webshop_dpo_ckpt_fromselftrain_e1 / trainer_state.json
xiaoxiaolin's picture
Upload folder using huggingface_hub
11d14d8 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 2.992,
"eval_steps": 500,
"global_step": 561,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02666666666666667,
"grad_norm": 18.322052001953125,
"learning_rate": 8.771929824561403e-08,
"logits/chosen": -0.8889046907424927,
"logits/rejected": -0.8736551403999329,
"logps/chosen": -81.54634094238281,
"logps/rejected": -75.81903839111328,
"loss": 0.6912,
"rewards/accuracies": 0.3499999940395355,
"rewards/chosen": 0.0006732392939738929,
"rewards/margins": 0.005079272203147411,
"rewards/rejected": -0.004406032618135214,
"step": 5
},
{
"epoch": 0.05333333333333334,
"grad_norm": 19.94486427307129,
"learning_rate": 1.7543859649122805e-07,
"logits/chosen": -0.9008264541625977,
"logits/rejected": -0.8997122645378113,
"logps/chosen": -56.543792724609375,
"logps/rejected": -70.56910705566406,
"loss": 0.6741,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.040357910096645355,
"rewards/margins": 0.049730073660612106,
"rewards/rejected": -0.009372168220579624,
"step": 10
},
{
"epoch": 0.08,
"grad_norm": 16.07955551147461,
"learning_rate": 2.631578947368421e-07,
"logits/chosen": -0.9306875467300415,
"logits/rejected": -0.9334227442741394,
"logps/chosen": -70.51209259033203,
"logps/rejected": -84.35716247558594,
"loss": 0.6422,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.06031479313969612,
"rewards/margins": 0.09791518747806549,
"rewards/rejected": -0.03760039433836937,
"step": 15
},
{
"epoch": 0.10666666666666667,
"grad_norm": 16.454225540161133,
"learning_rate": 3.508771929824561e-07,
"logits/chosen": -0.8976303339004517,
"logits/rejected": -0.8699060678482056,
"logps/chosen": -72.55310821533203,
"logps/rejected": -87.03462219238281,
"loss": 0.6069,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.05961589142680168,
"rewards/margins": 0.28127092123031616,
"rewards/rejected": -0.2216549664735794,
"step": 20
},
{
"epoch": 0.13333333333333333,
"grad_norm": 14.617012977600098,
"learning_rate": 4.3859649122807013e-07,
"logits/chosen": -0.9105884432792664,
"logits/rejected": -0.9121261835098267,
"logps/chosen": -81.52780151367188,
"logps/rejected": -91.52813720703125,
"loss": 0.6019,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.12029141187667847,
"rewards/margins": 0.3524962365627289,
"rewards/rejected": -0.2322048395872116,
"step": 25
},
{
"epoch": 0.16,
"grad_norm": 13.667928695678711,
"learning_rate": 5.263157894736842e-07,
"logits/chosen": -0.8984580039978027,
"logits/rejected": -0.9042614102363586,
"logps/chosen": -74.46806335449219,
"logps/rejected": -76.2645034790039,
"loss": 0.5363,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 0.09522397816181183,
"rewards/margins": 0.6430191993713379,
"rewards/rejected": -0.547795295715332,
"step": 30
},
{
"epoch": 0.18666666666666668,
"grad_norm": 16.28107452392578,
"learning_rate": 6.140350877192982e-07,
"logits/chosen": -0.8978121876716614,
"logits/rejected": -0.8999565243721008,
"logps/chosen": -82.0078353881836,
"logps/rejected": -80.48029327392578,
"loss": 0.5021,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.27644580602645874,
"rewards/margins": 0.4465056359767914,
"rewards/rejected": -0.17005985975265503,
"step": 35
},
{
"epoch": 0.21333333333333335,
"grad_norm": 16.186918258666992,
"learning_rate": 7.017543859649122e-07,
"logits/chosen": -0.8643107414245605,
"logits/rejected": -0.8585313558578491,
"logps/chosen": -103.23567962646484,
"logps/rejected": -90.96733856201172,
"loss": 0.4986,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.10163059085607529,
"rewards/margins": 0.47456789016723633,
"rewards/rejected": -0.5761984586715698,
"step": 40
},
{
"epoch": 0.24,
"grad_norm": 14.479534149169922,
"learning_rate": 7.894736842105263e-07,
"logits/chosen": -0.9103773236274719,
"logits/rejected": -0.9051542282104492,
"logps/chosen": -67.92506408691406,
"logps/rejected": -77.37336730957031,
"loss": 0.4875,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.29286879301071167,
"rewards/margins": 0.8011099100112915,
"rewards/rejected": -0.5082410573959351,
"step": 45
},
{
"epoch": 0.26666666666666666,
"grad_norm": 12.070276260375977,
"learning_rate": 8.771929824561403e-07,
"logits/chosen": -0.8517910242080688,
"logits/rejected": -0.8563734889030457,
"logps/chosen": -92.3311767578125,
"logps/rejected": -102.72621154785156,
"loss": 0.4479,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.06940027326345444,
"rewards/margins": 1.23981511592865,
"rewards/rejected": -1.3092153072357178,
"step": 50
},
{
"epoch": 0.29333333333333333,
"grad_norm": 15.573699951171875,
"learning_rate": 9.649122807017545e-07,
"logits/chosen": -0.9419733285903931,
"logits/rejected": -0.9363008737564087,
"logps/chosen": -76.72134399414062,
"logps/rejected": -82.520751953125,
"loss": 0.527,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.09638428688049316,
"rewards/margins": 0.900995135307312,
"rewards/rejected": -0.8046110272407532,
"step": 55
},
{
"epoch": 0.32,
"grad_norm": 13.533390998840332,
"learning_rate": 1e-06,
"logits/chosen": -0.9142201542854309,
"logits/rejected": -0.9173645973205566,
"logps/chosen": -79.86390686035156,
"logps/rejected": -71.9579086303711,
"loss": 0.5225,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.173385351896286,
"rewards/margins": 1.1095057725906372,
"rewards/rejected": -0.9361203908920288,
"step": 60
},
{
"epoch": 0.3466666666666667,
"grad_norm": 12.57950210571289,
"learning_rate": 1e-06,
"logits/chosen": -0.920464038848877,
"logits/rejected": -0.9172080159187317,
"logps/chosen": -80.74758911132812,
"logps/rejected": -76.41374969482422,
"loss": 0.5248,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.2707827687263489,
"rewards/margins": 0.8655017614364624,
"rewards/rejected": -0.5947191119194031,
"step": 65
},
{
"epoch": 0.37333333333333335,
"grad_norm": 17.20355224609375,
"learning_rate": 1e-06,
"logits/chosen": -0.8898583650588989,
"logits/rejected": -0.8803698420524597,
"logps/chosen": -80.18648529052734,
"logps/rejected": -109.63679504394531,
"loss": 0.3912,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.030266547575592995,
"rewards/margins": 1.6781041622161865,
"rewards/rejected": -1.7083708047866821,
"step": 70
},
{
"epoch": 0.4,
"grad_norm": 14.076622009277344,
"learning_rate": 1e-06,
"logits/chosen": -0.9825113415718079,
"logits/rejected": -0.9686470031738281,
"logps/chosen": -72.63887786865234,
"logps/rejected": -98.7930908203125,
"loss": 0.4875,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.3350667357444763,
"rewards/margins": 1.8856172561645508,
"rewards/rejected": -1.5505504608154297,
"step": 75
},
{
"epoch": 0.4266666666666667,
"grad_norm": 13.67685317993164,
"learning_rate": 1e-06,
"logits/chosen": -0.9678022265434265,
"logits/rejected": -0.9671600461006165,
"logps/chosen": -47.99736785888672,
"logps/rejected": -77.59354400634766,
"loss": 0.4304,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.5903635621070862,
"rewards/margins": 1.1374304294586182,
"rewards/rejected": -0.547066867351532,
"step": 80
},
{
"epoch": 0.4533333333333333,
"grad_norm": 12.650314331054688,
"learning_rate": 1e-06,
"logits/chosen": -0.9299849271774292,
"logits/rejected": -0.9215339422225952,
"logps/chosen": -84.56604766845703,
"logps/rejected": -96.4300308227539,
"loss": 0.3804,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.6257888078689575,
"rewards/margins": 1.4266914129257202,
"rewards/rejected": -0.8009026646614075,
"step": 85
},
{
"epoch": 0.48,
"grad_norm": 12.806818962097168,
"learning_rate": 1e-06,
"logits/chosen": -0.9175389409065247,
"logits/rejected": -0.9051604270935059,
"logps/chosen": -96.42292022705078,
"logps/rejected": -106.83795166015625,
"loss": 0.4908,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.018072601407766342,
"rewards/margins": 1.4671040773391724,
"rewards/rejected": -1.4851768016815186,
"step": 90
},
{
"epoch": 0.5066666666666667,
"grad_norm": 15.225826263427734,
"learning_rate": 1e-06,
"logits/chosen": -0.9506580233573914,
"logits/rejected": -0.9402367472648621,
"logps/chosen": -83.20651245117188,
"logps/rejected": -106.87049865722656,
"loss": 0.4715,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.14173248410224915,
"rewards/margins": 1.5761946439743042,
"rewards/rejected": -1.434462308883667,
"step": 95
},
{
"epoch": 0.5333333333333333,
"grad_norm": 17.279712677001953,
"learning_rate": 1e-06,
"logits/chosen": -0.8913367390632629,
"logits/rejected": -0.8918878436088562,
"logps/chosen": -46.1912956237793,
"logps/rejected": -58.76585006713867,
"loss": 0.4605,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.3738502860069275,
"rewards/margins": 1.1125816106796265,
"rewards/rejected": -0.7387313842773438,
"step": 100
},
{
"epoch": 0.56,
"grad_norm": 15.407678604125977,
"learning_rate": 1e-06,
"logits/chosen": -0.9308265447616577,
"logits/rejected": -0.9138677716255188,
"logps/chosen": -89.54922485351562,
"logps/rejected": -90.88710021972656,
"loss": 0.4469,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.10435737669467926,
"rewards/margins": 1.630505919456482,
"rewards/rejected": -1.5261485576629639,
"step": 105
},
{
"epoch": 0.5866666666666667,
"grad_norm": 14.938887596130371,
"learning_rate": 1e-06,
"logits/chosen": -0.965082049369812,
"logits/rejected": -0.9715268015861511,
"logps/chosen": -42.88313674926758,
"logps/rejected": -61.1027717590332,
"loss": 0.4947,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.0917281061410904,
"rewards/margins": 0.6923992037773132,
"rewards/rejected": -0.6006711721420288,
"step": 110
},
{
"epoch": 0.6133333333333333,
"grad_norm": 15.150486946105957,
"learning_rate": 1e-06,
"logits/chosen": -1.0348211526870728,
"logits/rejected": -1.0350416898727417,
"logps/chosen": -58.58684158325195,
"logps/rejected": -69.80925750732422,
"loss": 0.4301,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 0.741683304309845,
"rewards/margins": 1.3792873620986938,
"rewards/rejected": -0.6376041173934937,
"step": 115
},
{
"epoch": 0.64,
"grad_norm": 15.18857192993164,
"learning_rate": 1e-06,
"logits/chosen": -0.9605352282524109,
"logits/rejected": -0.9632300138473511,
"logps/chosen": -66.8796615600586,
"logps/rejected": -71.91226959228516,
"loss": 0.4272,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.334277480840683,
"rewards/margins": 1.3439124822616577,
"rewards/rejected": -1.0096348524093628,
"step": 120
},
{
"epoch": 0.6666666666666666,
"grad_norm": 17.82939338684082,
"learning_rate": 1e-06,
"logits/chosen": -0.9578378796577454,
"logits/rejected": -0.9434881210327148,
"logps/chosen": -91.96797943115234,
"logps/rejected": -118.94937896728516,
"loss": 0.469,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.11757056415081024,
"rewards/margins": 2.0075435638427734,
"rewards/rejected": -1.8899730443954468,
"step": 125
},
{
"epoch": 0.6933333333333334,
"grad_norm": 14.026658058166504,
"learning_rate": 1e-06,
"logits/chosen": -0.9356282353401184,
"logits/rejected": -0.9223386645317078,
"logps/chosen": -84.6489486694336,
"logps/rejected": -97.97027587890625,
"loss": 0.4071,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.1822044849395752,
"rewards/margins": 1.0983351469039917,
"rewards/rejected": -1.2805397510528564,
"step": 130
},
{
"epoch": 0.72,
"grad_norm": 16.48579978942871,
"learning_rate": 1e-06,
"logits/chosen": -0.9593449831008911,
"logits/rejected": -0.9539397358894348,
"logps/chosen": -87.32046508789062,
"logps/rejected": -104.6070327758789,
"loss": 0.4735,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 0.11158242076635361,
"rewards/margins": 1.4765504598617554,
"rewards/rejected": -1.3649680614471436,
"step": 135
},
{
"epoch": 0.7466666666666667,
"grad_norm": 15.402965545654297,
"learning_rate": 1e-06,
"logits/chosen": -0.8937503695487976,
"logits/rejected": -0.8833414912223816,
"logps/chosen": -56.55735397338867,
"logps/rejected": -64.20155334472656,
"loss": 0.444,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.04381413012742996,
"rewards/margins": 1.1586054563522339,
"rewards/rejected": -1.202419638633728,
"step": 140
},
{
"epoch": 0.7733333333333333,
"grad_norm": 13.747047424316406,
"learning_rate": 1e-06,
"logits/chosen": -0.9597466588020325,
"logits/rejected": -0.9614435434341431,
"logps/chosen": -63.605621337890625,
"logps/rejected": -67.96479797363281,
"loss": 0.3972,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.4212762713432312,
"rewards/margins": 0.9387065768241882,
"rewards/rejected": -0.5174302458763123,
"step": 145
},
{
"epoch": 0.8,
"grad_norm": 12.082706451416016,
"learning_rate": 1e-06,
"logits/chosen": -0.8831542730331421,
"logits/rejected": -0.8782299757003784,
"logps/chosen": -89.6635971069336,
"logps/rejected": -102.15290832519531,
"loss": 0.4195,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.23645827174186707,
"rewards/margins": 1.130458116531372,
"rewards/rejected": -1.3669164180755615,
"step": 150
},
{
"epoch": 0.8266666666666667,
"grad_norm": 11.634166717529297,
"learning_rate": 1e-06,
"logits/chosen": -0.9502588510513306,
"logits/rejected": -0.9416471719741821,
"logps/chosen": -68.0828857421875,
"logps/rejected": -100.38713073730469,
"loss": 0.3315,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.19413085281848907,
"rewards/margins": 2.068560838699341,
"rewards/rejected": -1.8744302988052368,
"step": 155
},
{
"epoch": 0.8533333333333334,
"grad_norm": 15.321958541870117,
"learning_rate": 1e-06,
"logits/chosen": -1.0011286735534668,
"logits/rejected": -1.0069797039031982,
"logps/chosen": -58.62431716918945,
"logps/rejected": -81.33189392089844,
"loss": 0.4751,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.06445435434579849,
"rewards/margins": 1.0574184656143188,
"rewards/rejected": -0.9929640889167786,
"step": 160
},
{
"epoch": 0.88,
"grad_norm": 13.829776763916016,
"learning_rate": 1e-06,
"logits/chosen": -0.9930332899093628,
"logits/rejected": -0.9925543069839478,
"logps/chosen": -65.17711639404297,
"logps/rejected": -79.74466705322266,
"loss": 0.4647,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.07586468756198883,
"rewards/margins": 0.9637776613235474,
"rewards/rejected": -0.8879130482673645,
"step": 165
},
{
"epoch": 0.9066666666666666,
"grad_norm": 15.529192924499512,
"learning_rate": 1e-06,
"logits/chosen": -0.9089356660842896,
"logits/rejected": -0.9082063436508179,
"logps/chosen": -73.97811126708984,
"logps/rejected": -95.45903778076172,
"loss": 0.4392,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.33137935400009155,
"rewards/margins": 1.3767789602279663,
"rewards/rejected": -1.7081581354141235,
"step": 170
},
{
"epoch": 0.9333333333333333,
"grad_norm": 16.684486389160156,
"learning_rate": 1e-06,
"logits/chosen": -0.97789067029953,
"logits/rejected": -0.9672166705131531,
"logps/chosen": -68.05913543701172,
"logps/rejected": -76.677490234375,
"loss": 0.3687,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.15937092900276184,
"rewards/margins": 1.4713166952133179,
"rewards/rejected": -1.3119456768035889,
"step": 175
},
{
"epoch": 0.96,
"grad_norm": 10.531698226928711,
"learning_rate": 1e-06,
"logits/chosen": -0.9809746742248535,
"logits/rejected": -0.963492214679718,
"logps/chosen": -82.75611877441406,
"logps/rejected": -115.19696044921875,
"loss": 0.3942,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1854589432477951,
"rewards/margins": 2.5827653408050537,
"rewards/rejected": -2.397306442260742,
"step": 180
},
{
"epoch": 0.9866666666666667,
"grad_norm": 13.4787015914917,
"learning_rate": 1e-06,
"logits/chosen": -0.96388179063797,
"logits/rejected": -0.9607783555984497,
"logps/chosen": -73.0042953491211,
"logps/rejected": -70.76305389404297,
"loss": 0.3791,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.26272085309028625,
"rewards/margins": 1.4707610607147217,
"rewards/rejected": -1.2080401182174683,
"step": 185
},
{
"epoch": 1.0133333333333334,
"grad_norm": 10.95159912109375,
"learning_rate": 1e-06,
"logits/chosen": -0.952096164226532,
"logits/rejected": -0.9481242895126343,
"logps/chosen": -65.6482162475586,
"logps/rejected": -94.6727523803711,
"loss": 0.3602,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.6763356924057007,
"rewards/margins": 1.9549248218536377,
"rewards/rejected": -1.2785890102386475,
"step": 190
},
{
"epoch": 1.04,
"grad_norm": 4.499619960784912,
"learning_rate": 1e-06,
"logits/chosen": -0.9216955900192261,
"logits/rejected": -0.9177412986755371,
"logps/chosen": -57.536109924316406,
"logps/rejected": -89.177734375,
"loss": 0.1967,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.7885057330131531,
"rewards/margins": 1.7107608318328857,
"rewards/rejected": -0.9222550392150879,
"step": 195
},
{
"epoch": 1.0666666666666667,
"grad_norm": 7.151574611663818,
"learning_rate": 1e-06,
"logits/chosen": -0.9569188952445984,
"logits/rejected": -0.9455480575561523,
"logps/chosen": -53.17787551879883,
"logps/rejected": -78.4444351196289,
"loss": 0.2083,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.8187260627746582,
"rewards/margins": 2.6014370918273926,
"rewards/rejected": -1.7827112674713135,
"step": 200
},
{
"epoch": 1.0933333333333333,
"grad_norm": 7.790274620056152,
"learning_rate": 1e-06,
"logits/chosen": -0.9871414303779602,
"logits/rejected": -0.9741198420524597,
"logps/chosen": -88.66045379638672,
"logps/rejected": -130.34814453125,
"loss": 0.1628,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.14811792969703674,
"rewards/margins": 3.1860337257385254,
"rewards/rejected": -3.0379159450531006,
"step": 205
},
{
"epoch": 1.12,
"grad_norm": 5.487607955932617,
"learning_rate": 1e-06,
"logits/chosen": -1.034203052520752,
"logits/rejected": -1.020320177078247,
"logps/chosen": -71.07328796386719,
"logps/rejected": -107.5151138305664,
"loss": 0.1747,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.050754476338624954,
"rewards/margins": 3.4751205444335938,
"rewards/rejected": -3.4243664741516113,
"step": 210
},
{
"epoch": 1.1466666666666667,
"grad_norm": 4.493232727050781,
"learning_rate": 1e-06,
"logits/chosen": -0.9926775693893433,
"logits/rejected": -0.9901949167251587,
"logps/chosen": -85.27278900146484,
"logps/rejected": -123.84268951416016,
"loss": 0.1556,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.030883288010954857,
"rewards/margins": 3.6463749408721924,
"rewards/rejected": -3.677258253097534,
"step": 215
},
{
"epoch": 1.1733333333333333,
"grad_norm": 8.204482078552246,
"learning_rate": 1e-06,
"logits/chosen": -1.012113332748413,
"logits/rejected": -1.012290596961975,
"logps/chosen": -85.76399993896484,
"logps/rejected": -119.85150146484375,
"loss": 0.1645,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.43645215034484863,
"rewards/margins": 3.1804490089416504,
"rewards/rejected": -3.616900682449341,
"step": 220
},
{
"epoch": 1.2,
"grad_norm": 8.91741943359375,
"learning_rate": 1e-06,
"logits/chosen": -0.9512453079223633,
"logits/rejected": -0.9447379112243652,
"logps/chosen": -99.74169921875,
"logps/rejected": -125.4989013671875,
"loss": 0.2412,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -1.6342785358428955,
"rewards/margins": 3.1997177600860596,
"rewards/rejected": -4.833995819091797,
"step": 225
},
{
"epoch": 1.2266666666666666,
"grad_norm": 8.655539512634277,
"learning_rate": 1e-06,
"logits/chosen": -0.9638906717300415,
"logits/rejected": -0.9550336599349976,
"logps/chosen": -90.74691009521484,
"logps/rejected": -129.67747497558594,
"loss": 0.1931,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.6765635013580322,
"rewards/margins": 3.924309492111206,
"rewards/rejected": -4.600872993469238,
"step": 230
},
{
"epoch": 1.2533333333333334,
"grad_norm": 4.054169654846191,
"learning_rate": 1e-06,
"logits/chosen": -1.030977487564087,
"logits/rejected": -1.0393321514129639,
"logps/chosen": -54.60747528076172,
"logps/rejected": -86.23637390136719,
"loss": 0.2012,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 0.5984514951705933,
"rewards/margins": 3.5396370887756348,
"rewards/rejected": -2.941185712814331,
"step": 235
},
{
"epoch": 1.28,
"grad_norm": 6.45543909072876,
"learning_rate": 1e-06,
"logits/chosen": -1.000688910484314,
"logits/rejected": -0.9959505200386047,
"logps/chosen": -51.59619140625,
"logps/rejected": -78.5389633178711,
"loss": 0.1828,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.38332298398017883,
"rewards/margins": 2.686915874481201,
"rewards/rejected": -2.3035929203033447,
"step": 240
},
{
"epoch": 1.3066666666666666,
"grad_norm": 7.938249111175537,
"learning_rate": 1e-06,
"logits/chosen": -0.9679718017578125,
"logits/rejected": -0.9519965052604675,
"logps/chosen": -91.22569274902344,
"logps/rejected": -120.7100601196289,
"loss": 0.1685,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.6173192262649536,
"rewards/margins": 3.579179286956787,
"rewards/rejected": -4.196498870849609,
"step": 245
},
{
"epoch": 1.3333333333333333,
"grad_norm": 5.844974040985107,
"learning_rate": 1e-06,
"logits/chosen": -0.9622223973274231,
"logits/rejected": -0.9677648544311523,
"logps/chosen": -92.24385070800781,
"logps/rejected": -129.18057250976562,
"loss": 0.1531,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.8045636415481567,
"rewards/margins": 2.904844284057617,
"rewards/rejected": -3.7094082832336426,
"step": 250
},
{
"epoch": 1.3599999999999999,
"grad_norm": 8.192586898803711,
"learning_rate": 1e-06,
"logits/chosen": -0.9564617276191711,
"logits/rejected": -0.953482985496521,
"logps/chosen": -92.21900939941406,
"logps/rejected": -115.93299865722656,
"loss": 0.1831,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.9339162707328796,
"rewards/margins": 3.1692094802856445,
"rewards/rejected": -4.103126049041748,
"step": 255
},
{
"epoch": 1.3866666666666667,
"grad_norm": 11.612703323364258,
"learning_rate": 1e-06,
"logits/chosen": -0.9484894871711731,
"logits/rejected": -0.9465099573135376,
"logps/chosen": -88.50215148925781,
"logps/rejected": -122.58497619628906,
"loss": 0.2373,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.8389456868171692,
"rewards/margins": 4.004569053649902,
"rewards/rejected": -4.843514442443848,
"step": 260
},
{
"epoch": 1.4133333333333333,
"grad_norm": 3.950273036956787,
"learning_rate": 1e-06,
"logits/chosen": -1.0333240032196045,
"logits/rejected": -1.026308298110962,
"logps/chosen": -81.88277435302734,
"logps/rejected": -115.6543960571289,
"loss": 0.1738,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.251314640045166,
"rewards/margins": 3.181288242340088,
"rewards/rejected": -3.432603120803833,
"step": 265
},
{
"epoch": 1.44,
"grad_norm": 9.259121894836426,
"learning_rate": 1e-06,
"logits/chosen": -0.9529396891593933,
"logits/rejected": -0.9476019740104675,
"logps/chosen": -93.74283599853516,
"logps/rejected": -117.82435607910156,
"loss": 0.1717,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.17003877460956573,
"rewards/margins": 3.552433490753174,
"rewards/rejected": -3.7224719524383545,
"step": 270
},
{
"epoch": 1.4666666666666668,
"grad_norm": 17.85709571838379,
"learning_rate": 1e-06,
"logits/chosen": -1.0109868049621582,
"logits/rejected": -1.0117450952529907,
"logps/chosen": -93.20274353027344,
"logps/rejected": -124.9032211303711,
"loss": 0.2349,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.15422244369983673,
"rewards/margins": 3.6027817726135254,
"rewards/rejected": -3.757004499435425,
"step": 275
},
{
"epoch": 1.4933333333333334,
"grad_norm": 12.522117614746094,
"learning_rate": 1e-06,
"logits/chosen": -1.0297268629074097,
"logits/rejected": -1.0234237909317017,
"logps/chosen": -50.927391052246094,
"logps/rejected": -76.98394775390625,
"loss": 0.2888,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 0.37279123067855835,
"rewards/margins": 2.9437782764434814,
"rewards/rejected": -2.5709872245788574,
"step": 280
},
{
"epoch": 1.52,
"grad_norm": 18.62285804748535,
"learning_rate": 1e-06,
"logits/chosen": -0.97401362657547,
"logits/rejected": -0.9802001714706421,
"logps/chosen": -63.81083297729492,
"logps/rejected": -86.34494018554688,
"loss": 0.2402,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.40913981199264526,
"rewards/margins": 2.3060526847839355,
"rewards/rejected": -2.7151923179626465,
"step": 285
},
{
"epoch": 1.5466666666666666,
"grad_norm": 14.160591125488281,
"learning_rate": 1e-06,
"logits/chosen": -1.0169212818145752,
"logits/rejected": -1.0096107721328735,
"logps/chosen": -66.8814926147461,
"logps/rejected": -97.15755462646484,
"loss": 0.1656,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 0.21899671852588654,
"rewards/margins": 3.427858352661133,
"rewards/rejected": -3.2088615894317627,
"step": 290
},
{
"epoch": 1.5733333333333333,
"grad_norm": 5.416359901428223,
"learning_rate": 1e-06,
"logits/chosen": -0.9776760935783386,
"logits/rejected": -0.9770166277885437,
"logps/chosen": -67.19379425048828,
"logps/rejected": -101.93242645263672,
"loss": 0.2106,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.027063405141234398,
"rewards/margins": 3.1595966815948486,
"rewards/rejected": -3.1866602897644043,
"step": 295
},
{
"epoch": 1.6,
"grad_norm": 7.018192291259766,
"learning_rate": 1e-06,
"logits/chosen": -1.023803949356079,
"logits/rejected": -1.0103000402450562,
"logps/chosen": -63.242576599121094,
"logps/rejected": -96.89302062988281,
"loss": 0.1886,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 0.034027088433504105,
"rewards/margins": 3.0590438842773438,
"rewards/rejected": -3.0250167846679688,
"step": 300
},
{
"epoch": 1.6266666666666667,
"grad_norm": 7.443521499633789,
"learning_rate": 1e-06,
"logits/chosen": -1.0116432905197144,
"logits/rejected": -1.023240327835083,
"logps/chosen": -84.08271026611328,
"logps/rejected": -159.3538360595703,
"loss": 0.1687,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.561992347240448,
"rewards/margins": 4.012616157531738,
"rewards/rejected": -4.57460880279541,
"step": 305
},
{
"epoch": 1.6533333333333333,
"grad_norm": 18.155902862548828,
"learning_rate": 1e-06,
"logits/chosen": -1.025331974029541,
"logits/rejected": -1.021927833557129,
"logps/chosen": -73.14906311035156,
"logps/rejected": -121.94499206542969,
"loss": 0.1807,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.2425353080034256,
"rewards/margins": 3.504972457885742,
"rewards/rejected": -3.7475082874298096,
"step": 310
},
{
"epoch": 1.6800000000000002,
"grad_norm": 11.773829460144043,
"learning_rate": 1e-06,
"logits/chosen": -0.9748228192329407,
"logits/rejected": -0.9767504930496216,
"logps/chosen": -88.07221984863281,
"logps/rejected": -147.24069213867188,
"loss": 0.1454,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.8688488006591797,
"rewards/margins": 4.263859748840332,
"rewards/rejected": -5.1327080726623535,
"step": 315
},
{
"epoch": 1.7066666666666666,
"grad_norm": 11.361552238464355,
"learning_rate": 1e-06,
"logits/chosen": -0.9369309544563293,
"logits/rejected": -0.9380790591239929,
"logps/chosen": -96.55718231201172,
"logps/rejected": -119.29121398925781,
"loss": 0.2044,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.5562452077865601,
"rewards/margins": 3.295253038406372,
"rewards/rejected": -3.8514983654022217,
"step": 320
},
{
"epoch": 1.7333333333333334,
"grad_norm": 10.781094551086426,
"learning_rate": 1e-06,
"logits/chosen": -0.99322509765625,
"logits/rejected": -0.9963001012802124,
"logps/chosen": -94.03239440917969,
"logps/rejected": -129.18307495117188,
"loss": 0.228,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.0290601253509521,
"rewards/margins": 3.1235461235046387,
"rewards/rejected": -4.152606010437012,
"step": 325
},
{
"epoch": 1.76,
"grad_norm": 13.973287582397461,
"learning_rate": 1e-06,
"logits/chosen": -0.9588006138801575,
"logits/rejected": -0.9579130411148071,
"logps/chosen": -93.3379135131836,
"logps/rejected": -145.30789184570312,
"loss": 0.1606,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.4644792675971985,
"rewards/margins": 4.343870639801025,
"rewards/rejected": -4.808349609375,
"step": 330
},
{
"epoch": 1.7866666666666666,
"grad_norm": 5.7010931968688965,
"learning_rate": 1e-06,
"logits/chosen": -0.9664214253425598,
"logits/rejected": -0.9622098207473755,
"logps/chosen": -77.8204574584961,
"logps/rejected": -139.20196533203125,
"loss": 0.1036,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.5566782355308533,
"rewards/margins": 4.256915092468262,
"rewards/rejected": -4.81359338760376,
"step": 335
},
{
"epoch": 1.8133333333333335,
"grad_norm": 15.839357376098633,
"learning_rate": 1e-06,
"logits/chosen": -0.9637653231620789,
"logits/rejected": -0.9616791009902954,
"logps/chosen": -97.48753356933594,
"logps/rejected": -127.541259765625,
"loss": 0.1682,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.562965452671051,
"rewards/margins": 3.636152744293213,
"rewards/rejected": -4.199118614196777,
"step": 340
},
{
"epoch": 1.8399999999999999,
"grad_norm": 7.101598739624023,
"learning_rate": 1e-06,
"logits/chosen": -1.0025413036346436,
"logits/rejected": -1.0012590885162354,
"logps/chosen": -79.85330963134766,
"logps/rejected": -98.39595794677734,
"loss": 0.1832,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 0.4088008403778076,
"rewards/margins": 3.5112171173095703,
"rewards/rejected": -3.102416515350342,
"step": 345
},
{
"epoch": 1.8666666666666667,
"grad_norm": 12.390422821044922,
"learning_rate": 1e-06,
"logits/chosen": -1.0001270771026611,
"logits/rejected": -1.0041121244430542,
"logps/chosen": -47.137535095214844,
"logps/rejected": -87.8708267211914,
"loss": 0.1838,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.07667568325996399,
"rewards/margins": 3.5933804512023926,
"rewards/rejected": -3.670056104660034,
"step": 350
},
{
"epoch": 1.8933333333333333,
"grad_norm": 6.624851226806641,
"learning_rate": 1e-06,
"logits/chosen": -1.0326136350631714,
"logits/rejected": -1.0377240180969238,
"logps/chosen": -78.13581848144531,
"logps/rejected": -121.1558837890625,
"loss": 0.1175,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.22754593193531036,
"rewards/margins": 3.7420687675476074,
"rewards/rejected": -3.9696147441864014,
"step": 355
},
{
"epoch": 1.92,
"grad_norm": 9.035065650939941,
"learning_rate": 1e-06,
"logits/chosen": -0.959603488445282,
"logits/rejected": -0.9709904789924622,
"logps/chosen": -76.04480743408203,
"logps/rejected": -128.50241088867188,
"loss": 0.1593,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.2988315522670746,
"rewards/margins": 3.922844648361206,
"rewards/rejected": -4.221675395965576,
"step": 360
},
{
"epoch": 1.9466666666666668,
"grad_norm": 5.306190013885498,
"learning_rate": 1e-06,
"logits/chosen": -0.9443675875663757,
"logits/rejected": -0.9363471865653992,
"logps/chosen": -99.31537628173828,
"logps/rejected": -136.64498901367188,
"loss": 0.1425,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -1.223646879196167,
"rewards/margins": 3.9859280586242676,
"rewards/rejected": -5.209575653076172,
"step": 365
},
{
"epoch": 1.9733333333333334,
"grad_norm": 12.788728713989258,
"learning_rate": 1e-06,
"logits/chosen": -0.9819198846817017,
"logits/rejected": -0.9810832142829895,
"logps/chosen": -79.12177276611328,
"logps/rejected": -109.90110778808594,
"loss": 0.209,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.3604089915752411,
"rewards/margins": 3.0035877227783203,
"rewards/rejected": -3.3639965057373047,
"step": 370
},
{
"epoch": 2.0,
"grad_norm": 13.587727546691895,
"learning_rate": 1e-06,
"logits/chosen": -0.9736833572387695,
"logits/rejected": -0.9709518551826477,
"logps/chosen": -111.94571685791016,
"logps/rejected": -123.80615234375,
"loss": 0.1791,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.4601728916168213,
"rewards/margins": 3.195746660232544,
"rewards/rejected": -4.655919551849365,
"step": 375
},
{
"epoch": 2.026666666666667,
"grad_norm": 2.3324151039123535,
"learning_rate": 1e-06,
"logits/chosen": -0.9637060165405273,
"logits/rejected": -0.9625332951545715,
"logps/chosen": -64.67164611816406,
"logps/rejected": -107.06028747558594,
"loss": 0.0579,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.13231408596038818,
"rewards/margins": 4.237712383270264,
"rewards/rejected": -4.370026588439941,
"step": 380
},
{
"epoch": 2.0533333333333332,
"grad_norm": 2.0872855186462402,
"learning_rate": 1e-06,
"logits/chosen": -0.9391244649887085,
"logits/rejected": -0.9333323240280151,
"logps/chosen": -53.344078063964844,
"logps/rejected": -118.3141098022461,
"loss": 0.0571,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.09113900363445282,
"rewards/margins": 4.982771873474121,
"rewards/rejected": -5.073911190032959,
"step": 385
},
{
"epoch": 2.08,
"grad_norm": 3.842839479446411,
"learning_rate": 1e-06,
"logits/chosen": -1.0063703060150146,
"logits/rejected": -1.0035226345062256,
"logps/chosen": -85.92240905761719,
"logps/rejected": -147.6622314453125,
"loss": 0.0788,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.7325506210327148,
"rewards/margins": 5.368863105773926,
"rewards/rejected": -6.101413249969482,
"step": 390
},
{
"epoch": 2.1066666666666665,
"grad_norm": 2.944322347640991,
"learning_rate": 1e-06,
"logits/chosen": -0.9857255816459656,
"logits/rejected": -0.9889955520629883,
"logps/chosen": -86.60405731201172,
"logps/rejected": -146.21925354003906,
"loss": 0.0604,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.44693589210510254,
"rewards/margins": 5.308945655822754,
"rewards/rejected": -5.755881309509277,
"step": 395
},
{
"epoch": 2.1333333333333333,
"grad_norm": 1.7155674695968628,
"learning_rate": 1e-06,
"logits/chosen": -0.994691014289856,
"logits/rejected": -0.9987820386886597,
"logps/chosen": -74.14541625976562,
"logps/rejected": -137.02389526367188,
"loss": 0.0644,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.06674417108297348,
"rewards/margins": 6.20205020904541,
"rewards/rejected": -6.268795013427734,
"step": 400
},
{
"epoch": 2.16,
"grad_norm": 1.7242580652236938,
"learning_rate": 1e-06,
"logits/chosen": -0.9776423573493958,
"logits/rejected": -0.9695422053337097,
"logps/chosen": -79.72328186035156,
"logps/rejected": -142.98239135742188,
"loss": 0.0439,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.8770833015441895,
"rewards/margins": 6.5918288230896,
"rewards/rejected": -7.4689130783081055,
"step": 405
},
{
"epoch": 2.1866666666666665,
"grad_norm": 2.5285656452178955,
"learning_rate": 1e-06,
"logits/chosen": -0.943740963935852,
"logits/rejected": -0.9352688789367676,
"logps/chosen": -92.97003936767578,
"logps/rejected": -145.69248962402344,
"loss": 0.0871,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.433349370956421,
"rewards/margins": 5.817093372344971,
"rewards/rejected": -7.250443458557129,
"step": 410
},
{
"epoch": 2.2133333333333334,
"grad_norm": 6.4492597579956055,
"learning_rate": 1e-06,
"logits/chosen": -0.9524329304695129,
"logits/rejected": -0.9578531384468079,
"logps/chosen": -79.1036148071289,
"logps/rejected": -126.85892486572266,
"loss": 0.0692,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -1.3942753076553345,
"rewards/margins": 5.110877990722656,
"rewards/rejected": -6.505153656005859,
"step": 415
},
{
"epoch": 2.24,
"grad_norm": 3.9530835151672363,
"learning_rate": 1e-06,
"logits/chosen": -0.9325224757194519,
"logits/rejected": -0.9292308688163757,
"logps/chosen": -92.8379135131836,
"logps/rejected": -124.98873138427734,
"loss": 0.0744,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.4230420589447021,
"rewards/margins": 4.427516937255859,
"rewards/rejected": -5.850558757781982,
"step": 420
},
{
"epoch": 2.2666666666666666,
"grad_norm": 3.244112491607666,
"learning_rate": 1e-06,
"logits/chosen": -0.9953921437263489,
"logits/rejected": -0.9901760220527649,
"logps/chosen": -90.95811462402344,
"logps/rejected": -153.2058563232422,
"loss": 0.0505,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.1016261577606201,
"rewards/margins": 5.2358317375183105,
"rewards/rejected": -6.33745813369751,
"step": 425
},
{
"epoch": 2.2933333333333334,
"grad_norm": 3.9486782550811768,
"learning_rate": 1e-06,
"logits/chosen": -0.9572321176528931,
"logits/rejected": -0.963973343372345,
"logps/chosen": -107.50956726074219,
"logps/rejected": -146.10031127929688,
"loss": 0.0639,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.6813853979110718,
"rewards/margins": 5.3139801025390625,
"rewards/rejected": -5.995365619659424,
"step": 430
},
{
"epoch": 2.32,
"grad_norm": 3.333291530609131,
"learning_rate": 1e-06,
"logits/chosen": -0.9470894932746887,
"logits/rejected": -0.9312503933906555,
"logps/chosen": -92.68052673339844,
"logps/rejected": -140.1178436279297,
"loss": 0.0572,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.656975507736206,
"rewards/margins": 5.908856391906738,
"rewards/rejected": -6.565831661224365,
"step": 435
},
{
"epoch": 2.3466666666666667,
"grad_norm": 9.420110702514648,
"learning_rate": 1e-06,
"logits/chosen": -0.905168354511261,
"logits/rejected": -0.905125617980957,
"logps/chosen": -82.85682678222656,
"logps/rejected": -134.2998046875,
"loss": 0.1159,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -1.2471545934677124,
"rewards/margins": 5.145517826080322,
"rewards/rejected": -6.392672538757324,
"step": 440
},
{
"epoch": 2.3733333333333335,
"grad_norm": 2.3797969818115234,
"learning_rate": 1e-06,
"logits/chosen": -0.9869009256362915,
"logits/rejected": -0.9733754396438599,
"logps/chosen": -124.99552917480469,
"logps/rejected": -176.72329711914062,
"loss": 0.0451,
"rewards/accuracies": 1.0,
"rewards/chosen": -2.0458180904388428,
"rewards/margins": 6.536404609680176,
"rewards/rejected": -8.582223892211914,
"step": 445
},
{
"epoch": 2.4,
"grad_norm": 6.350354194641113,
"learning_rate": 1e-06,
"logits/chosen": -0.9659305810928345,
"logits/rejected": -0.9653702974319458,
"logps/chosen": -71.59037017822266,
"logps/rejected": -125.0820083618164,
"loss": 0.0761,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.7029816508293152,
"rewards/margins": 5.654323577880859,
"rewards/rejected": -6.35730504989624,
"step": 450
},
{
"epoch": 2.4266666666666667,
"grad_norm": 3.5830423831939697,
"learning_rate": 1e-06,
"logits/chosen": -0.963516891002655,
"logits/rejected": -0.9658119082450867,
"logps/chosen": -102.05826568603516,
"logps/rejected": -157.795654296875,
"loss": 0.0456,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.9112430810928345,
"rewards/margins": 5.601228713989258,
"rewards/rejected": -6.512471675872803,
"step": 455
},
{
"epoch": 2.453333333333333,
"grad_norm": 1.1380105018615723,
"learning_rate": 1e-06,
"logits/chosen": -0.9547368884086609,
"logits/rejected": -0.9605048298835754,
"logps/chosen": -108.07603454589844,
"logps/rejected": -171.28262329101562,
"loss": 0.0595,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.361283779144287,
"rewards/margins": 6.42099142074585,
"rewards/rejected": -7.782275199890137,
"step": 460
},
{
"epoch": 2.48,
"grad_norm": 1.1063487529754639,
"learning_rate": 1e-06,
"logits/chosen": -0.914827823638916,
"logits/rejected": -0.9137645959854126,
"logps/chosen": -85.6763687133789,
"logps/rejected": -160.23040771484375,
"loss": 0.0519,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -1.18534255027771,
"rewards/margins": 6.746701240539551,
"rewards/rejected": -7.932042598724365,
"step": 465
},
{
"epoch": 2.506666666666667,
"grad_norm": 1.4177899360656738,
"learning_rate": 1e-06,
"logits/chosen": -0.941451370716095,
"logits/rejected": -0.9365523457527161,
"logps/chosen": -77.06974029541016,
"logps/rejected": -124.00151062011719,
"loss": 0.055,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.9638670086860657,
"rewards/margins": 4.646106719970703,
"rewards/rejected": -5.609973907470703,
"step": 470
},
{
"epoch": 2.533333333333333,
"grad_norm": 3.090573787689209,
"learning_rate": 1e-06,
"logits/chosen": -0.9199577569961548,
"logits/rejected": -0.9237100481987,
"logps/chosen": -86.69384765625,
"logps/rejected": -112.1557846069336,
"loss": 0.0502,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.1065205335617065,
"rewards/margins": 4.314031600952148,
"rewards/rejected": -5.4205522537231445,
"step": 475
},
{
"epoch": 2.56,
"grad_norm": 7.529189109802246,
"learning_rate": 1e-06,
"logits/chosen": -0.9499224424362183,
"logits/rejected": -0.9461296200752258,
"logps/chosen": -98.30276489257812,
"logps/rejected": -169.53909301757812,
"loss": 0.0692,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.8648740649223328,
"rewards/margins": 7.417494773864746,
"rewards/rejected": -8.282369613647461,
"step": 480
},
{
"epoch": 2.586666666666667,
"grad_norm": 4.119349479675293,
"learning_rate": 1e-06,
"logits/chosen": -0.9638736844062805,
"logits/rejected": -0.9807848930358887,
"logps/chosen": -82.62672424316406,
"logps/rejected": -177.9151611328125,
"loss": 0.0633,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.7524802684783936,
"rewards/margins": 7.268258094787598,
"rewards/rejected": -9.02073860168457,
"step": 485
},
{
"epoch": 2.6133333333333333,
"grad_norm": 3.560368299484253,
"learning_rate": 1e-06,
"logits/chosen": -0.9251823425292969,
"logits/rejected": -0.9247922897338867,
"logps/chosen": -87.5182876586914,
"logps/rejected": -169.91879272460938,
"loss": 0.1587,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.2214720249176025,
"rewards/margins": 6.715635776519775,
"rewards/rejected": -7.937108039855957,
"step": 490
},
{
"epoch": 2.64,
"grad_norm": 1.6422832012176514,
"learning_rate": 1e-06,
"logits/chosen": -0.9349163174629211,
"logits/rejected": -0.9354310035705566,
"logps/chosen": -70.24932861328125,
"logps/rejected": -146.46348571777344,
"loss": 0.0691,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.6841042637825012,
"rewards/margins": 6.8183794021606445,
"rewards/rejected": -7.502483367919922,
"step": 495
},
{
"epoch": 2.6666666666666665,
"grad_norm": 2.153717041015625,
"learning_rate": 1e-06,
"logits/chosen": -0.9372892379760742,
"logits/rejected": -0.9442557096481323,
"logps/chosen": -74.71461486816406,
"logps/rejected": -149.56219482421875,
"loss": 0.0355,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.513080894947052,
"rewards/margins": 6.21486759185791,
"rewards/rejected": -6.727948188781738,
"step": 500
},
{
"epoch": 2.6933333333333334,
"grad_norm": 2.593008279800415,
"learning_rate": 1e-06,
"logits/chosen": -0.9294622540473938,
"logits/rejected": -0.9304571151733398,
"logps/chosen": -90.94770050048828,
"logps/rejected": -147.8937225341797,
"loss": 0.0517,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.548829197883606,
"rewards/margins": 6.368669033050537,
"rewards/rejected": -7.917498588562012,
"step": 505
},
{
"epoch": 2.7199999999999998,
"grad_norm": 12.588436126708984,
"learning_rate": 1e-06,
"logits/chosen": -0.9092562794685364,
"logits/rejected": -0.9060165286064148,
"logps/chosen": -83.18638610839844,
"logps/rejected": -130.56942749023438,
"loss": 0.0565,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.80876886844635,
"rewards/margins": 5.5646467208862305,
"rewards/rejected": -7.373415946960449,
"step": 510
},
{
"epoch": 2.7466666666666666,
"grad_norm": 3.6610798835754395,
"learning_rate": 1e-06,
"logits/chosen": -1.00751531124115,
"logits/rejected": -1.0096744298934937,
"logps/chosen": -91.13447570800781,
"logps/rejected": -148.82164001464844,
"loss": 0.0285,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.2526090145111084,
"rewards/margins": 6.453673362731934,
"rewards/rejected": -7.706282138824463,
"step": 515
},
{
"epoch": 2.7733333333333334,
"grad_norm": 1.0663306713104248,
"learning_rate": 1e-06,
"logits/chosen": -0.9299656748771667,
"logits/rejected": -0.9350908994674683,
"logps/chosen": -94.07073211669922,
"logps/rejected": -164.70330810546875,
"loss": 0.0384,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.654295563697815,
"rewards/margins": 6.949212074279785,
"rewards/rejected": -8.603507995605469,
"step": 520
},
{
"epoch": 2.8,
"grad_norm": 6.753617286682129,
"learning_rate": 1e-06,
"logits/chosen": -0.9451917409896851,
"logits/rejected": -0.9450856447219849,
"logps/chosen": -101.37158203125,
"logps/rejected": -149.12391662597656,
"loss": 0.0611,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -2.0039877891540527,
"rewards/margins": 6.462770938873291,
"rewards/rejected": -8.46675968170166,
"step": 525
},
{
"epoch": 2.8266666666666667,
"grad_norm": 9.86711597442627,
"learning_rate": 1e-06,
"logits/chosen": -0.9280322194099426,
"logits/rejected": -0.9299230575561523,
"logps/chosen": -94.41612243652344,
"logps/rejected": -168.96092224121094,
"loss": 0.0731,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.973374366760254,
"rewards/margins": 6.930324554443359,
"rewards/rejected": -8.90369987487793,
"step": 530
},
{
"epoch": 2.8533333333333335,
"grad_norm": 10.411066055297852,
"learning_rate": 1e-06,
"logits/chosen": -0.9307094812393188,
"logits/rejected": -0.9296162724494934,
"logps/chosen": -87.57542419433594,
"logps/rejected": -134.86257934570312,
"loss": 0.0633,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.4023369550704956,
"rewards/margins": 5.565145969390869,
"rewards/rejected": -6.967482566833496,
"step": 535
},
{
"epoch": 2.88,
"grad_norm": 1.9066485166549683,
"learning_rate": 1e-06,
"logits/chosen": -0.9522913098335266,
"logits/rejected": -0.9543063044548035,
"logps/chosen": -90.59437561035156,
"logps/rejected": -152.72274780273438,
"loss": 0.1226,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -2.1479926109313965,
"rewards/margins": 6.265242576599121,
"rewards/rejected": -8.41323471069336,
"step": 540
},
{
"epoch": 2.9066666666666667,
"grad_norm": 1.3979381322860718,
"learning_rate": 1e-06,
"logits/chosen": -0.9681941866874695,
"logits/rejected": -0.9737881422042847,
"logps/chosen": -101.29634094238281,
"logps/rejected": -180.4349822998047,
"loss": 0.0586,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -2.112703323364258,
"rewards/margins": 6.271733283996582,
"rewards/rejected": -8.38443660736084,
"step": 545
},
{
"epoch": 2.9333333333333336,
"grad_norm": 7.896867275238037,
"learning_rate": 1e-06,
"logits/chosen": -0.9267778396606445,
"logits/rejected": -0.9226524233818054,
"logps/chosen": -86.86418914794922,
"logps/rejected": -140.9725341796875,
"loss": 0.0609,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6871792078018188,
"rewards/margins": 5.857264995574951,
"rewards/rejected": -7.544443607330322,
"step": 550
},
{
"epoch": 2.96,
"grad_norm": 2.1880979537963867,
"learning_rate": 1e-06,
"logits/chosen": -0.9265161752700806,
"logits/rejected": -0.9284146428108215,
"logps/chosen": -86.45133972167969,
"logps/rejected": -158.9849853515625,
"loss": 0.0429,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.7444225549697876,
"rewards/margins": 7.387899875640869,
"rewards/rejected": -9.132322311401367,
"step": 555
},
{
"epoch": 2.986666666666667,
"grad_norm": 0.5332115888595581,
"learning_rate": 1e-06,
"logits/chosen": -0.9316242933273315,
"logits/rejected": -0.9294928312301636,
"logps/chosen": -89.85722351074219,
"logps/rejected": -164.3010711669922,
"loss": 0.0779,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.4647209644317627,
"rewards/margins": 7.412230491638184,
"rewards/rejected": -8.876951217651367,
"step": 560
},
{
"epoch": 2.992,
"step": 561,
"total_flos": 0.0,
"train_loss": 0.24264521944982376,
"train_runtime": 1521.7244,
"train_samples_per_second": 11.827,
"train_steps_per_second": 0.369
}
],
"logging_steps": 5,
"max_steps": 561,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": false,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 2,
"trial_name": null,
"trial_params": null
}