07964052259922028,
"rewards/margins": 0.3218533396720886,
"rewards/rejected": -0.4014938771724701,
"step": 1000
},
{
"epoch": 0.54,
"learning_rate": 2.6000684608614594e-06,
"logits/chosen": -1.3095166683197021,
"logits/rejected": -0.6045408248901367,
"logps/chosen": -703.9786376953125,
"logps/rejected": -1905.917236328125,
"loss": 0.1328,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.08165670931339264,
"rewards/margins": 0.39877375960350037,
"rewards/rejected": -0.4804305136203766,
"step": 1010
},
{
"epoch": 0.54,
"learning_rate": 2.5535352015020338e-06,
"logits/chosen": -1.4772388935089111,
"logits/rejected": -0.006365871522575617,
"logps/chosen": -647.6913452148438,
"logps/rejected": -1643.7066650390625,
"loss": 0.1533,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.08445201814174652,
"rewards/margins": 0.3088130056858063,
"rewards/rejected": -0.3932650089263916,
"step": 1020
},
{
"epoch": 0.55,
"learning_rate": 2.506983377077741e-06,
"logits/chosen": -1.480623483657837,
"logits/rejected": -0.7142573595046997,
"logps/chosen": -687.6082763671875,
"logps/rejected": -1855.853515625,
"loss": 0.1823,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.10191044956445694,
"rewards/margins": 0.38636595010757446,
"rewards/rejected": -0.48827633261680603,
"step": 1030
},
{
"epoch": 0.55,
"learning_rate": 2.460429130941289e-06,
"logits/chosen": -1.3591618537902832,
"logits/rejected": -0.5336927771568298,
"logps/chosen": -549.4283447265625,
"logps/rejected": -1634.71240234375,
"loss": 0.1329,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.05249984189867973,
"rewards/margins": 0.3793482184410095,
"rewards/rejected": -0.43184810876846313,
"step": 1040
},
{
"epoch": 0.56,
"learning_rate": 2.413888607285192e-06,
"logits/chosen": -1.6825790405273438,
"logits/rejected": -0.8754396438598633,
"logps/chosen": -707.4894409179688,
"logps/rejected": -1693.762939453125,
"loss": 0.1264,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.05124876648187637,
"rewards/margins": 0.3190692365169525,
"rewards/rejected": -0.3703179955482483,
"step": 1050
},
{
"epoch": 0.57,
"learning_rate": 2.367377945543249e-06,
"logits/chosen": -1.4010018110275269,
"logits/rejected": -0.7923344373703003,
"logps/chosen": -623.4701538085938,
"logps/rejected": -1506.981689453125,
"loss": 0.1948,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.061522554606199265,
"rewards/margins": 0.23432263731956482,
"rewards/rejected": -0.2958451807498932,
"step": 1060
},
{
"epoch": 0.57,
"learning_rate": 2.320913274793676e-06,
"logits/chosen": -1.5388412475585938,
"logits/rejected": -0.185507133603096,
"logps/chosen": -622.5823974609375,
"logps/rejected": -1469.159912109375,
"loss": 0.2035,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.07265051454305649,
"rewards/margins": 0.2589751183986664,
"rewards/rejected": -0.33162569999694824,
"step": 1070
},
{
"epoch": 0.58,
"learning_rate": 2.27451070816582e-06,
"logits/chosen": -1.4987528324127197,
"logits/rejected": -0.6492592096328735,
"logps/chosen": -629.9580078125,
"logps/rejected": -1712.423095703125,
"loss": 0.1277,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.05187363177537918,
"rewards/margins": 0.3656329810619354,
"rewards/rejected": -0.417506605386734,
"step": 1080
},
{
"epoch": 0.58,
"learning_rate": 2.228186337252414e-06,
"logits/chosen": -1.4696351289749146,
"logits/rejected": -0.2756325900554657,
"logps/chosen": -624.1102294921875,
"logps/rejected": -1571.719482421875,
"loss": 0.1663,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.03262772411108017,
"rewards/margins": 0.2939130663871765,
"rewards/rejected": -0.3265407681465149,
"step": 1090
},
{
"epoch": 0.59,
"learning_rate": 2.1819562265292946e-06,
"logits/chosen": -1.5407931804656982,
"logits/rejected": -0.6150112152099609,
"logps/chosen": -606.6552734375,
"logps/rejected": -1499.91455078125,
"loss": 0.2052,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.047359392046928406,
"rewards/margins": 0.25734394788742065,
"rewards/rejected": -0.30470338463783264,
"step": 1100
},
{
"epoch": 0.59,
"learning_rate": 2.1358364077845236e-06,
"logits/chosen": -1.5111992359161377,
"logits/rejected": -0.6662728786468506,
"logps/chosen": -630.7785034179688,
"logps/rejected": -1606.4742431640625,
"loss": 0.1755,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.04623417183756828,
"rewards/margins": 0.32483187317848206,
"rewards/rejected": -0.37106603384017944,
"step": 1110
},
{
"epoch": 0.6,
"learning_rate": 2.089842874558849e-06,
"logits/chosen": -1.3981125354766846,
"logits/rejected": -0.36660924553871155,
"logps/chosen": -675.8720092773438,
"logps/rejected": -1751.074462890625,
"loss": 0.1254,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.03159493952989578,
"rewards/margins": 0.329929918050766,
"rewards/rejected": -0.3615248203277588,
"step": 1120
},
{
"epoch": 0.6,
"learning_rate": 2.0439915765994242e-06,
"logits/chosen": -1.5726184844970703,
"logits/rejected": -0.76716148853302,
"logps/chosen": -569.0755004882812,
"logps/rejected": -1726.5882568359375,
"loss": 0.1363,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.014021654613316059,
"rewards/margins": 0.37932100892066956,
"rewards/rejected": -0.39334264397621155,
"step": 1130
},
{
"epoch": 0.61,
"learning_rate": 1.9982984143287186e-06,
"logits/chosen": -1.3213517665863037,
"logits/rejected": -0.7754068374633789,
"logps/chosen": -576.7313232421875,
"logps/rejected": -1728.904541015625,
"loss": 0.1591,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.05365558713674545,
"rewards/margins": 0.3717700242996216,
"rewards/rejected": -0.42542558908462524,
"step": 1140
},
{
"epoch": 0.61,
"learning_rate": 1.95277923333053e-06,
"logits/chosen": -1.4281871318817139,
"logits/rejected": -0.8301714658737183,
"logps/chosen": -587.6362915039062,
"logps/rejected": -1682.287109375,
"loss": 0.1537,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.04957905039191246,
"rewards/margins": 0.3476232886314392,
"rewards/rejected": -0.39720237255096436,
"step": 1150
},
{
"epoch": 0.62,
"learning_rate": 1.9074498188550156e-06,
"logits/chosen": -1.2723270654678345,
"logits/rejected": -0.2873464524745941,
"logps/chosen": -679.1395263671875,
"logps/rejected": -1777.858154296875,
"loss": 0.1868,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.02069082297384739,
"rewards/margins": 0.3992442190647125,
"rewards/rejected": -0.41993504762649536,
"step": 1160
},
{
"epoch": 0.62,
"learning_rate": 1.862325890344643e-06,
"logits/chosen": -1.1864560842514038,
"logits/rejected": -0.4859984815120697,
"logps/chosen": -683.1911010742188,
"logps/rejected": -1602.1019287109375,
"loss": 0.1687,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.018726080656051636,
"rewards/margins": 0.29654327034950256,
"rewards/rejected": -0.3152693212032318,
"step": 1170
},
{
"epoch": 0.63,
"learning_rate": 1.817423095982972e-06,
"logits/chosen": -1.0037411451339722,
"logits/rejected": 0.02554410696029663,
"logps/chosen": -579.2996826171875,
"logps/rejected": -1649.9722900390625,
"loss": 0.1673,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.02113654837012291,
"rewards/margins": 0.3451305031776428,
"rewards/rejected": -0.36626705527305603,
"step": 1180
},
{
"epoch": 0.63,
"learning_rate": 1.7727570072681293e-06,
"logits/chosen": -1.3622825145721436,
"logits/rejected": -0.4110015332698822,
"logps/chosen": -589.6905517578125,
"logps/rejected": -1514.8289794921875,
"loss": 0.1964,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.05109922215342522,
"rewards/margins": 0.2517136335372925,
"rewards/rejected": -0.3028128147125244,
"step": 1190
},
{
"epoch": 0.64,
"learning_rate": 1.7283431136128961e-06,
"logits/chosen": -1.5205695629119873,
"logits/rejected": -0.16487915813922882,
"logps/chosen": -619.509521484375,
"logps/rejected": -1595.480712890625,
"loss": 0.1832,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.04752761498093605,
"rewards/margins": 0.3229573965072632,
"rewards/rejected": -0.37048497796058655,
"step": 1200
},
{
"epoch": 0.65,
"learning_rate": 1.6841968169732478e-06,
"logits/chosen": -1.4372893571853638,
"logits/rejected": -0.820128321647644,
"logps/chosen": -563.7713623046875,
"logps/rejected": -1725.425048828125,
"loss": 0.1289,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.05251551419496536,
"rewards/margins": 0.3404791057109833,
"rewards/rejected": -0.3929946720600128,
"step": 1210
},
{
"epoch": 0.65,
"learning_rate": 1.6403334265072284e-06,
"logits/chosen": -1.4803403615951538,
"logits/rejected": -0.7152812480926514,
"logps/chosen": -526.8184814453125,
"logps/rejected": -1638.043212890625,
"loss": 0.1768,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.02356002852320671,
"rewards/margins": 0.35318613052368164,
"rewards/rejected": -0.37674617767333984,
"step": 1220
},
{
"epoch": 0.66,
"learning_rate": 1.5967681532660066e-06,
"logits/chosen": -1.5741603374481201,
"logits/rejected": -0.42631644010543823,
"logps/chosen": -597.0006103515625,
"logps/rejected": -1670.453369140625,
"loss": 0.1572,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.043170906603336334,
"rewards/margins": 0.3142518401145935,
"rewards/rejected": -0.35742273926734924,
"step": 1230
},
{
"epoch": 0.66,
"learning_rate": 1.5535161049189463e-06,
"logits/chosen": -1.4529565572738647,
"logits/rejected": -0.5520704984664917,
"logps/chosen": -587.2677001953125,
"logps/rejected": -1800.4755859375,
"loss": 0.1516,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.017983034253120422,
"rewards/margins": 0.37644895911216736,
"rewards/rejected": -0.39443203806877136,
"step": 1240
},
{
"epoch": 0.67,
"learning_rate": 1.5105922805145356e-06,
"logits/chosen": -1.5385555028915405,
"logits/rejected": -0.7931126952171326,
"logps/chosen": -644.555419921875,
"logps/rejected": -1826.6881103515625,
"loss": 0.1694,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.06908350437879562,
"rewards/margins": 0.3630077540874481,
"rewards/rejected": -0.43209123611450195,
"step": 1250
},
{
"epoch": 0.67,
"learning_rate": 1.4680115652789823e-06,
"logits/chosen": -1.5401626825332642,
"logits/rejected": -0.22812063992023468,
"logps/chosen": -567.6744384765625,
"logps/rejected": -1582.8587646484375,
"loss": 0.1796,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.0517122745513916,
"rewards/margins": 0.33719173073768616,
"rewards/rejected": -0.38890397548675537,
"step": 1260
},
{
"epoch": 0.68,
"learning_rate": 1.4257887254542767e-06,
"logits/chosen": -1.204032301902771,
"logits/rejected": -0.2607814073562622,
"logps/chosen": -608.66015625,
"logps/rejected": -1682.341796875,
"loss": 0.1114,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.014287752099335194,
"rewards/margins": 0.3475797176361084,
"rewards/rejected": -0.3618674874305725,
"step": 1270
},
{
"epoch": 0.68,
"learning_rate": 1.3839384031775227e-06,
"logits/chosen": -1.3903391361236572,
"logits/rejected": -0.7587147951126099,
"logps/chosen": -676.6446533203125,
"logps/rejected": -1709.818603515625,
"loss": 0.1956,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.056112587451934814,
"rewards/margins": 0.281891405582428,
"rewards/rejected": -0.3380039930343628,
"step": 1280
},
{
"epoch": 0.69,
"learning_rate": 1.342475111403298e-06,
"logits/chosen": -1.519988775253296,
"logits/rejected": -1.1464245319366455,
"logps/chosen": -668.7023315429688,
"logps/rejected": -1754.447265625,
"loss": 0.1628,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.07075043022632599,
"rewards/margins": 0.3179227113723755,
"rewards/rejected": -0.38867321610450745,
"step": 1290
},
{
"epoch": 0.69,
"learning_rate": 1.3014132288708209e-06,
"logits/chosen": -1.45747971534729,
"logits/rejected": -0.840915858745575,
"logps/chosen": -672.7967529296875,
"logps/rejected": -1663.6207275390625,
"loss": 0.1768,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.05158332735300064,
"rewards/margins": 0.32626426219940186,
"rewards/rejected": -0.3778475821018219,
"step": 1300
},
{
"epoch": 0.7,
"learning_rate": 1.2607669951176549e-06,
"logits/chosen": -1.3014637231826782,
"logits/rejected": -0.45695143938064575,
"logps/chosen": -718.7424926757812,
"logps/rejected": -1485.810302734375,
"loss": 0.1979,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.08161523938179016,
"rewards/margins": 0.20484736561775208,
"rewards/rejected": -0.28646257519721985,
"step": 1310
},
{
"epoch": 0.7,
"learning_rate": 1.2205505055416891e-06,
"logits/chosen": -1.557997465133667,
"logits/rejected": 0.1897609382867813,
"logps/chosen": -638.6654052734375,
"logps/rejected": -1694.9169921875,
"loss": 0.1563,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.0795014351606369,
"rewards/margins": 0.33878716826438904,
"rewards/rejected": -0.41828861832618713,
"step": 1320
},
{
"epoch": 0.71,
"learning_rate": 1.1807777065131002e-06,
"logits/chosen": -1.5720739364624023,
"logits/rejected": -0.3039505183696747,
"logps/chosen": -629.5435791015625,
"logps/rejected": -1699.1048583984375,
"loss": 0.1449,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.02949734404683113,
"rewards/margins": 0.3422970175743103,
"rewards/rejected": -0.37179437279701233,
"step": 1330
},
{
"epoch": 0.71,
"learning_rate": 1.1414623905380012e-06,
"logits/chosen": -1.4923521280288696,
"logits/rejected": -0.543735682964325,
"logps/chosen": -737.5084838867188,
"logps/rejected": -1615.936279296875,
"loss": 0.1457,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.060771964490413666,
"rewards/margins": 0.2899934649467468,
"rewards/rejected": -0.3507654070854187,
"step": 1340
},
{
"epoch": 0.72,
"learning_rate": 1.1026181914754388e-06,
"logits/chosen": -1.5359987020492554,
"logits/rejected": -0.2435634583234787,
"logps/chosen": -701.9688720703125,
"logps/rejected": -1746.5748291015625,
"loss": 0.1456,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.04527285322546959,
"rewards/margins": 0.3345581889152527,
"rewards/rejected": -0.3798310160636902,
"step": 1350
},
{
"epoch": 0.73,
"learning_rate": 1.0642585798094136e-06,
"logits/chosen": -1.4265598058700562,
"logits/rejected": -0.9259235262870789,
"logps/chosen": -662.4346923828125,
"logps/rejected": -1895.0230712890625,
"loss": 0.1388,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.036455560475587845,
"rewards/margins": 0.42975935339927673,
"rewards/rejected": -0.46621495485305786,
"step": 1360
},
{
"epoch": 0.73,
"learning_rate": 1.0263968579775522e-06,
"logits/chosen": -1.4489424228668213,
"logits/rejected": -0.3130696713924408,
"logps/chosen": -677.5928955078125,
"logps/rejected": -1693.2991943359375,
"loss": 0.1653,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.07123799622058868,
"rewards/margins": 0.33964475989341736,
"rewards/rejected": -0.41088271141052246,
"step": 1370
},
{
"epoch": 0.74,
"learning_rate": 9.89046155758058e-07,
"logits/chosen": -1.332637071609497,
"logits/rejected": -0.8693612217903137,
"logps/chosen": -744.2261962890625,
"logps/rejected": -1811.7154541015625,
"loss": 0.1795,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.06084796041250229,
"rewards/margins": 0.39476925134658813,
"rewards/rejected": -0.45561712980270386,
"step": 1380
},
{
"epoch": 0.74,
"learning_rate": 9.52219425716534e-07,
"logits/chosen": -1.3288590908050537,
"logits/rejected": -0.16629138588905334,
"logps/chosen": -530.2454833984375,
"logps/rejected": -1626.1544189453125,
"loss": 0.1609,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.031980741769075394,
"rewards/margins": 0.3823274075984955,
"rewards/rejected": -0.4143081605434418,
"step": 1390
},
{
"epoch": 0.75,
"learning_rate": 9.15929438714262e-07,
"logits/chosen": -1.291416883468628,
"logits/rejected": -0.6190372705459595,
"logps/chosen": -561.5321044921875,
"logps/rejected": -1744.47265625,
"loss": 0.1245,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.05614073947072029,
"rewards/margins": 0.41081467270851135,
"rewards/rejected": -0.46695536375045776,
"step": 1400
},
{
"epoch": 0.75,
"learning_rate": 8.801887794794911e-07,
"logits/chosen": -1.505129098892212,
"logits/rejected": -0.4506317973136902,
"logps/chosen": -740.1429443359375,
"logps/rejected": -1584.6307373046875,
"loss": 0.1815,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.08272895216941833,
"rewards/margins": 0.3045782446861267,
"rewards/rejected": -0.38730719685554504,
"step": 1410
},
{
"epoch": 0.76,
"learning_rate": 8.450098422432787e-07,
"logits/chosen": -1.6013845205307007,
"logits/rejected": -0.6014512777328491,
"logps/chosen": -796.7876586914062,
"logps/rejected": -1698.9300537109375,
"loss": 0.1551,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.09957234561443329,
"rewards/margins": 0.32322657108306885,
"rewards/rejected": -0.42279893159866333,
"step": 1420
},
{
"epoch": 0.76,
"learning_rate": 8.104048264413858e-07,
"logits/chosen": -1.514638900756836,
"logits/rejected": -0.8343914747238159,
"logps/chosen": -563.3963012695312,
"logps/rejected": -1506.22509765625,
"loss": 0.154,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.046018943190574646,
"rewards/margins": 0.339202880859375,
"rewards/rejected": -0.38522180914878845,
"step": 1430
},
{
"epoch": 0.77,
"learning_rate": 7.763857324837321e-07,
"logits/chosen": -1.4705476760864258,
"logits/rejected": -0.28175801038742065,
"logps/chosen": -615.4827880859375,
"logps/rejected": -1487.244873046875,
"loss": 0.1992,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.07384703308343887,
"rewards/margins": 0.31356969475746155,
"rewards/rejected": -0.38741669058799744,
"step": 1440
},
{
"epoch": 0.77,
"learning_rate": 7.429643575928605e-07,
"logits/chosen": -1.1992992162704468,
"logits/rejected": -0.007973253726959229,
"logps/chosen": -689.2235107421875,
"logps/rejected": -1613.805419921875,
"loss": 0.1647,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.06208137422800064,
"rewards/margins": 0.2548282742500305,
"rewards/rejected": -0.31690964102745056,
"step": 1450
},
{
"epoch": 0.78,
"learning_rate": 7.101522917128709e-07,
"logits/chosen": -1.626381278038025,
"logits/rejected": -0.5168036222457886,
"logps/chosen": -666.3912963867188,
"logps/rejected": -1561.929443359375,
"loss": 0.1865,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.033760882914066315,
"rewards/margins": 0.2637539803981781,
"rewards/rejected": -0.2975148558616638,
"step": 1460
},
{
"epoch": 0.78,
"learning_rate": 6.779609134902312e-07,
"logits/chosen": -1.593224287033081,
"logits/rejected": -0.7308230400085449,
"logps/chosen": -683.6494750976562,
"logps/rejected": -1443.498046875,
"loss": 0.1364,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.04548930376768112,
"rewards/margins": 0.2794992923736572,
"rewards/rejected": -0.32498857378959656,
"step": 1470
},
{
"epoch": 0.79,
"learning_rate": 6.464013863278629e-07,
"logits/chosen": -1.511319637298584,
"logits/rejected": -0.5918843150138855,
"logps/chosen": -622.5799560546875,
"logps/rejected": -1675.676513671875,
"loss": 0.1608,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.02713722549378872,
"rewards/margins": 0.3606105446815491,
"rewards/rejected": -0.38774776458740234,
"step": 1480
},
{
"epoch": 0.79,
"learning_rate": 6.154846545138696e-07,
"logits/chosen": -1.368607997894287,
"logits/rejected": -0.45999422669410706,
"logps/chosen": -757.634033203125,
"logps/rejected": -1727.1351318359375,
"loss": 0.1844,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.09356950223445892,
"rewards/margins": 0.31011396646499634,
"rewards/rejected": -0.40368348360061646,
"step": 1490
},
{
"epoch": 0.8,
"learning_rate": 5.852214394262515e-07,
"logits/chosen": -1.2475618124008179,
"logits/rejected": -0.4017241895198822,
"logps/chosen": -602.7681884765625,
"logps/rejected": -1880.391357421875,
"loss": 0.1404,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.045240744948387146,
"rewards/margins": 0.4169200360774994,
"rewards/rejected": -0.46216076612472534,
"step": 1500
},
{
"epoch": 0.81,
"learning_rate": 5.556222358149191e-07,
"logits/chosen": -1.4531586170196533,
"logits/rejected": -0.3959062695503235,
"logps/chosen": -570.5740966796875,
"logps/rejected": -1546.8111572265625,
"loss": 0.1931,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.03957132622599602,
"rewards/margins": 0.31827348470687866,
"rewards/rejected": -0.3578447699546814,
"step": 1510
},
{
"epoch": 0.81,
"learning_rate": 5.266973081622992e-07,
"logits/chosen": -1.5894256830215454,
"logits/rejected": -0.623005747795105,
"logps/chosen": -632.9057006835938,
"logps/rejected": -1672.952392578125,
"loss": 0.1696,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.04810899868607521,
"rewards/margins": 0.34946009516716003,
"rewards/rejected": -0.39756911993026733,
"step": 1520
},
{
"epoch": 0.82,
"learning_rate": 4.984566871237942e-07,
"logits/chosen": -1.3502144813537598,
"logits/rejected": -0.1807008981704712,
"logps/chosen": -697.3865356445312,
"logps/rejected": -1577.202392578125,
"loss": 0.2037,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.07229883968830109,
"rewards/margins": 0.25227028131484985,
"rewards/rejected": -0.32456910610198975,
"step": 1530
},
{
"epoch": 0.82,
"learning_rate": 4.709101660493251e-07,
"logits/chosen": -1.3591182231903076,
"logits/rejected": 0.20267681777477264,
"logps/chosen": -638.6739501953125,
"logps/rejected": -1786.9075927734375,
"loss": 0.1373,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.047617390751838684,
"rewards/margins": 0.37514036893844604,
"rewards/rejected": -0.42275771498680115,
"step": 1540
},
{
"epoch": 0.83,
"learning_rate": 4.440672975871743e-07,
"logits/chosen": -1.493379831314087,
"logits/rejected": -0.6154463887214661,
"logps/chosen": -659.9649658203125,
"logps/rejected": -1666.797607421875,
"loss": 0.1751,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.07380831241607666,
"rewards/margins": 0.35080844163894653,
"rewards/rejected": -0.4246167242527008,
"step": 1550
},
{
"epoch": 0.83,
"learning_rate": 4.1793739037129134e-07,
"logits/chosen": -1.4797322750091553,
"logits/rejected": -0.10699132829904556,
"logps/chosen": -754.7605590820312,
"logps/rejected": -1794.226318359375,
"loss": 0.2233,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.12164326012134552,
"rewards/margins": 0.3239572048187256,
"rewards/rejected": -0.4456004500389099,
"step": 1560
},
{
"epoch": 0.84,
"learning_rate": 3.9252950579322405e-07,
"logits/chosen": -1.5567706823349,
"logits/rejected": -0.7062613368034363,
"logps/chosen": -536.9071044921875,
"logps/rejected": -1462.166015625,
"loss": 0.1451,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.053076934069395065,
"rewards/margins": 0.29014062881469727,
"rewards/rejected": -0.3432175815105438,
"step": 1570
},
{
"epoch": 0.84,
"learning_rate": 3.6785245485978864e-07,
"logits/chosen": -1.2004096508026123,
"logits/rejected": 0.3028511106967926,
"logps/chosen": -590.0845947265625,
"logps/rejected": -1635.791259765625,
"loss": 0.1189,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.05340021103620529,
"rewards/margins": 0.3109533488750458,
"rewards/rejected": -0.36435359716415405,
"step": 1580
},
{
"epoch": 0.85,
"learning_rate": 3.43914795137566e-07,
"logits/chosen": -1.3777801990509033,
"logits/rejected": -0.9794471859931946,
"logps/chosen": -539.28125,
"logps/rejected": -1819.6597900390625,
"loss": 0.1304,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.03970905393362045,
"rewards/margins": 0.4206715524196625,
"rewards/rejected": -0.4603806436061859,
"step": 1590
},
{
"epoch": 0.85,
"learning_rate": 3.207248277852901e-07,
"logits/chosen": -1.5900253057479858,
"logits/rejected": -0.42086416482925415,
"logps/chosen": -595.1795043945312,
"logps/rejected": -1580.691650390625,
"loss": 0.1431,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.05671688914299011,
"rewards/margins": 0.2946508824825287,
"rewards/rejected": -0.3513677716255188,
"step": 1600
},
{
"epoch": 0.86,
"learning_rate": 2.9829059467515074e-07,
"logits/chosen": -1.5492194890975952,
"logits/rejected": -0.7254107594490051,
"logps/chosen": -581.7901000976562,
"logps/rejected": -1662.1865234375,
"loss": 0.1538,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.062152404338121414,
"rewards/margins": 0.3534555435180664,
"rewards/rejected": -0.41560792922973633,
"step": 1610
},
{
"epoch": 0.86,
"learning_rate": 2.766198756040153e-07,
"logits/chosen": -1.5322610139846802,
"logits/rejected": -0.7641068696975708,
"logps/chosen": -629.3922119140625,
"logps/rejected": -1690.0902099609375,
"loss": 0.1418,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.07075804471969604,
"rewards/margins": 0.3472130596637726,
"rewards/rejected": -0.417971134185791,
"step": 1620
},
{
"epoch": 0.87,
"learning_rate": 2.5572018559553155e-07,
"logits/chosen": -1.4690418243408203,
"logits/rejected": 0.6738583445549011,
"logps/chosen": -743.3259887695312,
"logps/rejected": -1512.999267578125,
"loss": 0.1463,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.09778312593698502,
"rewards/margins": 0.22856023907661438,
"rewards/rejected": -0.326343297958374,
"step": 1630
},
{
"epoch": 0.87,
"learning_rate": 2.3559877229404864e-07,
"logits/chosen": -1.1898796558380127,
"logits/rejected": -0.6296781301498413,
"logps/chosen": -518.149658203125,
"logps/rejected": -1747.592041015625,
"loss": 0.1682,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.04969874396920204,
"rewards/margins": 0.3809133768081665,
"rewards/rejected": -0.43061214685440063,
"step": 1640
},
{
"epoch": 0.88,
"learning_rate": 2.1626261345126576e-07,
"logits/chosen": -1.4317132234573364,
"logits/rejected": -0.8821698427200317,
"logps/chosen": -600.8050537109375,
"logps/rejected": -1730.260986328125,
"loss": 0.1383,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.05417864769697189,
"rewards/margins": 0.30406099557876587,
"rewards/rejected": -0.35823965072631836,
"step": 1650
},
{
"epoch": 0.89,
"learning_rate": 1.9771841450646505e-07,
"logits/chosen": -1.2985970973968506,
"logits/rejected": -0.2940817177295685,
"logps/chosen": -608.9534912109375,
"logps/rejected": -1598.25341796875,
"loss": 0.136,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.019482949748635292,
"rewards/margins": 0.32808607816696167,
"rewards/rejected": -0.3475690186023712,
"step": 1660
},
{
"epoch": 0.89,
"learning_rate": 1.7997260626118758e-07,
"logits/chosen": -1.2966607809066772,
"logits/rejected": -0.3843843936920166,
"logps/chosen": -576.9232788085938,
"logps/rejected": -1571.2894287109375,
"loss": 0.146,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.06475577503442764,
"rewards/margins": 0.31269264221191406,
"rewards/rejected": -0.3774484097957611,
"step": 1670
},
{
"epoch": 0.9,
"learning_rate": 1.6303134264914365e-07,
"logits/chosen": -1.3799948692321777,
"logits/rejected": -0.3981359302997589,
"logps/chosen": -604.2813720703125,
"logps/rejected": -1830.979248046875,
"loss": 0.1423,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.0683603435754776,
"rewards/margins": 0.379177987575531,
"rewards/rejected": -0.4475383162498474,
"step": 1680
},
{
"epoch": 0.9,
"learning_rate": 1.469004986021355e-07,
"logits/chosen": -1.4023206233978271,
"logits/rejected": -0.3547496497631073,
"logps/chosen": -787.7603149414062,
"logps/rejected": -1988.9703369140625,
"loss": 0.1179,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.04920545220375061,
"rewards/margins": 0.41018810868263245,
"rewards/rejected": -0.45939356088638306,
"step": 1690
},
{
"epoch": 0.91,
"learning_rate": 1.315856680127367e-07,
"logits/chosen": -1.5553520917892456,
"logits/rejected": -0.3330727219581604,
"logps/chosen": -664.6277465820312,
"logps/rejected": -1760.999755859375,
"loss": 0.1363,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.040573738515377045,
"rewards/margins": 0.36900395154953003,
"rewards/rejected": -0.4095776677131653,
"step": 1700
},
{
"epoch": 0.91,
"learning_rate": 1.1709216179442817e-07,
"logits/chosen": -1.4014623165130615,
"logits/rejected": -0.6149917840957642,
"logps/chosen": -655.154541015625,
"logps/rejected": -1673.4769287109375,
"loss": 0.1396,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.08275039494037628,
"rewards/margins": 0.34146302938461304,
"rewards/rejected": -0.4242134690284729,
"step": 1710
},
{
"epoch": 0.92,
"learning_rate": 1.0342500603986421e-07,
"logits/chosen": -1.4432154893875122,
"logits/rejected": -0.2376745045185089,
"logps/chosen": -578.2418823242188,
"logps/rejected": -1667.619384765625,
"loss": 0.122,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.06366704404354095,
"rewards/margins": 0.3130820393562317,
"rewards/rejected": -0.37674909830093384,
"step": 1720
},
{
"epoch": 0.92,
"learning_rate": 9.058894027791643e-08,
"logits/chosen": -1.3630592823028564,
"logits/rejected": -0.21227815747261047,
"logps/chosen": -532.6649169921875,
"logps/rejected": -1589.3013916015625,
"loss": 0.1758,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.03975344076752663,
"rewards/margins": 0.3338061273097992,
"rewards/rejected": -0.37355953454971313,
"step": 1730
},
{
"epoch": 0.93,
"learning_rate": 7.858841583008592e-08,
"logits/chosen": -1.54172945022583,
"logits/rejected": -0.21348455548286438,
"logps/chosen": -780.7069702148438,
"logps/rejected": -1701.29296875,
"loss": 0.186,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.08455196768045425,
"rewards/margins": 0.28723907470703125,
"rewards/rejected": -0.3717910051345825,
"step": 1740
},
{
"epoch": 0.93,
"learning_rate": 6.742759426686313e-08,
"logits/chosen": -1.3796813488006592,
"logits/rejected": 0.4848087430000305,
"logps/chosen": -621.2150268554688,
"logps/rejected": -1688.8896484375,
"loss": 0.117,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.057764243334531784,
"rewards/margins": 0.32786962389945984,
"rewards/rejected": -0.3856338858604431,
"step": 1750
},
{
"epoch": 0.94,
"learning_rate": 5.7110345964571104e-08,
"logits/chosen": -1.475338101387024,
"logits/rejected": -0.41871923208236694,
"logps/chosen": -615.502685546875,
"logps/rejected": -1646.4254150390625,
"loss": 0.1425,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.07283326983451843,
"rewards/margins": 0.3077198565006256,
"rewards/rejected": -0.38055309653282166,
"step": 1760
},
{
"epoch": 0.94,
"learning_rate": 4.764024876318357e-08,
"logits/chosen": -1.38540518283844,
"logits/rejected": -0.8147419095039368,
"logps/chosen": -595.4188232421875,
"logps/rejected": -1683.2845458984375,
"loss": 0.1805,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.05248613283038139,
"rewards/margins": 0.3240980803966522,
"rewards/rejected": -0.3765842318534851,
"step": 1770
},
{
"epoch": 0.95,
"learning_rate": 3.902058672559633e-08,
"logits/chosen": -1.3075553178787231,
"logits/rejected": -0.05613366514444351,
"logps/chosen": -711.91357421875,
"logps/rejected": -1785.665283203125,
"loss": 0.1703,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.08939728140830994,
"rewards/margins": 0.3216257691383362,
"rewards/rejected": -0.4110230505466461,
"step": 1780
},
{
"epoch": 0.95,
"learning_rate": 3.125434899876933e-08,
"logits/chosen": -1.4797890186309814,
"logits/rejected": 0.0037591219879686832,
"logps/chosen": -605.4852905273438,
"logps/rejected": -1533.518310546875,
"loss": 0.1475,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.05776011943817139,
"rewards/margins": 0.2979200482368469,
"rewards/rejected": -0.3556801676750183,
"step": 1790
},
{
"epoch": 0.96,
"learning_rate": 2.4344228777145873e-08,
"logits/chosen": -1.4468438625335693,
"logits/rejected": -0.6082602739334106,
"logps/chosen": -594.7930908203125,
"logps/rejected": -1680.8695068359375,
"loss": 0.1701,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.04611823707818985,
"rewards/margins": 0.32163435220718384,
"rewards/rejected": -0.3677525818347931,
"step": 1800
},
{
"epoch": 0.97,
"learning_rate": 1.829262236869772e-08,
"logits/chosen": -1.326982855796814,
"logits/rejected": -0.7402676343917847,
"logps/chosen": -599.7637939453125,
"logps/rejected": -1530.202392578125,
"loss": 0.1555,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.056166619062423706,
"rewards/margins": 0.27239522337913513,
"rewards/rejected": -0.32856184244155884,
"step": 1810
},
{
"epoch": 0.97,
"learning_rate": 1.3101628363929586e-08,
"logits/chosen": -1.4013383388519287,
"logits/rejected": -0.5074108839035034,
"logps/chosen": -622.0450439453125,
"logps/rejected": -1639.010498046875,
"loss": 0.1346,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.10727888345718384,
"rewards/margins": 0.3056369721889496,
"rewards/rejected": -0.4129158854484558,
"step": 1820
},
{
"epoch": 0.98,
"learning_rate": 8.773046908123195e-09,
"logits/chosen": -1.490671992301941,
"logits/rejected": -0.3455984890460968,
"logps/chosen": -638.0214233398438,
"logps/rejected": -1695.2183837890625,
"loss": 0.175,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.05133664608001709,
"rewards/margins": 0.35318809747695923,
"rewards/rejected": -0.4045247435569763,
"step": 1830
},
{
"epoch": 0.98,
"learning_rate": 5.308379077080817e-09,
"logits/chosen": -1.2905375957489014,
"logits/rejected": -0.5379937887191772,
"logps/chosen": -673.1795654296875,
"logps/rejected": -1559.984619140625,
"loss": 0.1556,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.07675811648368835,
"rewards/margins": 0.2697897255420685,
"rewards/rejected": -0.34654784202575684,
"step": 1840
},
{
"epoch": 0.99,
"learning_rate": 2.7088263565760996e-09,
"logits/chosen": -1.4194942712783813,
"logits/rejected": -0.6771696209907532,
"logps/chosen": -676.0364379882812,
"logps/rejected": -1739.635498046875,
"loss": 0.1438,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.07915320247411728,
"rewards/margins": 0.31377238035202026,
"rewards/rejected": -0.39292556047439575,
"step": 1850
},
{
"epoch": 0.99,
"learning_rate": 9.752902257023633e-10,
"logits/chosen": -1.5524300336837769,
"logits/rejected": -0.5318177342414856,
"logps/chosen": -544.7901611328125,
"logps/rejected": -1705.066162109375,
"loss": 0.1563,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.04058445617556572,
"rewards/margins": 0.38282185792922974,
"rewards/rejected": -0.4234062731266022,
"step": 1860
},
{
"epoch": 1.0,
"learning_rate": 1.083718442532189e-10,
"logits/chosen": -1.2659026384353638,
"logits/rejected": -0.41987451910972595,
"logps/chosen": -756.4010009765625,
"logps/rejected": -1763.924560546875,
"loss": 0.1601,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.0818880945444107,
"rewards/margins": 0.3468380570411682,
"rewards/rejected": -0.4287261962890625,
"step": 1870
},
{
"epoch": 1.0,
"step": 1875,
"total_flos": 0.0,
"train_loss": 0.18001604646046956,
"train_runtime": 15472.4916,
"train_samples_per_second": 0.969,
"train_steps_per_second": 0.121
}
],
"logging_steps": 10,
"max_steps": 1875,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 100,
"total_flos": 0.0,
"train_batch_size": 2,
"trial_name": null,
"trial_params": null
}