| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 19.37219730941704, | |
| "eval_steps": 500, | |
| "global_step": 540, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.35874439461883406, | |
| "grad_norm": 5.573123455047607, | |
| "learning_rate": 9.259259259259259e-07, | |
| "logits/chosen": -2.3576977252960205, | |
| "logits/rejected": -2.3553850650787354, | |
| "logps/chosen": -61.0846061706543, | |
| "logps/rejected": -126.1152572631836, | |
| "loss": 0.6942, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": -0.0012381849810481071, | |
| "rewards/margins": -0.010831715539097786, | |
| "rewards/rejected": 0.009593529626727104, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.7174887892376681, | |
| "grad_norm": 3.8175244331359863, | |
| "learning_rate": 1.8518518518518519e-06, | |
| "logits/chosen": -2.3618054389953613, | |
| "logits/rejected": -2.384873867034912, | |
| "logps/chosen": -63.2865104675293, | |
| "logps/rejected": -81.8724365234375, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": -0.002874236088246107, | |
| "rewards/margins": -0.003056168556213379, | |
| "rewards/rejected": 0.00018193255527876318, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.0762331838565022, | |
| "grad_norm": 4.383662700653076, | |
| "learning_rate": 2.7777777777777783e-06, | |
| "logits/chosen": -2.3333091735839844, | |
| "logits/rejected": -2.3545360565185547, | |
| "logps/chosen": -76.02193450927734, | |
| "logps/rejected": -88.19218444824219, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.008687756024301052, | |
| "rewards/margins": -0.0016246589366346598, | |
| "rewards/rejected": -0.007063096854835749, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.4349775784753362, | |
| "grad_norm": 4.095139503479004, | |
| "learning_rate": 3.7037037037037037e-06, | |
| "logits/chosen": -2.3483197689056396, | |
| "logits/rejected": -2.363131523132324, | |
| "logps/chosen": -66.08761596679688, | |
| "logps/rejected": -78.22222900390625, | |
| "loss": 0.6889, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.014177674427628517, | |
| "rewards/margins": 0.012472175993025303, | |
| "rewards/rejected": -0.026649847626686096, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.7937219730941703, | |
| "grad_norm": 4.034746170043945, | |
| "learning_rate": 4.62962962962963e-06, | |
| "logits/chosen": -2.31459903717041, | |
| "logits/rejected": -2.335706949234009, | |
| "logps/chosen": -69.31814575195312, | |
| "logps/rejected": -77.91527557373047, | |
| "loss": 0.6823, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": -0.04294499754905701, | |
| "rewards/margins": 0.007987136952579021, | |
| "rewards/rejected": -0.050932131707668304, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 2.1524663677130045, | |
| "grad_norm": 4.620233535766602, | |
| "learning_rate": 4.998119881260576e-06, | |
| "logits/chosen": -2.364583969116211, | |
| "logits/rejected": -2.3598217964172363, | |
| "logps/chosen": -69.84207153320312, | |
| "logps/rejected": -79.67817687988281, | |
| "loss": 0.6717, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.06739393621683121, | |
| "rewards/margins": 0.050740111619234085, | |
| "rewards/rejected": -0.118134044110775, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 2.5112107623318387, | |
| "grad_norm": 4.450082302093506, | |
| "learning_rate": 4.9866405060165044e-06, | |
| "logits/chosen": -2.298379421234131, | |
| "logits/rejected": -2.3169469833374023, | |
| "logps/chosen": -79.13768005371094, | |
| "logps/rejected": -79.24192810058594, | |
| "loss": 0.6499, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.1282506138086319, | |
| "rewards/margins": 0.08815944194793701, | |
| "rewards/rejected": -0.21641004085540771, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.8699551569506725, | |
| "grad_norm": 4.7328877449035645, | |
| "learning_rate": 4.964774158361991e-06, | |
| "logits/chosen": -2.3363795280456543, | |
| "logits/rejected": -2.3331849575042725, | |
| "logps/chosen": -77.52214050292969, | |
| "logps/rejected": -86.90077209472656, | |
| "loss": 0.6346, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.22365322709083557, | |
| "rewards/margins": 0.1937221735715866, | |
| "rewards/rejected": -0.4173754155635834, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 3.2286995515695067, | |
| "grad_norm": 4.016219615936279, | |
| "learning_rate": 4.93261217644956e-06, | |
| "logits/chosen": -2.30871844291687, | |
| "logits/rejected": -2.3049044609069824, | |
| "logps/chosen": -65.6957778930664, | |
| "logps/rejected": -77.02860260009766, | |
| "loss": 0.6004, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.23905189335346222, | |
| "rewards/margins": 0.2932591438293457, | |
| "rewards/rejected": -0.5323110222816467, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 3.587443946188341, | |
| "grad_norm": 4.553714752197266, | |
| "learning_rate": 4.8902889044347e-06, | |
| "logits/chosen": -2.3068125247955322, | |
| "logits/rejected": -2.294462203979492, | |
| "logps/chosen": -85.46375274658203, | |
| "logps/rejected": -79.72854614257812, | |
| "loss": 0.5816, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.3506511151790619, | |
| "rewards/margins": 0.2997886538505554, | |
| "rewards/rejected": -0.6504397392272949, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 3.9461883408071747, | |
| "grad_norm": 4.804734230041504, | |
| "learning_rate": 4.837981131305475e-06, | |
| "logits/chosen": -2.3317933082580566, | |
| "logits/rejected": -2.3493664264678955, | |
| "logps/chosen": -68.45785522460938, | |
| "logps/rejected": -88.28421020507812, | |
| "loss": 0.5672, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.4280410706996918, | |
| "rewards/margins": 0.3461475670337677, | |
| "rewards/rejected": -0.7741886377334595, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 4.304932735426009, | |
| "grad_norm": 4.650998592376709, | |
| "learning_rate": 4.775907352415367e-06, | |
| "logits/chosen": -2.3039748668670654, | |
| "logits/rejected": -2.301835775375366, | |
| "logps/chosen": -76.26728820800781, | |
| "logps/rejected": -87.93345642089844, | |
| "loss": 0.5258, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.4848151206970215, | |
| "rewards/margins": 0.6179562211036682, | |
| "rewards/rejected": -1.102771282196045, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 4.663677130044843, | |
| "grad_norm": 4.17809534072876, | |
| "learning_rate": 4.70432685680402e-06, | |
| "logits/chosen": -2.299182653427124, | |
| "logits/rejected": -2.291717290878296, | |
| "logps/chosen": -77.27394104003906, | |
| "logps/rejected": -76.96590423583984, | |
| "loss": 0.5141, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.5012012720108032, | |
| "rewards/margins": 0.39593780040740967, | |
| "rewards/rejected": -0.8971391916275024, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 5.022421524663677, | |
| "grad_norm": 4.4158735275268555, | |
| "learning_rate": 4.623538644118244e-06, | |
| "logits/chosen": -2.267500400543213, | |
| "logits/rejected": -2.271239995956421, | |
| "logps/chosen": -77.53614807128906, | |
| "logps/rejected": -105.53443908691406, | |
| "loss": 0.4929, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.47719526290893555, | |
| "rewards/margins": 0.7529892921447754, | |
| "rewards/rejected": -1.230184555053711, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 5.381165919282511, | |
| "grad_norm": 4.5556640625, | |
| "learning_rate": 4.533880175657419e-06, | |
| "logits/chosen": -2.2397818565368652, | |
| "logits/rejected": -2.2796878814697266, | |
| "logps/chosen": -75.30777740478516, | |
| "logps/rejected": -97.59407043457031, | |
| "loss": 0.4545, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.3897199034690857, | |
| "rewards/margins": 0.8326279520988464, | |
| "rewards/rejected": -1.2223479747772217, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 5.739910313901345, | |
| "grad_norm": 5.067012786865234, | |
| "learning_rate": 4.435725964760331e-06, | |
| "logits/chosen": -2.284294843673706, | |
| "logits/rejected": -2.264681339263916, | |
| "logps/chosen": -87.51654052734375, | |
| "logps/rejected": -82.73191833496094, | |
| "loss": 0.4392, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.42664116621017456, | |
| "rewards/margins": 0.7053943872451782, | |
| "rewards/rejected": -1.132035493850708, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 6.098654708520179, | |
| "grad_norm": 4.425904273986816, | |
| "learning_rate": 4.329486012421531e-06, | |
| "logits/chosen": -2.296356201171875, | |
| "logits/rejected": -2.2917532920837402, | |
| "logps/chosen": -76.75779724121094, | |
| "logps/rejected": -90.15029907226562, | |
| "loss": 0.4153, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.5271193385124207, | |
| "rewards/margins": 0.7608373761177063, | |
| "rewards/rejected": -1.2879568338394165, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 6.457399103139013, | |
| "grad_norm": 4.75029182434082, | |
| "learning_rate": 4.215604094671835e-06, | |
| "logits/chosen": -2.2561912536621094, | |
| "logits/rejected": -2.253948450088501, | |
| "logps/chosen": -68.14691162109375, | |
| "logps/rejected": -85.39097595214844, | |
| "loss": 0.3785, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.5065070390701294, | |
| "rewards/margins": 1.024714708328247, | |
| "rewards/rejected": -1.531221628189087, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 6.816143497757848, | |
| "grad_norm": 5.048355579376221, | |
| "learning_rate": 4.094555908876765e-06, | |
| "logits/chosen": -2.323347568511963, | |
| "logits/rejected": -2.284219741821289, | |
| "logps/chosen": -74.8355484008789, | |
| "logps/rejected": -85.9789047241211, | |
| "loss": 0.3697, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.533603310585022, | |
| "rewards/margins": 1.2120137214660645, | |
| "rewards/rejected": -1.7456169128417969, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 7.174887892376682, | |
| "grad_norm": 5.01900053024292, | |
| "learning_rate": 3.966847086696045e-06, | |
| "logits/chosen": -2.300300121307373, | |
| "logits/rejected": -2.3180909156799316, | |
| "logps/chosen": -86.38873291015625, | |
| "logps/rejected": -89.72538757324219, | |
| "loss": 0.3611, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.7510987520217896, | |
| "rewards/margins": 0.9809296727180481, | |
| "rewards/rejected": -1.7320283651351929, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 7.533632286995516, | |
| "grad_norm": 4.92667818069458, | |
| "learning_rate": 3.833011082004229e-06, | |
| "logits/chosen": -2.297994375228882, | |
| "logits/rejected": -2.288382053375244, | |
| "logps/chosen": -71.64348602294922, | |
| "logps/rejected": -84.01541900634766, | |
| "loss": 0.3248, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.8383395075798035, | |
| "rewards/margins": 1.047420859336853, | |
| "rewards/rejected": -1.8857605457305908, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 7.8923766816143495, | |
| "grad_norm": 5.767988204956055, | |
| "learning_rate": 3.693606942594873e-06, | |
| "logits/chosen": -2.194643259048462, | |
| "logits/rejected": -2.1788828372955322, | |
| "logps/chosen": -83.70150756835938, | |
| "logps/rejected": -101.75688171386719, | |
| "loss": 0.3004, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.8993238210678101, | |
| "rewards/margins": 1.521676778793335, | |
| "rewards/rejected": -2.4210009574890137, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 8.251121076233185, | |
| "grad_norm": 4.738667964935303, | |
| "learning_rate": 3.549216974976073e-06, | |
| "logits/chosen": -2.2021820545196533, | |
| "logits/rejected": -2.234846591949463, | |
| "logps/chosen": -85.80267333984375, | |
| "logps/rejected": -126.21333312988281, | |
| "loss": 0.2709, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.1428090333938599, | |
| "rewards/margins": 1.669283151626587, | |
| "rewards/rejected": -2.8120923042297363, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 8.609865470852018, | |
| "grad_norm": 5.022470951080322, | |
| "learning_rate": 3.400444312011776e-06, | |
| "logits/chosen": -2.2339494228363037, | |
| "logits/rejected": -2.259722948074341, | |
| "logps/chosen": -79.16600036621094, | |
| "logps/rejected": -100.4195327758789, | |
| "loss": 0.2653, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.070847988128662, | |
| "rewards/margins": 1.8333218097686768, | |
| "rewards/rejected": -2.9041695594787598, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 8.968609865470851, | |
| "grad_norm": 6.2118096351623535, | |
| "learning_rate": 3.2479103935691047e-06, | |
| "logits/chosen": -2.197798490524292, | |
| "logits/rejected": -2.2084178924560547, | |
| "logps/chosen": -85.23640441894531, | |
| "logps/rejected": -116.28971862792969, | |
| "loss": 0.2426, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.2635414600372314, | |
| "rewards/margins": 1.8116687536239624, | |
| "rewards/rejected": -3.0752100944519043, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 9.327354260089686, | |
| "grad_norm": 5.798530101776123, | |
| "learning_rate": 3.092252370695298e-06, | |
| "logits/chosen": -2.1891565322875977, | |
| "logits/rejected": -2.2006757259368896, | |
| "logps/chosen": -77.15342712402344, | |
| "logps/rejected": -129.3516845703125, | |
| "loss": 0.2356, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -1.1600430011749268, | |
| "rewards/margins": 2.2219691276550293, | |
| "rewards/rejected": -3.382011890411377, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 9.68609865470852, | |
| "grad_norm": 6.839400768280029, | |
| "learning_rate": 2.9341204441673267e-06, | |
| "logits/chosen": -2.209928035736084, | |
| "logits/rejected": -2.2113702297210693, | |
| "logps/chosen": -79.22918701171875, | |
| "logps/rejected": -105.42140197753906, | |
| "loss": 0.2016, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.5536689758300781, | |
| "rewards/margins": 1.9507955312728882, | |
| "rewards/rejected": -3.504464626312256, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 10.044843049327355, | |
| "grad_norm": 4.958732604980469, | |
| "learning_rate": 2.7741751485313295e-06, | |
| "logits/chosen": -2.1307220458984375, | |
| "logits/rejected": -2.1803646087646484, | |
| "logps/chosen": -81.51334381103516, | |
| "logps/rejected": -134.2724151611328, | |
| "loss": 0.1909, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.3520208597183228, | |
| "rewards/margins": 2.556453227996826, | |
| "rewards/rejected": -3.9084744453430176, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 10.403587443946188, | |
| "grad_norm": 5.86897611618042, | |
| "learning_rate": 2.6130845929767662e-06, | |
| "logits/chosen": -2.139816999435425, | |
| "logits/rejected": -2.14668607711792, | |
| "logps/chosen": -87.5683822631836, | |
| "logps/rejected": -116.4664535522461, | |
| "loss": 0.1731, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.6710402965545654, | |
| "rewards/margins": 2.187368869781494, | |
| "rewards/rejected": -3.8584091663360596, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 10.762331838565022, | |
| "grad_norm": 4.753023147583008, | |
| "learning_rate": 2.4515216705704396e-06, | |
| "logits/chosen": -2.1322758197784424, | |
| "logits/rejected": -2.1539082527160645, | |
| "logps/chosen": -74.43647766113281, | |
| "logps/rejected": -104.6937255859375, | |
| "loss": 0.1715, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.6632559299468994, | |
| "rewards/margins": 2.3588404655456543, | |
| "rewards/rejected": -4.022096157073975, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 11.121076233183857, | |
| "grad_norm": 4.878518104553223, | |
| "learning_rate": 2.290161247507733e-06, | |
| "logits/chosen": -2.0836901664733887, | |
| "logits/rejected": -2.145113229751587, | |
| "logps/chosen": -102.97503662109375, | |
| "logps/rejected": -232.6587371826172, | |
| "loss": 0.1618, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.2652416229248047, | |
| "rewards/margins": 5.368926048278809, | |
| "rewards/rejected": -7.634167671203613, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 11.47982062780269, | |
| "grad_norm": 5.1058244705200195, | |
| "learning_rate": 2.129677344121879e-06, | |
| "logits/chosen": -2.157466173171997, | |
| "logits/rejected": -2.155546188354492, | |
| "logps/chosen": -95.01036834716797, | |
| "logps/rejected": -123.38541412353516, | |
| "loss": 0.1352, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.074333906173706, | |
| "rewards/margins": 2.805738925933838, | |
| "rewards/rejected": -4.880073070526123, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 11.838565022421525, | |
| "grad_norm": 5.019484043121338, | |
| "learning_rate": 1.970740319426474e-06, | |
| "logits/chosen": -2.1018729209899902, | |
| "logits/rejected": -2.116931438446045, | |
| "logps/chosen": -106.68685150146484, | |
| "logps/rejected": -124.8704833984375, | |
| "loss": 0.134, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.079122304916382, | |
| "rewards/margins": 2.727034568786621, | |
| "rewards/rejected": -4.806157112121582, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 12.197309417040358, | |
| "grad_norm": 4.403121471405029, | |
| "learning_rate": 1.8140140709517467e-06, | |
| "logits/chosen": -2.116088628768921, | |
| "logits/rejected": -2.1317386627197266, | |
| "logps/chosen": -98.20851135253906, | |
| "logps/rejected": -127.1669921875, | |
| "loss": 0.1183, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.1554946899414062, | |
| "rewards/margins": 2.7777180671691895, | |
| "rewards/rejected": -4.9332122802734375, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 12.556053811659194, | |
| "grad_norm": 4.341458797454834, | |
| "learning_rate": 1.6601532615711452e-06, | |
| "logits/chosen": -2.0657103061676025, | |
| "logits/rejected": -2.106078863143921, | |
| "logps/chosen": -97.9491958618164, | |
| "logps/rejected": -138.72988891601562, | |
| "loss": 0.1135, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.296207904815674, | |
| "rewards/margins": 3.2302253246307373, | |
| "rewards/rejected": -5.526432991027832, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 12.914798206278027, | |
| "grad_norm": 4.83184814453125, | |
| "learning_rate": 1.509800584902108e-06, | |
| "logits/chosen": -2.0779662132263184, | |
| "logits/rejected": -2.1079659461975098, | |
| "logps/chosen": -104.03936767578125, | |
| "logps/rejected": -128.2595672607422, | |
| "loss": 0.1083, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.526491165161133, | |
| "rewards/margins": 3.116278886795044, | |
| "rewards/rejected": -5.642770290374756, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 13.27354260089686, | |
| "grad_norm": 4.208549976348877, | |
| "learning_rate": 1.3635840807037487e-06, | |
| "logits/chosen": -2.0661022663116455, | |
| "logits/rejected": -2.1056418418884277, | |
| "logps/chosen": -83.17430114746094, | |
| "logps/rejected": -128.30245971679688, | |
| "loss": 0.0921, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.5337133407592773, | |
| "rewards/margins": 3.501122236251831, | |
| "rewards/rejected": -6.034835338592529, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 13.632286995515695, | |
| "grad_norm": 6.424919128417969, | |
| "learning_rate": 1.2221145114853172e-06, | |
| "logits/chosen": -2.0403847694396973, | |
| "logits/rejected": -2.0700008869171143, | |
| "logps/chosen": -105.04042053222656, | |
| "logps/rejected": -136.50611877441406, | |
| "loss": 0.1013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.4945261478424072, | |
| "rewards/margins": 3.4043514728546143, | |
| "rewards/rejected": -5.898877143859863, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 13.991031390134529, | |
| "grad_norm": 4.117794036865234, | |
| "learning_rate": 1.085982811283654e-06, | |
| "logits/chosen": -2.041865587234497, | |
| "logits/rejected": -2.0910420417785645, | |
| "logps/chosen": -114.09271240234375, | |
| "logps/rejected": -143.93826293945312, | |
| "loss": 0.0921, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.0714075565338135, | |
| "rewards/margins": 3.2793126106262207, | |
| "rewards/rejected": -6.350719451904297, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 14.349775784753364, | |
| "grad_norm": 4.791493892669678, | |
| "learning_rate": 9.557576172663577e-07, | |
| "logits/chosen": -2.048828601837158, | |
| "logits/rejected": -2.064530372619629, | |
| "logps/chosen": -101.74618530273438, | |
| "logps/rejected": -145.79129028320312, | |
| "loss": 0.0828, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.637481212615967, | |
| "rewards/margins": 3.572800397872925, | |
| "rewards/rejected": -6.2102813720703125, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 14.708520179372197, | |
| "grad_norm": 3.2735021114349365, | |
| "learning_rate": 8.319828944714508e-07, | |
| "logits/chosen": -2.0613853931427, | |
| "logits/rejected": -2.0846283435821533, | |
| "logps/chosen": -105.1708755493164, | |
| "logps/rejected": -135.4827117919922, | |
| "loss": 0.078, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.1742587089538574, | |
| "rewards/margins": 3.548724412918091, | |
| "rewards/rejected": -6.722983360290527, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 15.067264573991032, | |
| "grad_norm": 3.9526467323303223, | |
| "learning_rate": 7.151756636052529e-07, | |
| "logits/chosen": -2.0549488067626953, | |
| "logits/rejected": -2.0416836738586426, | |
| "logps/chosen": -111.19111633300781, | |
| "logps/rejected": -137.0021209716797, | |
| "loss": 0.0759, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.218317747116089, | |
| "rewards/margins": 3.6524665355682373, | |
| "rewards/rejected": -6.870783805847168, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 15.426008968609866, | |
| "grad_norm": 3.010849952697754, | |
| "learning_rate": 6.058238413897052e-07, | |
| "logits/chosen": -2.02756404876709, | |
| "logits/rejected": -2.025636672973633, | |
| "logps/chosen": -118.96656799316406, | |
| "logps/rejected": -152.67581176757812, | |
| "loss": 0.0776, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.3787834644317627, | |
| "rewards/margins": 3.612543821334839, | |
| "rewards/rejected": -6.991326808929443, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 15.784753363228699, | |
| "grad_norm": 3.2514679431915283, | |
| "learning_rate": 5.043842024802675e-07, | |
| "logits/chosen": -2.0224342346191406, | |
| "logits/rejected": -2.024209499359131, | |
| "logps/chosen": -100.25440216064453, | |
| "logps/rejected": -148.87582397460938, | |
| "loss": 0.0707, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.8083739280700684, | |
| "rewards/margins": 3.6432766914367676, | |
| "rewards/rejected": -6.451650142669678, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 16.143497757847534, | |
| "grad_norm": 4.400411605834961, | |
| "learning_rate": 4.1128047146765936e-07, | |
| "logits/chosen": -2.0213570594787598, | |
| "logits/rejected": -2.0214767456054688, | |
| "logps/chosen": -99.02058410644531, | |
| "logps/rejected": -140.37637329101562, | |
| "loss": 0.0631, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.552924394607544, | |
| "rewards/margins": 3.7184109687805176, | |
| "rewards/rejected": -7.271335601806641, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 16.50224215246637, | |
| "grad_norm": 4.746975898742676, | |
| "learning_rate": 3.269015529333805e-07, | |
| "logits/chosen": -1.9746768474578857, | |
| "logits/rejected": -2.0172688961029053, | |
| "logps/chosen": -108.14969635009766, | |
| "logps/rejected": -155.90933227539062, | |
| "loss": 0.0699, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.199385404586792, | |
| "rewards/margins": 4.21014928817749, | |
| "rewards/rejected": -7.4095354080200195, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 16.8609865470852, | |
| "grad_norm": 5.0379319190979, | |
| "learning_rate": 2.515999069522676e-07, | |
| "logits/chosen": -2.031390428543091, | |
| "logits/rejected": -2.0676589012145996, | |
| "logps/chosen": -118.65937805175781, | |
| "logps/rejected": -143.47872924804688, | |
| "loss": 0.0624, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.263296604156494, | |
| "rewards/margins": 3.919255018234253, | |
| "rewards/rejected": -7.182551383972168, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 17.219730941704036, | |
| "grad_norm": 3.773524284362793, | |
| "learning_rate": 1.8569007682777417e-07, | |
| "logits/chosen": -2.0223498344421387, | |
| "logits/rejected": -2.028245687484741, | |
| "logps/chosen": -98.20159149169922, | |
| "logps/rejected": -126.34019470214844, | |
| "loss": 0.0658, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.06457781791687, | |
| "rewards/margins": 3.4116806983947754, | |
| "rewards/rejected": -6.476258277893066, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 17.57847533632287, | |
| "grad_norm": 4.083519458770752, | |
| "learning_rate": 1.2944737520980883e-07, | |
| "logits/chosen": -1.9950075149536133, | |
| "logits/rejected": -2.0348432064056396, | |
| "logps/chosen": -115.1516342163086, | |
| "logps/rejected": -160.4219970703125, | |
| "loss": 0.066, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.625664234161377, | |
| "rewards/margins": 3.8759894371032715, | |
| "rewards/rejected": -7.501654148101807, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 17.937219730941703, | |
| "grad_norm": 4.096585273742676, | |
| "learning_rate": 8.310673408334496e-08, | |
| "logits/chosen": -1.9964863061904907, | |
| "logits/rejected": -1.9901469945907593, | |
| "logps/chosen": -107.2447509765625, | |
| "logps/rejected": -144.03562927246094, | |
| "loss": 0.0564, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.5319015979766846, | |
| "rewards/margins": 3.939161777496338, | |
| "rewards/rejected": -7.471064567565918, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 18.295964125560538, | |
| "grad_norm": 4.045906066894531, | |
| "learning_rate": 4.6861723431538273e-08, | |
| "logits/chosen": -1.9917402267456055, | |
| "logits/rejected": -2.024575710296631, | |
| "logps/chosen": -106.6629638671875, | |
| "logps/rejected": -165.69158935546875, | |
| "loss": 0.0627, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.547753095626831, | |
| "rewards/margins": 4.24500036239624, | |
| "rewards/rejected": -7.79275369644165, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 18.654708520179373, | |
| "grad_norm": 4.705228328704834, | |
| "learning_rate": 2.0863742672497244e-08, | |
| "logits/chosen": -1.9936788082122803, | |
| "logits/rejected": -2.010652780532837, | |
| "logps/chosen": -105.31771087646484, | |
| "logps/rejected": -149.80661010742188, | |
| "loss": 0.0576, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.252545118331909, | |
| "rewards/margins": 4.177481651306152, | |
| "rewards/rejected": -7.430027008056641, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 19.013452914798208, | |
| "grad_norm": 4.246400833129883, | |
| "learning_rate": 5.221388247169945e-09, | |
| "logits/chosen": -2.0077335834503174, | |
| "logits/rejected": -2.0036730766296387, | |
| "logps/chosen": -115.3812026977539, | |
| "logps/rejected": -141.77145385742188, | |
| "loss": 0.0655, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.551786422729492, | |
| "rewards/margins": 3.6455719470977783, | |
| "rewards/rejected": -7.197358131408691, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 19.37219730941704, | |
| "grad_norm": 3.7989847660064697, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -2.0463144779205322, | |
| "logits/rejected": -2.069537401199341, | |
| "logps/chosen": -103.3205795288086, | |
| "logps/rejected": -152.32081604003906, | |
| "loss": 0.054, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.5541598796844482, | |
| "rewards/margins": 4.153356552124023, | |
| "rewards/rejected": -7.707517147064209, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 19.37219730941704, | |
| "step": 540, | |
| "total_flos": 1.981675043968516e+18, | |
| "train_loss": 0.2879459043343862, | |
| "train_runtime": 4187.4777, | |
| "train_samples_per_second": 8.521, | |
| "train_steps_per_second": 0.129 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 540, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.981675043968516e+18, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |