seablue's picture
Upload trainer_state.json with huggingface_hub
0b5cde7 verified
raw
history blame
51.5 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.998691442030882,
"eval_steps": 500,
"global_step": 477,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.010468463752944255,
"grad_norm": 3.7953430363493075,
"learning_rate": 5.208333333333333e-08,
"logits/chosen": -2.994403839111328,
"logits/rejected": -2.932447671890259,
"logps/chosen": -1.0262531042099,
"logps/rejected": -1.1902999877929688,
"loss": -0.0244,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -1.0262531042099,
"rewards/margins": 0.16404692828655243,
"rewards/rejected": -1.1902999877929688,
"step": 5
},
{
"epoch": 0.02093692750588851,
"grad_norm": 4.19613852299559,
"learning_rate": 1.0416666666666667e-07,
"logits/chosen": -3.0355896949768066,
"logits/rejected": -2.978191375732422,
"logps/chosen": -1.1014703512191772,
"logps/rejected": -1.1318297386169434,
"loss": -0.0274,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -1.1014703512191772,
"rewards/margins": 0.030359352007508278,
"rewards/rejected": -1.1318297386169434,
"step": 10
},
{
"epoch": 0.031405391258832765,
"grad_norm": 4.396342901267451,
"learning_rate": 1.5624999999999999e-07,
"logits/chosen": -3.0005362033843994,
"logits/rejected": -2.9955382347106934,
"logps/chosen": -1.0661541223526,
"logps/rejected": -1.0591233968734741,
"loss": -0.0311,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -1.0661541223526,
"rewards/margins": -0.007030715234577656,
"rewards/rejected": -1.0591233968734741,
"step": 15
},
{
"epoch": 0.04187385501177702,
"grad_norm": 4.0816701499978265,
"learning_rate": 2.0833333333333333e-07,
"logits/chosen": -3.015394687652588,
"logits/rejected": -2.977292537689209,
"logps/chosen": -1.0042058229446411,
"logps/rejected": -1.0408815145492554,
"loss": -0.0327,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": -1.0042058229446411,
"rewards/margins": 0.03667566925287247,
"rewards/rejected": -1.0408815145492554,
"step": 20
},
{
"epoch": 0.05234231876472128,
"grad_norm": 4.229656757392526,
"learning_rate": 2.604166666666667e-07,
"logits/chosen": -3.0237503051757812,
"logits/rejected": -2.973411798477173,
"logps/chosen": -0.9534778594970703,
"logps/rejected": -1.1941173076629639,
"loss": -0.0414,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.9534778594970703,
"rewards/margins": 0.24063941836357117,
"rewards/rejected": -1.1941173076629639,
"step": 25
},
{
"epoch": 0.06281078251766553,
"grad_norm": 5.6524920417618585,
"learning_rate": 3.1249999999999997e-07,
"logits/chosen": -3.0303757190704346,
"logits/rejected": -2.9622457027435303,
"logps/chosen": -0.875920295715332,
"logps/rejected": -0.9086316227912903,
"loss": -0.0368,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.875920295715332,
"rewards/margins": 0.03271136432886124,
"rewards/rejected": -0.9086316227912903,
"step": 30
},
{
"epoch": 0.07327924627060979,
"grad_norm": 5.177058208668614,
"learning_rate": 3.645833333333333e-07,
"logits/chosen": -3.0142247676849365,
"logits/rejected": -2.9961647987365723,
"logps/chosen": -0.86571204662323,
"logps/rejected": -1.1118712425231934,
"loss": -0.0475,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -0.86571204662323,
"rewards/margins": 0.24615919589996338,
"rewards/rejected": -1.1118712425231934,
"step": 35
},
{
"epoch": 0.08374771002355404,
"grad_norm": 4.595347219497465,
"learning_rate": 4.1666666666666667e-07,
"logits/chosen": -2.9783272743225098,
"logits/rejected": -2.948113203048706,
"logps/chosen": -0.8789661526679993,
"logps/rejected": -0.9445465803146362,
"loss": -0.0484,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.8789661526679993,
"rewards/margins": 0.06558056175708771,
"rewards/rejected": -0.9445465803146362,
"step": 40
},
{
"epoch": 0.0942161737764983,
"grad_norm": 4.777810219545724,
"learning_rate": 4.6874999999999996e-07,
"logits/chosen": -3.0266075134277344,
"logits/rejected": -3.020308494567871,
"logps/chosen": -0.8326341509819031,
"logps/rejected": -0.9679509401321411,
"loss": -0.0617,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -0.8326341509819031,
"rewards/margins": 0.13531669974327087,
"rewards/rejected": -0.9679509401321411,
"step": 45
},
{
"epoch": 0.10468463752944256,
"grad_norm": 5.443655913072939,
"learning_rate": 4.999731868769026e-07,
"logits/chosen": -2.88401198387146,
"logits/rejected": -2.9223647117614746,
"logps/chosen": -0.9018585085868835,
"logps/rejected": -1.0319406986236572,
"loss": -0.0525,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.9018585085868835,
"rewards/margins": 0.13008219003677368,
"rewards/rejected": -1.0319406986236572,
"step": 50
},
{
"epoch": 0.11515310128238682,
"grad_norm": 8.004493874473818,
"learning_rate": 4.996716052911017e-07,
"logits/chosen": -3.0439066886901855,
"logits/rejected": -3.038985013961792,
"logps/chosen": -0.9202953577041626,
"logps/rejected": -1.087965726852417,
"loss": -0.0482,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.9202953577041626,
"rewards/margins": 0.1676703691482544,
"rewards/rejected": -1.087965726852417,
"step": 55
},
{
"epoch": 0.12562156503533106,
"grad_norm": 8.438457497880151,
"learning_rate": 4.990353313429303e-07,
"logits/chosen": -2.9756836891174316,
"logits/rejected": -2.904886245727539,
"logps/chosen": -0.9683998227119446,
"logps/rejected": -1.149547815322876,
"loss": -0.0681,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.9683998227119446,
"rewards/margins": 0.181147962808609,
"rewards/rejected": -1.149547815322876,
"step": 60
},
{
"epoch": 0.1360900287882753,
"grad_norm": 6.673228848159463,
"learning_rate": 4.980652179769217e-07,
"logits/chosen": -2.966130495071411,
"logits/rejected": -2.9898931980133057,
"logps/chosen": -0.9524378776550293,
"logps/rejected": -1.1887619495391846,
"loss": -0.0605,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.9524378776550293,
"rewards/margins": 0.23632416129112244,
"rewards/rejected": -1.1887619495391846,
"step": 65
},
{
"epoch": 0.14655849254121958,
"grad_norm": 7.12442296176354,
"learning_rate": 4.967625656594781e-07,
"logits/chosen": -3.0336520671844482,
"logits/rejected": -2.9966013431549072,
"logps/chosen": -1.0305348634719849,
"logps/rejected": -1.2037804126739502,
"loss": -0.0808,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -1.0305348634719849,
"rewards/margins": 0.1732456386089325,
"rewards/rejected": -1.2037804126739502,
"step": 70
},
{
"epoch": 0.15702695629416383,
"grad_norm": 8.805089068512533,
"learning_rate": 4.951291206355559e-07,
"logits/chosen": -2.9939000606536865,
"logits/rejected": -2.958610773086548,
"logps/chosen": -1.0724363327026367,
"logps/rejected": -1.3205912113189697,
"loss": -0.0632,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -1.0724363327026367,
"rewards/margins": 0.2481549233198166,
"rewards/rejected": -1.3205912113189697,
"step": 75
},
{
"epoch": 0.16749542004710807,
"grad_norm": 6.848905163137727,
"learning_rate": 4.93167072587771e-07,
"logits/chosen": -2.9966750144958496,
"logits/rejected": -2.9649088382720947,
"logps/chosen": -1.0316729545593262,
"logps/rejected": -1.4772520065307617,
"loss": -0.0784,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -1.0316729545593262,
"rewards/margins": 0.4455792009830475,
"rewards/rejected": -1.4772520065307617,
"step": 80
},
{
"epoch": 0.17796388380005235,
"grad_norm": 8.38726120157511,
"learning_rate": 4.908790517010636e-07,
"logits/chosen": -2.995182514190674,
"logits/rejected": -2.9591526985168457,
"logps/chosen": -0.920505166053772,
"logps/rejected": -1.3156445026397705,
"loss": -0.0829,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.920505166053772,
"rewards/margins": 0.39513930678367615,
"rewards/rejected": -1.3156445026397705,
"step": 85
},
{
"epoch": 0.1884323475529966,
"grad_norm": 5.335037541677766,
"learning_rate": 4.882681251368548e-07,
"logits/chosen": -2.9661142826080322,
"logits/rejected": -2.9613051414489746,
"logps/chosen": -1.0484893321990967,
"logps/rejected": -1.3956043720245361,
"loss": -0.0851,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -1.0484893321990967,
"rewards/margins": 0.34711501002311707,
"rewards/rejected": -1.3956043720245361,
"step": 90
},
{
"epoch": 0.19890081130594087,
"grad_norm": 5.9994342157674705,
"learning_rate": 4.853377929214243e-07,
"logits/chosen": -3.0459396839141846,
"logits/rejected": -3.03053617477417,
"logps/chosen": -1.004883050918579,
"logps/rejected": -1.2524257898330688,
"loss": -0.0726,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": -1.004883050918579,
"rewards/margins": 0.24754276871681213,
"rewards/rejected": -1.2524257898330688,
"step": 95
},
{
"epoch": 0.2093692750588851,
"grad_norm": 7.84898491024868,
"learning_rate": 4.820919832540181e-07,
"logits/chosen": -3.0662217140197754,
"logits/rejected": -3.080054521560669,
"logps/chosen": -1.0252565145492554,
"logps/rejected": -1.3465287685394287,
"loss": -0.0815,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -1.0252565145492554,
"rewards/margins": 0.32127222418785095,
"rewards/rejected": -1.3465287685394287,
"step": 100
},
{
"epoch": 0.21983773881182936,
"grad_norm": 6.574476192964948,
"learning_rate": 4.785350472409791e-07,
"logits/chosen": -3.0290255546569824,
"logits/rejected": -3.0409884452819824,
"logps/chosen": -1.2733252048492432,
"logps/rejected": -1.5657093524932861,
"loss": -0.1053,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.2733252048492432,
"rewards/margins": 0.29238420724868774,
"rewards/rejected": -1.5657093524932861,
"step": 105
},
{
"epoch": 0.23030620256477363,
"grad_norm": 10.583840544155157,
"learning_rate": 4.7467175306295647e-07,
"logits/chosen": -2.9230265617370605,
"logits/rejected": -2.9670557975769043,
"logps/chosen": -1.0758285522460938,
"logps/rejected": -1.6084600687026978,
"loss": -0.105,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.0758285522460938,
"rewards/margins": 0.5326316356658936,
"rewards/rejected": -1.6084600687026978,
"step": 110
},
{
"epoch": 0.24077466631771788,
"grad_norm": 8.422749604707546,
"learning_rate": 4.70507279583015e-07,
"logits/chosen": -3.0686347484588623,
"logits/rejected": -3.062587022781372,
"logps/chosen": -1.0952173471450806,
"logps/rejected": -1.419394612312317,
"loss": -0.0899,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -1.0952173471450806,
"rewards/margins": 0.32417723536491394,
"rewards/rejected": -1.419394612312317,
"step": 115
},
{
"epoch": 0.2512431300706621,
"grad_norm": 7.245428385744917,
"learning_rate": 4.6604720940421207e-07,
"logits/chosen": -2.9976067543029785,
"logits/rejected": -2.9936585426330566,
"logps/chosen": -1.1029393672943115,
"logps/rejected": -1.669960379600525,
"loss": -0.0782,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -1.1029393672943115,
"rewards/margins": 0.5670207738876343,
"rewards/rejected": -1.669960379600525,
"step": 120
},
{
"epoch": 0.26171159382360637,
"grad_norm": 9.260649226954074,
"learning_rate": 4.612975213859487e-07,
"logits/chosen": -2.89277982711792,
"logits/rejected": -2.8418478965759277,
"logps/chosen": -1.1914057731628418,
"logps/rejected": -1.5322601795196533,
"loss": -0.0806,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -1.1914057731628418,
"rewards/margins": 0.3408544659614563,
"rewards/rejected": -1.5322601795196533,
"step": 125
},
{
"epoch": 0.2721800575765506,
"grad_norm": 7.897383530777984,
"learning_rate": 4.5626458262912735e-07,
"logits/chosen": -2.9802322387695312,
"logits/rejected": -3.0263705253601074,
"logps/chosen": -1.0792829990386963,
"logps/rejected": -1.620160698890686,
"loss": -0.0969,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -1.0792829990386963,
"rewards/margins": 0.540877640247345,
"rewards/rejected": -1.620160698890686,
"step": 130
},
{
"epoch": 0.2826485213294949,
"grad_norm": 8.56684631151293,
"learning_rate": 4.5095513994085974e-07,
"logits/chosen": -2.9500975608825684,
"logits/rejected": -2.930224895477295,
"logps/chosen": -1.156005859375,
"logps/rejected": -1.5593708753585815,
"loss": -0.0941,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -1.156005859375,
"rewards/margins": 0.40336498618125916,
"rewards/rejected": -1.5593708753585815,
"step": 135
},
{
"epoch": 0.29311698508243916,
"grad_norm": 11.777411857127742,
"learning_rate": 4.453763107901675e-07,
"logits/chosen": -2.9826929569244385,
"logits/rejected": -2.996265411376953,
"logps/chosen": -1.2722768783569336,
"logps/rejected": -1.6690384149551392,
"loss": -0.0801,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -1.2722768783569336,
"rewards/margins": 0.39676159620285034,
"rewards/rejected": -1.6690384149551392,
"step": 140
},
{
"epoch": 0.3035854488353834,
"grad_norm": 13.441666638709599,
"learning_rate": 4.395355737667985e-07,
"logits/chosen": -2.8897225856781006,
"logits/rejected": -2.924546241760254,
"logps/chosen": -1.1064995527267456,
"logps/rejected": -1.5362865924835205,
"loss": -0.1008,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -1.1064995527267456,
"rewards/margins": 0.4297870695590973,
"rewards/rejected": -1.5362865924835205,
"step": 145
},
{
"epoch": 0.31405391258832765,
"grad_norm": 13.628880017668104,
"learning_rate": 4.3344075855595097e-07,
"logits/chosen": -2.99322247505188,
"logits/rejected": -2.963665246963501,
"logps/chosen": -1.135622501373291,
"logps/rejected": -1.865152359008789,
"loss": -0.0913,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -1.135622501373291,
"rewards/margins": 0.729529857635498,
"rewards/rejected": -1.865152359008789,
"step": 150
},
{
"epoch": 0.3245223763412719,
"grad_norm": 18.47300522097184,
"learning_rate": 4.271000354423425e-07,
"logits/chosen": -2.86858868598938,
"logits/rejected": -2.850736141204834,
"logps/chosen": -1.3002798557281494,
"logps/rejected": -1.8037567138671875,
"loss": -0.1109,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.3002798557281494,
"rewards/margins": 0.5034767985343933,
"rewards/rejected": -1.8037567138671875,
"step": 155
},
{
"epoch": 0.33499084009421615,
"grad_norm": 13.245417836958321,
"learning_rate": 4.2052190435769554e-07,
"logits/chosen": -2.8188085556030273,
"logits/rejected": -2.847963333129883,
"logps/chosen": -1.3339486122131348,
"logps/rejected": -2.279085636138916,
"loss": -0.1074,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -1.3339486122131348,
"rewards/margins": 0.9451370239257812,
"rewards/rejected": -2.279085636138916,
"step": 160
},
{
"epoch": 0.34545930384716045,
"grad_norm": 16.119605348885486,
"learning_rate": 4.137151834863213e-07,
"logits/chosen": -2.949324369430542,
"logits/rejected": -2.931403875350952,
"logps/chosen": -1.4816526174545288,
"logps/rejected": -1.9713685512542725,
"loss": -0.1042,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -1.4816526174545288,
"rewards/margins": 0.4897160530090332,
"rewards/rejected": -1.9713685512542725,
"step": 165
},
{
"epoch": 0.3559277676001047,
"grad_norm": 9.715633120289555,
"learning_rate": 4.0668899744407567e-07,
"logits/chosen": -2.9010634422302246,
"logits/rejected": -2.8985071182250977,
"logps/chosen": -1.2567403316497803,
"logps/rejected": -1.9025369882583618,
"loss": -0.1039,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -1.2567403316497803,
"rewards/margins": 0.6457964777946472,
"rewards/rejected": -1.9025369882583618,
"step": 170
},
{
"epoch": 0.36639623135304894,
"grad_norm": 17.275270203000964,
"learning_rate": 3.994527650465352e-07,
"logits/chosen": -2.82110595703125,
"logits/rejected": -2.841233491897583,
"logps/chosen": -1.009877324104309,
"logps/rejected": -1.7530879974365234,
"loss": -0.1124,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -1.009877324104309,
"rewards/margins": 0.7432108521461487,
"rewards/rejected": -1.7530879974365234,
"step": 175
},
{
"epoch": 0.3768646951059932,
"grad_norm": 14.01202945114004,
"learning_rate": 3.920161866827889e-07,
"logits/chosen": -2.848515272140503,
"logits/rejected": -2.8287739753723145,
"logps/chosen": -1.2261879444122314,
"logps/rejected": -1.877882719039917,
"loss": -0.1098,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -1.2261879444122314,
"rewards/margins": 0.6516948342323303,
"rewards/rejected": -1.877882719039917,
"step": 180
},
{
"epoch": 0.38733315885893743,
"grad_norm": 12.209017183487612,
"learning_rate": 3.8438923131177237e-07,
"logits/chosen": -2.842813491821289,
"logits/rejected": -2.884312152862549,
"logps/chosen": -1.138585090637207,
"logps/rejected": -1.9834768772125244,
"loss": -0.122,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.138585090637207,
"rewards/margins": 0.8448915481567383,
"rewards/rejected": -1.9834768772125244,
"step": 185
},
{
"epoch": 0.39780162261188173,
"grad_norm": 20.510663419119325,
"learning_rate": 3.765821230985757e-07,
"logits/chosen": -2.815495252609253,
"logits/rejected": -2.8491616249084473,
"logps/chosen": -1.116502285003662,
"logps/rejected": -2.221632480621338,
"loss": -0.1159,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -1.116502285003662,
"rewards/margins": 1.1051301956176758,
"rewards/rejected": -2.221632480621338,
"step": 190
},
{
"epoch": 0.408270086364826,
"grad_norm": 10.576124534133774,
"learning_rate": 3.6860532770864005e-07,
"logits/chosen": -2.8467681407928467,
"logits/rejected": -2.814634323120117,
"logps/chosen": -1.2795817852020264,
"logps/rejected": -2.408722400665283,
"loss": -0.098,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -1.2795817852020264,
"rewards/margins": 1.1291409730911255,
"rewards/rejected": -2.408722400665283,
"step": 195
},
{
"epoch": 0.4187385501177702,
"grad_norm": 9.7795437319874,
"learning_rate": 3.604695382782159e-07,
"logits/chosen": -2.747274398803711,
"logits/rejected": -2.6726021766662598,
"logps/chosen": -1.3564021587371826,
"logps/rejected": -1.8328683376312256,
"loss": -0.1253,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.3564021587371826,
"rewards/margins": 0.47646617889404297,
"rewards/rejected": -1.8328683376312256,
"step": 200
},
{
"epoch": 0.42920701387071447,
"grad_norm": 10.854270078977583,
"learning_rate": 3.5218566107988867e-07,
"logits/chosen": -2.751420021057129,
"logits/rejected": -2.7701942920684814,
"logps/chosen": -1.4730335474014282,
"logps/rejected": -2.165682315826416,
"loss": -0.1223,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -1.4730335474014282,
"rewards/margins": 0.6926488280296326,
"rewards/rejected": -2.165682315826416,
"step": 205
},
{
"epoch": 0.4396754776236587,
"grad_norm": 8.55442674152993,
"learning_rate": 3.4376480090239047e-07,
"logits/chosen": -2.741508960723877,
"logits/rejected": -2.699432849884033,
"logps/chosen": -1.2727248668670654,
"logps/rejected": -1.861159086227417,
"loss": -0.1068,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -1.2727248668670654,
"rewards/margins": 0.5884342193603516,
"rewards/rejected": -1.861159086227417,
"step": 210
},
{
"epoch": 0.45014394137660296,
"grad_norm": 11.243462810743234,
"learning_rate": 3.3521824616429284e-07,
"logits/chosen": -2.6302266120910645,
"logits/rejected": -2.601926326751709,
"logps/chosen": -1.2547543048858643,
"logps/rejected": -2.3449490070343018,
"loss": -0.1052,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -1.2547543048858643,
"rewards/margins": 1.090194582939148,
"rewards/rejected": -2.3449490070343018,
"step": 215
},
{
"epoch": 0.46061240512954726,
"grad_norm": 54.27098259781432,
"learning_rate": 3.265574537815398e-07,
"logits/chosen": -2.580867052078247,
"logits/rejected": -2.5696752071380615,
"logps/chosen": -1.3456079959869385,
"logps/rejected": -2.1853251457214355,
"loss": -0.1262,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -1.3456079959869385,
"rewards/margins": 0.8397170901298523,
"rewards/rejected": -2.1853251457214355,
"step": 220
},
{
"epoch": 0.4710808688824915,
"grad_norm": 48.88998141840299,
"learning_rate": 3.1779403380910425e-07,
"logits/chosen": -2.6540842056274414,
"logits/rejected": -2.501495838165283,
"logps/chosen": -1.7648048400878906,
"logps/rejected": -2.8733654022216797,
"loss": -0.1162,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -1.7648048400878906,
"rewards/margins": 1.1085606813430786,
"rewards/rejected": -2.8733654022216797,
"step": 225
},
{
"epoch": 0.48154933263543576,
"grad_norm": 14.602355735002494,
"learning_rate": 3.0893973387735683e-07,
"logits/chosen": -2.587488889694214,
"logits/rejected": -2.5237855911254883,
"logps/chosen": -1.4629151821136475,
"logps/rejected": -2.598280191421509,
"loss": -0.1311,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -1.4629151821136475,
"rewards/margins": 1.1353652477264404,
"rewards/rejected": -2.598280191421509,
"step": 230
},
{
"epoch": 0.49201779638838,
"grad_norm": 11.479507096481315,
"learning_rate": 3.000064234440111e-07,
"logits/chosen": -2.4094467163085938,
"logits/rejected": -2.3350281715393066,
"logps/chosen": -1.1608326435089111,
"logps/rejected": -2.2014541625976562,
"loss": -0.1266,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -1.1608326435089111,
"rewards/margins": 1.040621280670166,
"rewards/rejected": -2.2014541625976562,
"step": 235
},
{
"epoch": 0.5024862601413242,
"grad_norm": 10.608303165039153,
"learning_rate": 2.910060778827554e-07,
"logits/chosen": -2.469841718673706,
"logits/rejected": -2.4375293254852295,
"logps/chosen": -1.57926344871521,
"logps/rejected": -2.4959769248962402,
"loss": -0.1232,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -1.57926344871521,
"rewards/margins": 0.9167135953903198,
"rewards/rejected": -2.4959769248962402,
"step": 240
},
{
"epoch": 0.5129547238942685,
"grad_norm": 25.353094981844627,
"learning_rate": 2.8195076242990116e-07,
"logits/chosen": -2.5656850337982178,
"logits/rejected": -2.5005340576171875,
"logps/chosen": -1.3019143342971802,
"logps/rejected": -2.8710713386535645,
"loss": -0.1347,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.3019143342971802,
"rewards/margins": 1.5691570043563843,
"rewards/rejected": -2.8710713386535645,
"step": 245
},
{
"epoch": 0.5234231876472127,
"grad_norm": 21.49026192198408,
"learning_rate": 2.7285261601056697e-07,
"logits/chosen": -2.4698336124420166,
"logits/rejected": -2.3579485416412354,
"logps/chosen": -1.7306108474731445,
"logps/rejected": -3.098844051361084,
"loss": -0.1102,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -1.7306108474731445,
"rewards/margins": 1.368233323097229,
"rewards/rejected": -3.098844051361084,
"step": 250
},
{
"epoch": 0.533891651400157,
"grad_norm": 21.368862546983063,
"learning_rate": 2.6372383496608186e-07,
"logits/chosen": -2.613272190093994,
"logits/rejected": -2.5582284927368164,
"logps/chosen": -1.365275502204895,
"logps/rejected": -2.1760246753692627,
"loss": -0.1128,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -1.365275502204895,
"rewards/margins": 0.8107490539550781,
"rewards/rejected": -2.1760246753692627,
"step": 255
},
{
"epoch": 0.5443601151531012,
"grad_norm": 10.208688557762956,
"learning_rate": 2.5457665670441937e-07,
"logits/chosen": -2.587008237838745,
"logits/rejected": -2.517343521118164,
"logps/chosen": -1.1947706937789917,
"logps/rejected": -2.122412919998169,
"loss": -0.1285,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.1947706937789917,
"rewards/margins": 0.9276420474052429,
"rewards/rejected": -2.122412919998169,
"step": 260
},
{
"epoch": 0.5548285789060455,
"grad_norm": 17.908677113828368,
"learning_rate": 2.454233432955807e-07,
"logits/chosen": -2.5146946907043457,
"logits/rejected": -2.5216221809387207,
"logps/chosen": -1.1950175762176514,
"logps/rejected": -2.243485927581787,
"loss": -0.1212,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -1.1950175762176514,
"rewards/margins": 1.0484683513641357,
"rewards/rejected": -2.243485927581787,
"step": 265
},
{
"epoch": 0.5652970426589898,
"grad_norm": 10.277763557702748,
"learning_rate": 2.3627616503391812e-07,
"logits/chosen": -2.4133479595184326,
"logits/rejected": -2.4332380294799805,
"logps/chosen": -2.0134479999542236,
"logps/rejected": -2.5305628776550293,
"loss": -0.1209,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -2.0134479999542236,
"rewards/margins": 0.5171147584915161,
"rewards/rejected": -2.5305628776550293,
"step": 270
},
{
"epoch": 0.575765506411934,
"grad_norm": 20.809963361604478,
"learning_rate": 2.2714738398943308e-07,
"logits/chosen": -2.465207815170288,
"logits/rejected": -2.3378491401672363,
"logps/chosen": -1.4752328395843506,
"logps/rejected": -3.626649856567383,
"loss": -0.1508,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.4752328395843506,
"rewards/margins": 2.1514172554016113,
"rewards/rejected": -3.626649856567383,
"step": 275
},
{
"epoch": 0.5862339701648783,
"grad_norm": 14.238466786958266,
"learning_rate": 2.1804923757009882e-07,
"logits/chosen": -2.537182331085205,
"logits/rejected": -2.421999454498291,
"logps/chosen": -1.1645643711090088,
"logps/rejected": -2.6537137031555176,
"loss": -0.127,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.1645643711090088,
"rewards/margins": 1.4891494512557983,
"rewards/rejected": -2.6537137031555176,
"step": 280
},
{
"epoch": 0.5967024339178225,
"grad_norm": 16.046921334576403,
"learning_rate": 2.089939221172446e-07,
"logits/chosen": -2.5481698513031006,
"logits/rejected": -2.5069780349731445,
"logps/chosen": -1.4470568895339966,
"logps/rejected": -2.9359803199768066,
"loss": -0.1148,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -1.4470568895339966,
"rewards/margins": 1.4889236688613892,
"rewards/rejected": -2.9359803199768066,
"step": 285
},
{
"epoch": 0.6071708976707668,
"grad_norm": 18.892819573440462,
"learning_rate": 1.9999357655598891e-07,
"logits/chosen": -2.431614398956299,
"logits/rejected": -2.342071056365967,
"logps/chosen": -1.7402465343475342,
"logps/rejected": -2.2307562828063965,
"loss": -0.1091,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.7402465343475342,
"rewards/margins": 0.49050965905189514,
"rewards/rejected": -2.2307562828063965,
"step": 290
},
{
"epoch": 0.6176393614237111,
"grad_norm": 12.912710021152682,
"learning_rate": 1.9106026612264315e-07,
"logits/chosen": -2.553067684173584,
"logits/rejected": -2.392414093017578,
"logps/chosen": -1.4181064367294312,
"logps/rejected": -2.5548157691955566,
"loss": -0.1331,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -1.4181064367294312,
"rewards/margins": 1.136709451675415,
"rewards/rejected": -2.5548157691955566,
"step": 295
},
{
"epoch": 0.6281078251766553,
"grad_norm": 11.978919101167111,
"learning_rate": 1.8220596619089573e-07,
"logits/chosen": -2.3048555850982666,
"logits/rejected": -2.2152676582336426,
"logps/chosen": -1.5337917804718018,
"logps/rejected": -2.5993523597717285,
"loss": -0.1186,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -1.5337917804718018,
"rewards/margins": 1.0655608177185059,
"rewards/rejected": -2.5993523597717285,
"step": 300
},
{
"epoch": 0.6385762889295996,
"grad_norm": 27.362413259228795,
"learning_rate": 1.7344254621846017e-07,
"logits/chosen": -2.394150733947754,
"logits/rejected": -2.315016269683838,
"logps/chosen": -1.6578819751739502,
"logps/rejected": -2.4831995964050293,
"loss": -0.1346,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.6578819751739502,
"rewards/margins": 0.8253179788589478,
"rewards/rejected": -2.4831995964050293,
"step": 305
},
{
"epoch": 0.6490447526825438,
"grad_norm": 12.794098747918724,
"learning_rate": 1.647817538357072e-07,
"logits/chosen": -2.384284496307373,
"logits/rejected": -2.131286382675171,
"logps/chosen": -1.2339116334915161,
"logps/rejected": -2.9859485626220703,
"loss": -0.1415,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -1.2339116334915161,
"rewards/margins": 1.7520368099212646,
"rewards/rejected": -2.9859485626220703,
"step": 310
},
{
"epoch": 0.6595132164354881,
"grad_norm": 17.663997235638817,
"learning_rate": 1.562351990976095e-07,
"logits/chosen": -2.458538293838501,
"logits/rejected": -2.3150718212127686,
"logps/chosen": -1.5998607873916626,
"logps/rejected": -2.276341199874878,
"loss": -0.1284,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -1.5998607873916626,
"rewards/margins": 0.6764806509017944,
"rewards/rejected": -2.276341199874878,
"step": 315
},
{
"epoch": 0.6699816801884323,
"grad_norm": 17.859774700477008,
"learning_rate": 1.478143389201113e-07,
"logits/chosen": -2.4031972885131836,
"logits/rejected": -2.2891368865966797,
"logps/chosen": -1.5045208930969238,
"logps/rejected": -2.6329269409179688,
"loss": -0.136,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -1.5045208930969238,
"rewards/margins": 1.128406286239624,
"rewards/rejected": -2.6329269409179688,
"step": 320
},
{
"epoch": 0.6804501439413766,
"grad_norm": 22.79098412535109,
"learning_rate": 1.3953046172178413e-07,
"logits/chosen": -2.346923828125,
"logits/rejected": -2.2279601097106934,
"logps/chosen": -1.3641127347946167,
"logps/rejected": -2.919264316558838,
"loss": -0.1149,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -1.3641127347946167,
"rewards/margins": 1.5551520586013794,
"rewards/rejected": -2.919264316558838,
"step": 325
},
{
"epoch": 0.6909186076943209,
"grad_norm": 19.92126992190782,
"learning_rate": 1.3139467229135998e-07,
"logits/chosen": -2.4331743717193604,
"logits/rejected": -2.319657802581787,
"logps/chosen": -1.1333062648773193,
"logps/rejected": -2.6917026042938232,
"loss": -0.1293,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -1.1333062648773193,
"rewards/margins": 1.558396577835083,
"rewards/rejected": -2.6917026042938232,
"step": 330
},
{
"epoch": 0.7013870714472651,
"grad_norm": 66.91823491595629,
"learning_rate": 1.2341787690142435e-07,
"logits/chosen": -2.1722617149353027,
"logits/rejected": -2.239929676055908,
"logps/chosen": -1.8514454364776611,
"logps/rejected": -2.463306427001953,
"loss": -0.1216,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -1.8514454364776611,
"rewards/margins": 0.6118608713150024,
"rewards/rejected": -2.463306427001953,
"step": 335
},
{
"epoch": 0.7118555352002094,
"grad_norm": 31.938913094618993,
"learning_rate": 1.1561076868822755e-07,
"logits/chosen": -2.334263563156128,
"logits/rejected": -2.202540159225464,
"logps/chosen": -1.2181419134140015,
"logps/rejected": -3.0234227180480957,
"loss": -0.1432,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.2181419134140015,
"rewards/margins": 1.8052809238433838,
"rewards/rejected": -3.0234227180480957,
"step": 340
},
{
"epoch": 0.7223239989531536,
"grad_norm": 16.20254395042608,
"learning_rate": 1.0798381331721107e-07,
"logits/chosen": -2.3701541423797607,
"logits/rejected": -2.2050604820251465,
"logps/chosen": -1.4401299953460693,
"logps/rejected": -2.4977269172668457,
"loss": -0.114,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -1.4401299953460693,
"rewards/margins": 1.0575966835021973,
"rewards/rejected": -2.4977269172668457,
"step": 345
},
{
"epoch": 0.7327924627060979,
"grad_norm": 15.428080471435864,
"learning_rate": 1.0054723495346482e-07,
"logits/chosen": -2.2697322368621826,
"logits/rejected": -1.9184463024139404,
"logps/chosen": -1.9560050964355469,
"logps/rejected": -2.635880947113037,
"loss": -0.1343,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -1.9560050964355469,
"rewards/margins": 0.6798759698867798,
"rewards/rejected": -2.635880947113037,
"step": 350
},
{
"epoch": 0.7432609264590422,
"grad_norm": 42.2232530553668,
"learning_rate": 9.331100255592436e-08,
"logits/chosen": -2.2260258197784424,
"logits/rejected": -2.0879032611846924,
"logps/chosen": -1.52364182472229,
"logps/rejected": -3.164949655532837,
"loss": -0.1182,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -1.52364182472229,
"rewards/margins": 1.6413074731826782,
"rewards/rejected": -3.164949655532837,
"step": 355
},
{
"epoch": 0.7537293902119864,
"grad_norm": 15.626689472317864,
"learning_rate": 8.628481651367875e-08,
"logits/chosen": -2.132807731628418,
"logits/rejected": -2.0273263454437256,
"logps/chosen": -1.9577980041503906,
"logps/rejected": -2.6277079582214355,
"loss": -0.1234,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -1.9577980041503906,
"rewards/margins": 0.6699098348617554,
"rewards/rejected": -2.6277079582214355,
"step": 360
},
{
"epoch": 0.7641978539649307,
"grad_norm": 41.37098781296867,
"learning_rate": 7.947809564230445e-08,
"logits/chosen": -2.2800426483154297,
"logits/rejected": -2.0097782611846924,
"logps/chosen": -1.793992042541504,
"logps/rejected": -3.1277503967285156,
"loss": -0.139,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -1.793992042541504,
"rewards/margins": 1.3337581157684326,
"rewards/rejected": -3.1277503967285156,
"step": 365
},
{
"epoch": 0.7746663177178749,
"grad_norm": 62.29595045789601,
"learning_rate": 7.289996455765748e-08,
"logits/chosen": -2.3463916778564453,
"logits/rejected": -2.204310655593872,
"logps/chosen": -1.4694790840148926,
"logps/rejected": -2.4598724842071533,
"loss": -0.1446,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.4694790840148926,
"rewards/margins": 0.990393340587616,
"rewards/rejected": -2.4598724842071533,
"step": 370
},
{
"epoch": 0.7851347814708192,
"grad_norm": 26.402921223340687,
"learning_rate": 6.655924144404906e-08,
"logits/chosen": -2.1613121032714844,
"logits/rejected": -2.173642158508301,
"logps/chosen": -1.5132877826690674,
"logps/rejected": -2.912444591522217,
"loss": -0.1292,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -1.5132877826690674,
"rewards/margins": 1.399156928062439,
"rewards/rejected": -2.912444591522217,
"step": 375
},
{
"epoch": 0.7956032452237635,
"grad_norm": 30.264254092211626,
"learning_rate": 6.046442623320145e-08,
"logits/chosen": -2.3032054901123047,
"logits/rejected": -2.158975839614868,
"logps/chosen": -1.456990361213684,
"logps/rejected": -2.298990488052368,
"loss": -0.1205,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.456990361213684,
"rewards/margins": 0.8420003056526184,
"rewards/rejected": -2.298990488052368,
"step": 380
},
{
"epoch": 0.8060717089767077,
"grad_norm": 14.190868550898228,
"learning_rate": 5.4623689209832484e-08,
"logits/chosen": -2.239023447036743,
"logits/rejected": -2.178297758102417,
"logps/chosen": -1.2550222873687744,
"logps/rejected": -2.4818851947784424,
"loss": -0.1307,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -1.2550222873687744,
"rewards/margins": 1.226862907409668,
"rewards/rejected": -2.4818851947784424,
"step": 385
},
{
"epoch": 0.816540172729652,
"grad_norm": 23.083356140391576,
"learning_rate": 4.904486005914027e-08,
"logits/chosen": -2.3200623989105225,
"logits/rejected": -2.108736515045166,
"logps/chosen": -1.408341646194458,
"logps/rejected": -2.439392566680908,
"loss": -0.1144,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -1.408341646194458,
"rewards/margins": 1.0310509204864502,
"rewards/rejected": -2.439392566680908,
"step": 390
},
{
"epoch": 0.8270086364825961,
"grad_norm": 22.17340356091122,
"learning_rate": 4.373541737087263e-08,
"logits/chosen": -2.2692439556121826,
"logits/rejected": -2.039247751235962,
"logps/chosen": -1.2795946598052979,
"logps/rejected": -2.8815693855285645,
"loss": -0.1234,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.2795946598052979,
"rewards/margins": 1.6019747257232666,
"rewards/rejected": -2.8815693855285645,
"step": 395
},
{
"epoch": 0.8374771002355405,
"grad_norm": 16.463230317307886,
"learning_rate": 3.8702478614051345e-08,
"logits/chosen": -2.365762710571289,
"logits/rejected": -2.099416971206665,
"logps/chosen": -1.3784708976745605,
"logps/rejected": -2.2646727561950684,
"loss": -0.1132,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -1.3784708976745605,
"rewards/margins": 0.8862019777297974,
"rewards/rejected": -2.2646727561950684,
"step": 400
},
{
"epoch": 0.8479455639884846,
"grad_norm": 18.03293576484585,
"learning_rate": 3.3952790595787986e-08,
"logits/chosen": -2.3217966556549072,
"logits/rejected": -1.9924516677856445,
"logps/chosen": -1.5933209657669067,
"logps/rejected": -3.219921112060547,
"loss": -0.1356,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -1.5933209657669067,
"rewards/margins": 1.626599907875061,
"rewards/rejected": -3.219921112060547,
"step": 405
},
{
"epoch": 0.8584140277414289,
"grad_norm": 10.746995288687499,
"learning_rate": 2.9492720416985e-08,
"logits/chosen": -2.2354073524475098,
"logits/rejected": -2.0534746646881104,
"logps/chosen": -1.3867294788360596,
"logps/rejected": -2.943979024887085,
"loss": -0.1385,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -1.3867294788360596,
"rewards/margins": 1.5572497844696045,
"rewards/rejected": -2.943979024887085,
"step": 410
},
{
"epoch": 0.8688824914943732,
"grad_norm": 33.58919538932842,
"learning_rate": 2.5328246937043525e-08,
"logits/chosen": -2.1925342082977295,
"logits/rejected": -2.1190714836120605,
"logps/chosen": -1.6617501974105835,
"logps/rejected": -2.4904022216796875,
"loss": -0.1213,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -1.6617501974105835,
"rewards/margins": 0.828652024269104,
"rewards/rejected": -2.4904022216796875,
"step": 415
},
{
"epoch": 0.8793509552473174,
"grad_norm": 21.053333844149208,
"learning_rate": 2.1464952759020856e-08,
"logits/chosen": -2.1517724990844727,
"logits/rejected": -2.0602805614471436,
"logps/chosen": -1.5017045736312866,
"logps/rejected": -2.838300943374634,
"loss": -0.1388,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -1.5017045736312866,
"rewards/margins": 1.3365962505340576,
"rewards/rejected": -2.838300943374634,
"step": 420
},
{
"epoch": 0.8898194190002617,
"grad_norm": 22.480002670218727,
"learning_rate": 1.7908016745981856e-08,
"logits/chosen": -2.3761887550354004,
"logits/rejected": -2.0966172218322754,
"logps/chosen": -1.2259721755981445,
"logps/rejected": -2.8700621128082275,
"loss": -0.1468,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -1.2259721755981445,
"rewards/margins": 1.644089937210083,
"rewards/rejected": -2.8700621128082275,
"step": 425
},
{
"epoch": 0.9002878827532059,
"grad_norm": 21.22984719288559,
"learning_rate": 1.4662207078575684e-08,
"logits/chosen": -2.2501072883605957,
"logits/rejected": -1.9055511951446533,
"logps/chosen": -1.4240307807922363,
"logps/rejected": -2.8532519340515137,
"loss": -0.1372,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -1.4240307807922363,
"rewards/margins": 1.4292211532592773,
"rewards/rejected": -2.8532519340515137,
"step": 430
},
{
"epoch": 0.9107563465061502,
"grad_norm": 40.370663526287096,
"learning_rate": 1.1731874863145142e-08,
"logits/chosen": -2.192682981491089,
"logits/rejected": -2.085170030593872,
"logps/chosen": -1.2839529514312744,
"logps/rejected": -2.737715721130371,
"loss": -0.1369,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -1.2839529514312744,
"rewards/margins": 1.4537627696990967,
"rewards/rejected": -2.737715721130371,
"step": 435
},
{
"epoch": 0.9212248102590945,
"grad_norm": 18.579655203147492,
"learning_rate": 9.12094829893642e-09,
"logits/chosen": -2.225111484527588,
"logits/rejected": -2.090142250061035,
"logps/chosen": -1.4594054222106934,
"logps/rejected": -2.5845236778259277,
"loss": -0.1324,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -1.4594054222106934,
"rewards/margins": 1.1251181364059448,
"rewards/rejected": -2.5845236778259277,
"step": 440
},
{
"epoch": 0.9316932740120387,
"grad_norm": 36.396751960507515,
"learning_rate": 6.832927412229017e-09,
"logits/chosen": -2.068014144897461,
"logits/rejected": -1.9256622791290283,
"logps/chosen": -1.551655650138855,
"logps/rejected": -2.514082193374634,
"loss": -0.1397,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -1.551655650138855,
"rewards/margins": 0.9624261856079102,
"rewards/rejected": -2.514082193374634,
"step": 445
},
{
"epoch": 0.942161737764983,
"grad_norm": 14.063862533994572,
"learning_rate": 4.8708793644441086e-09,
"logits/chosen": -2.364987373352051,
"logits/rejected": -2.075181245803833,
"logps/chosen": -1.2103084325790405,
"logps/rejected": -2.0531344413757324,
"loss": -0.1367,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -1.2103084325790405,
"rewards/margins": 0.8428258895874023,
"rewards/rejected": -2.0531344413757324,
"step": 450
},
{
"epoch": 0.9526302015179272,
"grad_norm": 25.580016907055246,
"learning_rate": 3.2374343405217884e-09,
"logits/chosen": -2.308936357498169,
"logits/rejected": -2.066481590270996,
"logps/chosen": -1.6806007623672485,
"logps/rejected": -2.6968746185302734,
"loss": -0.1109,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -1.6806007623672485,
"rewards/margins": 1.016274094581604,
"rewards/rejected": -2.6968746185302734,
"step": 455
},
{
"epoch": 0.9630986652708715,
"grad_norm": 15.882341955307094,
"learning_rate": 1.9347820230782295e-09,
"logits/chosen": -2.4680800437927246,
"logits/rejected": -1.8986313343048096,
"logps/chosen": -1.2367839813232422,
"logps/rejected": -3.457548141479492,
"loss": -0.1378,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -1.2367839813232422,
"rewards/margins": 2.22076416015625,
"rewards/rejected": -3.457548141479492,
"step": 460
},
{
"epoch": 0.9735671290238157,
"grad_norm": 26.022689528234064,
"learning_rate": 9.64668657069706e-10,
"logits/chosen": -2.198855400085449,
"logits/rejected": -1.8760795593261719,
"logps/chosen": -1.3434503078460693,
"logps/rejected": -3.093702793121338,
"loss": -0.1446,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -1.3434503078460693,
"rewards/margins": 1.750252366065979,
"rewards/rejected": -3.093702793121338,
"step": 465
},
{
"epoch": 0.98403559277676,
"grad_norm": 10.255280078989255,
"learning_rate": 3.2839470889836627e-10,
"logits/chosen": -2.194026470184326,
"logits/rejected": -2.0016260147094727,
"logps/chosen": -1.472402811050415,
"logps/rejected": -2.952603340148926,
"loss": -0.1302,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.472402811050415,
"rewards/margins": 1.4802005290985107,
"rewards/rejected": -2.952603340148926,
"step": 470
},
{
"epoch": 0.9945040565297043,
"grad_norm": 22.001525465495153,
"learning_rate": 2.6813123097352287e-11,
"logits/chosen": -2.2689156532287598,
"logits/rejected": -2.0389671325683594,
"logps/chosen": -1.387656807899475,
"logps/rejected": -2.36881947517395,
"loss": -0.1428,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.387656807899475,
"rewards/margins": 0.9811626672744751,
"rewards/rejected": -2.36881947517395,
"step": 475
},
{
"epoch": 0.998691442030882,
"step": 477,
"total_flos": 0.0,
"train_loss": -0.10732274213551975,
"train_runtime": 21833.2592,
"train_samples_per_second": 2.8,
"train_steps_per_second": 0.022
}
],
"logging_steps": 5,
"max_steps": 477,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 239,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}