| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9996020692399522, | |
| "eval_steps": 1000, | |
| "global_step": 628, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0015917230401910067, | |
| "grad_norm": 3.414436464235707, | |
| "learning_rate": 7.936507936507936e-09, | |
| "logits/chosen": -2.458916187286377, | |
| "logits/rejected": -2.50296688079834, | |
| "logps/chosen": -112.41877746582031, | |
| "logps/rejected": -153.82745361328125, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01591723040191007, | |
| "grad_norm": 3.47500960837261, | |
| "learning_rate": 7.936507936507936e-08, | |
| "logits/chosen": -2.5214052200317383, | |
| "logits/rejected": -2.508702039718628, | |
| "logps/chosen": -143.01979064941406, | |
| "logps/rejected": -133.09210205078125, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.4166666567325592, | |
| "rewards/chosen": -0.00031594105530530214, | |
| "rewards/margins": -0.00045727533870376647, | |
| "rewards/rejected": 0.00014133438526187092, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03183446080382014, | |
| "grad_norm": 3.343704534649061, | |
| "learning_rate": 1.5873015873015872e-07, | |
| "logits/chosen": -2.4521286487579346, | |
| "logits/rejected": -2.448809862136841, | |
| "logps/chosen": -151.37095642089844, | |
| "logps/rejected": -162.80328369140625, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.43437498807907104, | |
| "rewards/chosen": 4.1505660192342475e-05, | |
| "rewards/margins": -0.00036320873186923563, | |
| "rewards/rejected": 0.000404714432079345, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.0477516912057302, | |
| "grad_norm": 3.366996896296212, | |
| "learning_rate": 2.3809523809523806e-07, | |
| "logits/chosen": -2.4904391765594482, | |
| "logits/rejected": -2.479889392852783, | |
| "logps/chosen": -147.56060791015625, | |
| "logps/rejected": -150.96810913085938, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.003838698612526059, | |
| "rewards/margins": 0.002006606664508581, | |
| "rewards/rejected": 0.001832091948017478, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.06366892160764027, | |
| "grad_norm": 3.3714607115747497, | |
| "learning_rate": 3.1746031746031743e-07, | |
| "logits/chosen": -2.503981590270996, | |
| "logits/rejected": -2.486074447631836, | |
| "logps/chosen": -153.0246124267578, | |
| "logps/rejected": -145.2753448486328, | |
| "loss": 0.6899, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.007592098321765661, | |
| "rewards/margins": 0.00429057702422142, | |
| "rewards/rejected": 0.0033015217632055283, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.07958615200955034, | |
| "grad_norm": 4.06423931986909, | |
| "learning_rate": 3.968253968253968e-07, | |
| "logits/chosen": -2.4848971366882324, | |
| "logits/rejected": -2.4830713272094727, | |
| "logps/chosen": -146.40438842773438, | |
| "logps/rejected": -152.903564453125, | |
| "loss": 0.6849, | |
| "rewards/accuracies": 0.5531250238418579, | |
| "rewards/chosen": -0.01682356372475624, | |
| "rewards/margins": 0.0171933826059103, | |
| "rewards/rejected": -0.03401694819331169, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.0955033824114604, | |
| "grad_norm": 3.933215748835345, | |
| "learning_rate": 4.761904761904761e-07, | |
| "logits/chosen": -2.4494073390960693, | |
| "logits/rejected": -2.425264835357666, | |
| "logps/chosen": -164.67039489746094, | |
| "logps/rejected": -159.42379760742188, | |
| "loss": 0.6811, | |
| "rewards/accuracies": 0.528124988079071, | |
| "rewards/chosen": -0.13071545958518982, | |
| "rewards/margins": 0.01678677648305893, | |
| "rewards/rejected": -0.14750224351882935, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.11142061281337047, | |
| "grad_norm": 6.465696654064857, | |
| "learning_rate": 4.998106548810311e-07, | |
| "logits/chosen": -2.4344656467437744, | |
| "logits/rejected": -2.3922276496887207, | |
| "logps/chosen": -171.0955352783203, | |
| "logps/rejected": -158.62640380859375, | |
| "loss": 0.6666, | |
| "rewards/accuracies": 0.596875011920929, | |
| "rewards/chosen": -0.17716214060783386, | |
| "rewards/margins": 0.07567773759365082, | |
| "rewards/rejected": -0.2528398633003235, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.12733784321528055, | |
| "grad_norm": 9.086502750868021, | |
| "learning_rate": 4.988839406031596e-07, | |
| "logits/chosen": -2.384453058242798, | |
| "logits/rejected": -2.3677101135253906, | |
| "logps/chosen": -153.10414123535156, | |
| "logps/rejected": -182.54360961914062, | |
| "loss": 0.6529, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": -0.20188739895820618, | |
| "rewards/margins": 0.12667986750602722, | |
| "rewards/rejected": -0.3285672962665558, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.14325507361719061, | |
| "grad_norm": 14.480905623348013, | |
| "learning_rate": 4.971879403278432e-07, | |
| "logits/chosen": -2.2432703971862793, | |
| "logits/rejected": -2.2313826084136963, | |
| "logps/chosen": -185.1287384033203, | |
| "logps/rejected": -201.3049774169922, | |
| "loss": 0.6398, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.39208894968032837, | |
| "rewards/margins": 0.144916832447052, | |
| "rewards/rejected": -0.5370057821273804, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.15917230401910068, | |
| "grad_norm": 12.626265646656822, | |
| "learning_rate": 4.947278962947386e-07, | |
| "logits/chosen": -1.7865173816680908, | |
| "logits/rejected": -1.7619308233261108, | |
| "logps/chosen": -234.49526977539062, | |
| "logps/rejected": -266.22235107421875, | |
| "loss": 0.6186, | |
| "rewards/accuracies": 0.659375011920929, | |
| "rewards/chosen": -0.8486677408218384, | |
| "rewards/margins": 0.2785636782646179, | |
| "rewards/rejected": -1.127231478691101, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.17508953442101075, | |
| "grad_norm": 11.440443186979932, | |
| "learning_rate": 4.915114123589732e-07, | |
| "logits/chosen": -1.477772831916809, | |
| "logits/rejected": -1.394840955734253, | |
| "logps/chosen": -206.17166137695312, | |
| "logps/rejected": -243.156005859375, | |
| "loss": 0.6156, | |
| "rewards/accuracies": 0.684374988079071, | |
| "rewards/chosen": -0.719870388507843, | |
| "rewards/margins": 0.34280186891555786, | |
| "rewards/rejected": -1.0626722574234009, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.1910067648229208, | |
| "grad_norm": 14.957458897670607, | |
| "learning_rate": 4.875484304880629e-07, | |
| "logits/chosen": -0.8948208093643188, | |
| "logits/rejected": -0.773229718208313, | |
| "logps/chosen": -234.4495086669922, | |
| "logps/rejected": -270.4801330566406, | |
| "loss": 0.5931, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.8048633337020874, | |
| "rewards/margins": 0.34693923592567444, | |
| "rewards/rejected": -1.151802659034729, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.20692399522483088, | |
| "grad_norm": 13.673290760971415, | |
| "learning_rate": 4.828512000318616e-07, | |
| "logits/chosen": -0.20325860381126404, | |
| "logits/rejected": -0.022473735734820366, | |
| "logps/chosen": -233.56399536132812, | |
| "logps/rejected": -278.225830078125, | |
| "loss": 0.5842, | |
| "rewards/accuracies": 0.7093750238418579, | |
| "rewards/chosen": -0.7881739139556885, | |
| "rewards/margins": 0.4328451156616211, | |
| "rewards/rejected": -1.2210190296173096, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.22284122562674094, | |
| "grad_norm": 15.252755388662429, | |
| "learning_rate": 4.774342398605221e-07, | |
| "logits/chosen": -0.4290715754032135, | |
| "logits/rejected": -0.26301848888397217, | |
| "logps/chosen": -245.796142578125, | |
| "logps/rejected": -278.6325988769531, | |
| "loss": 0.5803, | |
| "rewards/accuracies": 0.6781250238418579, | |
| "rewards/chosen": -0.9634370803833008, | |
| "rewards/margins": 0.42050686478614807, | |
| "rewards/rejected": -1.383944034576416, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.238758456028651, | |
| "grad_norm": 15.98257839847235, | |
| "learning_rate": 4.713142934875005e-07, | |
| "logits/chosen": 0.04155537486076355, | |
| "logits/rejected": 0.4105464518070221, | |
| "logps/chosen": -294.55078125, | |
| "logps/rejected": -322.9659118652344, | |
| "loss": 0.5928, | |
| "rewards/accuracies": 0.684374988079071, | |
| "rewards/chosen": -1.3449639081954956, | |
| "rewards/margins": 0.41515955328941345, | |
| "rewards/rejected": -1.7601234912872314, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.2546756864305611, | |
| "grad_norm": 16.296020720117806, | |
| "learning_rate": 4.64510277316316e-07, | |
| "logits/chosen": 0.07806523889303207, | |
| "logits/rejected": 0.4625118672847748, | |
| "logps/chosen": -242.23171997070312, | |
| "logps/rejected": -306.7259216308594, | |
| "loss": 0.5619, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -1.0831429958343506, | |
| "rewards/margins": 0.6153541803359985, | |
| "rewards/rejected": -1.6984970569610596, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.27059291683247116, | |
| "grad_norm": 15.29787072223291, | |
| "learning_rate": 4.570432221710314e-07, | |
| "logits/chosen": 0.8908351063728333, | |
| "logits/rejected": 1.1234996318817139, | |
| "logps/chosen": -285.02685546875, | |
| "logps/rejected": -332.22576904296875, | |
| "loss": 0.5968, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -1.369902491569519, | |
| "rewards/margins": 0.42481645941734314, | |
| "rewards/rejected": -1.7947190999984741, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.28651014723438123, | |
| "grad_norm": 12.440638905265972, | |
| "learning_rate": 4.4893620829118124e-07, | |
| "logits/chosen": 0.45191067457199097, | |
| "logits/rejected": 0.8097447156906128, | |
| "logps/chosen": -267.55511474609375, | |
| "logps/rejected": -303.0255432128906, | |
| "loss": 0.5782, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": -1.130518913269043, | |
| "rewards/margins": 0.47403573989868164, | |
| "rewards/rejected": -1.6045547723770142, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3024273776362913, | |
| "grad_norm": 14.662664269581565, | |
| "learning_rate": 4.40214293992074e-07, | |
| "logits/chosen": 0.07046165317296982, | |
| "logits/rejected": 0.5015997886657715, | |
| "logps/chosen": -276.8005676269531, | |
| "logps/rejected": -317.6512451171875, | |
| "loss": 0.5684, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.328792929649353, | |
| "rewards/margins": 0.5068601369857788, | |
| "rewards/rejected": -1.835653305053711, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.31834460803820136, | |
| "grad_norm": 13.757436126080059, | |
| "learning_rate": 4.3090443821097566e-07, | |
| "logits/chosen": 0.2187742292881012, | |
| "logits/rejected": 0.5610963106155396, | |
| "logps/chosen": -268.94757080078125, | |
| "logps/rejected": -333.0284729003906, | |
| "loss": 0.5596, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.2326629161834717, | |
| "rewards/margins": 0.5554476976394653, | |
| "rewards/rejected": -1.7881107330322266, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.3342618384401114, | |
| "grad_norm": 12.638546225207325, | |
| "learning_rate": 4.210354171785795e-07, | |
| "logits/chosen": 0.5555538535118103, | |
| "logits/rejected": 0.9611243009567261, | |
| "logps/chosen": -258.8009338378906, | |
| "logps/rejected": -315.17242431640625, | |
| "loss": 0.5627, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.1692590713500977, | |
| "rewards/margins": 0.5407182574272156, | |
| "rewards/rejected": -1.709977388381958, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.3501790688420215, | |
| "grad_norm": 12.533508219187622, | |
| "learning_rate": 4.1063773547332584e-07, | |
| "logits/chosen": 0.4853738248348236, | |
| "logits/rejected": 0.7825923562049866, | |
| "logps/chosen": -255.5019073486328, | |
| "logps/rejected": -307.07843017578125, | |
| "loss": 0.5684, | |
| "rewards/accuracies": 0.6656249761581421, | |
| "rewards/chosen": -1.0966681241989136, | |
| "rewards/margins": 0.5020471215248108, | |
| "rewards/rejected": -1.59871506690979, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.36609629924393156, | |
| "grad_norm": 13.14652988888945, | |
| "learning_rate": 3.997435317334988e-07, | |
| "logits/chosen": 1.3910820484161377, | |
| "logits/rejected": 1.7489607334136963, | |
| "logps/chosen": -285.7388916015625, | |
| "logps/rejected": -342.02557373046875, | |
| "loss": 0.5556, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.430323600769043, | |
| "rewards/margins": 0.5691004991531372, | |
| "rewards/rejected": -1.9994239807128906, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.3820135296458416, | |
| "grad_norm": 16.893312636051345, | |
| "learning_rate": 3.8838647931853684e-07, | |
| "logits/chosen": 0.9681981801986694, | |
| "logits/rejected": 1.3831474781036377, | |
| "logps/chosen": -274.5780029296875, | |
| "logps/rejected": -334.71624755859375, | |
| "loss": 0.565, | |
| "rewards/accuracies": 0.684374988079071, | |
| "rewards/chosen": -1.331866979598999, | |
| "rewards/margins": 0.6004709005355835, | |
| "rewards/rejected": -1.932337999343872, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.3979307600477517, | |
| "grad_norm": 16.93758119869443, | |
| "learning_rate": 3.7660168222660824e-07, | |
| "logits/chosen": -0.0617893747985363, | |
| "logits/rejected": 0.3982541561126709, | |
| "logps/chosen": -291.5325012207031, | |
| "logps/rejected": -349.10357666015625, | |
| "loss": 0.546, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -1.3192743062973022, | |
| "rewards/margins": 0.6018632650375366, | |
| "rewards/rejected": -1.9211375713348389, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.41384799044966175, | |
| "grad_norm": 13.594747823748454, | |
| "learning_rate": 3.6442556659016475e-07, | |
| "logits/chosen": -0.1596679985523224, | |
| "logits/rejected": 0.12459886074066162, | |
| "logps/chosen": -283.2134704589844, | |
| "logps/rejected": -335.0664367675781, | |
| "loss": 0.5677, | |
| "rewards/accuracies": 0.703125, | |
| "rewards/chosen": -1.3012964725494385, | |
| "rewards/margins": 0.5175098180770874, | |
| "rewards/rejected": -1.8188062906265259, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.4297652208515718, | |
| "grad_norm": 13.029006886811343, | |
| "learning_rate": 3.5189576808485404e-07, | |
| "logits/chosen": 0.6624835729598999, | |
| "logits/rejected": 0.867287278175354, | |
| "logps/chosen": -290.3371887207031, | |
| "logps/rejected": -335.45745849609375, | |
| "loss": 0.5603, | |
| "rewards/accuracies": 0.7093750238418579, | |
| "rewards/chosen": -1.3652312755584717, | |
| "rewards/margins": 0.44845834374427795, | |
| "rewards/rejected": -1.8136895895004272, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.4456824512534819, | |
| "grad_norm": 14.229518064110062, | |
| "learning_rate": 3.390510155998023e-07, | |
| "logits/chosen": 1.0853196382522583, | |
| "logits/rejected": 1.5426385402679443, | |
| "logps/chosen": -301.244873046875, | |
| "logps/rejected": -352.26214599609375, | |
| "loss": 0.5576, | |
| "rewards/accuracies": 0.721875011920929, | |
| "rewards/chosen": -1.4498573541641235, | |
| "rewards/margins": 0.5846196413040161, | |
| "rewards/rejected": -2.0344767570495605, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.46159968165539195, | |
| "grad_norm": 13.247814891025222, | |
| "learning_rate": 3.2593101152883795e-07, | |
| "logits/chosen": 0.6425064206123352, | |
| "logits/rejected": 1.0470354557037354, | |
| "logps/chosen": -304.13165283203125, | |
| "logps/rejected": -356.2009582519531, | |
| "loss": 0.5699, | |
| "rewards/accuracies": 0.6781250238418579, | |
| "rewards/chosen": -1.4935743808746338, | |
| "rewards/margins": 0.543701708316803, | |
| "rewards/rejected": -2.037276029586792, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.477516912057302, | |
| "grad_norm": 14.136317593633118, | |
| "learning_rate": 3.125763090526674e-07, | |
| "logits/chosen": 1.010498285293579, | |
| "logits/rejected": 1.5458686351776123, | |
| "logps/chosen": -312.71893310546875, | |
| "logps/rejected": -367.11956787109375, | |
| "loss": 0.5526, | |
| "rewards/accuracies": 0.6968749761581421, | |
| "rewards/chosen": -1.607622742652893, | |
| "rewards/margins": 0.5898184180259705, | |
| "rewards/rejected": -2.1974411010742188, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.4934341424592121, | |
| "grad_norm": 13.599580845693339, | |
| "learning_rate": 2.9902818679131775e-07, | |
| "logits/chosen": 1.556601881980896, | |
| "logits/rejected": 2.024049758911133, | |
| "logps/chosen": -317.42938232421875, | |
| "logps/rejected": -371.38623046875, | |
| "loss": 0.5571, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.6657898426055908, | |
| "rewards/margins": 0.5584100484848022, | |
| "rewards/rejected": -2.2241997718811035, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.5093513728611222, | |
| "grad_norm": 15.486325517579422, | |
| "learning_rate": 2.8532852121428733e-07, | |
| "logits/chosen": 2.424159526824951, | |
| "logits/rejected": 2.8681328296661377, | |
| "logps/chosen": -326.8099670410156, | |
| "logps/rejected": -373.8744201660156, | |
| "loss": 0.5411, | |
| "rewards/accuracies": 0.6656249761581421, | |
| "rewards/chosen": -1.7966935634613037, | |
| "rewards/margins": 0.5875706076622009, | |
| "rewards/rejected": -2.3842644691467285, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.5252686032630323, | |
| "grad_norm": 13.394275013231457, | |
| "learning_rate": 2.715196572027789e-07, | |
| "logits/chosen": 3.215583324432373, | |
| "logits/rejected": 3.725792646408081, | |
| "logps/chosen": -343.72344970703125, | |
| "logps/rejected": -432.32916259765625, | |
| "loss": 0.5414, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.9976279735565186, | |
| "rewards/margins": 0.806243896484375, | |
| "rewards/rejected": -2.8038716316223145, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.5411858336649423, | |
| "grad_norm": 13.373618178706888, | |
| "learning_rate": 2.5764427716409815e-07, | |
| "logits/chosen": 3.000930070877075, | |
| "logits/rejected": 3.6059279441833496, | |
| "logps/chosen": -347.54022216796875, | |
| "logps/rejected": -404.14654541015625, | |
| "loss": 0.5507, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.9271198511123657, | |
| "rewards/margins": 0.6116258502006531, | |
| "rewards/rejected": -2.538745880126953, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.5571030640668524, | |
| "grad_norm": 14.475581092351593, | |
| "learning_rate": 2.4374526910277886e-07, | |
| "logits/chosen": 2.816096782684326, | |
| "logits/rejected": 3.264390230178833, | |
| "logps/chosen": -353.548828125, | |
| "logps/rejected": -420.4794006347656, | |
| "loss": 0.5564, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -2.0463950634002686, | |
| "rewards/margins": 0.6025586128234863, | |
| "rewards/rejected": -2.648953914642334, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.5730202944687625, | |
| "grad_norm": 14.682353004178683, | |
| "learning_rate": 2.2986559405621886e-07, | |
| "logits/chosen": 2.8994641304016113, | |
| "logits/rejected": 3.608001232147217, | |
| "logps/chosen": -360.53143310546875, | |
| "logps/rejected": -417.3299865722656, | |
| "loss": 0.5525, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.981393814086914, | |
| "rewards/margins": 0.6567780375480652, | |
| "rewards/rejected": -2.638171911239624, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.5889375248706725, | |
| "grad_norm": 18.068831076710342, | |
| "learning_rate": 2.160481533045751e-07, | |
| "logits/chosen": 2.5152053833007812, | |
| "logits/rejected": 3.3655197620391846, | |
| "logps/chosen": -339.1783752441406, | |
| "logps/rejected": -407.2680358886719, | |
| "loss": 0.5377, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.932498574256897, | |
| "rewards/margins": 0.7526292204856873, | |
| "rewards/rejected": -2.6851277351379395, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.6048547552725826, | |
| "grad_norm": 13.903302227889153, | |
| "learning_rate": 2.0233565576536564e-07, | |
| "logits/chosen": 2.7033469676971436, | |
| "logits/rejected": 3.3517203330993652, | |
| "logps/chosen": -352.88543701171875, | |
| "logps/rejected": -414.46807861328125, | |
| "loss": 0.5464, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -2.0058979988098145, | |
| "rewards/margins": 0.6991748809814453, | |
| "rewards/rejected": -2.7050728797912598, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.6207719856744927, | |
| "grad_norm": 14.437387596932032, | |
| "learning_rate": 1.887704859826528e-07, | |
| "logits/chosen": 2.677304744720459, | |
| "logits/rejected": 3.3016350269317627, | |
| "logps/chosen": -345.10186767578125, | |
| "logps/rejected": -422.6465759277344, | |
| "loss": 0.5441, | |
| "rewards/accuracies": 0.753125011920929, | |
| "rewards/chosen": -1.9631824493408203, | |
| "rewards/margins": 0.7392277717590332, | |
| "rewards/rejected": -2.7024102210998535, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.6366892160764027, | |
| "grad_norm": 15.504689068952228, | |
| "learning_rate": 1.7539457311884675e-07, | |
| "logits/chosen": 2.6285808086395264, | |
| "logits/rejected": 3.402756452560425, | |
| "logps/chosen": -350.1034240722656, | |
| "logps/rejected": -420.3218688964844, | |
| "loss": 0.5387, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.9105831384658813, | |
| "rewards/margins": 0.7946080565452576, | |
| "rewards/rejected": -2.705191135406494, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.6526064464783128, | |
| "grad_norm": 12.730811962712194, | |
| "learning_rate": 1.6224926135406693e-07, | |
| "logits/chosen": 2.7364511489868164, | |
| "logits/rejected": 3.1644163131713867, | |
| "logps/chosen": -364.25469970703125, | |
| "logps/rejected": -419.5701599121094, | |
| "loss": 0.5618, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": -2.067072629928589, | |
| "rewards/margins": 0.5998597741127014, | |
| "rewards/rejected": -2.6669325828552246, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.6685236768802229, | |
| "grad_norm": 14.2356920381672, | |
| "learning_rate": 1.4937518209365108e-07, | |
| "logits/chosen": 2.5111005306243896, | |
| "logits/rejected": 3.2514469623565674, | |
| "logps/chosen": -381.5763244628906, | |
| "logps/rejected": -433.4600524902344, | |
| "loss": 0.5494, | |
| "rewards/accuracies": 0.778124988079071, | |
| "rewards/chosen": -2.099307060241699, | |
| "rewards/margins": 0.6568456888198853, | |
| "rewards/rejected": -2.756152629852295, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.6844409072821329, | |
| "grad_norm": 15.52832700397283, | |
| "learning_rate": 1.3681212837880977e-07, | |
| "logits/chosen": 2.709282398223877, | |
| "logits/rejected": 3.0862138271331787, | |
| "logps/chosen": -374.52288818359375, | |
| "logps/rejected": -447.3203125, | |
| "loss": 0.5539, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.311601161956787, | |
| "rewards/margins": 0.6582018136978149, | |
| "rewards/rejected": -2.9698030948638916, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.700358137684043, | |
| "grad_norm": 13.300949643909883, | |
| "learning_rate": 1.2459893188861613e-07, | |
| "logits/chosen": 2.896282196044922, | |
| "logits/rejected": 3.553002119064331, | |
| "logps/chosen": -382.75054931640625, | |
| "logps/rejected": -458.756591796875, | |
| "loss": 0.5426, | |
| "rewards/accuracies": 0.765625, | |
| "rewards/chosen": -2.2905333042144775, | |
| "rewards/margins": 0.7688990831375122, | |
| "rewards/rejected": -3.0594325065612793, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.716275368085953, | |
| "grad_norm": 14.064737088004772, | |
| "learning_rate": 1.1277334291351145e-07, | |
| "logits/chosen": 3.348231077194214, | |
| "logits/rejected": 4.010863780975342, | |
| "logps/chosen": -383.77569580078125, | |
| "logps/rejected": -465.36767578125, | |
| "loss": 0.5372, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -2.3260040283203125, | |
| "rewards/margins": 0.8284826278686523, | |
| "rewards/rejected": -3.1544864177703857, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.7321925984878631, | |
| "grad_norm": 13.977439772366749, | |
| "learning_rate": 1.0137191367132078e-07, | |
| "logits/chosen": 3.446108341217041, | |
| "logits/rejected": 3.9590706825256348, | |
| "logps/chosen": -375.79705810546875, | |
| "logps/rejected": -461.6592712402344, | |
| "loss": 0.5282, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.228874683380127, | |
| "rewards/margins": 0.7711628675460815, | |
| "rewards/rejected": -3.000037670135498, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.7481098288897732, | |
| "grad_norm": 15.466730145904025, | |
| "learning_rate": 9.042988532644249e-08, | |
| "logits/chosen": 3.792773485183716, | |
| "logits/rejected": 4.424340724945068, | |
| "logps/chosen": -395.2272033691406, | |
| "logps/rejected": -461.55352783203125, | |
| "loss": 0.5425, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": -2.369027614593506, | |
| "rewards/margins": 0.6728731393814087, | |
| "rewards/rejected": -3.041900157928467, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.7640270592916832, | |
| "grad_norm": 14.83331875533172, | |
| "learning_rate": 7.998107906142839e-08, | |
| "logits/chosen": 3.4053382873535156, | |
| "logits/rejected": 3.8911843299865723, | |
| "logps/chosen": -388.53155517578125, | |
| "logps/rejected": -455.40185546875, | |
| "loss": 0.5425, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -2.44258189201355, | |
| "rewards/margins": 0.6656660437583923, | |
| "rewards/rejected": -3.108247756958008, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.7799442896935933, | |
| "grad_norm": 15.620806793742632, | |
| "learning_rate": 7.005779153764682e-08, | |
| "logits/chosen": 3.4802298545837402, | |
| "logits/rejected": 3.920941114425659, | |
| "logps/chosen": -378.8732604980469, | |
| "logps/rejected": -466.740478515625, | |
| "loss": 0.5309, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.4023540019989014, | |
| "rewards/margins": 0.7670234441757202, | |
| "rewards/rejected": -3.169377565383911, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.7958615200955034, | |
| "grad_norm": 17.188173872991122, | |
| "learning_rate": 6.069069506815325e-08, | |
| "logits/chosen": 3.2549641132354736, | |
| "logits/rejected": 4.011484146118164, | |
| "logps/chosen": -376.5075988769531, | |
| "logps/rejected": -464.5885314941406, | |
| "loss": 0.5443, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.2434165477752686, | |
| "rewards/margins": 0.8864636421203613, | |
| "rewards/rejected": -3.129879951477051, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.8117787504974134, | |
| "grad_norm": 15.502718885907624, | |
| "learning_rate": 5.190874281132851e-08, | |
| "logits/chosen": 3.2985496520996094, | |
| "logits/rejected": 3.910979747772217, | |
| "logps/chosen": -390.97381591796875, | |
| "logps/rejected": -462.67816162109375, | |
| "loss": 0.5418, | |
| "rewards/accuracies": 0.7093750238418579, | |
| "rewards/chosen": -2.4295220375061035, | |
| "rewards/margins": 0.7420149445533752, | |
| "rewards/rejected": -3.171536922454834, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.8276959808993235, | |
| "grad_norm": 14.242739712121452, | |
| "learning_rate": 4.373907927832513e-08, | |
| "logits/chosen": 3.5766899585723877, | |
| "logits/rejected": 4.207441806793213, | |
| "logps/chosen": -382.2040710449219, | |
| "logps/rejected": -460.5889587402344, | |
| "loss": 0.5461, | |
| "rewards/accuracies": 0.6968749761581421, | |
| "rewards/chosen": -2.4501075744628906, | |
| "rewards/margins": 0.7706823348999023, | |
| "rewards/rejected": -3.220790147781372, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.8436132113012336, | |
| "grad_norm": 12.998296781925998, | |
| "learning_rate": 3.620695643093924e-08, | |
| "logits/chosen": 3.5845916271209717, | |
| "logits/rejected": 4.0032148361206055, | |
| "logps/chosen": -372.14898681640625, | |
| "logps/rejected": -455.68212890625, | |
| "loss": 0.5414, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": -2.2561497688293457, | |
| "rewards/margins": 0.7765234112739563, | |
| "rewards/rejected": -3.032672643661499, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.8595304417031436, | |
| "grad_norm": 17.292649632838344, | |
| "learning_rate": 2.9335655629243645e-08, | |
| "logits/chosen": 3.523092269897461, | |
| "logits/rejected": 3.9805259704589844, | |
| "logps/chosen": -388.76800537109375, | |
| "logps/rejected": -462.15625, | |
| "loss": 0.5454, | |
| "rewards/accuracies": 0.721875011920929, | |
| "rewards/chosen": -2.4221224784851074, | |
| "rewards/margins": 0.6886903643608093, | |
| "rewards/rejected": -3.1108126640319824, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.8754476721050537, | |
| "grad_norm": 13.631324448665103, | |
| "learning_rate": 2.31464156702382e-08, | |
| "logits/chosen": 3.492096424102783, | |
| "logits/rejected": 4.266209125518799, | |
| "logps/chosen": -393.3036804199219, | |
| "logps/rejected": -445.7330627441406, | |
| "loss": 0.5519, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -2.4094600677490234, | |
| "rewards/margins": 0.6608883738517761, | |
| "rewards/rejected": -3.0703485012054443, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.8913649025069638, | |
| "grad_norm": 15.338497216975412, | |
| "learning_rate": 1.7658367139945228e-08, | |
| "logits/chosen": 3.498994827270508, | |
| "logits/rejected": 3.9349846839904785, | |
| "logps/chosen": -387.2663269042969, | |
| "logps/rejected": -449.8990173339844, | |
| "loss": 0.5364, | |
| "rewards/accuracies": 0.659375011920929, | |
| "rewards/chosen": -2.4634361267089844, | |
| "rewards/margins": 0.5876127481460571, | |
| "rewards/rejected": -3.051048755645752, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.9072821329088738, | |
| "grad_norm": 13.324839693428412, | |
| "learning_rate": 1.2888473281864597e-08, | |
| "logits/chosen": 3.275439500808716, | |
| "logits/rejected": 3.773510694503784, | |
| "logps/chosen": -398.3496398925781, | |
| "logps/rejected": -472.8216247558594, | |
| "loss": 0.5281, | |
| "rewards/accuracies": 0.6968749761581421, | |
| "rewards/chosen": -2.447859287261963, | |
| "rewards/margins": 0.6992862820625305, | |
| "rewards/rejected": -3.1471457481384277, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.9231993633107839, | |
| "grad_norm": 16.191483693504896, | |
| "learning_rate": 8.851477564560061e-09, | |
| "logits/chosen": 3.820493221282959, | |
| "logits/rejected": 4.351908206939697, | |
| "logps/chosen": -389.7720947265625, | |
| "logps/rejected": -459.3812561035156, | |
| "loss": 0.55, | |
| "rewards/accuracies": 0.6968749761581421, | |
| "rewards/chosen": -2.404622793197632, | |
| "rewards/margins": 0.6956027746200562, | |
| "rewards/rejected": -3.1002252101898193, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.939116593712694, | |
| "grad_norm": 12.583560577719238, | |
| "learning_rate": 5.559858110443016e-09, | |
| "logits/chosen": 3.1959145069122314, | |
| "logits/rejected": 3.7669143676757812, | |
| "logps/chosen": -405.40618896484375, | |
| "logps/rejected": -479.1631774902344, | |
| "loss": 0.5251, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.4700753688812256, | |
| "rewards/margins": 0.7442026138305664, | |
| "rewards/rejected": -3.214277982711792, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.955033824114604, | |
| "grad_norm": 15.78612631007661, | |
| "learning_rate": 3.023789126611137e-09, | |
| "logits/chosen": 3.5097384452819824, | |
| "logits/rejected": 4.086105823516846, | |
| "logps/chosen": -395.22027587890625, | |
| "logps/rejected": -475.81787109375, | |
| "loss": 0.5376, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -2.4702706336975098, | |
| "rewards/margins": 0.8460418581962585, | |
| "rewards/rejected": -3.316312313079834, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.9709510545165141, | |
| "grad_norm": 15.913083563945913, | |
| "learning_rate": 1.2511094569571668e-09, | |
| "logits/chosen": 3.64544939994812, | |
| "logits/rejected": 4.29144811630249, | |
| "logps/chosen": -404.07501220703125, | |
| "logps/rejected": -467.01153564453125, | |
| "loss": 0.5567, | |
| "rewards/accuracies": 0.659375011920929, | |
| "rewards/chosen": -2.5605225563049316, | |
| "rewards/margins": 0.6661086082458496, | |
| "rewards/rejected": -3.2266311645507812, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.9868682849184242, | |
| "grad_norm": 15.29692820991952, | |
| "learning_rate": 2.4729835275189016e-10, | |
| "logits/chosen": 3.532379150390625, | |
| "logits/rejected": 3.938063859939575, | |
| "logps/chosen": -392.1960754394531, | |
| "logps/rejected": -481.100830078125, | |
| "loss": 0.5331, | |
| "rewards/accuracies": 0.7093750238418579, | |
| "rewards/chosen": -2.4987053871154785, | |
| "rewards/margins": 0.7970455288887024, | |
| "rewards/rejected": -3.295750379562378, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.9996020692399522, | |
| "step": 628, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5721431234080321, | |
| "train_runtime": 15529.5131, | |
| "train_samples_per_second": 10.354, | |
| "train_steps_per_second": 0.04 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 628, | |
| "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": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |