Qin Liu
Model save
2fc6103 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 5.0,
"eval_steps": 1000,
"global_step": 1145,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.004366812227074236,
"grad_norm": 0.49208763779092174,
"learning_rate": 4.347826086956522e-08,
"logits/chosen": -1.130352258682251,
"logits/rejected": -0.9433857798576355,
"logps/chosen": -272.3143005371094,
"logps/rejected": -290.848388671875,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.043668122270742356,
"grad_norm": 0.5373741627747036,
"learning_rate": 4.347826086956522e-07,
"logits/chosen": -1.0119584798812866,
"logits/rejected": -1.054663062095642,
"logps/chosen": -300.1561279296875,
"logps/rejected": -261.20513916015625,
"loss": 0.6928,
"rewards/accuracies": 0.4722222089767456,
"rewards/chosen": 0.0015893004601821303,
"rewards/margins": 0.001330445520579815,
"rewards/rejected": 0.0002588547649793327,
"step": 10
},
{
"epoch": 0.08733624454148471,
"grad_norm": 0.518120805736266,
"learning_rate": 8.695652173913044e-07,
"logits/chosen": -1.1330419778823853,
"logits/rejected": -1.0284273624420166,
"logps/chosen": -283.1369934082031,
"logps/rejected": -323.2412109375,
"loss": 0.6924,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.0005519447731785476,
"rewards/margins": 0.0006424236344173551,
"rewards/rejected": -0.0011943683493882418,
"step": 20
},
{
"epoch": 0.13100436681222707,
"grad_norm": 0.6136347394482697,
"learning_rate": 1.3043478260869566e-06,
"logits/chosen": -1.0650444030761719,
"logits/rejected": -1.107439637184143,
"logps/chosen": -290.29168701171875,
"logps/rejected": -270.15887451171875,
"loss": 0.6899,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.0012172337155789137,
"rewards/margins": 0.0061761606484651566,
"rewards/rejected": -0.007393394596874714,
"step": 30
},
{
"epoch": 0.17467248908296942,
"grad_norm": 0.5452485335548993,
"learning_rate": 1.7391304347826088e-06,
"logits/chosen": -1.0981855392456055,
"logits/rejected": -1.1204359531402588,
"logps/chosen": -288.19488525390625,
"logps/rejected": -279.60968017578125,
"loss": 0.6822,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.01171443797647953,
"rewards/margins": 0.023151502013206482,
"rewards/rejected": -0.03486593812704086,
"step": 40
},
{
"epoch": 0.2183406113537118,
"grad_norm": 0.5716667587131498,
"learning_rate": 2.173913043478261e-06,
"logits/chosen": -1.1297352313995361,
"logits/rejected": -1.0235356092453003,
"logps/chosen": -267.85931396484375,
"logps/rejected": -315.6715393066406,
"loss": 0.6681,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -0.03383685275912285,
"rewards/margins": 0.050773195922374725,
"rewards/rejected": -0.08461005985736847,
"step": 50
},
{
"epoch": 0.26200873362445415,
"grad_norm": 0.7397729774530086,
"learning_rate": 2.6086956521739132e-06,
"logits/chosen": -1.0748536586761475,
"logits/rejected": -1.096380352973938,
"logps/chosen": -298.35894775390625,
"logps/rejected": -278.65167236328125,
"loss": 0.6372,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.06509696692228317,
"rewards/margins": 0.13354261219501495,
"rewards/rejected": -0.19863960146903992,
"step": 60
},
{
"epoch": 0.3056768558951965,
"grad_norm": 0.9490199807243085,
"learning_rate": 3.043478260869566e-06,
"logits/chosen": -1.1220731735229492,
"logits/rejected": -1.1358802318572998,
"logps/chosen": -287.66119384765625,
"logps/rejected": -300.89215087890625,
"loss": 0.566,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -0.05601518228650093,
"rewards/margins": 0.2802619934082031,
"rewards/rejected": -0.33627718687057495,
"step": 70
},
{
"epoch": 0.34934497816593885,
"grad_norm": 1.1366581667233986,
"learning_rate": 3.4782608695652175e-06,
"logits/chosen": -1.1631652116775513,
"logits/rejected": -1.0547783374786377,
"logps/chosen": -273.6378173828125,
"logps/rejected": -378.0707702636719,
"loss": 0.4241,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.0387813039124012,
"rewards/margins": 0.6967908143997192,
"rewards/rejected": -0.7355720400810242,
"step": 80
},
{
"epoch": 0.3930131004366812,
"grad_norm": 0.699317306849894,
"learning_rate": 3.91304347826087e-06,
"logits/chosen": -1.13710355758667,
"logits/rejected": -1.1224480867385864,
"logps/chosen": -286.89630126953125,
"logps/rejected": -425.4947204589844,
"loss": 0.1966,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.010465627536177635,
"rewards/margins": 1.6259124279022217,
"rewards/rejected": -1.6154468059539795,
"step": 90
},
{
"epoch": 0.4366812227074236,
"grad_norm": 0.20225227555832068,
"learning_rate": 4.347826086956522e-06,
"logits/chosen": -1.162099003791809,
"logits/rejected": -1.0710939168930054,
"logps/chosen": -254.20108032226562,
"logps/rejected": -621.758544921875,
"loss": 0.0422,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.06020994111895561,
"rewards/margins": 3.4915261268615723,
"rewards/rejected": -3.431316375732422,
"step": 100
},
{
"epoch": 0.48034934497816595,
"grad_norm": 0.09010131899726684,
"learning_rate": 4.782608695652174e-06,
"logits/chosen": -1.0987495183944702,
"logits/rejected": -1.049612283706665,
"logps/chosen": -281.94390869140625,
"logps/rejected": -799.0904541015625,
"loss": 0.01,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1645246148109436,
"rewards/margins": 5.45929479598999,
"rewards/rejected": -5.29477071762085,
"step": 110
},
{
"epoch": 0.5240174672489083,
"grad_norm": 0.030005828505106115,
"learning_rate": 4.999709285361594e-06,
"logits/chosen": -1.1363656520843506,
"logits/rejected": -1.0334757566452026,
"logps/chosen": -270.98944091796875,
"logps/rejected": -936.4185791015625,
"loss": 0.0039,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.26095929741859436,
"rewards/margins": 6.916310787200928,
"rewards/rejected": -6.655351161956787,
"step": 120
},
{
"epoch": 0.5676855895196506,
"grad_norm": 0.019361577478786484,
"learning_rate": 4.997383973910631e-06,
"logits/chosen": -1.0980539321899414,
"logits/rejected": -0.9340838193893433,
"logps/chosen": -266.50250244140625,
"logps/rejected": -1020.0845947265625,
"loss": 0.002,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.24470993876457214,
"rewards/margins": 7.702264308929443,
"rewards/rejected": -7.457553863525391,
"step": 130
},
{
"epoch": 0.611353711790393,
"grad_norm": 0.012282637611773778,
"learning_rate": 4.9927355140895775e-06,
"logits/chosen": -1.1392152309417725,
"logits/rejected": -0.8957275152206421,
"logps/chosen": -242.6056365966797,
"logps/rejected": -1173.87158203125,
"loss": 0.0022,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.2704493999481201,
"rewards/margins": 8.908761978149414,
"rewards/rejected": -8.638312339782715,
"step": 140
},
{
"epoch": 0.6550218340611353,
"grad_norm": 0.04700706955431712,
"learning_rate": 4.985768230048011e-06,
"logits/chosen": -1.1246144771575928,
"logits/rejected": -0.9621224403381348,
"logps/chosen": -273.8687438964844,
"logps/rejected": -1090.0859375,
"loss": 0.0027,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.17607423663139343,
"rewards/margins": 8.613072395324707,
"rewards/rejected": -8.43699836730957,
"step": 150
},
{
"epoch": 0.6986899563318777,
"grad_norm": 0.01943896178498091,
"learning_rate": 4.976488602981748e-06,
"logits/chosen": -1.145508885383606,
"logits/rejected": -0.9405732154846191,
"logps/chosen": -249.9766387939453,
"logps/rejected": -1222.8212890625,
"loss": 0.0009,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.33843302726745605,
"rewards/margins": 9.41877269744873,
"rewards/rejected": -9.080339431762695,
"step": 160
},
{
"epoch": 0.74235807860262,
"grad_norm": 0.019746462175002255,
"learning_rate": 4.9649052651038255e-06,
"logits/chosen": -1.0897105932235718,
"logits/rejected": -0.9191669225692749,
"logps/chosen": -267.56549072265625,
"logps/rejected": -1116.246826171875,
"loss": 0.0007,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.28829699754714966,
"rewards/margins": 8.88953971862793,
"rewards/rejected": -8.601243019104004,
"step": 170
},
{
"epoch": 0.7860262008733624,
"grad_norm": 0.014841671172711,
"learning_rate": 4.9510289916145295e-06,
"logits/chosen": -1.0843827724456787,
"logits/rejected": -0.8706234693527222,
"logps/chosen": -254.61972045898438,
"logps/rejected": -1236.5650634765625,
"loss": 0.0006,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.22450891137123108,
"rewards/margins": 9.739435195922852,
"rewards/rejected": -9.514925956726074,
"step": 180
},
{
"epoch": 0.8296943231441049,
"grad_norm": 0.012319918167579146,
"learning_rate": 4.934872690677953e-06,
"logits/chosen": -1.0835936069488525,
"logits/rejected": -0.8566627502441406,
"logps/chosen": -267.7489013671875,
"logps/rejected": -1272.4853515625,
"loss": 0.0006,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1923988163471222,
"rewards/margins": 9.936999320983887,
"rewards/rejected": -9.744600296020508,
"step": 190
},
{
"epoch": 0.8733624454148472,
"grad_norm": 0.005430948770812638,
"learning_rate": 4.9164513914144005e-06,
"logits/chosen": -1.0162463188171387,
"logits/rejected": -0.7294031381607056,
"logps/chosen": -249.8229522705078,
"logps/rejected": -1294.401611328125,
"loss": 0.0005,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.20589938759803772,
"rewards/margins": 10.359135627746582,
"rewards/rejected": -10.153237342834473,
"step": 200
},
{
"epoch": 0.9170305676855895,
"grad_norm": 0.009500725575344962,
"learning_rate": 4.8957822299198045e-06,
"logits/chosen": -1.1154823303222656,
"logits/rejected": -0.8606049418449402,
"logps/chosen": -262.8265686035156,
"logps/rejected": -1394.8551025390625,
"loss": 0.0006,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.19020363688468933,
"rewards/margins": 11.131672859191895,
"rewards/rejected": -10.941469192504883,
"step": 210
},
{
"epoch": 0.9606986899563319,
"grad_norm": 0.02542499541949849,
"learning_rate": 4.872884433325169e-06,
"logits/chosen": -1.0547072887420654,
"logits/rejected": -0.8354827165603638,
"logps/chosen": -282.1262512207031,
"logps/rejected": -1277.065673828125,
"loss": 0.0003,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.10952471196651459,
"rewards/margins": 10.300054550170898,
"rewards/rejected": -10.190529823303223,
"step": 220
},
{
"epoch": 1.0043668122270741,
"grad_norm": 0.019502802530769667,
"learning_rate": 4.847779301910868e-06,
"logits/chosen": -1.088416576385498,
"logits/rejected": -0.8197258114814758,
"logps/chosen": -240.4183807373047,
"logps/rejected": -1398.536376953125,
"loss": 0.0004,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.14008644223213196,
"rewards/margins": 11.213750839233398,
"rewards/rejected": -11.073664665222168,
"step": 230
},
{
"epoch": 1.0480349344978166,
"grad_norm": 0.025635188925498188,
"learning_rate": 4.820490189292415e-06,
"logits/chosen": -1.019892930984497,
"logits/rejected": -0.7206239700317383,
"logps/chosen": -249.9215087890625,
"logps/rejected": -1368.528564453125,
"loss": 0.0003,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.11964831501245499,
"rewards/margins": 11.160282135009766,
"rewards/rejected": -11.040634155273438,
"step": 240
},
{
"epoch": 1.091703056768559,
"grad_norm": 0.0019862259316801,
"learning_rate": 4.791042480696179e-06,
"logits/chosen": -1.1091744899749756,
"logits/rejected": -0.8607629537582397,
"logps/chosen": -280.6003112792969,
"logps/rejected": -1413.723388671875,
"loss": 0.0004,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.14324969053268433,
"rewards/margins": 11.427762985229492,
"rewards/rejected": -11.284514427185059,
"step": 250
},
{
"epoch": 1.1353711790393013,
"grad_norm": 0.003461068195787667,
"learning_rate": 4.759463569345205e-06,
"logits/chosen": -1.0466573238372803,
"logits/rejected": -0.8044939041137695,
"logps/chosen": -261.8208312988281,
"logps/rejected": -1364.37255859375,
"loss": 0.0007,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.16325196623802185,
"rewards/margins": 11.088302612304688,
"rewards/rejected": -10.925050735473633,
"step": 260
},
{
"epoch": 1.1790393013100438,
"grad_norm": 0.005672203724021556,
"learning_rate": 4.725782830977145e-06,
"logits/chosen": -1.0772409439086914,
"logits/rejected": -0.7674180865287781,
"logps/chosen": -244.04861450195312,
"logps/rejected": -1468.4453125,
"loss": 0.0003,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1920471489429474,
"rewards/margins": 11.930693626403809,
"rewards/rejected": -11.738645553588867,
"step": 270
},
{
"epoch": 1.222707423580786,
"grad_norm": 0.00517545244251001,
"learning_rate": 4.690031596517984e-06,
"logits/chosen": -1.0256779193878174,
"logits/rejected": -0.7417846918106079,
"logps/chosen": -246.1106414794922,
"logps/rejected": -1458.6060791015625,
"loss": 0.0003,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.17811021208763123,
"rewards/margins": 11.871198654174805,
"rewards/rejected": -11.69308853149414,
"step": 280
},
{
"epoch": 1.2663755458515285,
"grad_norm": 0.0052311080547555655,
"learning_rate": 4.652243122936987e-06,
"logits/chosen": -1.0737969875335693,
"logits/rejected": -0.8021048307418823,
"logps/chosen": -262.6719665527344,
"logps/rejected": -1506.664306640625,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.19292786717414856,
"rewards/margins": 12.288152694702148,
"rewards/rejected": -12.09522533416748,
"step": 290
},
{
"epoch": 1.3100436681222707,
"grad_norm": 0.007845893197308343,
"learning_rate": 4.612452562309975e-06,
"logits/chosen": -1.0275319814682007,
"logits/rejected": -0.6741601228713989,
"logps/chosen": -233.83407592773438,
"logps/rejected": -1481.06640625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.20325830578804016,
"rewards/margins": 12.083272933959961,
"rewards/rejected": -11.88001537322998,
"step": 300
},
{
"epoch": 1.3537117903930131,
"grad_norm": 0.001639011685908484,
"learning_rate": 4.570696929119717e-06,
"logits/chosen": -1.0229878425598145,
"logits/rejected": -0.7648278474807739,
"logps/chosen": -259.880615234375,
"logps/rejected": -1439.419189453125,
"loss": 0.0004,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.13743802905082703,
"rewards/margins": 11.878623962402344,
"rewards/rejected": -11.741185188293457,
"step": 310
},
{
"epoch": 1.3973799126637554,
"grad_norm": 0.002036377670452987,
"learning_rate": 4.527015065823841e-06,
"logits/chosen": -1.013264536857605,
"logits/rejected": -0.7761918306350708,
"logps/chosen": -272.4781799316406,
"logps/rejected": -1432.2052001953125,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1656205952167511,
"rewards/margins": 11.849295616149902,
"rewards/rejected": -11.683676719665527,
"step": 320
},
{
"epoch": 1.4410480349344978,
"grad_norm": 0.00825121579599894,
"learning_rate": 4.481447606722309e-06,
"logits/chosen": -1.0379178524017334,
"logits/rejected": -0.7940360903739929,
"logps/chosen": -274.08929443359375,
"logps/rejected": -1472.6800537109375,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.19115516543388367,
"rewards/margins": 12.22935962677002,
"rewards/rejected": -12.03820514678955,
"step": 330
},
{
"epoch": 1.48471615720524,
"grad_norm": 0.006624246041316681,
"learning_rate": 4.434036940158062e-06,
"logits/chosen": -1.0243127346038818,
"logits/rejected": -0.6224126219749451,
"logps/chosen": -222.64013671875,
"logps/rejected": -1671.427001953125,
"loss": 0.0004,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.19593994319438934,
"rewards/margins": 13.684160232543945,
"rewards/rejected": -13.48822021484375,
"step": 340
},
{
"epoch": 1.5283842794759825,
"grad_norm": 0.00969023506429811,
"learning_rate": 4.384827169085993e-06,
"logits/chosen": -1.0768173933029175,
"logits/rejected": -0.9027138948440552,
"logps/chosen": -304.2896728515625,
"logps/rejected": -1465.677001953125,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.06724857538938522,
"rewards/margins": 12.110160827636719,
"rewards/rejected": -12.042912483215332,
"step": 350
},
{
"epoch": 1.572052401746725,
"grad_norm": 0.008449136082982325,
"learning_rate": 4.333864070046938e-06,
"logits/chosen": -0.9996398687362671,
"logits/rejected": -0.6732652187347412,
"logps/chosen": -264.0496520996094,
"logps/rejected": -1532.588134765625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.09769190102815628,
"rewards/margins": 12.633995056152344,
"rewards/rejected": -12.53630256652832,
"step": 360
},
{
"epoch": 1.6157205240174672,
"grad_norm": 0.0026627154109082445,
"learning_rate": 4.28119505058483e-06,
"logits/chosen": -1.0139929056167603,
"logits/rejected": -0.7123194336891174,
"logps/chosen": -275.62481689453125,
"logps/rejected": -1458.732421875,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.08639346808195114,
"rewards/margins": 12.07685661315918,
"rewards/rejected": -11.990463256835938,
"step": 370
},
{
"epoch": 1.6593886462882095,
"grad_norm": 0.004704630681169572,
"learning_rate": 4.226869105146658e-06,
"logits/chosen": -1.0525470972061157,
"logits/rejected": -0.7499625086784363,
"logps/chosen": -256.189453125,
"logps/rejected": -1665.451904296875,
"loss": 0.0005,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.15123505890369415,
"rewards/margins": 13.766995429992676,
"rewards/rejected": -13.615760803222656,
"step": 380
},
{
"epoch": 1.703056768558952,
"grad_norm": 0.008451726735303476,
"learning_rate": 4.170936769506222e-06,
"logits/chosen": -1.0494039058685303,
"logits/rejected": -0.7635387182235718,
"logps/chosen": -272.2103576660156,
"logps/rejected": -1517.40625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.11954255402088165,
"rewards/margins": 12.597132682800293,
"rewards/rejected": -12.477590560913086,
"step": 390
},
{
"epoch": 1.7467248908296944,
"grad_norm": 0.004007124213956819,
"learning_rate": 4.1134500737541026e-06,
"logits/chosen": -1.0403097867965698,
"logits/rejected": -0.7625882029533386,
"logps/chosen": -270.08642578125,
"logps/rejected": -1559.534423828125,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.14756450057029724,
"rewards/margins": 12.893702507019043,
"rewards/rejected": -12.746137619018555,
"step": 400
},
{
"epoch": 1.7903930131004366,
"grad_norm": 0.008071006947580056,
"learning_rate": 4.054462493897569e-06,
"logits/chosen": -1.0538218021392822,
"logits/rejected": -0.6631166934967041,
"logps/chosen": -230.2433319091797,
"logps/rejected": -1751.548095703125,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1488143503665924,
"rewards/margins": 14.362360000610352,
"rewards/rejected": -14.213546752929688,
"step": 410
},
{
"epoch": 1.8340611353711789,
"grad_norm": 0.0018837924584018762,
"learning_rate": 3.994028902115439e-06,
"logits/chosen": -1.0209208726882935,
"logits/rejected": -0.6195182204246521,
"logps/chosen": -235.20816040039062,
"logps/rejected": -1623.1568603515625,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.12359651178121567,
"rewards/margins": 13.363885879516602,
"rewards/rejected": -13.240289688110352,
"step": 420
},
{
"epoch": 1.8777292576419216,
"grad_norm": 0.0024505034148971385,
"learning_rate": 3.932205515714189e-06,
"logits/chosen": -1.0572965145111084,
"logits/rejected": -0.764691948890686,
"logps/chosen": -299.56732177734375,
"logps/rejected": -1628.056884765625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.033301644027233124,
"rewards/margins": 13.467303276062012,
"rewards/rejected": -13.500605583190918,
"step": 430
},
{
"epoch": 1.9213973799126638,
"grad_norm": 0.006019680096008564,
"learning_rate": 3.86904984483277e-06,
"logits/chosen": -1.0619957447052002,
"logits/rejected": -0.7923606634140015,
"logps/chosen": -298.5255432128906,
"logps/rejected": -1547.7294921875,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.007502234075218439,
"rewards/margins": 12.943402290344238,
"rewards/rejected": -12.93589973449707,
"step": 440
},
{
"epoch": 1.965065502183406,
"grad_norm": 0.0039079828979289005,
"learning_rate": 3.8046206389447916e-06,
"logits/chosen": -1.0510538816452026,
"logits/rejected": -0.7345406413078308,
"logps/chosen": -275.7720947265625,
"logps/rejected": -1645.208251953125,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.0038163154385983944,
"rewards/margins": 13.638483047485352,
"rewards/rejected": -13.634668350219727,
"step": 450
},
{
"epoch": 2.0087336244541483,
"grad_norm": 0.0017708249899463602,
"learning_rate": 3.738977832207839e-06,
"logits/chosen": -1.0177559852600098,
"logits/rejected": -0.7104192972183228,
"logps/chosen": -263.3643798828125,
"logps/rejected": -1611.4786376953125,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.02570909820497036,
"rewards/margins": 13.336071968078613,
"rewards/rejected": -13.3103609085083,
"step": 460
},
{
"epoch": 2.052401746724891,
"grad_norm": 0.000997282058931975,
"learning_rate": 3.6721824877107588e-06,
"logits/chosen": -1.0775004625320435,
"logits/rejected": -0.8090157508850098,
"logps/chosen": -305.11627197265625,
"logps/rejected": -1616.4625244140625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.022619858384132385,
"rewards/margins": 13.472871780395508,
"rewards/rejected": -13.450250625610352,
"step": 470
},
{
"epoch": 2.096069868995633,
"grad_norm": 0.0066682395247295335,
"learning_rate": 3.604296740670768e-06,
"logits/chosen": -1.0786068439483643,
"logits/rejected": -0.6820305585861206,
"logps/chosen": -257.0702819824219,
"logps/rejected": -1786.6517333984375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.09799555689096451,
"rewards/margins": 14.890355110168457,
"rewards/rejected": -14.7923583984375,
"step": 480
},
{
"epoch": 2.1397379912663754,
"grad_norm": 0.004948689561737302,
"learning_rate": 3.5353837406332464e-06,
"logits/chosen": -1.0740149021148682,
"logits/rejected": -0.7413342595100403,
"logps/chosen": -247.671630859375,
"logps/rejected": -1785.2659912109375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1334933489561081,
"rewards/margins": 14.81743335723877,
"rewards/rejected": -14.683941841125488,
"step": 490
},
{
"epoch": 2.183406113537118,
"grad_norm": 0.0032297799668046237,
"learning_rate": 3.4655075927279576e-06,
"logits/chosen": -1.0715500116348267,
"logits/rejected": -0.7507299184799194,
"logps/chosen": -257.148681640625,
"logps/rejected": -1643.170654296875,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.04553316906094551,
"rewards/margins": 13.757342338562012,
"rewards/rejected": -13.711812019348145,
"step": 500
},
{
"epoch": 2.2270742358078603,
"grad_norm": 0.0018344305263622298,
"learning_rate": 3.3947332980363552e-06,
"logits/chosen": -0.9531366229057312,
"logits/rejected": -0.5800803303718567,
"logps/chosen": -267.22320556640625,
"logps/rejected": -1683.7496337890625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.01091139018535614,
"rewards/margins": 14.13904094696045,
"rewards/rejected": -14.128130912780762,
"step": 510
},
{
"epoch": 2.2707423580786026,
"grad_norm": 0.0008743192695472521,
"learning_rate": 3.3231266931254546e-06,
"logits/chosen": -1.0591729879379272,
"logits/rejected": -0.7667123079299927,
"logps/chosen": -288.68048095703125,
"logps/rejected": -1706.4267578125,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.048968568444252014,
"rewards/margins": 14.315594673156738,
"rewards/rejected": -14.266626358032227,
"step": 520
},
{
"epoch": 2.314410480349345,
"grad_norm": 0.0006214259809055745,
"learning_rate": 3.250754388804495e-06,
"logits/chosen": -1.0545308589935303,
"logits/rejected": -0.6721210479736328,
"logps/chosen": -258.18707275390625,
"logps/rejected": -1773.6435546875,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.016913337633013725,
"rewards/margins": 14.813578605651855,
"rewards/rejected": -14.796663284301758,
"step": 530
},
{
"epoch": 2.3580786026200875,
"grad_norm": 0.0009986936165298707,
"learning_rate": 3.1776837081613893e-06,
"logits/chosen": -0.9897233247756958,
"logits/rejected": -0.6502217650413513,
"logps/chosen": -265.50994873046875,
"logps/rejected": -1707.3509521484375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.05818755552172661,
"rewards/margins": 14.338269233703613,
"rewards/rejected": -14.280080795288086,
"step": 540
},
{
"epoch": 2.4017467248908297,
"grad_norm": 0.00984202902471183,
"learning_rate": 3.1039826239365754e-06,
"logits/chosen": -1.026111364364624,
"logits/rejected": -0.7162208557128906,
"logps/chosen": -260.0814514160156,
"logps/rejected": -1701.0667724609375,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.060255538672208786,
"rewards/margins": 14.235310554504395,
"rewards/rejected": -14.175054550170898,
"step": 550
},
{
"epoch": 2.445414847161572,
"grad_norm": 0.0006428554350249238,
"learning_rate": 3.0297196952925533e-06,
"logits/chosen": -1.0254216194152832,
"logits/rejected": -0.7658672332763672,
"logps/chosen": -304.80194091796875,
"logps/rejected": -1645.176025390625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.019381849095225334,
"rewards/margins": 13.853315353393555,
"rewards/rejected": -13.872695922851562,
"step": 560
},
{
"epoch": 2.489082969432314,
"grad_norm": 0.0006532892925360383,
"learning_rate": 2.9549640040379043e-06,
"logits/chosen": -1.0421395301818848,
"logits/rejected": -0.7338224649429321,
"logps/chosen": -289.6170349121094,
"logps/rejected": -1806.13671875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.016583764925599098,
"rewards/margins": 15.094789505004883,
"rewards/rejected": -15.078204154968262,
"step": 570
},
{
"epoch": 2.532751091703057,
"grad_norm": 0.0006677213433183142,
"learning_rate": 2.8797850903651274e-06,
"logits/chosen": -1.0362173318862915,
"logits/rejected": -0.716796875,
"logps/chosen": -265.3944091796875,
"logps/rejected": -1734.1741943359375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.07199420034885406,
"rewards/margins": 14.672683715820312,
"rewards/rejected": -14.600687980651855,
"step": 580
},
{
"epoch": 2.576419213973799,
"grad_norm": 0.0017150618685269877,
"learning_rate": 2.804252888162079e-06,
"logits/chosen": -1.1411523818969727,
"logits/rejected": -0.7371311187744141,
"logps/chosen": -265.2336120605469,
"logps/rejected": -1891.736572265625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.006541900336742401,
"rewards/margins": 15.778755187988281,
"rewards/rejected": -15.772212028503418,
"step": 590
},
{
"epoch": 2.6200873362445414,
"grad_norm": 0.0014112958209940734,
"learning_rate": 2.7284376599571776e-06,
"logits/chosen": -1.054970622062683,
"logits/rejected": -0.7131912112236023,
"logps/chosen": -264.7176818847656,
"logps/rejected": -1720.33984375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.05957148224115372,
"rewards/margins": 14.534815788269043,
"rewards/rejected": -14.475244522094727,
"step": 600
},
{
"epoch": 2.6637554585152836,
"grad_norm": 0.0006698680618471773,
"learning_rate": 2.652409931558898e-06,
"logits/chosen": -1.053966760635376,
"logits/rejected": -0.7491308450698853,
"logps/chosen": -280.0963134765625,
"logps/rejected": -1777.6761474609375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.008257086388766766,
"rewards/margins": 14.989705085754395,
"rewards/rejected": -14.98144817352295,
"step": 610
},
{
"epoch": 2.7074235807860263,
"grad_norm": 0.00042568362060489706,
"learning_rate": 2.5762404264503538e-06,
"logits/chosen": -0.9900799989700317,
"logits/rejected": -0.6423364281654358,
"logps/chosen": -269.9592590332031,
"logps/rejected": -1832.5003662109375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.042371202260255814,
"rewards/margins": 15.435081481933594,
"rewards/rejected": -15.392709732055664,
"step": 620
},
{
"epoch": 2.7510917030567685,
"grad_norm": 0.006748202311280723,
"learning_rate": 2.5e-06,
"logits/chosen": -1.0268778800964355,
"logits/rejected": -0.7622432708740234,
"logps/chosen": -285.1914978027344,
"logps/rejected": -1769.4827880859375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.037803106009960175,
"rewards/margins": 14.83598804473877,
"rewards/rejected": -14.873788833618164,
"step": 630
},
{
"epoch": 2.7947598253275108,
"grad_norm": 0.0002702166687339878,
"learning_rate": 2.423759573549647e-06,
"logits/chosen": -1.0029267072677612,
"logits/rejected": -0.6316828727722168,
"logps/chosen": -260.8511962890625,
"logps/rejected": -1813.537841796875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.039741456508636475,
"rewards/margins": 15.310914993286133,
"rewards/rejected": -15.271173477172852,
"step": 640
},
{
"epoch": 2.8384279475982535,
"grad_norm": 0.0006646355698341602,
"learning_rate": 2.3475900684411027e-06,
"logits/chosen": -1.006449818611145,
"logits/rejected": -0.6019884943962097,
"logps/chosen": -250.7771453857422,
"logps/rejected": -1806.491455078125,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.015752162784337997,
"rewards/margins": 15.226987838745117,
"rewards/rejected": -15.2427396774292,
"step": 650
},
{
"epoch": 2.8820960698689957,
"grad_norm": 0.00047399361927177104,
"learning_rate": 2.2715623400428228e-06,
"logits/chosen": -1.0256198644638062,
"logits/rejected": -0.7082680463790894,
"logps/chosen": -287.2723083496094,
"logps/rejected": -1791.8154296875,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.028542857617139816,
"rewards/margins": 15.111543655395508,
"rewards/rejected": -15.14008617401123,
"step": 660
},
{
"epoch": 2.925764192139738,
"grad_norm": 0.054595514895715194,
"learning_rate": 2.1957471118379213e-06,
"logits/chosen": -1.1037609577178955,
"logits/rejected": -0.7577636241912842,
"logps/chosen": -303.9775390625,
"logps/rejected": -1865.5306396484375,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.036632128059864044,
"rewards/margins": 15.698193550109863,
"rewards/rejected": -15.734827041625977,
"step": 670
},
{
"epoch": 2.96943231441048,
"grad_norm": 0.0032068293343592976,
"learning_rate": 2.120214909634873e-06,
"logits/chosen": -1.0517921447753906,
"logits/rejected": -0.770795464515686,
"logps/chosen": -298.4246520996094,
"logps/rejected": -1795.1331787109375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.014761227183043957,
"rewards/margins": 15.1719970703125,
"rewards/rejected": -15.186758041381836,
"step": 680
},
{
"epoch": 3.013100436681223,
"grad_norm": 0.0011657499181104522,
"learning_rate": 2.045035995962097e-06,
"logits/chosen": -1.1193125247955322,
"logits/rejected": -0.7743921279907227,
"logps/chosen": -279.6126403808594,
"logps/rejected": -1916.1976318359375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.046258725225925446,
"rewards/margins": 16.126333236694336,
"rewards/rejected": -16.172592163085938,
"step": 690
},
{
"epoch": 3.056768558951965,
"grad_norm": 0.005031019960938767,
"learning_rate": 1.970280304707447e-06,
"logits/chosen": -1.0536220073699951,
"logits/rejected": -0.6994236707687378,
"logps/chosen": -274.2372131347656,
"logps/rejected": -1818.864013671875,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.030473018065094948,
"rewards/margins": 15.32104206085205,
"rewards/rejected": -15.351513862609863,
"step": 700
},
{
"epoch": 3.1004366812227073,
"grad_norm": 0.0003812225537618957,
"learning_rate": 1.8960173760634257e-06,
"logits/chosen": -1.1193227767944336,
"logits/rejected": -0.8343600034713745,
"logps/chosen": -317.20367431640625,
"logps/rejected": -1797.347900390625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.015081268735229969,
"rewards/margins": 15.233189582824707,
"rewards/rejected": -15.248270034790039,
"step": 710
},
{
"epoch": 3.14410480349345,
"grad_norm": 0.000800716771561942,
"learning_rate": 1.8223162918386122e-06,
"logits/chosen": -1.0911037921905518,
"logits/rejected": -0.7263038158416748,
"logps/chosen": -281.10723876953125,
"logps/rejected": -1893.2275390625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.07366519421339035,
"rewards/margins": 15.97996711730957,
"rewards/rejected": -16.053630828857422,
"step": 720
},
{
"epoch": 3.1877729257641922,
"grad_norm": 0.0033435888991239673,
"learning_rate": 1.7492456111955052e-06,
"logits/chosen": -1.0024917125701904,
"logits/rejected": -0.6330265998840332,
"logps/chosen": -266.6875915527344,
"logps/rejected": -1799.3509521484375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.05265193060040474,
"rewards/margins": 15.307687759399414,
"rewards/rejected": -15.255037307739258,
"step": 730
},
{
"epoch": 3.2314410480349345,
"grad_norm": 0.003033628662190874,
"learning_rate": 1.6768733068745468e-06,
"logits/chosen": -1.0493130683898926,
"logits/rejected": -0.7165160775184631,
"logps/chosen": -277.95294189453125,
"logps/rejected": -1927.5621337890625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.008759565651416779,
"rewards/margins": 16.3278865814209,
"rewards/rejected": -16.336645126342773,
"step": 740
},
{
"epoch": 3.2751091703056767,
"grad_norm": 0.00030739766784929366,
"learning_rate": 1.6052667019636462e-06,
"logits/chosen": -1.0858592987060547,
"logits/rejected": -0.7486026883125305,
"logps/chosen": -285.6336669921875,
"logps/rejected": -1904.0074462890625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.039479680359363556,
"rewards/margins": 15.980633735656738,
"rewards/rejected": -16.02011489868164,
"step": 750
},
{
"epoch": 3.3187772925764194,
"grad_norm": 0.0099847141974217,
"learning_rate": 1.5344924072720434e-06,
"logits/chosen": -1.0522048473358154,
"logits/rejected": -0.7434892058372498,
"logps/chosen": -294.0987548828125,
"logps/rejected": -1748.886962890625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.06646230816841125,
"rewards/margins": 14.771476745605469,
"rewards/rejected": -14.837939262390137,
"step": 760
},
{
"epoch": 3.3624454148471616,
"grad_norm": 0.00046452138658813734,
"learning_rate": 1.4646162593667535e-06,
"logits/chosen": -1.0780597925186157,
"logits/rejected": -0.8004827499389648,
"logps/chosen": -309.17120361328125,
"logps/rejected": -1823.7366943359375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.043299149721860886,
"rewards/margins": 15.575587272644043,
"rewards/rejected": -15.532289505004883,
"step": 770
},
{
"epoch": 3.406113537117904,
"grad_norm": 0.0018748318848776301,
"learning_rate": 1.3957032593292319e-06,
"logits/chosen": -1.042887568473816,
"logits/rejected": -0.6965152621269226,
"logps/chosen": -282.84088134765625,
"logps/rejected": -1889.387451171875,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.013044556602835655,
"rewards/margins": 15.971423149108887,
"rewards/rejected": -15.958378791809082,
"step": 780
},
{
"epoch": 3.449781659388646,
"grad_norm": 0.0009191121751634733,
"learning_rate": 1.3278175122892416e-06,
"logits/chosen": -1.029362678527832,
"logits/rejected": -0.6999324560165405,
"logps/chosen": -269.5941467285156,
"logps/rejected": -1812.1949462890625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.10188305377960205,
"rewards/margins": 15.274584770202637,
"rewards/rejected": -15.376466751098633,
"step": 790
},
{
"epoch": 3.493449781659389,
"grad_norm": 0.003126132491902847,
"learning_rate": 1.261022167792161e-06,
"logits/chosen": -1.0470654964447021,
"logits/rejected": -0.6794711351394653,
"logps/chosen": -278.49285888671875,
"logps/rejected": -1961.002197265625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.009503474459052086,
"rewards/margins": 16.650577545166016,
"rewards/rejected": -16.641077041625977,
"step": 800
},
{
"epoch": 3.537117903930131,
"grad_norm": 0.0059086305628955095,
"learning_rate": 1.195379361055209e-06,
"logits/chosen": -0.9826908111572266,
"logits/rejected": -0.6099307537078857,
"logps/chosen": -278.61492919921875,
"logps/rejected": -1975.528076171875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.03355104476213455,
"rewards/margins": 16.752765655517578,
"rewards/rejected": -16.78631591796875,
"step": 810
},
{
"epoch": 3.5807860262008733,
"grad_norm": 0.003101839804676522,
"learning_rate": 1.1309501551672303e-06,
"logits/chosen": -0.957851231098175,
"logits/rejected": -0.5697217583656311,
"logps/chosen": -272.66534423828125,
"logps/rejected": -1837.3662109375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.07275199890136719,
"rewards/margins": 15.55827522277832,
"rewards/rejected": -15.631025314331055,
"step": 820
},
{
"epoch": 3.6244541484716155,
"grad_norm": 0.006553990432039902,
"learning_rate": 1.0677944842858112e-06,
"logits/chosen": -1.0141226053237915,
"logits/rejected": -0.6257001757621765,
"logps/chosen": -280.70672607421875,
"logps/rejected": -1861.9078369140625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.05027519538998604,
"rewards/margins": 15.878219604492188,
"rewards/rejected": -15.928494453430176,
"step": 830
},
{
"epoch": 3.668122270742358,
"grad_norm": 0.0003416983892558372,
"learning_rate": 1.005971097884561e-06,
"logits/chosen": -1.0156476497650146,
"logits/rejected": -0.668128252029419,
"logps/chosen": -302.67767333984375,
"logps/rejected": -1935.9398193359375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.08607998490333557,
"rewards/margins": 16.394611358642578,
"rewards/rejected": -16.480690002441406,
"step": 840
},
{
"epoch": 3.7117903930131004,
"grad_norm": 0.0015327333742183899,
"learning_rate": 9.455375061024319e-07,
"logits/chosen": -1.0618703365325928,
"logits/rejected": -0.7237562537193298,
"logps/chosen": -283.15557861328125,
"logps/rejected": -1865.9681396484375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.054165177047252655,
"rewards/margins": 15.781875610351562,
"rewards/rejected": -15.836041450500488,
"step": 850
},
{
"epoch": 3.7554585152838427,
"grad_norm": 0.004269773041594158,
"learning_rate": 8.86549926245898e-07,
"logits/chosen": -1.060253381729126,
"logits/rejected": -0.6907247304916382,
"logps/chosen": -279.11090087890625,
"logps/rejected": -1889.4056396484375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.0650867372751236,
"rewards/margins": 15.91883659362793,
"rewards/rejected": -15.983922958374023,
"step": 860
},
{
"epoch": 3.7991266375545854,
"grad_norm": 0.0010427555714299119,
"learning_rate": 8.29063230493779e-07,
"logits/chosen": -1.0654280185699463,
"logits/rejected": -0.6975089907646179,
"logps/chosen": -275.31011962890625,
"logps/rejected": -1959.466796875,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.10659889876842499,
"rewards/margins": 16.469921112060547,
"rewards/rejected": -16.576520919799805,
"step": 870
},
{
"epoch": 3.8427947598253276,
"grad_norm": 0.0051211999521691845,
"learning_rate": 7.731308948533431e-07,
"logits/chosen": -1.0260651111602783,
"logits/rejected": -0.7374225854873657,
"logps/chosen": -309.5072937011719,
"logps/rejected": -1826.8656005859375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.08689963072538376,
"rewards/margins": 15.4946928024292,
"rewards/rejected": -15.58159351348877,
"step": 880
},
{
"epoch": 3.88646288209607,
"grad_norm": 0.002864032933131896,
"learning_rate": 7.188049494151703e-07,
"logits/chosen": -0.9869282841682434,
"logits/rejected": -0.6170369386672974,
"logps/chosen": -270.72308349609375,
"logps/rejected": -1809.623046875,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.03863278776407242,
"rewards/margins": 15.385714530944824,
"rewards/rejected": -15.424346923828125,
"step": 890
},
{
"epoch": 3.930131004366812,
"grad_norm": 0.0005105250492242013,
"learning_rate": 6.661359299530626e-07,
"logits/chosen": -1.0458041429519653,
"logits/rejected": -0.8353503346443176,
"logps/chosen": -325.7237548828125,
"logps/rejected": -1614.521484375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.15461190044879913,
"rewards/margins": 13.781686782836914,
"rewards/rejected": -13.936299324035645,
"step": 900
},
{
"epoch": 3.9737991266375547,
"grad_norm": 0.0019150602563612286,
"learning_rate": 6.151728309140071e-07,
"logits/chosen": -1.0450153350830078,
"logits/rejected": -0.7124982476234436,
"logps/chosen": -285.0978698730469,
"logps/rejected": -1913.976318359375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.026266271248459816,
"rewards/margins": 16.297687530517578,
"rewards/rejected": -16.32395362854004,
"step": 910
},
{
"epoch": 4.0174672489082965,
"grad_norm": 0.003658964581230378,
"learning_rate": 5.659630598419391e-07,
"logits/chosen": -1.0644596815109253,
"logits/rejected": -0.6967732906341553,
"logps/chosen": -284.8916320800781,
"logps/rejected": -1840.668212890625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.04237327724695206,
"rewards/margins": 15.650690078735352,
"rewards/rejected": -15.69306468963623,
"step": 920
},
{
"epoch": 4.06113537117904,
"grad_norm": 0.015359415097265612,
"learning_rate": 5.185523932776923e-07,
"logits/chosen": -1.0555784702301025,
"logits/rejected": -0.7345298528671265,
"logps/chosen": -284.2557067871094,
"logps/rejected": -1905.572509765625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.04300800338387489,
"rewards/margins": 16.220272064208984,
"rewards/rejected": -16.26327896118164,
"step": 930
},
{
"epoch": 4.104803493449782,
"grad_norm": 0.0043225257316477665,
"learning_rate": 4.7298493417616024e-07,
"logits/chosen": -1.024458646774292,
"logits/rejected": -0.7369360327720642,
"logps/chosen": -297.3896484375,
"logps/rejected": -1880.3125,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.06699924170970917,
"rewards/margins": 15.987092971801758,
"rewards/rejected": -16.05409049987793,
"step": 940
},
{
"epoch": 4.148471615720524,
"grad_norm": 0.0007120346767669291,
"learning_rate": 4.293030708802834e-07,
"logits/chosen": -1.0639699697494507,
"logits/rejected": -0.7229939699172974,
"logps/chosen": -270.7474060058594,
"logps/rejected": -1974.702392578125,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.030464956536889076,
"rewards/margins": 16.875165939331055,
"rewards/rejected": -16.84469985961914,
"step": 950
},
{
"epoch": 4.192139737991266,
"grad_norm": 0.0004859166750622356,
"learning_rate": 3.875474376900254e-07,
"logits/chosen": -0.9850033521652222,
"logits/rejected": -0.6558709144592285,
"logps/chosen": -287.7508544921875,
"logps/rejected": -1824.685791015625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.10610561072826385,
"rewards/margins": 15.522783279418945,
"rewards/rejected": -15.628889083862305,
"step": 960
},
{
"epoch": 4.235807860262009,
"grad_norm": 0.002799456551978829,
"learning_rate": 3.4775687706301437e-07,
"logits/chosen": -1.1057827472686768,
"logits/rejected": -0.7324444055557251,
"logps/chosen": -268.9285583496094,
"logps/rejected": -2027.9287109375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.012243424542248249,
"rewards/margins": 17.31172752380371,
"rewards/rejected": -17.299488067626953,
"step": 970
},
{
"epoch": 4.279475982532751,
"grad_norm": 0.0022373731136834133,
"learning_rate": 3.0996840348201717e-07,
"logits/chosen": -1.1121981143951416,
"logits/rejected": -0.8360017538070679,
"logps/chosen": -309.40155029296875,
"logps/rejected": -1771.3427734375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.12068891525268555,
"rewards/margins": 14.9802827835083,
"rewards/rejected": -15.100972175598145,
"step": 980
},
{
"epoch": 4.323144104803493,
"grad_norm": 0.00044941977383980663,
"learning_rate": 2.742171690228562e-07,
"logits/chosen": -1.0458042621612549,
"logits/rejected": -0.7098134756088257,
"logps/chosen": -300.3705749511719,
"logps/rejected": -1846.2822265625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.13983380794525146,
"rewards/margins": 15.580610275268555,
"rewards/rejected": -15.720443725585938,
"step": 990
},
{
"epoch": 4.366812227074236,
"grad_norm": 0.002386856484501028,
"learning_rate": 2.405364306547955e-07,
"logits/chosen": -1.082747459411621,
"logits/rejected": -0.8248895406723022,
"logps/chosen": -306.8914794921875,
"logps/rejected": -1777.709228515625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.0018977939616888762,
"rewards/margins": 15.146817207336426,
"rewards/rejected": -15.148715019226074,
"step": 1000
},
{
"epoch": 4.4104803493449785,
"grad_norm": 0.0029792452832107918,
"learning_rate": 2.0895751930382125e-07,
"logits/chosen": -1.0091092586517334,
"logits/rejected": -0.6796575784683228,
"logps/chosen": -295.8236389160156,
"logps/rejected": -1882.607421875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.007821875624358654,
"rewards/margins": 16.001935958862305,
"rewards/rejected": -16.00975799560547,
"step": 1010
},
{
"epoch": 4.454148471615721,
"grad_norm": 0.0002768349770005794,
"learning_rate": 1.7950981070758488e-07,
"logits/chosen": -1.0478919744491577,
"logits/rejected": -0.7594279050827026,
"logps/chosen": -287.52288818359375,
"logps/rejected": -1790.5562744140625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.0713457316160202,
"rewards/margins": 15.24449634552002,
"rewards/rejected": -15.315841674804688,
"step": 1020
},
{
"epoch": 4.497816593886463,
"grad_norm": 0.005069342731202753,
"learning_rate": 1.5222069808913303e-07,
"logits/chosen": -1.0174686908721924,
"logits/rejected": -0.683228611946106,
"logps/chosen": -301.92144775390625,
"logps/rejected": -1872.2666015625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.04044654220342636,
"rewards/margins": 15.973681449890137,
"rewards/rejected": -16.01412582397461,
"step": 1030
},
{
"epoch": 4.541484716157205,
"grad_norm": 0.0004959573429988225,
"learning_rate": 1.271155666748311e-07,
"logits/chosen": -1.0908715724945068,
"logits/rejected": -0.7660378217697144,
"logps/chosen": -306.63043212890625,
"logps/rejected": -1871.072509765625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.09966543316841125,
"rewards/margins": 15.866052627563477,
"rewards/rejected": -15.965716361999512,
"step": 1040
},
{
"epoch": 4.585152838427947,
"grad_norm": 0.00019462022413360382,
"learning_rate": 1.0421777008019663e-07,
"logits/chosen": -1.0952246189117432,
"logits/rejected": -0.709916353225708,
"logps/chosen": -295.17877197265625,
"logps/rejected": -2033.247314453125,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.07402383536100388,
"rewards/margins": 17.205215454101562,
"rewards/rejected": -17.279239654541016,
"step": 1050
},
{
"epoch": 4.62882096069869,
"grad_norm": 0.0036632249464871505,
"learning_rate": 8.354860858560021e-08,
"logits/chosen": -1.041784405708313,
"logits/rejected": -0.6548532247543335,
"logps/chosen": -293.025146484375,
"logps/rejected": -2029.672607421875,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.0596146360039711,
"rewards/margins": 17.3281307220459,
"rewards/rejected": -17.387744903564453,
"step": 1060
},
{
"epoch": 4.672489082969433,
"grad_norm": 0.00032895380717983924,
"learning_rate": 6.512730932204698e-08,
"logits/chosen": -1.0059609413146973,
"logits/rejected": -0.7132547497749329,
"logps/chosen": -317.78680419921875,
"logps/rejected": -1786.2691650390625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.03703427314758301,
"rewards/margins": 15.275805473327637,
"rewards/rejected": -15.312841415405273,
"step": 1070
},
{
"epoch": 4.716157205240175,
"grad_norm": 0.0031606785955010904,
"learning_rate": 4.897100838547081e-08,
"logits/chosen": -1.0450844764709473,
"logits/rejected": -0.7194000482559204,
"logps/chosen": -301.86541748046875,
"logps/rejected": -1955.65625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.05521082133054733,
"rewards/margins": 16.574329376220703,
"rewards/rejected": -16.62953758239746,
"step": 1080
},
{
"epoch": 4.759825327510917,
"grad_norm": 0.000347629698955204,
"learning_rate": 3.5094734896174985e-08,
"logits/chosen": -1.080841302871704,
"logits/rejected": -0.6831713914871216,
"logps/chosen": -268.0194091796875,
"logps/rejected": -1998.539306640625,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.04163534566760063,
"rewards/margins": 16.94226837158203,
"rewards/rejected": -16.98390007019043,
"step": 1090
},
{
"epoch": 4.8034934497816595,
"grad_norm": 0.004295417718120041,
"learning_rate": 2.351139701825267e-08,
"logits/chosen": -0.9642395973205566,
"logits/rejected": -0.6139574646949768,
"logps/chosen": -276.0238037109375,
"logps/rejected": -1780.542236328125,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.0755363255739212,
"rewards/margins": 15.09814167022705,
"rewards/rejected": -15.173677444458008,
"step": 1100
},
{
"epoch": 4.847161572052402,
"grad_norm": 0.00037246559243259353,
"learning_rate": 1.4231769951990326e-08,
"logits/chosen": -1.033734679222107,
"logits/rejected": -0.6827409267425537,
"logps/chosen": -287.67584228515625,
"logps/rejected": -1885.5576171875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.08407995104789734,
"rewards/margins": 16.050437927246094,
"rewards/rejected": -16.134517669677734,
"step": 1110
},
{
"epoch": 4.890829694323144,
"grad_norm": 0.0003852961387970997,
"learning_rate": 7.264485910423447e-09,
"logits/chosen": -0.9720686078071594,
"logits/rejected": -0.5935501456260681,
"logps/chosen": -265.44622802734375,
"logps/rejected": -1915.1038818359375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.024737151339650154,
"rewards/margins": 16.284820556640625,
"rewards/rejected": -16.309558868408203,
"step": 1120
},
{
"epoch": 4.934497816593886,
"grad_norm": 0.015272373299215733,
"learning_rate": 2.6160260893692833e-09,
"logits/chosen": -1.0101922750473022,
"logits/rejected": -0.6879107356071472,
"logps/chosen": -289.36138916015625,
"logps/rejected": -1796.4605712890625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.028253447264432907,
"rewards/margins": 15.371627807617188,
"rewards/rejected": -15.34337329864502,
"step": 1130
},
{
"epoch": 4.978165938864628,
"grad_norm": 0.0002424398537573296,
"learning_rate": 2.9071463840540936e-10,
"logits/chosen": -1.0956311225891113,
"logits/rejected": -0.8314415216445923,
"logps/chosen": -341.66766357421875,
"logps/rejected": -1909.0863037109375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.11481740325689316,
"rewards/margins": 16.255878448486328,
"rewards/rejected": -16.370695114135742,
"step": 1140
},
{
"epoch": 5.0,
"step": 1145,
"total_flos": 0.0,
"train_loss": 0.046525747915877304,
"train_runtime": 10909.6849,
"train_samples_per_second": 6.711,
"train_steps_per_second": 0.105
}
],
"logging_steps": 10,
"max_steps": 1145,
"num_input_tokens_seen": 0,
"num_train_epochs": 5,
"save_steps": 100,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 2,
"trial_name": null,
"trial_params": null
}