zephyr-7b-gpo-v3-1-i2 / trainer_state.json
lole25's picture
Model save
83e9d28 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": -1.5569357872009277,
"logits/rejected": -1.1471056938171387,
"logps/chosen": -378.52911376953125,
"logps/rejected": -854.833984375,
"loss": 0.2347,
"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": -1.6401379108428955,
"logits/rejected": -1.022291898727417,
"logps/chosen": -509.1824645996094,
"logps/rejected": -899.5430908203125,
"loss": 0.2181,
"rewards/accuracies": 0.4166666567325592,
"rewards/chosen": -0.00027799332747235894,
"rewards/margins": -5.683075141860172e-05,
"rewards/rejected": -0.00022116259788163006,
"step": 10
},
{
"epoch": 0.01,
"learning_rate": 5.319148936170213e-07,
"logits/chosen": -1.6371427774429321,
"logits/rejected": -1.1658873558044434,
"logps/chosen": -452.5133361816406,
"logps/rejected": -771.5206298828125,
"loss": 0.1959,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0010695771779865026,
"rewards/margins": 0.0016313500236719847,
"rewards/rejected": -0.0005617729621008039,
"step": 20
},
{
"epoch": 0.02,
"learning_rate": 7.97872340425532e-07,
"logits/chosen": -1.614117980003357,
"logits/rejected": -0.9896963834762573,
"logps/chosen": -505.353271484375,
"logps/rejected": -859.1238403320312,
"loss": 0.1969,
"rewards/accuracies": 0.875,
"rewards/chosen": 0.005224255379289389,
"rewards/margins": 0.010899050161242485,
"rewards/rejected": -0.005674795713275671,
"step": 30
},
{
"epoch": 0.02,
"learning_rate": 1.0638297872340427e-06,
"logits/chosen": -1.5090134143829346,
"logits/rejected": -1.2780263423919678,
"logps/chosen": -342.17144775390625,
"logps/rejected": -809.4546508789062,
"loss": 0.2029,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.01147081796079874,
"rewards/margins": 0.030131075531244278,
"rewards/rejected": -0.018660254776477814,
"step": 40
},
{
"epoch": 0.03,
"learning_rate": 1.3297872340425533e-06,
"logits/chosen": -1.7051512002944946,
"logits/rejected": -0.8763967752456665,
"logps/chosen": -585.2564697265625,
"logps/rejected": -909.919921875,
"loss": 0.1786,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.01587619259953499,
"rewards/margins": 0.05883794277906418,
"rewards/rejected": -0.04296175390481949,
"step": 50
},
{
"epoch": 0.03,
"learning_rate": 1.595744680851064e-06,
"logits/chosen": -1.5889899730682373,
"logits/rejected": -1.1295216083526611,
"logps/chosen": -424.6382751464844,
"logps/rejected": -895.8190307617188,
"loss": 0.1636,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.009497600607573986,
"rewards/margins": 0.08446536958217621,
"rewards/rejected": -0.07496776431798935,
"step": 60
},
{
"epoch": 0.04,
"learning_rate": 1.8617021276595745e-06,
"logits/chosen": -1.6142644882202148,
"logits/rejected": -1.1362119913101196,
"logps/chosen": -428.67236328125,
"logps/rejected": -934.1236572265625,
"loss": 0.1257,
"rewards/accuracies": 0.875,
"rewards/chosen": 0.003880441188812256,
"rewards/margins": 0.14223161339759827,
"rewards/rejected": -0.138351172208786,
"step": 70
},
{
"epoch": 0.04,
"learning_rate": 2.1276595744680853e-06,
"logits/chosen": -1.5118087530136108,
"logits/rejected": -1.1322299242019653,
"logps/chosen": -615.1442260742188,
"logps/rejected": -1068.8787841796875,
"loss": 0.1128,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.06628059595823288,
"rewards/margins": 0.14910638332366943,
"rewards/rejected": -0.2153869867324829,
"step": 80
},
{
"epoch": 0.05,
"learning_rate": 2.393617021276596e-06,
"logits/chosen": -1.748167634010315,
"logits/rejected": -1.1786916255950928,
"logps/chosen": -513.2774658203125,
"logps/rejected": -1070.9239501953125,
"loss": 0.113,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.058650992810726166,
"rewards/margins": 0.2291336953639984,
"rewards/rejected": -0.28778475522994995,
"step": 90
},
{
"epoch": 0.05,
"learning_rate": 2.6595744680851065e-06,
"logits/chosen": -1.7060625553131104,
"logits/rejected": -1.0770542621612549,
"logps/chosen": -546.6444091796875,
"logps/rejected": -1195.038818359375,
"loss": 0.1108,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.14496687054634094,
"rewards/margins": 0.21418336033821106,
"rewards/rejected": -0.359150230884552,
"step": 100
},
{
"epoch": 0.06,
"learning_rate": 2.9255319148936174e-06,
"logits/chosen": -1.6031602621078491,
"logits/rejected": -1.0294073820114136,
"logps/chosen": -659.802490234375,
"logps/rejected": -1333.0833740234375,
"loss": 0.0903,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.17693635821342468,
"rewards/margins": 0.28644508123397827,
"rewards/rejected": -0.4633813798427582,
"step": 110
},
{
"epoch": 0.06,
"learning_rate": 3.191489361702128e-06,
"logits/chosen": -1.7648108005523682,
"logits/rejected": -1.4070419073104858,
"logps/chosen": -478.43487548828125,
"logps/rejected": -950.3967895507812,
"loss": 0.1227,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.1012553796172142,
"rewards/margins": 0.1941373646259308,
"rewards/rejected": -0.2953927218914032,
"step": 120
},
{
"epoch": 0.07,
"learning_rate": 3.457446808510639e-06,
"logits/chosen": -1.5940220355987549,
"logits/rejected": -1.2087233066558838,
"logps/chosen": -476.9007873535156,
"logps/rejected": -1160.44580078125,
"loss": 0.0836,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.09528017044067383,
"rewards/margins": 0.253213495016098,
"rewards/rejected": -0.34849366545677185,
"step": 130
},
{
"epoch": 0.07,
"learning_rate": 3.723404255319149e-06,
"logits/chosen": -1.2940385341644287,
"logits/rejected": -0.8031653165817261,
"logps/chosen": -744.671142578125,
"logps/rejected": -1387.1114501953125,
"loss": 0.0972,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.26588675379753113,
"rewards/margins": 0.2445482313632965,
"rewards/rejected": -0.5104349851608276,
"step": 140
},
{
"epoch": 0.08,
"learning_rate": 3.98936170212766e-06,
"logits/chosen": -1.6005325317382812,
"logits/rejected": -1.2362384796142578,
"logps/chosen": -655.4281005859375,
"logps/rejected": -1130.398193359375,
"loss": 0.0971,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.196586474776268,
"rewards/margins": 0.21370963752269745,
"rewards/rejected": -0.41029611229896545,
"step": 150
},
{
"epoch": 0.09,
"learning_rate": 4.255319148936171e-06,
"logits/chosen": -1.7509543895721436,
"logits/rejected": -1.1238092184066772,
"logps/chosen": -569.5955810546875,
"logps/rejected": -1124.465576171875,
"loss": 0.102,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.12171213328838348,
"rewards/margins": 0.21310141682624817,
"rewards/rejected": -0.33481353521347046,
"step": 160
},
{
"epoch": 0.09,
"learning_rate": 4.521276595744681e-06,
"logits/chosen": -1.7424652576446533,
"logits/rejected": -1.2808568477630615,
"logps/chosen": -563.4480590820312,
"logps/rejected": -1213.89453125,
"loss": 0.0727,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.13533855974674225,
"rewards/margins": 0.23045413196086884,
"rewards/rejected": -0.3657926917076111,
"step": 170
},
{
"epoch": 0.1,
"learning_rate": 4.787234042553192e-06,
"logits/chosen": -1.6149215698242188,
"logits/rejected": -1.3036892414093018,
"logps/chosen": -604.5150146484375,
"logps/rejected": -1181.5762939453125,
"loss": 0.0986,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.1363724172115326,
"rewards/margins": 0.24034619331359863,
"rewards/rejected": -0.3767186105251312,
"step": 180
},
{
"epoch": 0.1,
"learning_rate": 4.999982660399688e-06,
"logits/chosen": -1.6364609003067017,
"logits/rejected": -0.9650686979293823,
"logps/chosen": -680.2525634765625,
"logps/rejected": -1388.20068359375,
"loss": 0.0492,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.17130868136882782,
"rewards/margins": 0.3032084107398987,
"rewards/rejected": -0.4745171070098877,
"step": 190
},
{
"epoch": 0.11,
"learning_rate": 4.99937579964398e-06,
"logits/chosen": -1.5488389730453491,
"logits/rejected": -1.0396603345870972,
"logps/chosen": -626.0272216796875,
"logps/rejected": -1298.0181884765625,
"loss": 0.0843,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.1749267876148224,
"rewards/margins": 0.25047367811203003,
"rewards/rejected": -0.42540043592453003,
"step": 200
},
{
"epoch": 0.11,
"learning_rate": 4.9979021993870645e-06,
"logits/chosen": -1.5960906744003296,
"logits/rejected": -1.0231754779815674,
"logps/chosen": -657.3213500976562,
"logps/rejected": -1247.23974609375,
"loss": 0.0739,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.20409782230854034,
"rewards/margins": 0.2475307732820511,
"rewards/rejected": -0.45162859559059143,
"step": 210
},
{
"epoch": 0.12,
"learning_rate": 4.995562370647553e-06,
"logits/chosen": -1.7341878414154053,
"logits/rejected": -1.1983798742294312,
"logps/chosen": -753.7733154296875,
"logps/rejected": -1359.67626953125,
"loss": 0.071,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.2541276514530182,
"rewards/margins": 0.2833201289176941,
"rewards/rejected": -0.5374478101730347,
"step": 220
},
{
"epoch": 0.12,
"learning_rate": 4.992357124836838e-06,
"logits/chosen": -1.4590599536895752,
"logits/rejected": -0.934046745300293,
"logps/chosen": -695.98046875,
"logps/rejected": -1367.89453125,
"loss": 0.0944,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.2104768306016922,
"rewards/margins": 0.27617698907852173,
"rewards/rejected": -0.48665380477905273,
"step": 230
},
{
"epoch": 0.13,
"learning_rate": 4.9882875734777044e-06,
"logits/chosen": -1.640160322189331,
"logits/rejected": -1.1560494899749756,
"logps/chosen": -578.0391845703125,
"logps/rejected": -1119.09814453125,
"loss": 0.1005,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.16664345562458038,
"rewards/margins": 0.2392512559890747,
"rewards/rejected": -0.4058946967124939,
"step": 240
},
{
"epoch": 0.13,
"learning_rate": 4.983355127818882e-06,
"logits/chosen": -1.4978324174880981,
"logits/rejected": -1.134303092956543,
"logps/chosen": -570.8470458984375,
"logps/rejected": -1243.4251708984375,
"loss": 0.0959,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.1637270152568817,
"rewards/margins": 0.2728789150714874,
"rewards/rejected": -0.43660593032836914,
"step": 250
},
{
"epoch": 0.14,
"learning_rate": 4.977561498345639e-06,
"logits/chosen": -1.788405179977417,
"logits/rejected": -1.0717966556549072,
"logps/chosen": -580.3256225585938,
"logps/rejected": -1118.1702880859375,
"loss": 0.0845,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.1016082763671875,
"rewards/margins": 0.24942278861999512,
"rewards/rejected": -0.3510310649871826,
"step": 260
},
{
"epoch": 0.14,
"learning_rate": 4.970908694186624e-06,
"logits/chosen": -1.5624245405197144,
"logits/rejected": -0.8945890665054321,
"logps/chosen": -625.0382080078125,
"logps/rejected": -1272.6126708984375,
"loss": 0.0823,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.15467065572738647,
"rewards/margins": 0.27355173230171204,
"rewards/rejected": -0.4282223582267761,
"step": 270
},
{
"epoch": 0.15,
"learning_rate": 4.9633990224171305e-06,
"logits/chosen": -1.5876206159591675,
"logits/rejected": -1.138290286064148,
"logps/chosen": -727.9985961914062,
"logps/rejected": -1291.621826171875,
"loss": 0.0646,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.23997028172016144,
"rewards/margins": 0.23828229308128357,
"rewards/rejected": -0.4782525897026062,
"step": 280
},
{
"epoch": 0.15,
"learning_rate": 4.955035087259046e-06,
"logits/chosen": -1.7230398654937744,
"logits/rejected": -1.4215495586395264,
"logps/chosen": -605.0308837890625,
"logps/rejected": -1185.82861328125,
"loss": 0.0736,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.17933842539787292,
"rewards/margins": 0.22512169182300568,
"rewards/rejected": -0.4044601321220398,
"step": 290
},
{
"epoch": 0.16,
"learning_rate": 4.945819789177756e-06,
"logits/chosen": -1.5644534826278687,
"logits/rejected": -1.0705270767211914,
"logps/chosen": -733.014404296875,
"logps/rejected": -1408.350830078125,
"loss": 0.0755,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.2445208728313446,
"rewards/margins": 0.2835485339164734,
"rewards/rejected": -0.5280693769454956,
"step": 300
},
{
"epoch": 0.17,
"learning_rate": 4.935756323876306e-06,
"logits/chosen": -1.6218621730804443,
"logits/rejected": -0.9748885035514832,
"logps/chosen": -564.2091064453125,
"logps/rejected": -1149.895263671875,
"loss": 0.0768,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.1630648970603943,
"rewards/margins": 0.2708643972873688,
"rewards/rejected": -0.43392929434776306,
"step": 310
},
{
"epoch": 0.17,
"learning_rate": 4.924848181187199e-06,
"logits/chosen": -1.7458088397979736,
"logits/rejected": -1.3406083583831787,
"logps/chosen": -561.4813232421875,
"logps/rejected": -1230.68017578125,
"loss": 0.0746,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.16397038102149963,
"rewards/margins": 0.27870216965675354,
"rewards/rejected": -0.44267258048057556,
"step": 320
},
{
"epoch": 0.18,
"learning_rate": 4.913099143862173e-06,
"logits/chosen": -1.6369669437408447,
"logits/rejected": -1.1729207038879395,
"logps/chosen": -627.0284423828125,
"logps/rejected": -1398.7330322265625,
"loss": 0.0666,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.18270769715309143,
"rewards/margins": 0.31160640716552734,
"rewards/rejected": -0.49431413412094116,
"step": 330
},
{
"epoch": 0.18,
"learning_rate": 4.900513286260416e-06,
"logits/chosen": -1.6523929834365845,
"logits/rejected": -1.0893100500106812,
"logps/chosen": -597.8865966796875,
"logps/rejected": -1325.775390625,
"loss": 0.0936,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.17563173174858093,
"rewards/margins": 0.269579142332077,
"rewards/rejected": -0.44521084427833557,
"step": 340
},
{
"epoch": 0.19,
"learning_rate": 4.887094972935645e-06,
"logits/chosen": -1.7033765316009521,
"logits/rejected": -1.280895709991455,
"logps/chosen": -710.4462280273438,
"logps/rejected": -1266.050048828125,
"loss": 0.1008,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.2460685521364212,
"rewards/margins": 0.2334461659193039,
"rewards/rejected": -0.4795146882534027,
"step": 350
},
{
"epoch": 0.19,
"learning_rate": 4.87284885712256e-06,
"logits/chosen": -1.630501389503479,
"logits/rejected": -0.9762773513793945,
"logps/chosen": -765.29248046875,
"logps/rejected": -1381.234619140625,
"loss": 0.0966,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.2656300961971283,
"rewards/margins": 0.25425031781196594,
"rewards/rejected": -0.5198804140090942,
"step": 360
},
{
"epoch": 0.2,
"learning_rate": 4.857779879123181e-06,
"logits/chosen": -1.5260308980941772,
"logits/rejected": -1.0575653314590454,
"logps/chosen": -580.313232421875,
"logps/rejected": -1192.9915771484375,
"loss": 0.0654,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.16813692450523376,
"rewards/margins": 0.25473329424858093,
"rewards/rejected": -0.4228702485561371,
"step": 370
},
{
"epoch": 0.2,
"learning_rate": 4.841893264593643e-06,
"logits/chosen": -1.7011750936508179,
"logits/rejected": -1.1821006536483765,
"logps/chosen": -703.5277099609375,
"logps/rejected": -1323.509033203125,
"loss": 0.0809,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.2287747859954834,
"rewards/margins": 0.27832740545272827,
"rewards/rejected": -0.5071021914482117,
"step": 380
},
{
"epoch": 0.21,
"learning_rate": 4.825194522732023e-06,
"logits/chosen": -1.4276914596557617,
"logits/rejected": -0.9805229306221008,
"logps/chosen": -592.0125122070312,
"logps/rejected": -1193.139892578125,
"loss": 0.064,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.18544098734855652,
"rewards/margins": 0.2668260931968689,
"rewards/rejected": -0.45226702094078064,
"step": 390
},
{
"epoch": 0.21,
"learning_rate": 4.807689444367853e-06,
"logits/chosen": -1.5613304376602173,
"logits/rejected": -0.8922286033630371,
"logps/chosen": -665.95556640625,
"logps/rejected": -1267.384765625,
"loss": 0.0621,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.19959279894828796,
"rewards/margins": 0.29206138849258423,
"rewards/rejected": -0.4916541576385498,
"step": 400
},
{
"epoch": 0.22,
"learning_rate": 4.78938409995396e-06,
"logits/chosen": -1.9131767749786377,
"logits/rejected": -1.3205538988113403,
"logps/chosen": -605.758056640625,
"logps/rejected": -1235.4910888671875,
"loss": 0.072,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.17551980912685394,
"rewards/margins": 0.2815834879875183,
"rewards/rejected": -0.45710331201553345,
"step": 410
},
{
"epoch": 0.22,
"learning_rate": 4.770284837461342e-06,
"logits/chosen": -1.8387680053710938,
"logits/rejected": -1.0919946432113647,
"logps/chosen": -603.5829467773438,
"logps/rejected": -1248.807861328125,
"loss": 0.0806,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.17225244641304016,
"rewards/margins": 0.29293113946914673,
"rewards/rejected": -0.4651835560798645,
"step": 420
},
{
"epoch": 0.23,
"learning_rate": 4.7503982801778015e-06,
"logits/chosen": -1.942421317100525,
"logits/rejected": -0.9755213856697083,
"logps/chosen": -629.6397094726562,
"logps/rejected": -1305.019775390625,
"loss": 0.0636,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.17476651072502136,
"rewards/margins": 0.336466521024704,
"rewards/rejected": -0.5112330317497253,
"step": 430
},
{
"epoch": 0.23,
"learning_rate": 4.729731324411104e-06,
"logits/chosen": -1.6092408895492554,
"logits/rejected": -1.0839555263519287,
"logps/chosen": -642.392333984375,
"logps/rejected": -1305.717041015625,
"loss": 0.0773,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.23566822707653046,
"rewards/margins": 0.27665117383003235,
"rewards/rejected": -0.5123193860054016,
"step": 440
},
{
"epoch": 0.24,
"learning_rate": 4.7082911370974645e-06,
"logits/chosen": -1.7550840377807617,
"logits/rejected": -1.3057055473327637,
"logps/chosen": -744.6429443359375,
"logps/rejected": -1384.8328857421875,
"loss": 0.0808,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.25410711765289307,
"rewards/margins": 0.2785530686378479,
"rewards/rejected": -0.5326601266860962,
"step": 450
},
{
"epoch": 0.25,
"learning_rate": 4.68608515331618e-06,
"logits/chosen": -1.651061773300171,
"logits/rejected": -1.0968434810638428,
"logps/chosen": -696.5147705078125,
"logps/rejected": -1150.5166015625,
"loss": 0.0887,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.24368226528167725,
"rewards/margins": 0.20816107094287872,
"rewards/rejected": -0.4518433213233948,
"step": 460
},
{
"epoch": 0.25,
"learning_rate": 4.663121073711269e-06,
"logits/chosen": -1.8011071681976318,
"logits/rejected": -1.0791230201721191,
"logps/chosen": -579.2171020507812,
"logps/rejected": -1126.716552734375,
"loss": 0.091,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.13108104467391968,
"rewards/margins": 0.2670387625694275,
"rewards/rejected": -0.3981197774410248,
"step": 470
},
{
"epoch": 0.26,
"learning_rate": 4.63940686182103e-06,
"logits/chosen": -1.5227606296539307,
"logits/rejected": -0.9799526929855347,
"logps/chosen": -755.5355834960938,
"logps/rejected": -1411.053955078125,
"loss": 0.1072,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.2255210429430008,
"rewards/margins": 0.28304630517959595,
"rewards/rejected": -0.5085673332214355,
"step": 480
},
{
"epoch": 0.26,
"learning_rate": 4.614950741316425e-06,
"logits/chosen": -1.7349796295166016,
"logits/rejected": -0.942231297492981,
"logps/chosen": -764.5320434570312,
"logps/rejected": -1312.6400146484375,
"loss": 0.0975,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.24115994572639465,
"rewards/margins": 0.24554534256458282,
"rewards/rejected": -0.4867052435874939,
"step": 490
},
{
"epoch": 0.27,
"learning_rate": 4.589761193149254e-06,
"logits/chosen": -1.5131117105484009,
"logits/rejected": -1.1059508323669434,
"logps/chosen": -780.0518798828125,
"logps/rejected": -1328.439208984375,
"loss": 0.0985,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3147870600223541,
"rewards/margins": 0.2081507444381714,
"rewards/rejected": -0.5229377746582031,
"step": 500
},
{
"epoch": 0.27,
"learning_rate": 4.563846952611112e-06,
"logits/chosen": -1.5837140083312988,
"logits/rejected": -0.9551746249198914,
"logps/chosen": -628.8701782226562,
"logps/rejected": -1158.2474365234375,
"loss": 0.0979,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.16708222031593323,
"rewards/margins": 0.24669913947582245,
"rewards/rejected": -0.41378140449523926,
"step": 510
},
{
"epoch": 0.28,
"learning_rate": 4.537217006304141e-06,
"logits/chosen": -1.6675571203231812,
"logits/rejected": -0.9593210220336914,
"logps/chosen": -682.0783081054688,
"logps/rejected": -1288.447509765625,
"loss": 0.0898,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.19303472340106964,
"rewards/margins": 0.25880172848701477,
"rewards/rejected": -0.4518364369869232,
"step": 520
},
{
"epoch": 0.28,
"learning_rate": 4.50988058902464e-06,
"logits/chosen": -1.6770139932632446,
"logits/rejected": -0.937613308429718,
"logps/chosen": -713.4398193359375,
"logps/rejected": -1368.5531005859375,
"loss": 0.0852,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.24874809384346008,
"rewards/margins": 0.315912663936615,
"rewards/rejected": -0.5646606683731079,
"step": 530
},
{
"epoch": 0.29,
"learning_rate": 4.481847180560593e-06,
"logits/chosen": -1.4941534996032715,
"logits/rejected": -0.8417922258377075,
"logps/chosen": -761.2758178710938,
"logps/rejected": -1347.4490966796875,
"loss": 0.0908,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.25566571950912476,
"rewards/margins": 0.26928967237472534,
"rewards/rejected": -0.5249553918838501,
"step": 540
},
{
"epoch": 0.29,
"learning_rate": 4.453126502404253e-06,
"logits/chosen": -1.638262391090393,
"logits/rejected": -0.953241229057312,
"logps/chosen": -680.6768188476562,
"logps/rejected": -1407.271240234375,
"loss": 0.0657,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.21453741192817688,
"rewards/margins": 0.332472026348114,
"rewards/rejected": -0.5470094084739685,
"step": 550
},
{
"epoch": 0.3,
"learning_rate": 4.423728514380892e-06,
"logits/chosen": -1.597510576248169,
"logits/rejected": -0.9470292925834656,
"logps/chosen": -748.9639892578125,
"logps/rejected": -1390.034912109375,
"loss": 0.0565,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.23396451771259308,
"rewards/margins": 0.3206896483898163,
"rewards/rejected": -0.5546542406082153,
"step": 560
},
{
"epoch": 0.3,
"learning_rate": 4.393663411194918e-06,
"logits/chosen": -1.550355076789856,
"logits/rejected": -1.0478793382644653,
"logps/chosen": -604.4364013671875,
"logps/rejected": -1344.87841796875,
"loss": 0.0637,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.2098585069179535,
"rewards/margins": 0.31462934613227844,
"rewards/rejected": -0.5244878530502319,
"step": 570
},
{
"epoch": 0.31,
"learning_rate": 4.362941618894523e-06,
"logits/chosen": -1.533578634262085,
"logits/rejected": -1.0312515497207642,
"logps/chosen": -538.365478515625,
"logps/rejected": -1111.39501953125,
"loss": 0.0999,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.18333378434181213,
"rewards/margins": 0.26067209243774414,
"rewards/rejected": -0.4440058767795563,
"step": 580
},
{
"epoch": 0.31,
"learning_rate": 4.331573791256116e-06,
"logits/chosen": -1.4897745847702026,
"logits/rejected": -1.0129917860031128,
"logps/chosen": -689.7890014648438,
"logps/rejected": -1376.696044921875,
"loss": 0.0581,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.2648084759712219,
"rewards/margins": 0.2891417145729065,
"rewards/rejected": -0.5539501905441284,
"step": 590
},
{
"epoch": 0.32,
"learning_rate": 4.299570806089786e-06,
"logits/chosen": -1.4845857620239258,
"logits/rejected": -0.9828746914863586,
"logps/chosen": -596.8510131835938,
"logps/rejected": -1313.0968017578125,
"loss": 0.0747,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.20095543563365936,
"rewards/margins": 0.307460755109787,
"rewards/rejected": -0.5084162354469299,
"step": 600
},
{
"epoch": 0.33,
"learning_rate": 4.266943761467057e-06,
"logits/chosen": -1.7313928604125977,
"logits/rejected": -1.3923299312591553,
"logps/chosen": -712.3792724609375,
"logps/rejected": -1243.148681640625,
"loss": 0.0924,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.24510283768177032,
"rewards/margins": 0.24176783859729767,
"rewards/rejected": -0.486870676279068,
"step": 610
},
{
"epoch": 0.33,
"learning_rate": 4.233703971872287e-06,
"logits/chosen": -1.5082043409347534,
"logits/rejected": -0.8806392550468445,
"logps/chosen": -682.8671875,
"logps/rejected": -1272.7772216796875,
"loss": 0.0522,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.18572933971881866,
"rewards/margins": 0.2751011252403259,
"rewards/rejected": -0.46083053946495056,
"step": 620
},
{
"epoch": 0.34,
"learning_rate": 4.1998629642789925e-06,
"logits/chosen": -1.6908938884735107,
"logits/rejected": -0.9967711567878723,
"logps/chosen": -673.7935180664062,
"logps/rejected": -1292.220458984375,
"loss": 0.0842,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.16402257978916168,
"rewards/margins": 0.3048457205295563,
"rewards/rejected": -0.46886831521987915,
"step": 630
},
{
"epoch": 0.34,
"learning_rate": 4.165432474152505e-06,
"logits/chosen": -1.3779823780059814,
"logits/rejected": -0.8789410591125488,
"logps/chosen": -565.2269287109375,
"logps/rejected": -1224.1700439453125,
"loss": 0.0805,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.11619114875793457,
"rewards/margins": 0.29433438181877136,
"rewards/rejected": -0.41052550077438354,
"step": 640
},
{
"epoch": 0.35,
"learning_rate": 4.130424441380308e-06,
"logits/chosen": -1.4996254444122314,
"logits/rejected": -1.0665868520736694,
"logps/chosen": -620.0261840820312,
"logps/rejected": -1192.489013671875,
"loss": 0.0758,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.1719478964805603,
"rewards/margins": 0.2514215111732483,
"rewards/rejected": -0.4233694076538086,
"step": 650
},
{
"epoch": 0.35,
"learning_rate": 4.09485100613151e-06,
"logits/chosen": -1.6758674383163452,
"logits/rejected": -1.0761094093322754,
"logps/chosen": -661.2818603515625,
"logps/rejected": -1286.9786376953125,
"loss": 0.043,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.1839846819639206,
"rewards/margins": 0.2680208086967468,
"rewards/rejected": -0.4520054757595062,
"step": 660
},
{
"epoch": 0.36,
"learning_rate": 4.058724504646834e-06,
"logits/chosen": -1.3314406871795654,
"logits/rejected": -0.9023712873458862,
"logps/chosen": -654.1249389648438,
"logps/rejected": -1329.398193359375,
"loss": 0.0722,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.1925685852766037,
"rewards/margins": 0.28446346521377563,
"rewards/rejected": -0.4770320951938629,
"step": 670
},
{
"epoch": 0.36,
"learning_rate": 4.022057464960632e-06,
"logits/chosen": -1.578636884689331,
"logits/rejected": -1.0915998220443726,
"logps/chosen": -740.0035400390625,
"logps/rejected": -1387.459228515625,
"loss": 0.0693,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.21566300094127655,
"rewards/margins": 0.3169075846672058,
"rewards/rejected": -0.5325705409049988,
"step": 680
},
{
"epoch": 0.37,
"learning_rate": 3.984862602556383e-06,
"logits/chosen": -1.5450822114944458,
"logits/rejected": -0.7816729545593262,
"logps/chosen": -626.339111328125,
"logps/rejected": -1215.27734375,
"loss": 0.0805,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.18071869015693665,
"rewards/margins": 0.2859996259212494,
"rewards/rejected": -0.46671828627586365,
"step": 690
},
{
"epoch": 0.37,
"learning_rate": 3.947152815957187e-06,
"logits/chosen": -1.3680191040039062,
"logits/rejected": -0.8944322466850281,
"logps/chosen": -557.0673828125,
"logps/rejected": -1440.092041015625,
"loss": 0.066,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.1859458088874817,
"rewards/margins": 0.374511182308197,
"rewards/rejected": -0.5604569911956787,
"step": 700
},
{
"epoch": 0.38,
"learning_rate": 3.908941182252785e-06,
"logits/chosen": -1.6122426986694336,
"logits/rejected": -0.6363096237182617,
"logps/chosen": -730.7781372070312,
"logps/rejected": -1428.550537109375,
"loss": 0.0613,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.2527967095375061,
"rewards/margins": 0.3397744297981262,
"rewards/rejected": -0.5925711393356323,
"step": 710
},
{
"epoch": 0.38,
"learning_rate": 3.8702409525646535e-06,
"logits/chosen": -1.4658266305923462,
"logits/rejected": -0.9103586077690125,
"logps/chosen": -749.474365234375,
"logps/rejected": -1346.693603515625,
"loss": 0.0691,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.3004884421825409,
"rewards/margins": 0.2709447145462036,
"rewards/rejected": -0.5714331865310669,
"step": 720
},
{
"epoch": 0.39,
"learning_rate": 3.8310655474507495e-06,
"logits/chosen": -1.651769995689392,
"logits/rejected": -0.9803410768508911,
"logps/chosen": -826.3870239257812,
"logps/rejected": -1311.737548828125,
"loss": 0.0961,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.3051601052284241,
"rewards/margins": 0.23282098770141602,
"rewards/rejected": -0.5379811525344849,
"step": 730
},
{
"epoch": 0.39,
"learning_rate": 3.7914285522515002e-06,
"logits/chosen": -1.6758012771606445,
"logits/rejected": -0.8760524988174438,
"logps/chosen": -703.2643432617188,
"logps/rejected": -1416.0147705078125,
"loss": 0.0536,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.2222570925951004,
"rewards/margins": 0.3216368556022644,
"rewards/rejected": -0.5438939332962036,
"step": 740
},
{
"epoch": 0.4,
"learning_rate": 3.751343712378639e-06,
"logits/chosen": -1.5657737255096436,
"logits/rejected": -0.6773009300231934,
"logps/chosen": -660.05224609375,
"logps/rejected": -1413.72412109375,
"loss": 0.0704,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.17005473375320435,
"rewards/margins": 0.3672229051589966,
"rewards/rejected": -0.5372776985168457,
"step": 750
},
{
"epoch": 0.41,
"learning_rate": 3.710824928548546e-06,
"logits/chosen": -1.5888159275054932,
"logits/rejected": -0.9240763783454895,
"logps/chosen": -703.791259765625,
"logps/rejected": -1287.2347412109375,
"loss": 0.0847,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.2002638578414917,
"rewards/margins": 0.2807930111885071,
"rewards/rejected": -0.4810568690299988,
"step": 760
},
{
"epoch": 0.41,
"learning_rate": 3.6698862519617225e-06,
"logits/chosen": -1.5560834407806396,
"logits/rejected": -0.8829728960990906,
"logps/chosen": -687.9392700195312,
"logps/rejected": -1264.13427734375,
"loss": 0.0614,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.23228220641613007,
"rewards/margins": 0.27504783868789673,
"rewards/rejected": -0.507330060005188,
"step": 770
},
{
"epoch": 0.42,
"learning_rate": 3.6285418794300793e-06,
"logits/chosen": -1.6791499853134155,
"logits/rejected": -0.868704617023468,
"logps/chosen": -739.4034423828125,
"logps/rejected": -1467.380859375,
"loss": 0.0709,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.26428860425949097,
"rewards/margins": 0.3761114478111267,
"rewards/rejected": -0.6404000520706177,
"step": 780
},
{
"epoch": 0.42,
"learning_rate": 3.5868061484537365e-06,
"logits/chosen": -1.6334638595581055,
"logits/rejected": -0.9652324914932251,
"logps/chosen": -696.2661743164062,
"logps/rejected": -1319.149169921875,
"loss": 0.0726,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.2676045298576355,
"rewards/margins": 0.2971785068511963,
"rewards/rejected": -0.5647830367088318,
"step": 790
},
{
"epoch": 0.43,
"learning_rate": 3.5446935322490285e-06,
"logits/chosen": -1.5483821630477905,
"logits/rejected": -0.752991795539856,
"logps/chosen": -802.7833251953125,
"logps/rejected": -1420.5225830078125,
"loss": 0.064,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.2800512909889221,
"rewards/margins": 0.28601035475730896,
"rewards/rejected": -0.5660616159439087,
"step": 800
},
{
"epoch": 0.43,
"learning_rate": 3.502218634729447e-06,
"logits/chosen": -1.4806568622589111,
"logits/rejected": -1.0608776807785034,
"logps/chosen": -572.4467163085938,
"logps/rejected": -1223.5535888671875,
"loss": 0.0826,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.19516542553901672,
"rewards/margins": 0.2695198655128479,
"rewards/rejected": -0.46468526124954224,
"step": 810
},
{
"epoch": 0.44,
"learning_rate": 3.459396185441265e-06,
"logits/chosen": -1.6573173999786377,
"logits/rejected": -0.9227333068847656,
"logps/chosen": -583.91357421875,
"logps/rejected": -1404.231201171875,
"loss": 0.0654,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.19101464748382568,
"rewards/margins": 0.35477685928344727,
"rewards/rejected": -0.5457914471626282,
"step": 820
},
{
"epoch": 0.44,
"learning_rate": 3.4162410344555834e-06,
"logits/chosen": -1.5851019620895386,
"logits/rejected": -1.1501173973083496,
"logps/chosen": -773.6009521484375,
"logps/rejected": -1467.8524169921875,
"loss": 0.0722,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.28683894872665405,
"rewards/margins": 0.3111644387245178,
"rewards/rejected": -0.5980033874511719,
"step": 830
},
{
"epoch": 0.45,
"learning_rate": 3.3727681472185937e-06,
"logits/chosen": -1.7856470346450806,
"logits/rejected": -0.7052544951438904,
"logps/chosen": -758.5326538085938,
"logps/rejected": -1374.1175537109375,
"loss": 0.0738,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.21995194256305695,
"rewards/margins": 0.3128163814544678,
"rewards/rejected": -0.5327683687210083,
"step": 840
},
{
"epoch": 0.45,
"learning_rate": 3.3289925993618217e-06,
"logits/chosen": -1.5121822357177734,
"logits/rejected": -0.6632939577102661,
"logps/chosen": -758.4606323242188,
"logps/rejected": -1379.579833984375,
"loss": 0.0696,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.24226677417755127,
"rewards/margins": 0.2715800404548645,
"rewards/rejected": -0.5138468146324158,
"step": 850
},
{
"epoch": 0.46,
"learning_rate": 3.2849295714741643e-06,
"logits/chosen": -1.4147189855575562,
"logits/rejected": -0.8505645990371704,
"logps/chosen": -685.36865234375,
"logps/rejected": -1207.5999755859375,
"loss": 0.0923,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.20475773513317108,
"rewards/margins": 0.26389437913894653,
"rewards/rejected": -0.4686520993709564,
"step": 860
},
{
"epoch": 0.46,
"learning_rate": 3.2405943438375287e-06,
"logits/chosen": -1.3422014713287354,
"logits/rejected": -1.0369704961776733,
"logps/chosen": -558.9657592773438,
"logps/rejected": -1227.389404296875,
"loss": 0.0724,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.14610688388347626,
"rewards/margins": 0.2603977918624878,
"rewards/rejected": -0.40650469064712524,
"step": 870
},
{
"epoch": 0.47,
"learning_rate": 3.1960022911279036e-06,
"logits/chosen": -1.413086175918579,
"logits/rejected": -0.8006247282028198,
"logps/chosen": -568.7772216796875,
"logps/rejected": -1355.8843994140625,
"loss": 0.066,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.1402299702167511,
"rewards/margins": 0.332205593585968,
"rewards/rejected": -0.4724355638027191,
"step": 880
},
{
"epoch": 0.47,
"learning_rate": 3.1511688770836844e-06,
"logits/chosen": -1.7916749715805054,
"logits/rejected": -1.1966052055358887,
"logps/chosen": -509.91522216796875,
"logps/rejected": -1213.2950439453125,
"loss": 0.0795,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.13929350674152374,
"rewards/margins": 0.29636120796203613,
"rewards/rejected": -0.43565472960472107,
"step": 890
},
{
"epoch": 0.48,
"learning_rate": 3.1061096491431307e-06,
"logits/chosen": -1.959776520729065,
"logits/rejected": -1.0398948192596436,
"logps/chosen": -717.4497680664062,
"logps/rejected": -1357.809814453125,
"loss": 0.0751,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.17033252120018005,
"rewards/margins": 0.3062911629676819,
"rewards/rejected": -0.47662368416786194,
"step": 900
},
{
"epoch": 0.49,
"learning_rate": 3.0608402330527796e-06,
"logits/chosen": -1.7569372653961182,
"logits/rejected": -0.9980508089065552,
"logps/chosen": -681.2301025390625,
"logps/rejected": -1178.5247802734375,
"loss": 0.0747,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.19839105010032654,
"rewards/margins": 0.2534715533256531,
"rewards/rejected": -0.4518625736236572,
"step": 910
},
{
"epoch": 0.49,
"learning_rate": 3.0153763274487176e-06,
"logits/chosen": -1.4997788667678833,
"logits/rejected": -1.0186471939086914,
"logps/chosen": -680.5430297851562,
"logps/rejected": -1420.2906494140625,
"loss": 0.0483,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -0.212947279214859,
"rewards/margins": 0.3238678574562073,
"rewards/rejected": -0.5368151664733887,
"step": 920
},
{
"epoch": 0.5,
"learning_rate": 2.9697336984125683e-06,
"logits/chosen": -1.8647575378417969,
"logits/rejected": -1.020261526107788,
"logps/chosen": -713.0181274414062,
"logps/rejected": -1391.114013671875,
"loss": 0.0696,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.1832994669675827,
"rewards/margins": 0.3290650248527527,
"rewards/rejected": -0.5123645067214966,
"step": 930
},
{
"epoch": 0.5,
"learning_rate": 2.923928174004094e-06,
"logits/chosen": -1.7081159353256226,
"logits/rejected": -0.9963458776473999,
"logps/chosen": -635.8792724609375,
"logps/rejected": -1234.2698974609375,
"loss": 0.0705,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.1768474280834198,
"rewards/margins": 0.28107231855392456,
"rewards/rejected": -0.4579196870326996,
"step": 940
},
{
"epoch": 0.51,
"learning_rate": 2.8779756387723036e-06,
"logits/chosen": -1.4116071462631226,
"logits/rejected": -0.8761368989944458,
"logps/chosen": -595.3892211914062,
"logps/rejected": -1141.560546875,
"loss": 0.0793,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.15147005021572113,
"rewards/margins": 0.25574880838394165,
"rewards/rejected": -0.407218873500824,
"step": 950
},
{
"epoch": 0.51,
"learning_rate": 2.831892028246968e-06,
"logits/chosen": -1.6526048183441162,
"logits/rejected": -1.0399147272109985,
"logps/chosen": -609.0975341796875,
"logps/rejected": -1123.380126953125,
"loss": 0.0792,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.18281634151935577,
"rewards/margins": 0.2393408715724945,
"rewards/rejected": -0.4221572279930115,
"step": 960
},
{
"epoch": 0.52,
"learning_rate": 2.7856933234124617e-06,
"logits/chosen": -1.5252645015716553,
"logits/rejected": -0.9064356088638306,
"logps/chosen": -705.7091064453125,
"logps/rejected": -1318.63525390625,
"loss": 0.0836,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.24514424800872803,
"rewards/margins": 0.25952351093292236,
"rewards/rejected": -0.5046677589416504,
"step": 970
},
{
"epoch": 0.52,
"learning_rate": 2.7393955451658387e-06,
"logits/chosen": -1.6756401062011719,
"logits/rejected": -1.1370999813079834,
"logps/chosen": -696.5738525390625,
"logps/rejected": -1390.819580078125,
"loss": 0.0636,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.19441962242126465,
"rewards/margins": 0.326028048992157,
"rewards/rejected": -0.5204476714134216,
"step": 980
},
{
"epoch": 0.53,
"learning_rate": 2.6930147487610667e-06,
"logits/chosen": -1.6806871891021729,
"logits/rejected": -0.9537825584411621,
"logps/chosen": -707.53564453125,
"logps/rejected": -1410.8720703125,
"loss": 0.0615,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.2122100293636322,
"rewards/margins": 0.30375686287879944,
"rewards/rejected": -0.5159668922424316,
"step": 990
},
{
"epoch": 0.53,
"learning_rate": 2.6465670182413487e-06,
"logits/chosen": -1.6937450170516968,
"logits/rejected": -0.8107245564460754,
"logps/chosen": -796.8629150390625,
"logps/rejected": -1336.6978759765625,
"loss": 0.0605,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.24894428253173828,
"rewards/margins": 0.2776329815387726,
"rewards/rejected": -0.5265771746635437,
"step": 1000
},
{
"epoch": 0.54,
"learning_rate": 2.6000684608614594e-06,
"logits/chosen": -1.3929378986358643,
"logits/rejected": -1.027085542678833,
"logps/chosen": -734.2984008789062,
"logps/rejected": -1459.5130615234375,
"loss": 0.0574,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.22584040462970734,
"rewards/margins": 0.3117733895778656,
"rewards/rejected": -0.5376137495040894,
"step": 1010
},
{
"epoch": 0.54,
"learning_rate": 2.5535352015020338e-06,
"logits/chosen": -1.5820205211639404,
"logits/rejected": -0.9477416276931763,
"logps/chosen": -645.0429077148438,
"logps/rejected": -1322.5606689453125,
"loss": 0.0627,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.2123514711856842,
"rewards/margins": 0.3156682550907135,
"rewards/rejected": -0.5280197262763977,
"step": 1020
},
{
"epoch": 0.55,
"learning_rate": 2.506983377077741e-06,
"logits/chosen": -1.5610507726669312,
"logits/rejected": -1.2100675106048584,
"logps/chosen": -643.7442626953125,
"logps/rejected": -1385.019287109375,
"loss": 0.0935,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.22349460422992706,
"rewards/margins": 0.3161535859107971,
"rewards/rejected": -0.539648175239563,
"step": 1030
},
{
"epoch": 0.55,
"learning_rate": 2.460429130941289e-06,
"logits/chosen": -1.4837318658828735,
"logits/rejected": -1.084937334060669,
"logps/chosen": -602.5631103515625,
"logps/rejected": -1196.161865234375,
"loss": 0.072,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.19741597771644592,
"rewards/margins": 0.2487364560365677,
"rewards/rejected": -0.44615238904953003,
"step": 1040
},
{
"epoch": 0.56,
"learning_rate": 2.413888607285192e-06,
"logits/chosen": -1.7739849090576172,
"logits/rejected": -1.2259210348129272,
"logps/chosen": -704.445068359375,
"logps/rejected": -1330.047119140625,
"loss": 0.0771,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.18476839363574982,
"rewards/margins": 0.2862476408481598,
"rewards/rejected": -0.4710160195827484,
"step": 1050
},
{
"epoch": 0.57,
"learning_rate": 2.367377945543249e-06,
"logits/chosen": -1.4854525327682495,
"logits/rejected": -0.9668752551078796,
"logps/chosen": -602.9412841796875,
"logps/rejected": -1244.30712890625,
"loss": 0.0693,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.1880866140127182,
"rewards/margins": 0.27652662992477417,
"rewards/rejected": -0.46461325883865356,
"step": 1060
},
{
"epoch": 0.57,
"learning_rate": 2.320913274793676e-06,
"logits/chosen": -1.6579099893569946,
"logits/rejected": -1.1884429454803467,
"logps/chosen": -621.8614501953125,
"logps/rejected": -1197.624267578125,
"loss": 0.0775,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.2010684311389923,
"rewards/margins": 0.24993105232715607,
"rewards/rejected": -0.4509994387626648,
"step": 1070
},
{
"epoch": 0.58,
"learning_rate": 2.27451070816582e-06,
"logits/chosen": -1.5667635202407837,
"logits/rejected": -1.0707104206085205,
"logps/chosen": -598.3284301757812,
"logps/rejected": -1302.890380859375,
"loss": 0.0719,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.1652025729417801,
"rewards/margins": 0.3089999854564667,
"rewards/rejected": -0.47420254349708557,
"step": 1080
},
{
"epoch": 0.58,
"learning_rate": 2.228186337252414e-06,
"logits/chosen": -1.6411329507827759,
"logits/rejected": -0.8932744860649109,
"logps/chosen": -635.66357421875,
"logps/rejected": -1338.457275390625,
"loss": 0.0729,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.19713297486305237,
"rewards/margins": 0.33209219574928284,
"rewards/rejected": -0.52922523021698,
"step": 1090
},
{
"epoch": 0.59,
"learning_rate": 2.1819562265292946e-06,
"logits/chosen": -1.6340553760528564,
"logits/rejected": -0.8918849229812622,
"logps/chosen": -700.44873046875,
"logps/rejected": -1321.0467529296875,
"loss": 0.1045,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.24557940661907196,
"rewards/margins": 0.28524118661880493,
"rewards/rejected": -0.5308206677436829,
"step": 1100
},
{
"epoch": 0.59,
"learning_rate": 2.1358364077845236e-06,
"logits/chosen": -1.6686341762542725,
"logits/rejected": -1.1630200147628784,
"logps/chosen": -622.9368286132812,
"logps/rejected": -1250.1239013671875,
"loss": 0.0765,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.188968688249588,
"rewards/margins": 0.2903675138950348,
"rewards/rejected": -0.4793361723423004,
"step": 1110
},
{
"epoch": 0.6,
"learning_rate": 2.089842874558849e-06,
"logits/chosen": -1.606802225112915,
"logits/rejected": -0.9541341066360474,
"logps/chosen": -725.2327270507812,
"logps/rejected": -1400.559326171875,
"loss": 0.0566,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.20348282158374786,
"rewards/margins": 0.2979651689529419,
"rewards/rejected": -0.5014479756355286,
"step": 1120
},
{
"epoch": 0.6,
"learning_rate": 2.0439915765994242e-06,
"logits/chosen": -1.7844717502593994,
"logits/rejected": -1.3071167469024658,
"logps/chosen": -600.643798828125,
"logps/rejected": -1224.3431396484375,
"loss": 0.0653,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.12222178280353546,
"rewards/margins": 0.28509068489074707,
"rewards/rejected": -0.4073124825954437,
"step": 1130
},
{
"epoch": 0.61,
"learning_rate": 1.9982984143287186e-06,
"logits/chosen": -1.5732018947601318,
"logits/rejected": -1.1557753086090088,
"logps/chosen": -515.3253173828125,
"logps/rejected": -1262.8011474609375,
"loss": 0.0652,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.1322193145751953,
"rewards/margins": 0.31077098846435547,
"rewards/rejected": -0.442990243434906,
"step": 1140
},
{
"epoch": 0.61,
"learning_rate": 1.95277923333053e-06,
"logits/chosen": -1.5410287380218506,
"logits/rejected": -1.1754248142242432,
"logps/chosen": -524.8809814453125,
"logps/rejected": -1286.816162109375,
"loss": 0.0595,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.14767605066299438,
"rewards/margins": 0.2989698350429535,
"rewards/rejected": -0.4466458857059479,
"step": 1150
},
{
"epoch": 0.62,
"learning_rate": 1.9074498188550156e-06,
"logits/chosen": -1.4306999444961548,
"logits/rejected": -0.7852525115013123,
"logps/chosen": -668.5032958984375,
"logps/rejected": -1384.123046875,
"loss": 0.0601,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.19471505284309387,
"rewards/margins": 0.3257969617843628,
"rewards/rejected": -0.5205119848251343,
"step": 1160
},
{
"epoch": 0.62,
"learning_rate": 1.862325890344643e-06,
"logits/chosen": -1.3971048593521118,
"logits/rejected": -0.927646279335022,
"logps/chosen": -705.084228515625,
"logps/rejected": -1310.335205078125,
"loss": 0.077,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.21002249419689178,
"rewards/margins": 0.27293241024017334,
"rewards/rejected": -0.4829549193382263,
"step": 1170
},
{
"epoch": 0.63,
"learning_rate": 1.817423095982972e-06,
"logits/chosen": -1.394499659538269,
"logits/rejected": -0.9891581535339355,
"logps/chosen": -564.6983642578125,
"logps/rejected": -1220.2957763671875,
"loss": 0.0915,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.16074185073375702,
"rewards/margins": 0.2923213839530945,
"rewards/rejected": -0.4530632495880127,
"step": 1180
},
{
"epoch": 0.63,
"learning_rate": 1.7727570072681293e-06,
"logits/chosen": -1.4173166751861572,
"logits/rejected": -1.0692417621612549,
"logps/chosen": -611.6875,
"logps/rejected": -1175.55419921875,
"loss": 0.0834,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.19331012666225433,
"rewards/margins": 0.24444107711315155,
"rewards/rejected": -0.4377512037754059,
"step": 1190
},
{
"epoch": 0.64,
"learning_rate": 1.7283431136128961e-06,
"logits/chosen": -1.5889681577682495,
"logits/rejected": -0.8274670839309692,
"logps/chosen": -562.2679443359375,
"logps/rejected": -1287.1878662109375,
"loss": 0.0698,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.14730304479599,
"rewards/margins": 0.3559080958366394,
"rewards/rejected": -0.5032111406326294,
"step": 1200
},
{
"epoch": 0.65,
"learning_rate": 1.6841968169732478e-06,
"logits/chosen": -1.555487871170044,
"logits/rejected": -1.0957095623016357,
"logps/chosen": -603.4686889648438,
"logps/rejected": -1382.740234375,
"loss": 0.0576,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.19595518708229065,
"rewards/margins": 0.31043919920921326,
"rewards/rejected": -0.5063944458961487,
"step": 1210
},
{
"epoch": 0.65,
"learning_rate": 1.6403334265072284e-06,
"logits/chosen": -1.5740302801132202,
"logits/rejected": -1.0648415088653564,
"logps/chosen": -545.7623291015625,
"logps/rejected": -1294.9000244140625,
"loss": 0.0593,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.16540229320526123,
"rewards/margins": 0.3433569669723511,
"rewards/rejected": -0.5087593197822571,
"step": 1220
},
{
"epoch": 0.66,
"learning_rate": 1.5967681532660066e-06,
"logits/chosen": -1.6975386142730713,
"logits/rejected": -0.972979724407196,
"logps/chosen": -599.3750610351562,
"logps/rejected": -1323.2943115234375,
"loss": 0.0697,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.17359238862991333,
"rewards/margins": 0.32719483971595764,
"rewards/rejected": -0.5007871985435486,
"step": 1230
},
{
"epoch": 0.66,
"learning_rate": 1.5535161049189463e-06,
"logits/chosen": -1.4844261407852173,
"logits/rejected": -1.1203126907348633,
"logps/chosen": -650.5365600585938,
"logps/rejected": -1473.968994140625,
"loss": 0.058,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.1938629448413849,
"rewards/margins": 0.3511131703853607,
"rewards/rejected": -0.5449761748313904,
"step": 1240
},
{
"epoch": 0.67,
"learning_rate": 1.5105922805145356e-06,
"logits/chosen": -1.6657180786132812,
"logits/rejected": -1.084463119506836,
"logps/chosen": -700.7429809570312,
"logps/rejected": -1405.9703369140625,
"loss": 0.079,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.23423318564891815,
"rewards/margins": 0.29485347867012024,
"rewards/rejected": -0.5290866494178772,
"step": 1250
},
{
"epoch": 0.67,
"learning_rate": 1.4680115652789823e-06,
"logits/chosen": -1.6721817255020142,
"logits/rejected": -1.380765676498413,
"logps/chosen": -583.6463623046875,
"logps/rejected": -1295.441650390625,
"loss": 0.0835,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.1985049545764923,
"rewards/margins": 0.2941189706325531,
"rewards/rejected": -0.49262386560440063,
"step": 1260
},
{
"epoch": 0.68,
"learning_rate": 1.4257887254542767e-06,
"logits/chosen": -1.4841744899749756,
"logits/rejected": -0.8630777597427368,
"logps/chosen": -643.0660400390625,
"logps/rejected": -1336.444580078125,
"loss": 0.0778,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.18584981560707092,
"rewards/margins": 0.3217209577560425,
"rewards/rejected": -0.507570743560791,
"step": 1270
},
{
"epoch": 0.68,
"learning_rate": 1.3839384031775227e-06,
"logits/chosen": -1.5687516927719116,
"logits/rejected": -1.116389513015747,
"logps/chosen": -692.3853759765625,
"logps/rejected": -1371.9207763671875,
"loss": 0.0825,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.21149520576000214,
"rewards/margins": 0.29092541337013245,
"rewards/rejected": -0.5024206042289734,
"step": 1280
},
{
"epoch": 0.69,
"learning_rate": 1.342475111403298e-06,
"logits/chosen": -1.6287143230438232,
"logits/rejected": -1.275514006614685,
"logps/chosen": -669.5472412109375,
"logps/rejected": -1387.4217529296875,
"loss": 0.0611,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.22967293858528137,
"rewards/margins": 0.29838404059410095,
"rewards/rejected": -0.5280569195747375,
"step": 1290
},
{
"epoch": 0.69,
"learning_rate": 1.3014132288708209e-06,
"logits/chosen": -1.565351128578186,
"logits/rejected": -1.1336759328842163,
"logps/chosen": -672.641357421875,
"logps/rejected": -1306.2833251953125,
"loss": 0.0709,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.20514118671417236,
"rewards/margins": 0.29974114894866943,
"rewards/rejected": -0.5048823356628418,
"step": 1300
},
{
"epoch": 0.7,
"learning_rate": 1.2607669951176549e-06,
"logits/chosen": -1.4452847242355347,
"logits/rejected": -1.112763524055481,
"logps/chosen": -749.3744506835938,
"logps/rejected": -1290.004150390625,
"loss": 0.077,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.24272188544273376,
"rewards/margins": 0.256221204996109,
"rewards/rejected": -0.498943030834198,
"step": 1310
},
{
"epoch": 0.7,
"learning_rate": 1.2205505055416891e-06,
"logits/chosen": -1.8478819131851196,
"logits/rejected": -0.9717103242874146,
"logps/chosen": -609.0137329101562,
"logps/rejected": -1343.593505859375,
"loss": 0.0584,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.15877112746238708,
"rewards/margins": 0.347220242023468,
"rewards/rejected": -0.5059913396835327,
"step": 1320
},
{
"epoch": 0.71,
"learning_rate": 1.1807777065131002e-06,
"logits/chosen": -1.8385093212127686,
"logits/rejected": -1.008526086807251,
"logps/chosen": -653.6995239257812,
"logps/rejected": -1285.072021484375,
"loss": 0.0666,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.17687949538230896,
"rewards/margins": 0.3174961805343628,
"rewards/rejected": -0.49437564611434937,
"step": 1330
},
{
"epoch": 0.71,
"learning_rate": 1.1414623905380012e-06,
"logits/chosen": -1.7583434581756592,
"logits/rejected": -1.0419925451278687,
"logps/chosen": -690.4033203125,
"logps/rejected": -1265.659423828125,
"loss": 0.0614,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.18742068111896515,
"rewards/margins": 0.3070259392261505,
"rewards/rejected": -0.49444660544395447,
"step": 1340
},
{
"epoch": 0.72,
"learning_rate": 1.1026181914754388e-06,
"logits/chosen": -1.6502761840820312,
"logits/rejected": -1.0335781574249268,
"logps/chosen": -684.777099609375,
"logps/rejected": -1356.904541015625,
"loss": 0.0649,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.1998976767063141,
"rewards/margins": 0.31228137016296387,
"rewards/rejected": -0.5121790170669556,
"step": 1350
},
{
"epoch": 0.73,
"learning_rate": 1.0642585798094136e-06,
"logits/chosen": -1.586370587348938,
"logits/rejected": -1.2236953973770142,
"logps/chosen": -682.054931640625,
"logps/rejected": -1410.645263671875,
"loss": 0.0772,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.16813689470291138,
"rewards/margins": 0.3100218176841736,
"rewards/rejected": -0.47815871238708496,
"step": 1360
},
{
"epoch": 0.73,
"learning_rate": 1.0263968579775522e-06,
"logits/chosen": -1.5936353206634521,
"logits/rejected": -0.8407376408576965,
"logps/chosen": -605.4591064453125,
"logps/rejected": -1212.3292236328125,
"loss": 0.0761,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.18272905051708221,
"rewards/margins": 0.28132033348083496,
"rewards/rejected": -0.46404939889907837,
"step": 1370
},
{
"epoch": 0.74,
"learning_rate": 9.89046155758058e-07,
"logits/chosen": -1.4739325046539307,
"logits/rejected": -1.010182499885559,
"logps/chosen": -632.5126953125,
"logps/rejected": -1296.5963134765625,
"loss": 0.0674,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.16441567242145538,
"rewards/margins": 0.30406633019447327,
"rewards/rejected": -0.46848195791244507,
"step": 1380
},
{
"epoch": 0.74,
"learning_rate": 9.52219425716534e-07,
"logits/chosen": -1.432307243347168,
"logits/rejected": -0.9531373977661133,
"logps/chosen": -556.3389892578125,
"logps/rejected": -1208.9351806640625,
"loss": 0.0902,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.15656593441963196,
"rewards/margins": 0.3002859652042389,
"rewards/rejected": -0.45685189962387085,
"step": 1390
},
{
"epoch": 0.75,
"learning_rate": 9.15929438714262e-07,
"logits/chosen": -1.424543857574463,
"logits/rejected": -1.2399370670318604,
"logps/chosen": -589.8561401367188,
"logps/rejected": -1314.0848388671875,
"loss": 0.0612,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.20633108913898468,
"rewards/margins": 0.29661285877227783,
"rewards/rejected": -0.5029439330101013,
"step": 1400
},
{
"epoch": 0.75,
"learning_rate": 8.801887794794911e-07,
"logits/chosen": -1.7470901012420654,
"logits/rejected": -0.9560185670852661,
"logps/chosen": -725.0481567382812,
"logps/rejected": -1290.740234375,
"loss": 0.0812,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.21728627383708954,
"rewards/margins": 0.2701825797557831,
"rewards/rejected": -0.4874688684940338,
"step": 1410
},
{
"epoch": 0.76,
"learning_rate": 8.450098422432787e-07,
"logits/chosen": -1.7515980005264282,
"logits/rejected": -1.0790961980819702,
"logps/chosen": -780.40625,
"logps/rejected": -1295.2103271484375,
"loss": 0.0703,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.2122478485107422,
"rewards/margins": 0.28053194284439087,
"rewards/rejected": -0.4927797317504883,
"step": 1420
},
{
"epoch": 0.76,
"learning_rate": 8.104048264413858e-07,
"logits/chosen": -1.7474523782730103,
"logits/rejected": -1.2634751796722412,
"logps/chosen": -568.755859375,
"logps/rejected": -1179.9197998046875,
"loss": 0.0782,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.16928540170192719,
"rewards/margins": 0.2702566385269165,
"rewards/rejected": -0.4395419955253601,
"step": 1430
},
{
"epoch": 0.77,
"learning_rate": 7.763857324837321e-07,
"logits/chosen": -1.5095304250717163,
"logits/rejected": -1.0166012048721313,
"logps/chosen": -598.3046264648438,
"logps/rejected": -1174.777099609375,
"loss": 0.0757,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.18509359657764435,
"rewards/margins": 0.2518666684627533,
"rewards/rejected": -0.43696027994155884,
"step": 1440
},
{
"epoch": 0.77,
"learning_rate": 7.429643575928605e-07,
"logits/chosen": -1.4101845026016235,
"logits/rejected": -1.092771291732788,
"logps/chosen": -667.2734375,
"logps/rejected": -1364.451416015625,
"loss": 0.0659,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.17928771674633026,
"rewards/margins": 0.2924876809120178,
"rewards/rejected": -0.4717753827571869,
"step": 1450
},
{
"epoch": 0.78,
"learning_rate": 7.101522917128709e-07,
"logits/chosen": -1.8168420791625977,
"logits/rejected": -1.1101696491241455,
"logps/chosen": -688.029052734375,
"logps/rejected": -1324.9195556640625,
"loss": 0.0877,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.1806691437959671,
"rewards/margins": 0.2972376346588135,
"rewards/rejected": -0.4779067933559418,
"step": 1460
},
{
"epoch": 0.78,
"learning_rate": 6.779609134902312e-07,
"logits/chosen": -1.6712579727172852,
"logits/rejected": -1.150111436843872,
"logps/chosen": -659.3943481445312,
"logps/rejected": -1223.517822265625,
"loss": 0.0633,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.17544549703598022,
"rewards/margins": 0.2652505934238434,
"rewards/rejected": -0.4406960904598236,
"step": 1470
},
{
"epoch": 0.79,
"learning_rate": 6.464013863278629e-07,
"logits/chosen": -1.6671737432479858,
"logits/rejected": -1.052958369255066,
"logps/chosen": -633.9606323242188,
"logps/rejected": -1298.9013671875,
"loss": 0.0725,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.14675958454608917,
"rewards/margins": 0.3194653391838074,
"rewards/rejected": -0.4662249982357025,
"step": 1480
},
{
"epoch": 0.79,
"learning_rate": 6.154846545138696e-07,
"logits/chosen": -1.67061448097229,
"logits/rejected": -0.9715604782104492,
"logps/chosen": -676.41748046875,
"logps/rejected": -1426.921875,
"loss": 0.0659,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -0.17396974563598633,
"rewards/margins": 0.3531404137611389,
"rewards/rejected": -0.5271100997924805,
"step": 1490
},
{
"epoch": 0.8,
"learning_rate": 5.852214394262515e-07,
"logits/chosen": -1.5575685501098633,
"logits/rejected": -1.0988106727600098,
"logps/chosen": -639.9945678710938,
"logps/rejected": -1345.9605712890625,
"loss": 0.0642,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.16887350380420685,
"rewards/margins": 0.3250221610069275,
"rewards/rejected": -0.49389567971229553,
"step": 1500
},
{
"epoch": 0.81,
"learning_rate": 5.556222358149191e-07,
"logits/chosen": -1.6690590381622314,
"logits/rejected": -1.222768783569336,
"logps/chosen": -570.43359375,
"logps/rejected": -1150.0074462890625,
"loss": 0.0805,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.14664652943611145,
"rewards/margins": 0.266937255859375,
"rewards/rejected": -0.41358375549316406,
"step": 1510
},
{
"epoch": 0.81,
"learning_rate": 5.266973081622992e-07,
"logits/chosen": -1.7312631607055664,
"logits/rejected": -1.1374391317367554,
"logps/chosen": -684.4765625,
"logps/rejected": -1306.3896484375,
"loss": 0.0647,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.18494835495948792,
"rewards/margins": 0.2902507185935974,
"rewards/rejected": -0.4751990735530853,
"step": 1520
},
{
"epoch": 0.82,
"learning_rate": 4.984566871237942e-07,
"logits/chosen": -1.5153663158416748,
"logits/rejected": -0.8972418904304504,
"logps/chosen": -728.5112915039062,
"logps/rejected": -1351.343994140625,
"loss": 0.0615,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.18961231410503387,
"rewards/margins": 0.2944989502429962,
"rewards/rejected": -0.4841112494468689,
"step": 1530
},
{
"epoch": 0.82,
"learning_rate": 4.709101660493251e-07,
"logits/chosen": -1.5436667203903198,
"logits/rejected": -1.0006486177444458,
"logps/chosen": -612.3607788085938,
"logps/rejected": -1392.751220703125,
"loss": 0.063,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.17778488993644714,
"rewards/margins": 0.3482268452644348,
"rewards/rejected": -0.5260116457939148,
"step": 1540
},
{
"epoch": 0.83,
"learning_rate": 4.440672975871743e-07,
"logits/chosen": -1.6940997838974,
"logits/rejected": -0.9330794215202332,
"logps/chosen": -640.0445556640625,
"logps/rejected": -1303.176025390625,
"loss": 0.0609,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.16365936398506165,
"rewards/margins": 0.3356683552265167,
"rewards/rejected": -0.4993276596069336,
"step": 1550
},
{
"epoch": 0.83,
"learning_rate": 4.1793739037129134e-07,
"logits/chosen": -1.6769466400146484,
"logits/rejected": -1.0806591510772705,
"logps/chosen": -695.0977783203125,
"logps/rejected": -1364.507568359375,
"loss": 0.0848,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.19976365566253662,
"rewards/margins": 0.30775851011276245,
"rewards/rejected": -0.5075221657752991,
"step": 1560
},
{
"epoch": 0.84,
"learning_rate": 3.9252950579322405e-07,
"logits/chosen": -1.6324169635772705,
"logits/rejected": -1.0680665969848633,
"logps/chosen": -531.390869140625,
"logps/rejected": -1179.87109375,
"loss": 0.0585,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.14503778517246246,
"rewards/margins": 0.3031737208366394,
"rewards/rejected": -0.4482114911079407,
"step": 1570
},
{
"epoch": 0.84,
"learning_rate": 3.6785245485978864e-07,
"logits/chosen": -1.3932682275772095,
"logits/rejected": -0.8974223136901855,
"logps/chosen": -596.6116333007812,
"logps/rejected": -1304.460205078125,
"loss": 0.0652,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.17924772202968597,
"rewards/margins": 0.32127711176872253,
"rewards/rejected": -0.5005248785018921,
"step": 1580
},
{
"epoch": 0.85,
"learning_rate": 3.43914795137566e-07,
"logits/chosen": -1.5574601888656616,
"logits/rejected": -1.1562343835830688,
"logps/chosen": -570.518798828125,
"logps/rejected": -1341.0030517578125,
"loss": 0.0601,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.16952624917030334,
"rewards/margins": 0.30771198868751526,
"rewards/rejected": -0.477238267660141,
"step": 1590
},
{
"epoch": 0.85,
"learning_rate": 3.207248277852901e-07,
"logits/chosen": -1.7783597707748413,
"logits/rejected": -1.0141966342926025,
"logps/chosen": -588.4285888671875,
"logps/rejected": -1261.8870849609375,
"loss": 0.0478,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.14779958128929138,
"rewards/margins": 0.32103490829467773,
"rewards/rejected": -0.4688344895839691,
"step": 1600
},
{
"epoch": 0.86,
"learning_rate": 2.9829059467515074e-07,
"logits/chosen": -1.584507942199707,
"logits/rejected": -1.1659561395645142,
"logps/chosen": -588.9163818359375,
"logps/rejected": -1310.104248046875,
"loss": 0.0611,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.15733642876148224,
"rewards/margins": 0.3178580403327942,
"rewards/rejected": -0.47519451379776,
"step": 1610
},
{
"epoch": 0.86,
"learning_rate": 2.766198756040153e-07,
"logits/chosen": -1.7042678594589233,
"logits/rejected": -0.97917640209198,
"logps/chosen": -641.7584228515625,
"logps/rejected": -1274.931640625,
"loss": 0.0733,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.1470673382282257,
"rewards/margins": 0.30876293778419495,
"rewards/rejected": -0.4558301866054535,
"step": 1620
},
{
"epoch": 0.87,
"learning_rate": 2.5572018559553155e-07,
"logits/chosen": -1.5918161869049072,
"logits/rejected": -0.6565667986869812,
"logps/chosen": -689.022705078125,
"logps/rejected": -1258.276123046875,
"loss": 0.057,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.19913257658481598,
"rewards/margins": 0.2890087962150574,
"rewards/rejected": -0.48814138770103455,
"step": 1630
},
{
"epoch": 0.87,
"learning_rate": 2.3559877229404864e-07,
"logits/chosen": -1.3946174383163452,
"logits/rejected": -0.9993401765823364,
"logps/chosen": -509.4546813964844,
"logps/rejected": -1304.57373046875,
"loss": 0.0567,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.14354419708251953,
"rewards/margins": 0.3309141993522644,
"rewards/rejected": -0.47445839643478394,
"step": 1640
},
{
"epoch": 0.88,
"learning_rate": 2.1626261345126576e-07,
"logits/chosen": -1.566758394241333,
"logits/rejected": -1.0273525714874268,
"logps/chosen": -628.7932739257812,
"logps/rejected": -1379.468017578125,
"loss": 0.0589,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.18863460421562195,
"rewards/margins": 0.322934091091156,
"rewards/rejected": -0.5115686655044556,
"step": 1650
},
{
"epoch": 0.89,
"learning_rate": 1.9771841450646505e-07,
"logits/chosen": -1.3873016834259033,
"logits/rejected": -1.0786285400390625,
"logps/chosen": -586.18994140625,
"logps/rejected": -1279.546630859375,
"loss": 0.0641,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.1676522046327591,
"rewards/margins": 0.30437177419662476,
"rewards/rejected": -0.47202396392822266,
"step": 1660
},
{
"epoch": 0.89,
"learning_rate": 1.7997260626118758e-07,
"logits/chosen": -1.485750436782837,
"logits/rejected": -1.0917704105377197,
"logps/chosen": -545.1902465820312,
"logps/rejected": -1163.88232421875,
"loss": 0.0665,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.15834537148475647,
"rewards/margins": 0.2890353500843048,
"rewards/rejected": -0.4473806917667389,
"step": 1670
},
{
"epoch": 0.9,
"learning_rate": 1.6303134264914365e-07,
"logits/chosen": -1.6250501871109009,
"logits/rejected": -0.9212428331375122,
"logps/chosen": -730.2625122070312,
"logps/rejected": -1425.178466796875,
"loss": 0.0554,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.21810364723205566,
"rewards/margins": 0.31878364086151123,
"rewards/rejected": -0.5368873476982117,
"step": 1680
},
{
"epoch": 0.9,
"learning_rate": 1.469004986021355e-07,
"logits/chosen": -1.6128244400024414,
"logits/rejected": -0.892177402973175,
"logps/chosen": -818.83154296875,
"logps/rejected": -1455.890625,
"loss": 0.0571,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.2139212191104889,
"rewards/margins": 0.3149806261062622,
"rewards/rejected": -0.5289018750190735,
"step": 1690
},
{
"epoch": 0.91,
"learning_rate": 1.315856680127367e-07,
"logits/chosen": -1.692974328994751,
"logits/rejected": -0.840363621711731,
"logps/chosen": -631.8255615234375,
"logps/rejected": -1351.36181640625,
"loss": 0.0581,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.16292592883110046,
"rewards/margins": 0.3375861942768097,
"rewards/rejected": -0.5005121231079102,
"step": 1700
},
{
"epoch": 0.91,
"learning_rate": 1.1709216179442817e-07,
"logits/chosen": -1.6714107990264893,
"logits/rejected": -0.9252397418022156,
"logps/chosen": -709.752685546875,
"logps/rejected": -1277.791748046875,
"loss": 0.0673,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.2090533971786499,
"rewards/margins": 0.2772943079471588,
"rewards/rejected": -0.4863477349281311,
"step": 1710
},
{
"epoch": 0.92,
"learning_rate": 1.0342500603986421e-07,
"logits/chosen": -1.5991243124008179,
"logits/rejected": -1.109691858291626,
"logps/chosen": -545.5228271484375,
"logps/rejected": -1269.11572265625,
"loss": 0.0427,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.14758913218975067,
"rewards/margins": 0.3135208189487457,
"rewards/rejected": -0.4611099660396576,
"step": 1720
},
{
"epoch": 0.92,
"learning_rate": 9.058894027791643e-08,
"logits/chosen": -1.5346282720565796,
"logits/rejected": -1.0807735919952393,
"logps/chosen": -547.136474609375,
"logps/rejected": -1270.7979736328125,
"loss": 0.0532,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.16198331117630005,
"rewards/margins": 0.3293561637401581,
"rewards/rejected": -0.4913395345211029,
"step": 1730
},
{
"epoch": 0.93,
"learning_rate": 7.858841583008592e-08,
"logits/chosen": -1.6227678060531616,
"logits/rejected": -1.0483043193817139,
"logps/chosen": -680.5284423828125,
"logps/rejected": -1272.281494140625,
"loss": 0.0837,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.1814839243888855,
"rewards/margins": 0.2653439939022064,
"rewards/rejected": -0.44682788848876953,
"step": 1740
},
{
"epoch": 0.93,
"learning_rate": 6.742759426686313e-08,
"logits/chosen": -1.5319443941116333,
"logits/rejected": -0.925460159778595,
"logps/chosen": -579.5850830078125,
"logps/rejected": -1320.3759765625,
"loss": 0.0402,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.17916153371334076,
"rewards/margins": 0.3212929368019104,
"rewards/rejected": -0.5004544854164124,
"step": 1750
},
{
"epoch": 0.94,
"learning_rate": 5.7110345964571104e-08,
"logits/chosen": -1.565772294998169,
"logits/rejected": -1.096321940422058,
"logps/chosen": -534.4620361328125,
"logps/rejected": -1283.698486328125,
"loss": 0.0594,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.16089969873428345,
"rewards/margins": 0.31563904881477356,
"rewards/rejected": -0.4765387177467346,
"step": 1760
},
{
"epoch": 0.94,
"learning_rate": 4.764024876318357e-08,
"logits/chosen": -1.536378026008606,
"logits/rejected": -1.0789868831634521,
"logps/chosen": -596.4452514648438,
"logps/rejected": -1328.3414306640625,
"loss": 0.065,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.1719946265220642,
"rewards/margins": 0.3378538489341736,
"rewards/rejected": -0.5098485350608826,
"step": 1770
},
{
"epoch": 0.95,
"learning_rate": 3.902058672559633e-08,
"logits/chosen": -1.4989714622497559,
"logits/rejected": -0.8038862347602844,
"logps/chosen": -660.2965087890625,
"logps/rejected": -1396.283203125,
"loss": 0.0495,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.187455415725708,
"rewards/margins": 0.32814398407936096,
"rewards/rejected": -0.5155993700027466,
"step": 1780
},
{
"epoch": 0.95,
"learning_rate": 3.125434899876933e-08,
"logits/chosen": -1.6087837219238281,
"logits/rejected": -1.2250018119812012,
"logps/chosen": -563.0875854492188,
"logps/rejected": -1169.5546875,
"loss": 0.0771,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.16033567488193512,
"rewards/margins": 0.2670869529247284,
"rewards/rejected": -0.4274226129055023,
"step": 1790
},
{
"epoch": 0.96,
"learning_rate": 2.4344228777145873e-08,
"logits/chosen": -1.6887807846069336,
"logits/rejected": -1.1670516729354858,
"logps/chosen": -587.798583984375,
"logps/rejected": -1290.9248046875,
"loss": 0.0879,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.1586194485425949,
"rewards/margins": 0.324545294046402,
"rewards/rejected": -0.48316478729248047,
"step": 1800
},
{
"epoch": 0.97,
"learning_rate": 1.829262236869772e-08,
"logits/chosen": -1.4965977668762207,
"logits/rejected": -1.1161562204360962,
"logps/chosen": -609.0550537109375,
"logps/rejected": -1276.494140625,
"loss": 0.0809,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.19842302799224854,
"rewards/margins": 0.28119295835494995,
"rewards/rejected": -0.4796159863471985,
"step": 1810
},
{
"epoch": 0.97,
"learning_rate": 1.3101628363929586e-08,
"logits/chosen": -1.492071270942688,
"logits/rejected": -1.0277793407440186,
"logps/chosen": -578.52734375,
"logps/rejected": -1304.014404296875,
"loss": 0.0619,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.19293084740638733,
"rewards/margins": 0.2969210743904114,
"rewards/rejected": -0.4898518919944763,
"step": 1820
},
{
"epoch": 0.98,
"learning_rate": 8.773046908123195e-09,
"logits/chosen": -1.704077124595642,
"logits/rejected": -0.9419189691543579,
"logps/chosen": -614.6487426757812,
"logps/rejected": -1268.572509765625,
"loss": 0.0591,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.1651904135942459,
"rewards/margins": 0.3228462338447571,
"rewards/rejected": -0.4880366325378418,
"step": 1830
},
{
"epoch": 0.98,
"learning_rate": 5.308379077080817e-09,
"logits/chosen": -1.511736273765564,
"logits/rejected": -0.8690618276596069,
"logps/chosen": -628.6026611328125,
"logps/rejected": -1221.356689453125,
"loss": 0.0798,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.18123619258403778,
"rewards/margins": 0.28583425283432007,
"rewards/rejected": -0.46707040071487427,
"step": 1840
},
{
"epoch": 0.99,
"learning_rate": 2.7088263565760996e-09,
"logits/chosen": -1.657123327255249,
"logits/rejected": -1.1370261907577515,
"logps/chosen": -642.603515625,
"logps/rejected": -1387.967529296875,
"loss": 0.0748,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.1762182116508484,
"rewards/margins": 0.3377261161804199,
"rewards/rejected": -0.5139442682266235,
"step": 1850
},
{
"epoch": 0.99,
"learning_rate": 9.752902257023633e-10,
"logits/chosen": -1.7564224004745483,
"logits/rejected": -0.7881779074668884,
"logps/chosen": -575.2108154296875,
"logps/rejected": -1265.4073486328125,
"loss": 0.0536,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.15589746832847595,
"rewards/margins": 0.3169409930706024,
"rewards/rejected": -0.47283846139907837,
"step": 1860
},
{
"epoch": 1.0,
"learning_rate": 1.083718442532189e-10,
"logits/chosen": -1.4379812479019165,
"logits/rejected": -0.6557310819625854,
"logps/chosen": -743.7650146484375,
"logps/rejected": -1402.105712890625,
"loss": 0.0572,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.20985308289527893,
"rewards/margins": 0.33815231919288635,
"rewards/rejected": -0.5480054020881653,
"step": 1870
},
{
"epoch": 1.0,
"step": 1875,
"total_flos": 0.0,
"train_loss": 0.06912825773557027,
"train_runtime": 9941.711,
"train_samples_per_second": 1.509,
"train_steps_per_second": 0.189
}
],
"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
}