cira-7b-dpo-lora-merge / trainer_state.json
David-Xu's picture
Training in progress, step 900
89bd182 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 100,
"global_step": 897,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 5.555555555555556e-08,
"logits/chosen": -0.5698858499526978,
"logits/rejected": -0.7608428597450256,
"logps/chosen": -628.0617065429688,
"logps/rejected": -619.150390625,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.01,
"learning_rate": 5.555555555555555e-07,
"logits/chosen": -0.6823484897613525,
"logits/rejected": -0.8546125888824463,
"logps/chosen": -777.4924926757812,
"logps/rejected": -634.3888549804688,
"loss": 0.6933,
"rewards/accuracies": 0.3611111044883728,
"rewards/chosen": -0.0005514526856131852,
"rewards/margins": -0.00033656222512945533,
"rewards/rejected": -0.00021489031496457756,
"step": 10
},
{
"epoch": 0.02,
"learning_rate": 1.111111111111111e-06,
"logits/chosen": -0.7127803564071655,
"logits/rejected": -0.8547202348709106,
"logps/chosen": -899.9298095703125,
"logps/rejected": -712.7205810546875,
"loss": 0.6912,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.005615489557385445,
"rewards/margins": 0.003956252709031105,
"rewards/rejected": 0.0016592368483543396,
"step": 20
},
{
"epoch": 0.03,
"learning_rate": 1.6666666666666667e-06,
"logits/chosen": -0.702410101890564,
"logits/rejected": -0.9075536727905273,
"logps/chosen": -799.2589111328125,
"logps/rejected": -717.34716796875,
"loss": 0.6903,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.01230540033429861,
"rewards/margins": 0.005792214069515467,
"rewards/rejected": 0.006513187196105719,
"step": 30
},
{
"epoch": 0.04,
"learning_rate": 2.222222222222222e-06,
"logits/chosen": -0.7552632093429565,
"logits/rejected": -0.9524158239364624,
"logps/chosen": -835.037109375,
"logps/rejected": -631.7636108398438,
"loss": 0.6904,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.025444606319069862,
"rewards/margins": 0.005691047292202711,
"rewards/rejected": 0.019753558561205864,
"step": 40
},
{
"epoch": 0.06,
"learning_rate": 2.7777777777777783e-06,
"logits/chosen": -0.7381051778793335,
"logits/rejected": -0.8419072031974792,
"logps/chosen": -788.6356811523438,
"logps/rejected": -730.46728515625,
"loss": 0.6901,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.04884253069758415,
"rewards/margins": 0.006506313569843769,
"rewards/rejected": 0.04233621805906296,
"step": 50
},
{
"epoch": 0.07,
"learning_rate": 3.3333333333333333e-06,
"logits/chosen": -0.6829635500907898,
"logits/rejected": -0.8546798825263977,
"logps/chosen": -803.8533325195312,
"logps/rejected": -685.8370361328125,
"loss": 0.684,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.07093875110149384,
"rewards/margins": 0.01889834739267826,
"rewards/rejected": 0.052040405571460724,
"step": 60
},
{
"epoch": 0.08,
"learning_rate": 3.88888888888889e-06,
"logits/chosen": -0.6931325197219849,
"logits/rejected": -0.8457029461860657,
"logps/chosen": -835.8101806640625,
"logps/rejected": -559.0534057617188,
"loss": 0.674,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.11979933828115463,
"rewards/margins": 0.04078071564435959,
"rewards/rejected": 0.07901863008737564,
"step": 70
},
{
"epoch": 0.09,
"learning_rate": 4.444444444444444e-06,
"logits/chosen": -0.7829106450080872,
"logits/rejected": -1.0052988529205322,
"logps/chosen": -857.0613403320312,
"logps/rejected": -581.3321533203125,
"loss": 0.6535,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.18766728043556213,
"rewards/margins": 0.08457548916339874,
"rewards/rejected": 0.10309179872274399,
"step": 80
},
{
"epoch": 0.1,
"learning_rate": 5e-06,
"logits/chosen": -0.8815785646438599,
"logits/rejected": -0.96449214220047,
"logps/chosen": -789.5969848632812,
"logps/rejected": -702.9019165039062,
"loss": 0.6794,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": 0.20566365122795105,
"rewards/margins": 0.033137936145067215,
"rewards/rejected": 0.17252573370933533,
"step": 90
},
{
"epoch": 0.11,
"learning_rate": 4.9981058784687895e-06,
"logits/chosen": -0.8663153648376465,
"logits/rejected": -1.0621612071990967,
"logps/chosen": -868.1552734375,
"logps/rejected": -664.0401611328125,
"loss": 0.6618,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.2395782470703125,
"rewards/margins": 0.06794992089271545,
"rewards/rejected": 0.17162834107875824,
"step": 100
},
{
"epoch": 0.11,
"eval_logits/chosen": -0.8082190752029419,
"eval_logits/rejected": -1.0028988122940063,
"eval_logps/chosen": -823.6102294921875,
"eval_logps/rejected": -665.3922729492188,
"eval_loss": 0.6664446592330933,
"eval_rewards/accuracies": 0.643216073513031,
"eval_rewards/chosen": 0.268514484167099,
"eval_rewards/margins": 0.061498455703258514,
"eval_rewards/rejected": 0.20701603591442108,
"eval_runtime": 181.7686,
"eval_samples_per_second": 2.19,
"eval_steps_per_second": 2.19,
"step": 100
},
{
"epoch": 0.12,
"learning_rate": 4.992426384032258e-06,
"logits/chosen": -0.8113915324211121,
"logits/rejected": -1.1199345588684082,
"logps/chosen": -766.8582763671875,
"logps/rejected": -536.8991088867188,
"loss": 0.6432,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": 0.2860961854457855,
"rewards/margins": 0.11079935729503632,
"rewards/rejected": 0.1752968281507492,
"step": 110
},
{
"epoch": 0.13,
"learning_rate": 4.982970122812566e-06,
"logits/chosen": -0.8754690289497375,
"logits/rejected": -0.9867092370986938,
"logps/chosen": -783.0337524414062,
"logps/rejected": -615.1319580078125,
"loss": 0.6622,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.3586369454860687,
"rewards/margins": 0.08097393810749054,
"rewards/rejected": 0.277662992477417,
"step": 120
},
{
"epoch": 0.14,
"learning_rate": 4.969751423856095e-06,
"logits/chosen": -0.8801660537719727,
"logits/rejected": -1.0285800695419312,
"logps/chosen": -857.1593017578125,
"logps/rejected": -575.841796875,
"loss": 0.6472,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.41047531366348267,
"rewards/margins": 0.11212178319692612,
"rewards/rejected": 0.29835352301597595,
"step": 130
},
{
"epoch": 0.16,
"learning_rate": 4.952790317420694e-06,
"logits/chosen": -0.9968269467353821,
"logits/rejected": -1.1242711544036865,
"logps/chosen": -922.9993286132812,
"logps/rejected": -759.6329345703125,
"loss": 0.6605,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.4612547755241394,
"rewards/margins": 0.084444060921669,
"rewards/rejected": 0.3768107295036316,
"step": 140
},
{
"epoch": 0.17,
"learning_rate": 4.932112504623876e-06,
"logits/chosen": -0.8811306953430176,
"logits/rejected": -1.0080692768096924,
"logps/chosen": -805.9840698242188,
"logps/rejected": -609.9173583984375,
"loss": 0.6289,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": 0.5022997260093689,
"rewards/margins": 0.16012665629386902,
"rewards/rejected": 0.34217312932014465,
"step": 150
},
{
"epoch": 0.18,
"learning_rate": 4.907749318497991e-06,
"logits/chosen": -0.9709693193435669,
"logits/rejected": -1.1685694456100464,
"logps/chosen": -752.7962646484375,
"logps/rejected": -603.7000122070312,
"loss": 0.6427,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.4435661733150482,
"rewards/margins": 0.13528670370578766,
"rewards/rejected": 0.30827948451042175,
"step": 160
},
{
"epoch": 0.19,
"learning_rate": 4.879737676511367e-06,
"logits/chosen": -0.9455941319465637,
"logits/rejected": -1.234893560409546,
"logps/chosen": -779.9085083007812,
"logps/rejected": -608.73291015625,
"loss": 0.6212,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": 0.5160066485404968,
"rewards/margins": 0.17969150841236115,
"rewards/rejected": 0.3363151252269745,
"step": 170
},
{
"epoch": 0.2,
"learning_rate": 4.848120024627372e-06,
"logits/chosen": -1.0280869007110596,
"logits/rejected": -1.0985267162322998,
"logps/chosen": -857.74609375,
"logps/rejected": -563.0875244140625,
"loss": 0.6145,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.5581072568893433,
"rewards/margins": 0.19067201018333435,
"rewards/rejected": 0.36743518710136414,
"step": 180
},
{
"epoch": 0.21,
"learning_rate": 4.812944272986166e-06,
"logits/chosen": -1.0628365278244019,
"logits/rejected": -1.268007516860962,
"logps/chosen": -884.31884765625,
"logps/rejected": -716.6842651367188,
"loss": 0.6429,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.5645692944526672,
"rewards/margins": 0.12494896352291107,
"rewards/rejected": 0.43962034583091736,
"step": 190
},
{
"epoch": 0.22,
"learning_rate": 4.774263723306599e-06,
"logits/chosen": -1.1383074522018433,
"logits/rejected": -1.194709062576294,
"logps/chosen": -809.2498168945312,
"logps/rejected": -548.3192138671875,
"loss": 0.6079,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.5740378499031067,
"rewards/margins": 0.20076331496238708,
"rewards/rejected": 0.3732745051383972,
"step": 200
},
{
"epoch": 0.22,
"eval_logits/chosen": -1.0529733896255493,
"eval_logits/rejected": -1.2187565565109253,
"eval_logps/chosen": -794.3279418945312,
"eval_logps/rejected": -642.638916015625,
"eval_loss": 0.6462566256523132,
"eval_rewards/accuracies": 0.6507537961006165,
"eval_rewards/chosen": 0.5613374710083008,
"eval_rewards/margins": 0.12678806483745575,
"eval_rewards/rejected": 0.4345494508743286,
"eval_runtime": 182.8948,
"eval_samples_per_second": 2.176,
"eval_steps_per_second": 2.176,
"step": 200
},
{
"epoch": 0.23,
"learning_rate": 4.732136988118259e-06,
"logits/chosen": -1.0591025352478027,
"logits/rejected": -1.1989389657974243,
"logps/chosen": -871.8209838867188,
"logps/rejected": -564.6995849609375,
"loss": 0.617,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.5726070404052734,
"rewards/margins": 0.1896977722644806,
"rewards/rejected": 0.38290935754776,
"step": 210
},
{
"epoch": 0.25,
"learning_rate": 4.6866279019460744e-06,
"logits/chosen": -1.0191378593444824,
"logits/rejected": -1.0795801877975464,
"logps/chosen": -749.4663696289062,
"logps/rejected": -598.59619140625,
"loss": 0.6369,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.5931397676467896,
"rewards/margins": 0.16538675129413605,
"rewards/rejected": 0.4277530312538147,
"step": 220
},
{
"epoch": 0.26,
"learning_rate": 4.637805424582033e-06,
"logits/chosen": -1.0304447412490845,
"logits/rejected": -1.1981227397918701,
"logps/chosen": -828.51171875,
"logps/rejected": -623.6228637695312,
"loss": 0.6275,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.6149128079414368,
"rewards/margins": 0.18098048865795135,
"rewards/rejected": 0.4339323043823242,
"step": 230
},
{
"epoch": 0.27,
"learning_rate": 4.585743536590599e-06,
"logits/chosen": -1.041161298751831,
"logits/rejected": -1.301321268081665,
"logps/chosen": -936.7018432617188,
"logps/rejected": -561.6809692382812,
"loss": 0.5616,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 0.6473914384841919,
"rewards/margins": 0.3042447566986084,
"rewards/rejected": 0.3431466221809387,
"step": 240
},
{
"epoch": 0.28,
"learning_rate": 4.530521127206173e-06,
"logits/chosen": -0.9962241053581238,
"logits/rejected": -1.1600590944290161,
"logps/chosen": -969.3317260742188,
"logps/rejected": -634.9935302734375,
"loss": 0.5807,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.6946474313735962,
"rewards/margins": 0.2879478335380554,
"rewards/rejected": 0.4066995680332184,
"step": 250
},
{
"epoch": 0.29,
"learning_rate": 4.472221874792454e-06,
"logits/chosen": -1.0495914220809937,
"logits/rejected": -1.218639612197876,
"logps/chosen": -753.663818359375,
"logps/rejected": -488.46539306640625,
"loss": 0.5902,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": 0.5778111219406128,
"rewards/margins": 0.25250428915023804,
"rewards/rejected": 0.32530683279037476,
"step": 260
},
{
"epoch": 0.3,
"learning_rate": 4.410934120044838e-06,
"logits/chosen": -1.087903618812561,
"logits/rejected": -1.208064317703247,
"logps/chosen": -869.251953125,
"logps/rejected": -616.2910766601562,
"loss": 0.617,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.6315535306930542,
"rewards/margins": 0.21421785652637482,
"rewards/rejected": 0.4173356592655182,
"step": 270
},
{
"epoch": 0.31,
"learning_rate": 4.346750732128023e-06,
"logits/chosen": -0.9508038759231567,
"logits/rejected": -1.128430724143982,
"logps/chosen": -715.2760009765625,
"logps/rejected": -632.925048828125,
"loss": 0.6425,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.6043787002563477,
"rewards/margins": 0.15197435021400452,
"rewards/rejected": 0.45240435004234314,
"step": 280
},
{
"epoch": 0.32,
"learning_rate": 4.279768967951605e-06,
"logits/chosen": -1.0877809524536133,
"logits/rejected": -1.196732997894287,
"logps/chosen": -732.0111694335938,
"logps/rejected": -645.47998046875,
"loss": 0.5899,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.5990191698074341,
"rewards/margins": 0.24647493660449982,
"rewards/rejected": 0.35254424810409546,
"step": 290
},
{
"epoch": 0.33,
"learning_rate": 4.210090324796965e-06,
"logits/chosen": -1.1224285364151,
"logits/rejected": -1.2125352621078491,
"logps/chosen": -701.78173828125,
"logps/rejected": -531.5074462890625,
"loss": 0.6029,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": 0.5454891324043274,
"rewards/margins": 0.2352382391691208,
"rewards/rejected": 0.3102509081363678,
"step": 300
},
{
"epoch": 0.33,
"eval_logits/chosen": -1.0366891622543335,
"eval_logits/rejected": -1.1964892148971558,
"eval_logps/chosen": -793.2078247070312,
"eval_logps/rejected": -644.8512573242188,
"eval_loss": 0.6359822750091553,
"eval_rewards/accuracies": 0.6557788848876953,
"eval_rewards/chosen": 0.5725387930870056,
"eval_rewards/margins": 0.16011320054531097,
"eval_rewards/rejected": 0.4124256670475006,
"eval_runtime": 182.1079,
"eval_samples_per_second": 2.186,
"eval_steps_per_second": 2.186,
"step": 300
},
{
"epoch": 0.35,
"learning_rate": 4.137820386518716e-06,
"logits/chosen": -1.007893681526184,
"logits/rejected": -1.2096047401428223,
"logps/chosen": -871.7268676757812,
"logps/rejected": -572.1101684570312,
"loss": 0.5544,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": 0.6873822808265686,
"rewards/margins": 0.35040804743766785,
"rewards/rejected": 0.33697420358657837,
"step": 310
},
{
"epoch": 0.36,
"learning_rate": 4.063068663553778e-06,
"logits/chosen": -1.0717604160308838,
"logits/rejected": -1.252614974975586,
"logps/chosen": -747.0122680664062,
"logps/rejected": -644.8168334960938,
"loss": 0.5943,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.6160334944725037,
"rewards/margins": 0.25227323174476624,
"rewards/rejected": 0.36376023292541504,
"step": 320
},
{
"epoch": 0.37,
"learning_rate": 3.9859484269805215e-06,
"logits/chosen": -1.1249014139175415,
"logits/rejected": -1.2282134294509888,
"logps/chosen": -852.0465087890625,
"logps/rejected": -594.2730712890625,
"loss": 0.6014,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": 0.6664583683013916,
"rewards/margins": 0.2528621554374695,
"rewards/rejected": 0.4135962128639221,
"step": 330
},
{
"epoch": 0.38,
"learning_rate": 3.906576536879416e-06,
"logits/chosen": -1.0270237922668457,
"logits/rejected": -1.242011308670044,
"logps/chosen": -827.2908325195312,
"logps/rejected": -647.3381958007812,
"loss": 0.5662,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.6910017728805542,
"rewards/margins": 0.3151417672634125,
"rewards/rejected": 0.3758600056171417,
"step": 340
},
{
"epoch": 0.39,
"learning_rate": 3.825073265255271e-06,
"logits/chosen": -1.1385161876678467,
"logits/rejected": -1.1466325521469116,
"logps/chosen": -724.5143432617188,
"logps/rejected": -692.8756103515625,
"loss": 0.6208,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.6377968192100525,
"rewards/margins": 0.20125746726989746,
"rewards/rejected": 0.43653935194015503,
"step": 350
},
{
"epoch": 0.4,
"learning_rate": 3.7415621137894055e-06,
"logits/chosen": -1.1019694805145264,
"logits/rejected": -1.3558331727981567,
"logps/chosen": -752.0270385742188,
"logps/rejected": -574.9373779296875,
"loss": 0.5653,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.6500095725059509,
"rewards/margins": 0.3266737759113312,
"rewards/rejected": 0.3233358561992645,
"step": 360
},
{
"epoch": 0.41,
"learning_rate": 3.656169626697889e-06,
"logits/chosen": -1.1742336750030518,
"logits/rejected": -1.2145332098007202,
"logps/chosen": -733.7666015625,
"logps/rejected": -577.3091430664062,
"loss": 0.6117,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.6057840585708618,
"rewards/margins": 0.2176964282989502,
"rewards/rejected": 0.3880876302719116,
"step": 370
},
{
"epoch": 0.42,
"learning_rate": 3.5690251989794443e-06,
"logits/chosen": -1.1451467275619507,
"logits/rejected": -1.4090936183929443,
"logps/chosen": -791.69287109375,
"logps/rejected": -664.351318359375,
"loss": 0.5472,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 0.7302919626235962,
"rewards/margins": 0.39029377698898315,
"rewards/rejected": 0.33999815583229065,
"step": 380
},
{
"epoch": 0.43,
"learning_rate": 3.480260880343565e-06,
"logits/chosen": -1.1370658874511719,
"logits/rejected": -1.268176555633545,
"logps/chosen": -704.5770263671875,
"logps/rejected": -549.7819213867188,
"loss": 0.5728,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.665199875831604,
"rewards/margins": 0.32814812660217285,
"rewards/rejected": 0.3370516896247864,
"step": 390
},
{
"epoch": 0.45,
"learning_rate": 3.390011175115956e-06,
"logits/chosen": -1.0938374996185303,
"logits/rejected": -1.208974003791809,
"logps/chosen": -662.2012939453125,
"logps/rejected": -570.6355590820312,
"loss": 0.6123,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.6198933124542236,
"rewards/margins": 0.2259746491909027,
"rewards/rejected": 0.39391860365867615,
"step": 400
},
{
"epoch": 0.45,
"eval_logits/chosen": -1.1220430135726929,
"eval_logits/rejected": -1.265822172164917,
"eval_logps/chosen": -787.7749633789062,
"eval_logps/rejected": -641.9633178710938,
"eval_loss": 0.6290538907051086,
"eval_rewards/accuracies": 0.660804033279419,
"eval_rewards/chosen": 0.6268669962882996,
"eval_rewards/margins": 0.18556177616119385,
"eval_rewards/rejected": 0.4413052201271057,
"eval_runtime": 182.0967,
"eval_samples_per_second": 2.186,
"eval_steps_per_second": 2.186,
"step": 400
},
{
"epoch": 0.46,
"learning_rate": 3.298412838424503e-06,
"logits/chosen": -1.1896421909332275,
"logits/rejected": -1.2548556327819824,
"logps/chosen": -818.2477416992188,
"logps/rejected": -607.974365234375,
"loss": 0.5652,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.6872069239616394,
"rewards/margins": 0.32436543703079224,
"rewards/rejected": 0.36284154653549194,
"step": 410
},
{
"epoch": 0.47,
"learning_rate": 3.205604668974607e-06,
"logits/chosen": -1.0528907775878906,
"logits/rejected": -1.1407417058944702,
"logps/chosen": -798.6719970703125,
"logps/rejected": -551.5494995117188,
"loss": 0.5562,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.7110682725906372,
"rewards/margins": 0.3393869996070862,
"rewards/rejected": 0.37168124318122864,
"step": 420
},
{
"epoch": 0.48,
"learning_rate": 3.111727298727888e-06,
"logits/chosen": -1.0838218927383423,
"logits/rejected": -1.3229072093963623,
"logps/chosen": -683.1719970703125,
"logps/rejected": -585.8663330078125,
"loss": 0.6278,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.5717045068740845,
"rewards/margins": 0.19018582999706268,
"rewards/rejected": 0.3815186619758606,
"step": 430
},
{
"epoch": 0.49,
"learning_rate": 3.0169229798029698e-06,
"logits/chosen": -1.1374232769012451,
"logits/rejected": -1.2010905742645264,
"logps/chosen": -860.7879638671875,
"logps/rejected": -513.562744140625,
"loss": 0.5396,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.7182751893997192,
"rewards/margins": 0.4120573103427887,
"rewards/rejected": 0.30621784925460815,
"step": 440
},
{
"epoch": 0.5,
"learning_rate": 2.9213353689212337e-06,
"logits/chosen": -1.1559637784957886,
"logits/rejected": -1.3046997785568237,
"logps/chosen": -728.3955078125,
"logps/rejected": -552.3156127929688,
"loss": 0.5476,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.6234858632087708,
"rewards/margins": 0.37362077832221985,
"rewards/rejected": 0.2498650997877121,
"step": 450
},
{
"epoch": 0.51,
"learning_rate": 2.8251093097241895e-06,
"logits/chosen": -1.0692123174667358,
"logits/rejected": -1.2266539335250854,
"logps/chosen": -792.2379150390625,
"logps/rejected": -661.500732421875,
"loss": 0.615,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.6842690706253052,
"rewards/margins": 0.26619952917099,
"rewards/rejected": 0.4180695414543152,
"step": 460
},
{
"epoch": 0.52,
"learning_rate": 2.7283906132923104e-06,
"logits/chosen": -1.1081640720367432,
"logits/rejected": -1.196122407913208,
"logps/chosen": -861.2401123046875,
"logps/rejected": -630.8099365234375,
"loss": 0.5369,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": 0.7281755208969116,
"rewards/margins": 0.4001568853855133,
"rewards/rejected": 0.3280186355113983,
"step": 470
},
{
"epoch": 0.54,
"learning_rate": 2.6313258371978996e-06,
"logits/chosen": -1.0956523418426514,
"logits/rejected": -1.2124046087265015,
"logps/chosen": -806.178466796875,
"logps/rejected": -556.8995361328125,
"loss": 0.5704,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.6232426762580872,
"rewards/margins": 0.3145415782928467,
"rewards/rejected": 0.3087010979652405,
"step": 480
},
{
"epoch": 0.55,
"learning_rate": 2.5340620634268167e-06,
"logits/chosen": -1.1129533052444458,
"logits/rejected": -1.3338663578033447,
"logps/chosen": -722.3619995117188,
"logps/rejected": -528.3717041015625,
"loss": 0.5459,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.6257763504981995,
"rewards/margins": 0.37175193428993225,
"rewards/rejected": 0.2540244162082672,
"step": 490
},
{
"epoch": 0.56,
"learning_rate": 2.436746675505545e-06,
"logits/chosen": -1.0671157836914062,
"logits/rejected": -1.2726842164993286,
"logps/chosen": -709.8563842773438,
"logps/rejected": -538.7957763671875,
"loss": 0.5596,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 0.6499138474464417,
"rewards/margins": 0.3205423951148987,
"rewards/rejected": 0.3293713927268982,
"step": 500
},
{
"epoch": 0.56,
"eval_logits/chosen": -1.085227131843567,
"eval_logits/rejected": -1.2330302000045776,
"eval_logps/chosen": -790.7927856445312,
"eval_logps/rejected": -646.7930297851562,
"eval_loss": 0.6230102181434631,
"eval_rewards/accuracies": 0.6683416962623596,
"eval_rewards/chosen": 0.5966897010803223,
"eval_rewards/margins": 0.20368172228336334,
"eval_rewards/rejected": 0.39300796389579773,
"eval_runtime": 182.1485,
"eval_samples_per_second": 2.185,
"eval_steps_per_second": 2.185,
"step": 500
},
{
"epoch": 0.57,
"learning_rate": 2.3395271351713515e-06,
"logits/chosen": -1.0355545282363892,
"logits/rejected": -1.2773510217666626,
"logps/chosen": -846.6865234375,
"logps/rejected": -515.6051635742188,
"loss": 0.5052,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 0.7559247016906738,
"rewards/margins": 0.48392027616500854,
"rewards/rejected": 0.2720043659210205,
"step": 510
},
{
"epoch": 0.58,
"learning_rate": 2.2425507589239154e-06,
"logits/chosen": -1.0901484489440918,
"logits/rejected": -1.2016217708587646,
"logps/chosen": -946.5360107421875,
"logps/rejected": -568.994384765625,
"loss": 0.519,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.8060197830200195,
"rewards/margins": 0.48728522658348083,
"rewards/rejected": 0.3187345862388611,
"step": 520
},
{
"epoch": 0.59,
"learning_rate": 2.145964494797051e-06,
"logits/chosen": -1.0998233556747437,
"logits/rejected": -1.2087761163711548,
"logps/chosen": -728.5558471679688,
"logps/rejected": -671.7276611328125,
"loss": 0.6116,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.6140581965446472,
"rewards/margins": 0.23174139857292175,
"rewards/rejected": 0.3823167681694031,
"step": 530
},
{
"epoch": 0.6,
"learning_rate": 2.049914699688762e-06,
"logits/chosen": -1.1990127563476562,
"logits/rejected": -1.2407617568969727,
"logps/chosen": -739.6600341796875,
"logps/rejected": -640.65380859375,
"loss": 0.5987,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": 0.6110137104988098,
"rewards/margins": 0.2484373301267624,
"rewards/rejected": 0.36257636547088623,
"step": 540
},
{
"epoch": 0.61,
"learning_rate": 1.954546917587033e-06,
"logits/chosen": -1.1063880920410156,
"logits/rejected": -1.2011101245880127,
"logps/chosen": -756.6212768554688,
"logps/rejected": -455.97979736328125,
"loss": 0.5123,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 0.7002598643302917,
"rewards/margins": 0.47388529777526855,
"rewards/rejected": 0.22637462615966797,
"step": 550
},
{
"epoch": 0.62,
"learning_rate": 1.8600056590274355e-06,
"logits/chosen": -1.095252513885498,
"logits/rejected": -1.1555767059326172,
"logps/chosen": -730.5572509765625,
"logps/rejected": -615.13134765625,
"loss": 0.619,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.6304242610931396,
"rewards/margins": 0.23751536011695862,
"rewards/rejected": 0.3929089605808258,
"step": 560
},
{
"epoch": 0.64,
"learning_rate": 1.766434182116708e-06,
"logits/chosen": -1.173472285270691,
"logits/rejected": -1.1808496713638306,
"logps/chosen": -754.0181884765625,
"logps/rejected": -660.9012451171875,
"loss": 0.6232,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.5536493062973022,
"rewards/margins": 0.21175837516784668,
"rewards/rejected": 0.34189099073410034,
"step": 570
},
{
"epoch": 0.65,
"learning_rate": 1.6739742754541515e-06,
"logits/chosen": -0.9952909350395203,
"logits/rejected": -1.2373504638671875,
"logps/chosen": -740.7738037109375,
"logps/rejected": -658.3294067382812,
"loss": 0.57,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.6436474323272705,
"rewards/margins": 0.3406828045845032,
"rewards/rejected": 0.30296462774276733,
"step": 580
},
{
"epoch": 0.66,
"learning_rate": 1.582766043279752e-06,
"logits/chosen": -1.0711066722869873,
"logits/rejected": -1.2044506072998047,
"logps/chosen": -700.0191650390625,
"logps/rejected": -573.0767211914062,
"loss": 0.5885,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": 0.5518115758895874,
"rewards/margins": 0.28940147161483765,
"rewards/rejected": 0.26241010427474976,
"step": 590
},
{
"epoch": 0.67,
"learning_rate": 1.4929476931746167e-06,
"logits/chosen": -0.9843810796737671,
"logits/rejected": -1.1629724502563477,
"logps/chosen": -815.2390747070312,
"logps/rejected": -556.5223999023438,
"loss": 0.5382,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 0.6965817213058472,
"rewards/margins": 0.40665000677108765,
"rewards/rejected": 0.2899317145347595,
"step": 600
},
{
"epoch": 0.67,
"eval_logits/chosen": -1.054700493812561,
"eval_logits/rejected": -1.2033686637878418,
"eval_logps/chosen": -793.2485961914062,
"eval_logps/rejected": -650.0925903320312,
"eval_loss": 0.6198973059654236,
"eval_rewards/accuracies": 0.6708542704582214,
"eval_rewards/chosen": 0.5721314549446106,
"eval_rewards/margins": 0.21211864054203033,
"eval_rewards/rejected": 0.36001279950141907,
"eval_runtime": 182.0552,
"eval_samples_per_second": 2.186,
"eval_steps_per_second": 2.186,
"step": 600
},
{
"epoch": 0.68,
"learning_rate": 1.4046553266354126e-06,
"logits/chosen": -1.1579875946044922,
"logits/rejected": -1.1627570390701294,
"logps/chosen": -710.740966796875,
"logps/rejected": -705.4386596679688,
"loss": 0.651,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.5279535055160522,
"rewards/margins": 0.15107524394989014,
"rewards/rejected": 0.3768783211708069,
"step": 610
},
{
"epoch": 0.69,
"learning_rate": 1.318022732840141e-06,
"logits/chosen": -1.082596778869629,
"logits/rejected": -1.2048799991607666,
"logps/chosen": -773.497802734375,
"logps/rejected": -556.2584838867188,
"loss": 0.5922,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.5189529657363892,
"rewards/margins": 0.2554023265838623,
"rewards/rejected": 0.26355063915252686,
"step": 620
},
{
"epoch": 0.7,
"learning_rate": 1.2331811859177722e-06,
"logits/chosen": -1.1044152975082397,
"logits/rejected": -1.3350199460983276,
"logps/chosen": -763.0442504882812,
"logps/rejected": -605.748046875,
"loss": 0.5383,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.6514481902122498,
"rewards/margins": 0.42245060205459595,
"rewards/rejected": 0.2289975881576538,
"step": 630
},
{
"epoch": 0.71,
"learning_rate": 1.150259246028921e-06,
"logits/chosen": -0.9752397537231445,
"logits/rejected": -1.1639258861541748,
"logps/chosen": -736.8223266601562,
"logps/rejected": -610.3005981445312,
"loss": 0.5711,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.6042425036430359,
"rewards/margins": 0.3154502809047699,
"rewards/rejected": 0.288792222738266,
"step": 640
},
{
"epoch": 0.72,
"learning_rate": 1.0693825645589887e-06,
"logits/chosen": -0.9845565557479858,
"logits/rejected": -1.205209732055664,
"logps/chosen": -778.3556518554688,
"logps/rejected": -546.72265625,
"loss": 0.5316,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.6856497526168823,
"rewards/margins": 0.4366929531097412,
"rewards/rejected": 0.24895676970481873,
"step": 650
},
{
"epoch": 0.74,
"learning_rate": 9.9067369371897e-07,
"logits/chosen": -1.1334224939346313,
"logits/rejected": -1.261704683303833,
"logps/chosen": -795.8439331054688,
"logps/rejected": -630.4856567382812,
"loss": 0.5529,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.6260807514190674,
"rewards/margins": 0.3732090890407562,
"rewards/rejected": 0.25287163257598877,
"step": 660
},
{
"epoch": 0.75,
"learning_rate": 9.14251900842432e-07,
"logits/chosen": -1.0554249286651611,
"logits/rejected": -1.1402459144592285,
"logps/chosen": -697.9842529296875,
"logps/rejected": -549.4150390625,
"loss": 0.5811,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.530659556388855,
"rewards/margins": 0.29245835542678833,
"rewards/rejected": 0.23820118606090546,
"step": 670
},
{
"epoch": 0.76,
"learning_rate": 8.402329876600462e-07,
"logits/chosen": -1.0341722965240479,
"logits/rejected": -1.126814603805542,
"logps/chosen": -864.65966796875,
"logps/rejected": -632.652099609375,
"loss": 0.5377,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.72022545337677,
"rewards/margins": 0.4036920964717865,
"rewards/rejected": 0.3165333867073059,
"step": 680
},
{
"epoch": 0.77,
"learning_rate": 7.687291148255527e-07,
"logits/chosen": -1.0247344970703125,
"logits/rejected": -1.1219924688339233,
"logps/chosen": -944.3568115234375,
"logps/rejected": -662.272216796875,
"loss": 0.5183,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.8120288848876953,
"rewards/margins": 0.4406498968601227,
"rewards/rejected": 0.37137895822525024,
"step": 690
},
{
"epoch": 0.78,
"learning_rate": 6.998486319590323e-07,
"logits/chosen": -1.0355504751205444,
"logits/rejected": -1.0869207382202148,
"logps/chosen": -768.5548095703125,
"logps/rejected": -638.7791748046875,
"loss": 0.5952,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": 0.6144096851348877,
"rewards/margins": 0.27622345089912415,
"rewards/rejected": 0.33818623423576355,
"step": 700
},
{
"epoch": 0.78,
"eval_logits/chosen": -1.0323765277862549,
"eval_logits/rejected": -1.182665228843689,
"eval_logps/chosen": -794.96044921875,
"eval_logps/rejected": -652.0419921875,
"eval_loss": 0.6185528635978699,
"eval_rewards/accuracies": 0.6783919334411621,
"eval_rewards/chosen": 0.5550140738487244,
"eval_rewards/margins": 0.21449486911296844,
"eval_rewards/rejected": 0.3405191898345947,
"eval_runtime": 182.2003,
"eval_samples_per_second": 2.184,
"eval_steps_per_second": 2.184,
"step": 700
},
{
"epoch": 0.79,
"learning_rate": 6.336959134650278e-07,
"logits/chosen": -1.1061818599700928,
"logits/rejected": -1.1679563522338867,
"logps/chosen": -823.1553955078125,
"logps/rejected": -556.3447875976562,
"loss": 0.5435,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.6446937322616577,
"rewards/margins": 0.3869627118110657,
"rewards/rejected": 0.2577310800552368,
"step": 710
},
{
"epoch": 0.8,
"learning_rate": 5.703712003742965e-07,
"logits/chosen": -1.013837456703186,
"logits/rejected": -1.2135231494903564,
"logps/chosen": -790.27490234375,
"logps/rejected": -591.6689453125,
"loss": 0.5499,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.6053417325019836,
"rewards/margins": 0.37499576807022095,
"rewards/rejected": 0.2303459644317627,
"step": 720
},
{
"epoch": 0.81,
"learning_rate": 5.099704484488569e-07,
"logits/chosen": -1.0917243957519531,
"logits/rejected": -1.1174938678741455,
"logps/chosen": -800.6257934570312,
"logps/rejected": -576.2686767578125,
"loss": 0.5762,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": 0.5678650736808777,
"rewards/margins": 0.30998459458351135,
"rewards/rejected": 0.25788044929504395,
"step": 730
},
{
"epoch": 0.82,
"learning_rate": 4.525851827804881e-07,
"logits/chosen": -1.0324584245681763,
"logits/rejected": -1.2507680654525757,
"logps/chosen": -849.1965942382812,
"logps/rejected": -620.4736328125,
"loss": 0.5563,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": 0.6641701459884644,
"rewards/margins": 0.3699476420879364,
"rewards/rejected": 0.2942224442958832,
"step": 740
},
{
"epoch": 0.84,
"learning_rate": 3.983023591030113e-07,
"logits/chosen": -1.0120137929916382,
"logits/rejected": -1.2262170314788818,
"logps/chosen": -834.7081909179688,
"logps/rejected": -564.6629638671875,
"loss": 0.5074,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.6688768267631531,
"rewards/margins": 0.48412173986434937,
"rewards/rejected": 0.1847551316022873,
"step": 750
},
{
"epoch": 0.85,
"learning_rate": 3.472042320285071e-07,
"logits/chosen": -1.0783493518829346,
"logits/rejected": -1.2198795080184937,
"logps/chosen": -753.7183837890625,
"logps/rejected": -542.9868774414062,
"loss": 0.5736,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.5568192005157471,
"rewards/margins": 0.3265761733055115,
"rewards/rejected": 0.2302430123090744,
"step": 760
},
{
"epoch": 0.86,
"learning_rate": 2.9936823040713464e-07,
"logits/chosen": -1.0335241556167603,
"logits/rejected": -1.1321046352386475,
"logps/chosen": -708.8296508789062,
"logps/rejected": -637.4175415039062,
"loss": 0.587,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.6139867901802063,
"rewards/margins": 0.3065374493598938,
"rewards/rejected": 0.3074493110179901,
"step": 770
},
{
"epoch": 0.87,
"learning_rate": 2.5486683999940335e-07,
"logits/chosen": -1.0177851915359497,
"logits/rejected": -1.2519065141677856,
"logps/chosen": -767.447021484375,
"logps/rejected": -714.9308471679688,
"loss": 0.5753,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.6363558769226074,
"rewards/margins": 0.33481666445732117,
"rewards/rejected": 0.30153924226760864,
"step": 780
},
{
"epoch": 0.88,
"learning_rate": 2.137674936387049e-07,
"logits/chosen": -1.1295353174209595,
"logits/rejected": -1.1564512252807617,
"logps/chosen": -883.1470947265625,
"logps/rejected": -631.2457275390625,
"loss": 0.5517,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.695419430732727,
"rewards/margins": 0.38906994462013245,
"rewards/rejected": 0.30634939670562744,
"step": 790
},
{
"epoch": 0.89,
"learning_rate": 1.7613246905052812e-07,
"logits/chosen": -1.0573675632476807,
"logits/rejected": -1.1950371265411377,
"logps/chosen": -760.9122314453125,
"logps/rejected": -630.8699340820312,
"loss": 0.5792,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.5912384390830994,
"rewards/margins": 0.32750827074050903,
"rewards/rejected": 0.2637301981449127,
"step": 800
},
{
"epoch": 0.89,
"eval_logits/chosen": -1.0307811498641968,
"eval_logits/rejected": -1.1811619997024536,
"eval_logps/chosen": -795.125,
"eval_logps/rejected": -652.2705078125,
"eval_loss": 0.6182109117507935,
"eval_rewards/accuracies": 0.6783919334411621,
"eval_rewards/chosen": 0.553367018699646,
"eval_rewards/margins": 0.2151338756084442,
"eval_rewards/rejected": 0.3382331430912018,
"eval_runtime": 182.0989,
"eval_samples_per_second": 2.186,
"eval_steps_per_second": 2.186,
"step": 800
},
{
"epoch": 0.9,
"learning_rate": 1.4201879448319356e-07,
"logits/chosen": -1.0675430297851562,
"logits/rejected": -1.1287426948547363,
"logps/chosen": -711.8330078125,
"logps/rejected": -551.1827392578125,
"loss": 0.5842,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.5683950185775757,
"rewards/margins": 0.2922239601612091,
"rewards/rejected": 0.2761710286140442,
"step": 810
},
{
"epoch": 0.91,
"learning_rate": 1.1147816229310549e-07,
"logits/chosen": -1.016808032989502,
"logits/rejected": -1.1210168600082397,
"logps/chosen": -691.964599609375,
"logps/rejected": -732.0343017578125,
"loss": 0.637,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.5624631643295288,
"rewards/margins": 0.19316788017749786,
"rewards/rejected": 0.3692953288555145,
"step": 820
},
{
"epoch": 0.93,
"learning_rate": 8.455685061547119e-08,
"logits/chosen": -1.1123645305633545,
"logits/rejected": -1.132232904434204,
"logps/chosen": -859.2161254882812,
"logps/rejected": -662.1595458984375,
"loss": 0.5684,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.7145462036132812,
"rewards/margins": 0.3542642295360565,
"rewards/rejected": 0.36028194427490234,
"step": 830
},
{
"epoch": 0.94,
"learning_rate": 6.129565323916814e-08,
"logits/chosen": -1.0635792016983032,
"logits/rejected": -1.1541740894317627,
"logps/chosen": -795.508056640625,
"logps/rejected": -634.9236450195312,
"loss": 0.5683,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.6695567965507507,
"rewards/margins": 0.3517037034034729,
"rewards/rejected": 0.3178531527519226,
"step": 840
},
{
"epoch": 0.95,
"learning_rate": 4.1729817792030004e-08,
"logits/chosen": -1.1519498825073242,
"logits/rejected": -1.1835066080093384,
"logps/chosen": -785.9249267578125,
"logps/rejected": -617.5560913085938,
"loss": 0.579,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.589630126953125,
"rewards/margins": 0.2917063534259796,
"rewards/rejected": 0.297923743724823,
"step": 850
},
{
"epoch": 0.96,
"learning_rate": 2.588899233021358e-08,
"logits/chosen": -1.0503435134887695,
"logits/rejected": -1.218450665473938,
"logps/chosen": -840.5927734375,
"logps/rejected": -520.814208984375,
"loss": 0.5205,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.6971001029014587,
"rewards/margins": 0.4537246823310852,
"rewards/rejected": 0.24337545037269592,
"step": 860
},
{
"epoch": 0.97,
"learning_rate": 1.3797180412583322e-08,
"logits/chosen": -1.0805847644805908,
"logits/rejected": -1.2713693380355835,
"logps/chosen": -800.3570556640625,
"logps/rejected": -632.8238525390625,
"loss": 0.5758,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.5660972595214844,
"rewards/margins": 0.31249552965164185,
"rewards/rejected": 0.25360172986984253,
"step": 870
},
{
"epoch": 0.98,
"learning_rate": 5.4727047281821764e-09,
"logits/chosen": -1.0479118824005127,
"logits/rejected": -1.3037327527999878,
"logps/chosen": -783.6592407226562,
"logps/rejected": -529.0799560546875,
"loss": 0.5171,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.6520611047744751,
"rewards/margins": 0.47154098749160767,
"rewards/rejected": 0.18052014708518982,
"step": 880
},
{
"epoch": 0.99,
"learning_rate": 9.281793319140808e-10,
"logits/chosen": -1.0751326084136963,
"logits/rejected": -1.2520333528518677,
"logps/chosen": -834.0614013671875,
"logps/rejected": -488.2327575683594,
"loss": 0.5113,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.6582953929901123,
"rewards/margins": 0.4814949631690979,
"rewards/rejected": 0.17680040001869202,
"step": 890
},
{
"epoch": 1.0,
"step": 897,
"total_flos": 0.0,
"train_loss": 0.06080360662445443,
"train_runtime": 395.6281,
"train_samples_per_second": 9.069,
"train_steps_per_second": 2.267
}
],
"logging_steps": 10,
"max_steps": 897,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 100,
"total_flos": 0.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}