| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9995419147961521, | |
| "eval_steps": 100, | |
| "global_step": 1091, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0009161704076958314, | |
| "grad_norm": 2.538351535797119, | |
| "learning_rate": 9.09090909090909e-09, | |
| "logits/chosen": -0.612647533416748, | |
| "logits/rejected": -0.43005383014678955, | |
| "logps/chosen": -269.1338195800781, | |
| "logps/rejected": -265.996826171875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.009161704076958314, | |
| "grad_norm": 3.99981427192688, | |
| "learning_rate": 9.09090909090909e-08, | |
| "logits/chosen": -0.7947888374328613, | |
| "logits/rejected": -0.8272799849510193, | |
| "logps/chosen": -172.9255828857422, | |
| "logps/rejected": -192.15106201171875, | |
| "loss": 0.6942, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.012904312461614609, | |
| "rewards/margins": -0.015133652836084366, | |
| "rewards/rejected": 0.002229340374469757, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.01832340815391663, | |
| "grad_norm": 1.3273310661315918, | |
| "learning_rate": 1.818181818181818e-07, | |
| "logits/chosen": -0.7671724557876587, | |
| "logits/rejected": -0.7643235921859741, | |
| "logps/chosen": -177.4793243408203, | |
| "logps/rejected": -201.07656860351562, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.006979703903198242, | |
| "rewards/margins": -0.00390571728348732, | |
| "rewards/rejected": -0.003073985455557704, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.027485112230874943, | |
| "grad_norm": 2.06701922416687, | |
| "learning_rate": 2.727272727272727e-07, | |
| "logits/chosen": -0.7477768659591675, | |
| "logits/rejected": -0.770847737789154, | |
| "logps/chosen": -219.7743377685547, | |
| "logps/rejected": -253.83798217773438, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.007790957577526569, | |
| "rewards/margins": 0.0053612953051924706, | |
| "rewards/rejected": -0.013152251951396465, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.03664681630783326, | |
| "grad_norm": 2.176619052886963, | |
| "learning_rate": 3.636363636363636e-07, | |
| "logits/chosen": -0.6840143203735352, | |
| "logits/rejected": -0.6969764828681946, | |
| "logps/chosen": -159.89796447753906, | |
| "logps/rejected": -218.1041717529297, | |
| "loss": 0.6902, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": 0.0005666827782988548, | |
| "rewards/margins": 0.012707608751952648, | |
| "rewards/rejected": -0.012140927836298943, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.04580852038479157, | |
| "grad_norm": 2.182387590408325, | |
| "learning_rate": 4.545454545454545e-07, | |
| "logits/chosen": -0.5698288083076477, | |
| "logits/rejected": -0.7136921286582947, | |
| "logps/chosen": -203.5851287841797, | |
| "logps/rejected": -231.3826446533203, | |
| "loss": 0.6878, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.01093050092458725, | |
| "rewards/margins": 0.021884005516767502, | |
| "rewards/rejected": -0.010953502729535103, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.054970224461749886, | |
| "grad_norm": 1.757957935333252, | |
| "learning_rate": 5.454545454545454e-07, | |
| "logits/chosen": -0.7015949487686157, | |
| "logits/rejected": -0.7727667689323425, | |
| "logps/chosen": -233.5804901123047, | |
| "logps/rejected": -278.2530212402344, | |
| "loss": 0.6876, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.021485041826963425, | |
| "rewards/margins": 0.0006370929768308997, | |
| "rewards/rejected": 0.020847950130701065, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.0641319285387082, | |
| "grad_norm": 2.0394017696380615, | |
| "learning_rate": 6.363636363636363e-07, | |
| "logits/chosen": -0.8689748644828796, | |
| "logits/rejected": -0.808820366859436, | |
| "logps/chosen": -147.0337371826172, | |
| "logps/rejected": -209.3056182861328, | |
| "loss": 0.6815, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.05839823558926582, | |
| "rewards/margins": 0.03576263412833214, | |
| "rewards/rejected": 0.022635603323578835, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.07329363261566652, | |
| "grad_norm": 3.2475733757019043, | |
| "learning_rate": 7.272727272727272e-07, | |
| "logits/chosen": -0.7972738742828369, | |
| "logits/rejected": -0.8733490109443665, | |
| "logps/chosen": -202.31976318359375, | |
| "logps/rejected": -230.676513671875, | |
| "loss": 0.6706, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.0555371418595314, | |
| "rewards/margins": 0.04019620642066002, | |
| "rewards/rejected": 0.015340929850935936, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.08245533669262482, | |
| "grad_norm": 2.1004042625427246, | |
| "learning_rate": 8.181818181818182e-07, | |
| "logits/chosen": -0.7756220698356628, | |
| "logits/rejected": -0.7883769869804382, | |
| "logps/chosen": -181.07911682128906, | |
| "logps/rejected": -253.35226440429688, | |
| "loss": 0.657, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.09145340323448181, | |
| "rewards/margins": 0.05497417598962784, | |
| "rewards/rejected": 0.03647923097014427, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.09161704076958314, | |
| "grad_norm": 1.711387276649475, | |
| "learning_rate": 9.09090909090909e-07, | |
| "logits/chosen": -0.6988880634307861, | |
| "logits/rejected": -0.695970356464386, | |
| "logps/chosen": -210.9344940185547, | |
| "logps/rejected": -292.63885498046875, | |
| "loss": 0.6437, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.2389262169599533, | |
| "rewards/margins": 0.13863424956798553, | |
| "rewards/rejected": 0.10029196739196777, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.09161704076958314, | |
| "eval_logits/chosen": -0.7206078171730042, | |
| "eval_logits/rejected": -0.7319283485412598, | |
| "eval_logps/chosen": -228.31155395507812, | |
| "eval_logps/rejected": -265.4963073730469, | |
| "eval_loss": 0.6128209829330444, | |
| "eval_rewards/accuracies": 0.7254335284233093, | |
| "eval_rewards/chosen": 0.30496734380722046, | |
| "eval_rewards/margins": 0.23110172152519226, | |
| "eval_rewards/rejected": 0.0738656222820282, | |
| "eval_runtime": 264.1865, | |
| "eval_samples_per_second": 10.459, | |
| "eval_steps_per_second": 1.31, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.10077874484654145, | |
| "grad_norm": 2.3657712936401367, | |
| "learning_rate": 1e-06, | |
| "logits/chosen": -0.9458168745040894, | |
| "logits/rejected": -0.912136435508728, | |
| "logps/chosen": -266.46087646484375, | |
| "logps/rejected": -255.1190643310547, | |
| "loss": 0.6226, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.24092476069927216, | |
| "rewards/margins": 0.15844842791557312, | |
| "rewards/rejected": 0.08247633278369904, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.10994044892349977, | |
| "grad_norm": 1.905785083770752, | |
| "learning_rate": 9.997436315234263e-07, | |
| "logits/chosen": -0.7388381958007812, | |
| "logits/rejected": -0.7873945236206055, | |
| "logps/chosen": -167.81883239746094, | |
| "logps/rejected": -176.68295288085938, | |
| "loss": 0.6136, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.3122307062149048, | |
| "rewards/margins": 0.20933406054973602, | |
| "rewards/rejected": 0.10289661586284637, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.11910215300045808, | |
| "grad_norm": 1.5484569072723389, | |
| "learning_rate": 9.989747889928883e-07, | |
| "logits/chosen": -0.7524275183677673, | |
| "logits/rejected": -0.8292320370674133, | |
| "logps/chosen": -197.0650177001953, | |
| "logps/rejected": -237.57839965820312, | |
| "loss": 0.5938, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.5554567575454712, | |
| "rewards/margins": 0.4092523455619812, | |
| "rewards/rejected": 0.14620442688465118, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.1282638570774164, | |
| "grad_norm": 1.4368153810501099, | |
| "learning_rate": 9.976942608363393e-07, | |
| "logits/chosen": -0.632690966129303, | |
| "logits/rejected": -0.7571207284927368, | |
| "logps/chosen": -173.34666442871094, | |
| "logps/rejected": -211.22317504882812, | |
| "loss": 0.5709, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.58498615026474, | |
| "rewards/margins": 0.40937310457229614, | |
| "rewards/rejected": 0.17561307549476624, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.1374255611543747, | |
| "grad_norm": 1.5536640882492065, | |
| "learning_rate": 9.9590336020199e-07, | |
| "logits/chosen": -0.6405803561210632, | |
| "logits/rejected": -0.7340038418769836, | |
| "logps/chosen": -182.8199462890625, | |
| "logps/rejected": -243.0964813232422, | |
| "loss": 0.5731, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.6845705509185791, | |
| "rewards/margins": 0.4665129780769348, | |
| "rewards/rejected": 0.2180575579404831, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.14658726523133303, | |
| "grad_norm": 1.1569141149520874, | |
| "learning_rate": 9.936039236117095e-07, | |
| "logits/chosen": -0.8644550442695618, | |
| "logits/rejected": -0.8035632967948914, | |
| "logps/chosen": -200.20794677734375, | |
| "logps/rejected": -239.16915893554688, | |
| "loss": 0.544, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.826056957244873, | |
| "rewards/margins": 0.5113142132759094, | |
| "rewards/rejected": 0.3147428035736084, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.15574896930829135, | |
| "grad_norm": 1.5735410451889038, | |
| "learning_rate": 9.907983090777206e-07, | |
| "logits/chosen": -0.7325465083122253, | |
| "logits/rejected": -0.7175777554512024, | |
| "logps/chosen": -195.422607421875, | |
| "logps/rejected": -211.19418334960938, | |
| "loss": 0.5275, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.7453492879867554, | |
| "rewards/margins": 0.558306872844696, | |
| "rewards/rejected": 0.1870425045490265, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.16491067338524965, | |
| "grad_norm": 1.4354184865951538, | |
| "learning_rate": 9.874893936845187e-07, | |
| "logits/chosen": -0.6600767374038696, | |
| "logits/rejected": -0.7184512615203857, | |
| "logps/chosen": -214.3909149169922, | |
| "logps/rejected": -291.69512939453125, | |
| "loss": 0.5309, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 0.8096879124641418, | |
| "rewards/margins": 0.5819981098175049, | |
| "rewards/rejected": 0.2276897132396698, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.17407237746220797, | |
| "grad_norm": 1.9378727674484253, | |
| "learning_rate": 9.836805706384983e-07, | |
| "logits/chosen": -0.796898603439331, | |
| "logits/rejected": -0.7903488874435425, | |
| "logps/chosen": -151.75074768066406, | |
| "logps/rejected": -193.89720153808594, | |
| "loss": 0.4979, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 0.9862753748893738, | |
| "rewards/margins": 0.6403971910476685, | |
| "rewards/rejected": 0.3458781838417053, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.1832340815391663, | |
| "grad_norm": 1.368870735168457, | |
| "learning_rate": 9.793757457883061e-07, | |
| "logits/chosen": -0.7233944535255432, | |
| "logits/rejected": -0.7641991972923279, | |
| "logps/chosen": -131.52737426757812, | |
| "logps/rejected": -177.9940185546875, | |
| "loss": 0.5175, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 0.935263454914093, | |
| "rewards/margins": 0.7129745483398438, | |
| "rewards/rejected": 0.2222888469696045, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.1832340815391663, | |
| "eval_logits/chosen": -0.7067926526069641, | |
| "eval_logits/rejected": -0.7133929133415222, | |
| "eval_logps/chosen": -211.8815460205078, | |
| "eval_logps/rejected": -261.1459655761719, | |
| "eval_loss": 0.4986709654331207, | |
| "eval_rewards/accuracies": 0.823699414730072, | |
| "eval_rewards/chosen": 1.1264668703079224, | |
| "eval_rewards/margins": 0.8350852131843567, | |
| "eval_rewards/rejected": 0.29138168692588806, | |
| "eval_runtime": 253.2388, | |
| "eval_samples_per_second": 10.911, | |
| "eval_steps_per_second": 1.366, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.1923957856161246, | |
| "grad_norm": 1.181227207183838, | |
| "learning_rate": 9.745793336194975e-07, | |
| "logits/chosen": -0.744937539100647, | |
| "logits/rejected": -0.753220796585083, | |
| "logps/chosen": -157.19473266601562, | |
| "logps/rejected": -238.6236114501953, | |
| "loss": 0.485, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 0.9487007260322571, | |
| "rewards/margins": 0.5822451710700989, | |
| "rewards/rejected": 0.366455614566803, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.2015574896930829, | |
| "grad_norm": 2.322333812713623, | |
| "learning_rate": 9.69296252727595e-07, | |
| "logits/chosen": -0.7686578035354614, | |
| "logits/rejected": -0.7208544015884399, | |
| "logps/chosen": -161.76914978027344, | |
| "logps/rejected": -203.6949005126953, | |
| "loss": 0.486, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 1.1700208187103271, | |
| "rewards/margins": 0.8085041046142578, | |
| "rewards/rejected": 0.36151671409606934, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.21071919377004122, | |
| "grad_norm": 3.2851109504699707, | |
| "learning_rate": 9.63531920774199e-07, | |
| "logits/chosen": -0.8623464703559875, | |
| "logits/rejected": -0.864261269569397, | |
| "logps/chosen": -132.91622924804688, | |
| "logps/rejected": -190.5310516357422, | |
| "loss": 0.4889, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.2793647050857544, | |
| "rewards/margins": 0.8540099859237671, | |
| "rewards/rejected": 0.4253547787666321, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.21988089784699955, | |
| "grad_norm": 2.092069387435913, | |
| "learning_rate": 9.572922489313142e-07, | |
| "logits/chosen": -0.8346965909004211, | |
| "logits/rejected": -0.8602925539016724, | |
| "logps/chosen": -177.9022216796875, | |
| "logps/rejected": -209.01220703125, | |
| "loss": 0.4294, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 1.2863993644714355, | |
| "rewards/margins": 0.5576584935188293, | |
| "rewards/rejected": 0.7287408709526062, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.22904260192395787, | |
| "grad_norm": 1.2932628393173218, | |
| "learning_rate": 9.505836358195993e-07, | |
| "logits/chosen": -0.7524776458740234, | |
| "logits/rejected": -0.8281770944595337, | |
| "logps/chosen": -144.74905395507812, | |
| "logps/rejected": -226.6658172607422, | |
| "loss": 0.4293, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.3644707202911377, | |
| "rewards/margins": 0.8554983139038086, | |
| "rewards/rejected": 0.5089724659919739, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.23820430600091616, | |
| "grad_norm": 1.3950133323669434, | |
| "learning_rate": 9.434129609467483e-07, | |
| "logits/chosen": -0.6876672506332397, | |
| "logits/rejected": -0.6876662969589233, | |
| "logps/chosen": -263.18585205078125, | |
| "logps/rejected": -264.4542541503906, | |
| "loss": 0.4518, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 1.4175243377685547, | |
| "rewards/margins": 0.8856536746025085, | |
| "rewards/rejected": 0.5318707227706909, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.24736601007787448, | |
| "grad_norm": 1.4819687604904175, | |
| "learning_rate": 9.357875776527333e-07, | |
| "logits/chosen": -0.6820667386054993, | |
| "logits/rejected": -0.6322587132453918, | |
| "logps/chosen": -173.28085327148438, | |
| "logps/rejected": -196.96194458007812, | |
| "loss": 0.4433, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.1709938049316406, | |
| "rewards/margins": 0.7361399531364441, | |
| "rewards/rejected": 0.4348538815975189, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.2565277141548328, | |
| "grad_norm": 1.2729604244232178, | |
| "learning_rate": 9.27715305569148e-07, | |
| "logits/chosen": -0.6576655507087708, | |
| "logits/rejected": -0.6413074731826782, | |
| "logps/chosen": -159.5121307373047, | |
| "logps/rejected": -207.47195434570312, | |
| "loss": 0.4029, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.8569538593292236, | |
| "rewards/margins": 1.4031217098236084, | |
| "rewards/rejected": 0.45383185148239136, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.2656894182317911, | |
| "grad_norm": 1.4309152364730835, | |
| "learning_rate": 9.192044226003788e-07, | |
| "logits/chosen": -0.7165778875350952, | |
| "logits/rejected": -0.7207854390144348, | |
| "logps/chosen": -171.72616577148438, | |
| "logps/rejected": -212.6331024169922, | |
| "loss": 0.4505, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 1.8912508487701416, | |
| "rewards/margins": 1.2686805725097656, | |
| "rewards/rejected": 0.6225701570510864, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.2748511223087494, | |
| "grad_norm": 1.5633419752120972, | |
| "learning_rate": 9.102636564348294e-07, | |
| "logits/chosen": -0.5978932976722717, | |
| "logits/rejected": -0.7183943390846252, | |
| "logps/chosen": -173.94984436035156, | |
| "logps/rejected": -202.29295349121094, | |
| "loss": 0.3903, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.945762276649475, | |
| "rewards/margins": 1.3720638751983643, | |
| "rewards/rejected": 0.5736981630325317, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.2748511223087494, | |
| "eval_logits/chosen": -0.6642194390296936, | |
| "eval_logits/rejected": -0.6700440049171448, | |
| "eval_logps/chosen": -199.8172607421875, | |
| "eval_logps/rejected": -257.196044921875, | |
| "eval_loss": 0.4279369115829468, | |
| "eval_rewards/accuracies": 0.8468208312988281, | |
| "eval_rewards/chosen": 1.7296818494796753, | |
| "eval_rewards/margins": 1.240803837776184, | |
| "eval_rewards/rejected": 0.48887789249420166, | |
| "eval_runtime": 253.1385, | |
| "eval_samples_per_second": 10.915, | |
| "eval_steps_per_second": 1.367, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.28401282638570774, | |
| "grad_norm": 1.3390766382217407, | |
| "learning_rate": 9.009021755949051e-07, | |
| "logits/chosen": -0.6982103586196899, | |
| "logits/rejected": -0.7174701690673828, | |
| "logps/chosen": -159.46409606933594, | |
| "logps/rejected": -160.19491577148438, | |
| "loss": 0.4083, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.7644752264022827, | |
| "rewards/margins": 1.0205423831939697, | |
| "rewards/rejected": 0.7439330220222473, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.29317453046266606, | |
| "grad_norm": 1.3620954751968384, | |
| "learning_rate": 8.911295800349314e-07, | |
| "logits/chosen": -0.6473032832145691, | |
| "logits/rejected": -0.6676048040390015, | |
| "logps/chosen": -232.72787475585938, | |
| "logps/rejected": -252.5911407470703, | |
| "loss": 0.4178, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.7409747838974, | |
| "rewards/margins": 1.134825348854065, | |
| "rewards/rejected": 0.6061495542526245, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.3023362345396244, | |
| "grad_norm": 0.9761648774147034, | |
| "learning_rate": 8.809558912966519e-07, | |
| "logits/chosen": -0.6019878387451172, | |
| "logits/rejected": -0.6780513525009155, | |
| "logps/chosen": -134.56356811523438, | |
| "logps/rejected": -185.90939331054688, | |
| "loss": 0.354, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.3775978088378906, | |
| "rewards/margins": 1.7786915302276611, | |
| "rewards/rejected": 0.5989062786102295, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.3114979386165827, | |
| "grad_norm": 1.6643030643463135, | |
| "learning_rate": 8.703915422323984e-07, | |
| "logits/chosen": -0.5226669907569885, | |
| "logits/rejected": -0.5020606517791748, | |
| "logps/chosen": -184.17132568359375, | |
| "logps/rejected": -203.7013397216797, | |
| "loss": 0.4022, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.763214111328125, | |
| "rewards/margins": 1.1016991138458252, | |
| "rewards/rejected": 0.661514937877655, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.320659642693541, | |
| "grad_norm": 1.045599102973938, | |
| "learning_rate": 8.594473663064734e-07, | |
| "logits/chosen": -0.7285621762275696, | |
| "logits/rejected": -0.7740557193756104, | |
| "logps/chosen": -133.10691833496094, | |
| "logps/rejected": -191.98646545410156, | |
| "loss": 0.3784, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.8301376104354858, | |
| "rewards/margins": 1.3312700986862183, | |
| "rewards/rejected": 0.49886733293533325, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.3298213467704993, | |
| "grad_norm": 2.110759973526001, | |
| "learning_rate": 8.481345864857146e-07, | |
| "logits/chosen": -0.5418592095375061, | |
| "logits/rejected": -0.588280975818634, | |
| "logps/chosen": -179.9706573486328, | |
| "logps/rejected": -242.3267822265625, | |
| "loss": 0.401, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 1.6830288171768188, | |
| "rewards/margins": 1.2477144002914429, | |
| "rewards/rejected": 0.43531447649002075, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.3389830508474576, | |
| "grad_norm": 1.335124135017395, | |
| "learning_rate": 8.36464803730636e-07, | |
| "logits/chosen": -0.8127607107162476, | |
| "logits/rejected": -0.8397665023803711, | |
| "logps/chosen": -143.43109130859375, | |
| "logps/rejected": -185.8011932373047, | |
| "loss": 0.3572, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.7833560705184937, | |
| "rewards/margins": 1.1723896265029907, | |
| "rewards/rejected": 0.6109665632247925, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.34814475492441593, | |
| "grad_norm": 1.2357772588729858, | |
| "learning_rate": 8.244499850989451e-07, | |
| "logits/chosen": -0.7481725811958313, | |
| "logits/rejected": -0.756425678730011, | |
| "logps/chosen": -117.90687561035156, | |
| "logps/rejected": -198.514892578125, | |
| "loss": 0.3691, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.9100993871688843, | |
| "rewards/margins": 1.3691816329956055, | |
| "rewards/rejected": 0.5409177541732788, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.35730645900137425, | |
| "grad_norm": 1.2337630987167358, | |
| "learning_rate": 8.121024514736377e-07, | |
| "logits/chosen": -0.5506582260131836, | |
| "logits/rejected": -0.6542531251907349, | |
| "logps/chosen": -116.30818176269531, | |
| "logps/rejected": -186.07508850097656, | |
| "loss": 0.3132, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 2.1005172729492188, | |
| "rewards/margins": 1.9795573949813843, | |
| "rewards/rejected": 0.12095997482538223, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.3664681630783326, | |
| "grad_norm": 1.261998176574707, | |
| "learning_rate": 7.994348649282532e-07, | |
| "logits/chosen": -0.6467943787574768, | |
| "logits/rejected": -0.6284207701683044, | |
| "logps/chosen": -167.55819702148438, | |
| "logps/rejected": -237.6490936279297, | |
| "loss": 0.3712, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 1.9129873514175415, | |
| "rewards/margins": 1.4685484170913696, | |
| "rewards/rejected": 0.4444388747215271, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.3664681630783326, | |
| "eval_logits/chosen": -0.6691383123397827, | |
| "eval_logits/rejected": -0.6756234765052795, | |
| "eval_logps/chosen": -199.86724853515625, | |
| "eval_logps/rejected": -262.46453857421875, | |
| "eval_loss": 0.37812188267707825, | |
| "eval_rewards/accuracies": 0.8468208312988281, | |
| "eval_rewards/chosen": 1.7271815538406372, | |
| "eval_rewards/margins": 1.5017303228378296, | |
| "eval_rewards/rejected": 0.22545117139816284, | |
| "eval_runtime": 253.3553, | |
| "eval_samples_per_second": 10.906, | |
| "eval_steps_per_second": 1.366, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.3756298671552909, | |
| "grad_norm": 1.2864240407943726, | |
| "learning_rate": 7.8646021574225e-07, | |
| "logits/chosen": -0.5721285343170166, | |
| "logits/rejected": -0.5845418572425842, | |
| "logps/chosen": -163.110595703125, | |
| "logps/rejected": -217.90377807617188, | |
| "loss": 0.3708, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 2.0738914012908936, | |
| "rewards/margins": 1.6767698526382446, | |
| "rewards/rejected": 0.397121787071228, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.3847915712322492, | |
| "grad_norm": 1.066874623298645, | |
| "learning_rate": 7.731918090798113e-07, | |
| "logits/chosen": -0.6550859212875366, | |
| "logits/rejected": -0.6960932016372681, | |
| "logps/chosen": -151.92108154296875, | |
| "logps/rejected": -187.91384887695312, | |
| "loss": 0.3284, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.930013656616211, | |
| "rewards/margins": 1.5344407558441162, | |
| "rewards/rejected": 0.39557284116744995, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.39395327530920754, | |
| "grad_norm": 1.4810634851455688, | |
| "learning_rate": 7.596432513457482e-07, | |
| "logits/chosen": -0.7274349927902222, | |
| "logits/rejected": -0.7099400162696838, | |
| "logps/chosen": -145.75698852539062, | |
| "logps/rejected": -185.64678955078125, | |
| "loss": 0.3359, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.9275859594345093, | |
| "rewards/margins": 1.4946165084838867, | |
| "rewards/rejected": 0.43296942114830017, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.4031149793861658, | |
| "grad_norm": 1.650376319885254, | |
| "learning_rate": 7.458284362324842e-07, | |
| "logits/chosen": -0.5438752174377441, | |
| "logits/rejected": -0.6026032567024231, | |
| "logps/chosen": -132.947021484375, | |
| "logps/rejected": -219.47055053710938, | |
| "loss": 0.3276, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.8638055324554443, | |
| "rewards/margins": 2.066131353378296, | |
| "rewards/rejected": -0.20232591032981873, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.4122766834631241, | |
| "grad_norm": 0.992289125919342, | |
| "learning_rate": 7.317615304724387e-07, | |
| "logits/chosen": -0.6501020193099976, | |
| "logits/rejected": -0.680503249168396, | |
| "logps/chosen": -157.24925231933594, | |
| "logps/rejected": -180.8325958251953, | |
| "loss": 0.3249, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.7174543142318726, | |
| "rewards/margins": 1.594322681427002, | |
| "rewards/rejected": 0.12313172966241837, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.42143838754008245, | |
| "grad_norm": 1.2098772525787354, | |
| "learning_rate": 7.174569593104108e-07, | |
| "logits/chosen": -0.7602720260620117, | |
| "logits/rejected": -0.7695900201797485, | |
| "logps/chosen": -174.65504455566406, | |
| "logps/rejected": -225.5537567138672, | |
| "loss": 0.3396, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.8735862970352173, | |
| "rewards/margins": 1.4029169082641602, | |
| "rewards/rejected": 0.47066912055015564, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.43060009161704077, | |
| "grad_norm": 1.896801471710205, | |
| "learning_rate": 7.029293917108677e-07, | |
| "logits/chosen": -0.6379343271255493, | |
| "logits/rejected": -0.629081130027771, | |
| "logps/chosen": -264.4695739746094, | |
| "logps/rejected": -247.133056640625, | |
| "loss": 0.3216, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.6404482126235962, | |
| "rewards/margins": 1.7144851684570312, | |
| "rewards/rejected": -0.07403700053691864, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.4397617956939991, | |
| "grad_norm": 2.521409273147583, | |
| "learning_rate": 6.881937253153051e-07, | |
| "logits/chosen": -0.7469202876091003, | |
| "logits/rejected": -0.7621224522590637, | |
| "logps/chosen": -165.92147827148438, | |
| "logps/rejected": -223.8385772705078, | |
| "loss": 0.3246, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 2.0734469890594482, | |
| "rewards/margins": 2.012188673019409, | |
| "rewards/rejected": 0.061258465051651, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.4489234997709574, | |
| "grad_norm": 1.2489752769470215, | |
| "learning_rate": 6.732650711651031e-07, | |
| "logits/chosen": -0.5696443319320679, | |
| "logits/rejected": -0.618468165397644, | |
| "logps/chosen": -189.3000030517578, | |
| "logps/rejected": -246.3811798095703, | |
| "loss": 0.2909, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.7483714818954468, | |
| "rewards/margins": 2.0019850730895996, | |
| "rewards/rejected": -0.2536138892173767, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.45808520384791573, | |
| "grad_norm": 1.3916680812835693, | |
| "learning_rate": 6.581587382055491e-07, | |
| "logits/chosen": -0.761835515499115, | |
| "logits/rejected": -0.7702199220657349, | |
| "logps/chosen": -151.83370971679688, | |
| "logps/rejected": -229.975341796875, | |
| "loss": 0.3064, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.8061788082122803, | |
| "rewards/margins": 1.7005802392959595, | |
| "rewards/rejected": 0.10559873282909393, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.45808520384791573, | |
| "eval_logits/chosen": -0.6487900018692017, | |
| "eval_logits/rejected": -0.6642398834228516, | |
| "eval_logps/chosen": -199.9703826904297, | |
| "eval_logps/rejected": -267.3388671875, | |
| "eval_loss": 0.347669780254364, | |
| "eval_rewards/accuracies": 0.8612716794013977, | |
| "eval_rewards/chosen": 1.7220263481140137, | |
| "eval_rewards/margins": 1.7402905225753784, | |
| "eval_rewards/rejected": -0.01826408877968788, | |
| "eval_runtime": 253.4005, | |
| "eval_samples_per_second": 10.904, | |
| "eval_steps_per_second": 1.365, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.467246907924874, | |
| "grad_norm": 1.3511228561401367, | |
| "learning_rate": 6.428902175869126e-07, | |
| "logits/chosen": -0.680508553981781, | |
| "logits/rejected": -0.6774734258651733, | |
| "logps/chosen": -177.41610717773438, | |
| "logps/rejected": -221.32138061523438, | |
| "loss": 0.3258, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 2.022366762161255, | |
| "rewards/margins": 2.05668044090271, | |
| "rewards/rejected": -0.03431398794054985, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.4764086120018323, | |
| "grad_norm": 0.9841328263282776, | |
| "learning_rate": 6.274751667786761e-07, | |
| "logits/chosen": -0.6339150667190552, | |
| "logits/rejected": -0.5782631635665894, | |
| "logps/chosen": -230.2734832763672, | |
| "logps/rejected": -311.7682800292969, | |
| "loss": 0.3229, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.7927030324935913, | |
| "rewards/margins": 1.7812907695770264, | |
| "rewards/rejected": 0.011412340216338634, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.48557031607879064, | |
| "grad_norm": 2.7643489837646484, | |
| "learning_rate": 6.119293935132075e-07, | |
| "logits/chosen": -0.6203776597976685, | |
| "logits/rejected": -0.6682835817337036, | |
| "logps/chosen": -151.23324584960938, | |
| "logps/rejected": -187.77114868164062, | |
| "loss": 0.3034, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.8519824743270874, | |
| "rewards/margins": 2.0718624591827393, | |
| "rewards/rejected": -0.21987971663475037, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.49473202015574896, | |
| "grad_norm": 1.3088020086288452, | |
| "learning_rate": 5.962688395753437e-07, | |
| "logits/chosen": -0.8648042678833008, | |
| "logits/rejected": -0.9114105105400085, | |
| "logps/chosen": -137.6197052001953, | |
| "logps/rejected": -205.31900024414062, | |
| "loss": 0.2938, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.5214287042617798, | |
| "rewards/margins": 1.6869144439697266, | |
| "rewards/rejected": -0.16548602283000946, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.5038937242327073, | |
| "grad_norm": 1.9471220970153809, | |
| "learning_rate": 5.80509564454506e-07, | |
| "logits/chosen": -0.7066371440887451, | |
| "logits/rejected": -0.6964636445045471, | |
| "logps/chosen": -106.696533203125, | |
| "logps/rejected": -200.95729064941406, | |
| "loss": 0.311, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 1.6337581872940063, | |
| "rewards/margins": 1.6309497356414795, | |
| "rewards/rejected": 0.0028083801735192537, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5130554283096656, | |
| "grad_norm": 1.3545409440994263, | |
| "learning_rate": 5.646677288761132e-07, | |
| "logits/chosen": -0.653856098651886, | |
| "logits/rejected": -0.7129830121994019, | |
| "logps/chosen": -153.89035034179688, | |
| "logps/rejected": -214.0993194580078, | |
| "loss": 0.3112, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.5801527500152588, | |
| "rewards/margins": 1.731884241104126, | |
| "rewards/rejected": -0.1517314463853836, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.5222171323866239, | |
| "grad_norm": 2.128307819366455, | |
| "learning_rate": 5.487595782291784e-07, | |
| "logits/chosen": -0.6990654468536377, | |
| "logits/rejected": -0.7342058420181274, | |
| "logps/chosen": -178.05380249023438, | |
| "logps/rejected": -225.2575225830078, | |
| "loss": 0.289, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 1.7597728967666626, | |
| "rewards/margins": 1.7312393188476562, | |
| "rewards/rejected": 0.02853367291390896, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.5313788364635822, | |
| "grad_norm": 1.2450644969940186, | |
| "learning_rate": 5.328014259070878e-07, | |
| "logits/chosen": -0.619064211845398, | |
| "logits/rejected": -0.6336346864700317, | |
| "logps/chosen": -188.67752075195312, | |
| "logps/rejected": -231.69503784179688, | |
| "loss": 0.3388, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.4915143251419067, | |
| "rewards/margins": 1.5344158411026, | |
| "rewards/rejected": -0.042901456356048584, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.5405405405405406, | |
| "grad_norm": 1.2011312246322632, | |
| "learning_rate": 5.168096365786402e-07, | |
| "logits/chosen": -0.7081841230392456, | |
| "logits/rejected": -0.7206417918205261, | |
| "logps/chosen": -163.11752319335938, | |
| "logps/rejected": -235.26065063476562, | |
| "loss": 0.311, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.9438568353652954, | |
| "rewards/margins": 2.347649574279785, | |
| "rewards/rejected": -0.4037927985191345, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.5497022446174988, | |
| "grad_norm": 1.7624154090881348, | |
| "learning_rate": 5.008006094065069e-07, | |
| "logits/chosen": -0.721166729927063, | |
| "logits/rejected": -0.7850608229637146, | |
| "logps/chosen": -165.5583953857422, | |
| "logps/rejected": -213.85855102539062, | |
| "loss": 0.3054, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.5962111949920654, | |
| "rewards/margins": 1.4221036434173584, | |
| "rewards/rejected": 0.174107626080513, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.5497022446174988, | |
| "eval_logits/chosen": -0.6406525373458862, | |
| "eval_logits/rejected": -0.6576036214828491, | |
| "eval_logps/chosen": -201.4723358154297, | |
| "eval_logps/rejected": -270.9281311035156, | |
| "eval_loss": 0.3270590603351593, | |
| "eval_rewards/accuracies": 0.8670520186424255, | |
| "eval_rewards/chosen": 1.6469277143478394, | |
| "eval_rewards/margins": 1.8446547985076904, | |
| "eval_rewards/rejected": -0.1977270245552063, | |
| "eval_runtime": 253.7625, | |
| "eval_samples_per_second": 10.888, | |
| "eval_steps_per_second": 1.363, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.5588639486944572, | |
| "grad_norm": 1.902051329612732, | |
| "learning_rate": 4.847907612303182e-07, | |
| "logits/chosen": -0.7130570411682129, | |
| "logits/rejected": -0.7378814816474915, | |
| "logps/chosen": -188.9144287109375, | |
| "logps/rejected": -259.7423400878906, | |
| "loss": 0.3009, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 1.5585817098617554, | |
| "rewards/margins": 1.3624755144119263, | |
| "rewards/rejected": 0.19610631465911865, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.5680256527714155, | |
| "grad_norm": 1.07755446434021, | |
| "learning_rate": 4.687965097316223e-07, | |
| "logits/chosen": -0.5912365317344666, | |
| "logits/rejected": -0.7409440875053406, | |
| "logps/chosen": -126.41873931884766, | |
| "logps/rejected": -238.76437377929688, | |
| "loss": 0.275, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.855369210243225, | |
| "rewards/margins": 2.4938321113586426, | |
| "rewards/rejected": -0.638463020324707, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.5771873568483737, | |
| "grad_norm": 1.3470137119293213, | |
| "learning_rate": 4.5283425659798175e-07, | |
| "logits/chosen": -0.8164669275283813, | |
| "logits/rejected": -0.8073943853378296, | |
| "logps/chosen": -201.78501892089844, | |
| "logps/rejected": -291.32183837890625, | |
| "loss": 0.3179, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.852560043334961, | |
| "rewards/margins": 1.8420225381851196, | |
| "rewards/rejected": 0.010537643916904926, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.5863490609253321, | |
| "grad_norm": 1.2435747385025024, | |
| "learning_rate": 4.3692037070347123e-07, | |
| "logits/chosen": -0.6459102630615234, | |
| "logits/rejected": -0.6654535531997681, | |
| "logps/chosen": -139.29067993164062, | |
| "logps/rejected": -212.83740234375, | |
| "loss": 0.3016, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.6537023782730103, | |
| "rewards/margins": 2.188789129257202, | |
| "rewards/rejected": -0.5350866913795471, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.5955107650022904, | |
| "grad_norm": 1.0110821723937988, | |
| "learning_rate": 4.21071171322823e-07, | |
| "logits/chosen": -0.6093601584434509, | |
| "logits/rejected": -0.5892384648323059, | |
| "logps/chosen": -275.14105224609375, | |
| "logps/rejected": -322.5516662597656, | |
| "loss": 0.3134, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 1.5108485221862793, | |
| "rewards/margins": 1.933189034461975, | |
| "rewards/rejected": -0.4223404824733734, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6046724690792488, | |
| "grad_norm": 0.9678570628166199, | |
| "learning_rate": 4.0530291139643755e-07, | |
| "logits/chosen": -0.8226197957992554, | |
| "logits/rejected": -0.8122960925102234, | |
| "logps/chosen": -143.47821044921875, | |
| "logps/rejected": -207.33187866210938, | |
| "loss": 0.2829, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.0050506591796875, | |
| "rewards/margins": 2.1615443229675293, | |
| "rewards/rejected": -0.1564939320087433, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.613834173156207, | |
| "grad_norm": 1.5877900123596191, | |
| "learning_rate": 3.8963176086341727e-07, | |
| "logits/chosen": -0.6404609084129333, | |
| "logits/rejected": -0.7193113565444946, | |
| "logps/chosen": -163.66485595703125, | |
| "logps/rejected": -219.49063110351562, | |
| "loss": 0.2704, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.613854169845581, | |
| "rewards/margins": 2.1832098960876465, | |
| "rewards/rejected": -0.5693557858467102, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.6229958772331654, | |
| "grad_norm": 1.301003098487854, | |
| "learning_rate": 3.7407379007971506e-07, | |
| "logits/chosen": -0.6765211820602417, | |
| "logits/rejected": -0.6541970372200012, | |
| "logps/chosen": -199.01051330566406, | |
| "logps/rejected": -279.47454833984375, | |
| "loss": 0.2896, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.5939905643463135, | |
| "rewards/margins": 2.2719438076019287, | |
| "rewards/rejected": -0.6779531240463257, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.6321575813101237, | |
| "grad_norm": 2.2288615703582764, | |
| "learning_rate": 3.586449533384048e-07, | |
| "logits/chosen": -0.6064814925193787, | |
| "logits/rejected": -0.6105560660362244, | |
| "logps/chosen": -137.80624389648438, | |
| "logps/rejected": -186.99191284179688, | |
| "loss": 0.3041, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.5289627313613892, | |
| "rewards/margins": 1.8546040058135986, | |
| "rewards/rejected": -0.32564133405685425, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.641319285387082, | |
| "grad_norm": 5.320913314819336, | |
| "learning_rate": 3.433610725089692e-07, | |
| "logits/chosen": -0.7031580209732056, | |
| "logits/rejected": -0.6630051136016846, | |
| "logps/chosen": -185.18197631835938, | |
| "logps/rejected": -280.9058532714844, | |
| "loss": 0.2919, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.8430522680282593, | |
| "rewards/margins": 1.9076077938079834, | |
| "rewards/rejected": -0.06455531716346741, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.641319285387082, | |
| "eval_logits/chosen": -0.6672143936157227, | |
| "eval_logits/rejected": -0.6753049492835999, | |
| "eval_logps/chosen": -199.65896606445312, | |
| "eval_logps/rejected": -273.0414123535156, | |
| "eval_loss": 0.3144252896308899, | |
| "eval_rewards/accuracies": 0.8641618490219116, | |
| "eval_rewards/chosen": 1.7375967502593994, | |
| "eval_rewards/margins": 2.0409882068634033, | |
| "eval_rewards/rejected": -0.30339136719703674, | |
| "eval_runtime": 253.1657, | |
| "eval_samples_per_second": 10.914, | |
| "eval_steps_per_second": 1.367, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.6504809894640403, | |
| "grad_norm": 1.6359617710113525, | |
| "learning_rate": 3.2823782081238555e-07, | |
| "logits/chosen": -0.7354801893234253, | |
| "logits/rejected": -0.7596527338027954, | |
| "logps/chosen": -145.60165405273438, | |
| "logps/rejected": -207.4108123779297, | |
| "loss": 0.3059, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.8920338153839111, | |
| "rewards/margins": 1.903387427330017, | |
| "rewards/rejected": -0.011353528127074242, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.6596426935409986, | |
| "grad_norm": 1.6316208839416504, | |
| "learning_rate": 3.132907067486471e-07, | |
| "logits/chosen": -0.6749696135520935, | |
| "logits/rejected": -0.7063171863555908, | |
| "logps/chosen": -158.86273193359375, | |
| "logps/rejected": -211.850341796875, | |
| "loss": 0.2837, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.6359964609146118, | |
| "rewards/margins": 1.919615387916565, | |
| "rewards/rejected": -0.2836189270019531, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.668804397617957, | |
| "grad_norm": 2.493192672729492, | |
| "learning_rate": 2.985350581932005e-07, | |
| "logits/chosen": -0.7640320062637329, | |
| "logits/rejected": -0.7546281814575195, | |
| "logps/chosen": -185.2474822998047, | |
| "logps/rejected": -231.47119140625, | |
| "loss": 0.3068, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 2.423882007598877, | |
| "rewards/margins": 2.5569610595703125, | |
| "rewards/rejected": -0.13307929039001465, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.6779661016949152, | |
| "grad_norm": 1.0160990953445435, | |
| "learning_rate": 2.839860066786103e-07, | |
| "logits/chosen": -0.7123746871948242, | |
| "logits/rejected": -0.7270351648330688, | |
| "logps/chosen": -131.58505249023438, | |
| "logps/rejected": -189.05264282226562, | |
| "loss": 0.2778, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.8099651336669922, | |
| "rewards/margins": 2.3398964405059814, | |
| "rewards/rejected": -0.5299314260482788, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.6871278057718736, | |
| "grad_norm": 1.4675796031951904, | |
| "learning_rate": 2.6965847187756553e-07, | |
| "logits/chosen": -0.7339153289794922, | |
| "logits/rejected": -0.6953274011611938, | |
| "logps/chosen": -144.21389770507812, | |
| "logps/rejected": -171.4956512451172, | |
| "loss": 0.2746, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 2.1637752056121826, | |
| "rewards/margins": 2.3990566730499268, | |
| "rewards/rejected": -0.2352815419435501, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.6962895098488319, | |
| "grad_norm": 2.0078845024108887, | |
| "learning_rate": 2.5556714630314613e-07, | |
| "logits/chosen": -0.783401370048523, | |
| "logits/rejected": -0.854455828666687, | |
| "logps/chosen": -105.17523193359375, | |
| "logps/rejected": -193.60110473632812, | |
| "loss": 0.2727, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 2.012260913848877, | |
| "rewards/margins": 2.8885557651519775, | |
| "rewards/rejected": -0.8762944340705872, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.7054512139257902, | |
| "grad_norm": 1.2648169994354248, | |
| "learning_rate": 2.417264802420343e-07, | |
| "logits/chosen": -0.6513649225234985, | |
| "logits/rejected": -0.6898313760757446, | |
| "logps/chosen": -141.83282470703125, | |
| "logps/rejected": -238.6311492919922, | |
| "loss": 0.3154, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.7510058879852295, | |
| "rewards/margins": 2.0821216106414795, | |
| "rewards/rejected": -0.33111587166786194, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.7146129180027485, | |
| "grad_norm": 2.6904611587524414, | |
| "learning_rate": 2.2815066693612117e-07, | |
| "logits/chosen": -0.7128955721855164, | |
| "logits/rejected": -0.7633499503135681, | |
| "logps/chosen": -155.6310577392578, | |
| "logps/rejected": -210.39285278320312, | |
| "loss": 0.2793, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.7025461196899414, | |
| "rewards/margins": 2.1338300704956055, | |
| "rewards/rejected": -0.4312838613986969, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.7237746220797068, | |
| "grad_norm": 2.1170003414154053, | |
| "learning_rate": 2.1485362802770862e-07, | |
| "logits/chosen": -0.6805752515792847, | |
| "logits/rejected": -0.7278204560279846, | |
| "logps/chosen": -207.82858276367188, | |
| "logps/rejected": -320.7551574707031, | |
| "loss": 0.2457, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.8095613718032837, | |
| "rewards/margins": 2.516162157058716, | |
| "rewards/rejected": -0.7066007852554321, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.7329363261566652, | |
| "grad_norm": 1.6859688758850098, | |
| "learning_rate": 2.018489992832283e-07, | |
| "logits/chosen": -0.7043607831001282, | |
| "logits/rejected": -0.6326649785041809, | |
| "logps/chosen": -189.71005249023438, | |
| "logps/rejected": -247.50723266601562, | |
| "loss": 0.314, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.972821593284607, | |
| "rewards/margins": 2.5004470348358154, | |
| "rewards/rejected": -0.5276254415512085, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.7329363261566652, | |
| "eval_logits/chosen": -0.6573572754859924, | |
| "eval_logits/rejected": -0.6684801578521729, | |
| "eval_logps/chosen": -200.33786010742188, | |
| "eval_logps/rejected": -275.4322814941406, | |
| "eval_loss": 0.3056153357028961, | |
| "eval_rewards/accuracies": 0.8670520186424255, | |
| "eval_rewards/chosen": 1.703651785850525, | |
| "eval_rewards/margins": 2.126586437225342, | |
| "eval_rewards/rejected": -0.4229348599910736, | |
| "eval_runtime": 253.4624, | |
| "eval_samples_per_second": 10.901, | |
| "eval_steps_per_second": 1.365, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.7420980302336234, | |
| "grad_norm": 1.46702241897583, | |
| "learning_rate": 1.891501166101187e-07, | |
| "logits/chosen": -0.8299296498298645, | |
| "logits/rejected": -0.8194143176078796, | |
| "logps/chosen": -139.70791625976562, | |
| "logps/rejected": -179.2274932861328, | |
| "loss": 0.2988, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.740025281906128, | |
| "rewards/margins": 2.0235095024108887, | |
| "rewards/rejected": -0.28348422050476074, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.7512597343105818, | |
| "grad_norm": 2.791546583175659, | |
| "learning_rate": 1.767700023812e-07, | |
| "logits/chosen": -0.7000871896743774, | |
| "logits/rejected": -0.7028741240501404, | |
| "logps/chosen": -174.82821655273438, | |
| "logps/rejected": -264.07470703125, | |
| "loss": 0.272, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 2.0946412086486816, | |
| "rewards/margins": 2.783123016357422, | |
| "rewards/rejected": -0.6884818077087402, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.7604214383875401, | |
| "grad_norm": 1.3349977731704712, | |
| "learning_rate": 1.6472135208057125e-07, | |
| "logits/chosen": -0.6306108832359314, | |
| "logits/rejected": -0.698475182056427, | |
| "logps/chosen": -155.9135284423828, | |
| "logps/rejected": -198.2060546875, | |
| "loss": 0.3032, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.55352783203125, | |
| "rewards/margins": 2.126960515975952, | |
| "rewards/rejected": -0.5734325647354126, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.7695831424644984, | |
| "grad_norm": 2.1777162551879883, | |
| "learning_rate": 1.530165212847217e-07, | |
| "logits/chosen": -0.7774958610534668, | |
| "logits/rejected": -0.7312562465667725, | |
| "logps/chosen": -141.79075622558594, | |
| "logps/rejected": -190.49310302734375, | |
| "loss": 0.2807, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.8885326385498047, | |
| "rewards/margins": 2.229785442352295, | |
| "rewards/rejected": -0.3412528932094574, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.7787448465414567, | |
| "grad_norm": 1.2950890064239502, | |
| "learning_rate": 1.4166751299221003e-07, | |
| "logits/chosen": -0.6935632228851318, | |
| "logits/rejected": -0.681289553642273, | |
| "logps/chosen": -166.78561401367188, | |
| "logps/rejected": -229.1016387939453, | |
| "loss": 0.2802, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.8389393091201782, | |
| "rewards/margins": 3.066588878631592, | |
| "rewards/rejected": -1.2276496887207031, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.7879065506184151, | |
| "grad_norm": 1.1921981573104858, | |
| "learning_rate": 1.306859653149025e-07, | |
| "logits/chosen": -0.7318634390830994, | |
| "logits/rejected": -0.7256627082824707, | |
| "logps/chosen": -171.42092895507812, | |
| "logps/rejected": -247.11831665039062, | |
| "loss": 0.2934, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.7239230871200562, | |
| "rewards/margins": 2.5621562004089355, | |
| "rewards/rejected": -0.8382335901260376, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.7970682546953733, | |
| "grad_norm": 1.4422613382339478, | |
| "learning_rate": 1.2008313954339305e-07, | |
| "logits/chosen": -0.586450457572937, | |
| "logits/rejected": -0.5815967321395874, | |
| "logps/chosen": -215.48007202148438, | |
| "logps/rejected": -242.58511352539062, | |
| "loss": 0.2549, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 1.6724342107772827, | |
| "rewards/margins": 1.982318639755249, | |
| "rewards/rejected": -0.3098844885826111, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.8062299587723316, | |
| "grad_norm": 1.141564130783081, | |
| "learning_rate": 1.098699085988432e-07, | |
| "logits/chosen": -0.7615233659744263, | |
| "logits/rejected": -0.8405082821846008, | |
| "logps/chosen": -163.58689880371094, | |
| "logps/rejected": -256.6318054199219, | |
| "loss": 0.2998, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.7078421115875244, | |
| "rewards/margins": 1.9117094278335571, | |
| "rewards/rejected": -0.20386750996112823, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.81539166284929, | |
| "grad_norm": 1.329354166984558, | |
| "learning_rate": 1.0005674588308566e-07, | |
| "logits/chosen": -0.7380274534225464, | |
| "logits/rejected": -0.75429767370224, | |
| "logps/chosen": -136.90902709960938, | |
| "logps/rejected": -224.83151245117188, | |
| "loss": 0.2231, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.7314188480377197, | |
| "rewards/margins": 2.723328113555908, | |
| "rewards/rejected": -0.9919096231460571, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.8245533669262483, | |
| "grad_norm": 1.717936396598816, | |
| "learning_rate": 9.065371453842358e-08, | |
| "logits/chosen": -0.670835018157959, | |
| "logits/rejected": -0.6879727244377136, | |
| "logps/chosen": -138.87889099121094, | |
| "logps/rejected": -184.62257385253906, | |
| "loss": 0.3014, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.0553038120269775, | |
| "rewards/margins": 2.568469524383545, | |
| "rewards/rejected": -0.5131659507751465, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.8245533669262483, | |
| "eval_logits/chosen": -0.6641213893890381, | |
| "eval_logits/rejected": -0.670166015625, | |
| "eval_logps/chosen": -200.79710388183594, | |
| "eval_logps/rejected": -276.2374267578125, | |
| "eval_loss": 0.3019951581954956, | |
| "eval_rewards/accuracies": 0.8699421882629395, | |
| "eval_rewards/chosen": 1.6806890964508057, | |
| "eval_rewards/margins": 2.1438791751861572, | |
| "eval_rewards/rejected": -0.4631901979446411, | |
| "eval_runtime": 253.6623, | |
| "eval_samples_per_second": 10.892, | |
| "eval_steps_per_second": 1.364, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.8337150710032066, | |
| "grad_norm": 1.479974627494812, | |
| "learning_rate": 8.167045712814108e-08, | |
| "logits/chosen": -0.6214216351509094, | |
| "logits/rejected": -0.641666054725647, | |
| "logps/chosen": -177.58828735351562, | |
| "logps/rejected": -273.30755615234375, | |
| "loss": 0.2533, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.7214215993881226, | |
| "rewards/margins": 1.9738502502441406, | |
| "rewards/rejected": -0.2524286210536957, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.8428767750801649, | |
| "grad_norm": 2.062375783920288, | |
| "learning_rate": 7.311618574830569e-08, | |
| "logits/chosen": -0.5943703651428223, | |
| "logits/rejected": -0.6128894686698914, | |
| "logps/chosen": -163.90029907226562, | |
| "logps/rejected": -248.54415893554688, | |
| "loss": 0.2868, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 2.0122509002685547, | |
| "rewards/margins": 2.265479803085327, | |
| "rewards/rejected": -0.25322893261909485, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.8520384791571233, | |
| "grad_norm": 1.7314789295196533, | |
| "learning_rate": 6.499967258100514e-08, | |
| "logits/chosen": -0.6944621801376343, | |
| "logits/rejected": -0.7938731908798218, | |
| "logps/chosen": -163.20828247070312, | |
| "logps/rejected": -237.55078125, | |
| "loss": 0.2645, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.7433500289916992, | |
| "rewards/margins": 2.302280902862549, | |
| "rewards/rejected": -0.5589307546615601, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.8612001832340815, | |
| "grad_norm": 0.9841431379318237, | |
| "learning_rate": 5.732924089870245e-08, | |
| "logits/chosen": -0.46664732694625854, | |
| "logits/rejected": -0.6183963418006897, | |
| "logps/chosen": -208.5711669921875, | |
| "logps/rejected": -259.8545837402344, | |
| "loss": 0.2725, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.87375009059906, | |
| "rewards/margins": 2.2145848274230957, | |
| "rewards/rejected": -0.34083452820777893, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.8703618873110398, | |
| "grad_norm": 1.0470377206802368, | |
| "learning_rate": 5.011275652893782e-08, | |
| "logits/chosen": -0.6211365461349487, | |
| "logits/rejected": -0.6796506643295288, | |
| "logps/chosen": -152.32177734375, | |
| "logps/rejected": -218.6947479248047, | |
| "loss": 0.2395, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.6267896890640259, | |
| "rewards/margins": 2.076831102371216, | |
| "rewards/rejected": -0.45004144310951233, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.8795235913879982, | |
| "grad_norm": 1.149141550064087, | |
| "learning_rate": 4.3357619788127634e-08, | |
| "logits/chosen": -0.6848306655883789, | |
| "logits/rejected": -0.7928074598312378, | |
| "logps/chosen": -200.9459228515625, | |
| "logps/rejected": -254.53952026367188, | |
| "loss": 0.2671, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 1.663435697555542, | |
| "rewards/margins": 1.9824724197387695, | |
| "rewards/rejected": -0.3190363943576813, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 0.8886852954649564, | |
| "grad_norm": 1.5327177047729492, | |
| "learning_rate": 3.707075789273306e-08, | |
| "logits/chosen": -0.6709850430488586, | |
| "logits/rejected": -0.6786261796951294, | |
| "logps/chosen": -146.47488403320312, | |
| "logps/rejected": -247.5612335205078, | |
| "loss": 0.2843, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.8335565328598022, | |
| "rewards/margins": 2.448019504547119, | |
| "rewards/rejected": -0.6144627332687378, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 0.8978469995419148, | |
| "grad_norm": 1.4160970449447632, | |
| "learning_rate": 3.125861785558015e-08, | |
| "logits/chosen": -0.7669690251350403, | |
| "logits/rejected": -0.8644639849662781, | |
| "logps/chosen": -169.50274658203125, | |
| "logps/rejected": -304.95098876953125, | |
| "loss": 0.2834, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.8083531856536865, | |
| "rewards/margins": 2.9228787422180176, | |
| "rewards/rejected": -1.114525556564331, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 0.9070087036188731, | |
| "grad_norm": 2.2089922428131104, | |
| "learning_rate": 2.592715987461702e-08, | |
| "logits/chosen": -0.7596947550773621, | |
| "logits/rejected": -0.7801377177238464, | |
| "logps/chosen": -217.3747100830078, | |
| "logps/rejected": -258.47796630859375, | |
| "loss": 0.3283, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.5211076736450195, | |
| "rewards/margins": 1.6776663064956665, | |
| "rewards/rejected": -0.15655846893787384, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 0.9161704076958315, | |
| "grad_norm": 1.1835498809814453, | |
| "learning_rate": 2.108185122088546e-08, | |
| "logits/chosen": -0.7457928657531738, | |
| "logits/rejected": -0.6937334537506104, | |
| "logps/chosen": -169.40447998046875, | |
| "logps/rejected": -247.85061645507812, | |
| "loss": 0.268, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.8612730503082275, | |
| "rewards/margins": 2.29799747467041, | |
| "rewards/rejected": -0.43672457337379456, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.9161704076958315, | |
| "eval_logits/chosen": -0.6635109186172485, | |
| "eval_logits/rejected": -0.6690148711204529, | |
| "eval_logps/chosen": -200.81570434570312, | |
| "eval_logps/rejected": -276.8312072753906, | |
| "eval_loss": 0.29993191361427307, | |
| "eval_rewards/accuracies": 0.884393036365509, | |
| "eval_rewards/chosen": 1.6797596216201782, | |
| "eval_rewards/margins": 2.1726391315460205, | |
| "eval_rewards/rejected": -0.4928795397281647, | |
| "eval_runtime": 253.6729, | |
| "eval_samples_per_second": 10.892, | |
| "eval_steps_per_second": 1.364, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.9253321117727897, | |
| "grad_norm": 1.9540088176727295, | |
| "learning_rate": 1.672766063197789e-08, | |
| "logits/chosen": -0.6848675608634949, | |
| "logits/rejected": -0.6749913692474365, | |
| "logps/chosen": -182.75892639160156, | |
| "logps/rejected": -234.47232055664062, | |
| "loss": 0.2621, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.729029655456543, | |
| "rewards/margins": 2.2015597820281982, | |
| "rewards/rejected": -0.47253018617630005, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 0.934493815849748, | |
| "grad_norm": 1.5091657638549805, | |
| "learning_rate": 1.286905321672621e-08, | |
| "logits/chosen": -0.6353659629821777, | |
| "logits/rejected": -0.6381145715713501, | |
| "logps/chosen": -118.5170669555664, | |
| "logps/rejected": -205.03689575195312, | |
| "loss": 0.2602, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.9368598461151123, | |
| "rewards/margins": 2.6814284324645996, | |
| "rewards/rejected": -0.744568407535553, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 0.9436555199267064, | |
| "grad_norm": 1.7073416709899902, | |
| "learning_rate": 9.509985876349491e-09, | |
| "logits/chosen": -0.61668461561203, | |
| "logits/rejected": -0.6262907385826111, | |
| "logps/chosen": -143.79531860351562, | |
| "logps/rejected": -229.5824737548828, | |
| "loss": 0.2564, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.7922176122665405, | |
| "rewards/margins": 2.186096429824829, | |
| "rewards/rejected": -0.3938787579536438, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 0.9528172240036646, | |
| "grad_norm": 1.6124757528305054, | |
| "learning_rate": 6.6539032467546885e-09, | |
| "logits/chosen": -0.790496289730072, | |
| "logits/rejected": -0.7677043080329895, | |
| "logps/chosen": -256.6092834472656, | |
| "logps/rejected": -325.1084289550781, | |
| "loss": 0.262, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 1.504821538925171, | |
| "rewards/margins": 2.0350840091705322, | |
| "rewards/rejected": -0.5302623510360718, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 0.961978928080623, | |
| "grad_norm": 1.0290203094482422, | |
| "learning_rate": 4.303734166152706e-09, | |
| "logits/chosen": -0.7079821825027466, | |
| "logits/rejected": -0.6711887121200562, | |
| "logps/chosen": -177.19007873535156, | |
| "logps/rejected": -210.00613403320312, | |
| "loss": 0.3015, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.794146180152893, | |
| "rewards/margins": 2.358898639678955, | |
| "rewards/rejected": -0.564752459526062, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.9711406321575813, | |
| "grad_norm": 1.0389478206634521, | |
| "learning_rate": 2.4618886716110676e-09, | |
| "logits/chosen": -0.6768301725387573, | |
| "logits/rejected": -0.6700726747512817, | |
| "logps/chosen": -167.974609375, | |
| "logps/rejected": -247.9666290283203, | |
| "loss": 0.2928, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 1.5335534811019897, | |
| "rewards/margins": 1.9949595928192139, | |
| "rewards/rejected": -0.4614059329032898, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 0.9803023362345397, | |
| "grad_norm": 0.8039052486419678, | |
| "learning_rate": 1.1302555276238579e-09, | |
| "logits/chosen": -0.6699908375740051, | |
| "logits/rejected": -0.7130982279777527, | |
| "logps/chosen": -109.6313247680664, | |
| "logps/rejected": -177.20127868652344, | |
| "loss": 0.232, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.9235477447509766, | |
| "rewards/margins": 2.2871975898742676, | |
| "rewards/rejected": -0.3636501729488373, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 0.9894640403114979, | |
| "grad_norm": 2.4051105976104736, | |
| "learning_rate": 3.102002892329536e-10, | |
| "logits/chosen": -0.6233320832252502, | |
| "logits/rejected": -0.6594210267066956, | |
| "logps/chosen": -142.9510040283203, | |
| "logps/rejected": -230.3099365234375, | |
| "loss": 0.2467, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.8195747137069702, | |
| "rewards/margins": 2.5644664764404297, | |
| "rewards/rejected": -0.7448917031288147, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 0.9986257443884563, | |
| "grad_norm": 1.9035027027130127, | |
| "learning_rate": 2.5639016871248366e-12, | |
| "logits/chosen": -0.608766496181488, | |
| "logits/rejected": -0.5473885536193848, | |
| "logps/chosen": -232.18392944335938, | |
| "logps/rejected": -249.3759002685547, | |
| "loss": 0.2929, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.7724415063858032, | |
| "rewards/margins": 2.1429972648620605, | |
| "rewards/rejected": -0.3705558478832245, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 0.9995419147961521, | |
| "step": 1091, | |
| "total_flos": 0.0, | |
| "train_loss": 0.37405444834616947, | |
| "train_runtime": 8738.9435, | |
| "train_samples_per_second": 3.996, | |
| "train_steps_per_second": 0.125 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1091, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |