| { | |
| "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": 1.083263489032715, | |
| "learning_rate": 7.936507936507936e-09, | |
| "logits/chosen": -1.870174765586853, | |
| "logits/rejected": -1.8014098405838013, | |
| "logps/chosen": -158.94805908203125, | |
| "logps/rejected": -194.90118408203125, | |
| "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": 1.7486374056371567, | |
| "learning_rate": 7.936507936507936e-08, | |
| "logits/chosen": -1.8175890445709229, | |
| "logits/rejected": -1.8403409719467163, | |
| "logps/chosen": -188.36404418945312, | |
| "logps/rejected": -179.35687255859375, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.3576388955116272, | |
| "rewards/chosen": -0.0004040856729261577, | |
| "rewards/margins": -0.0008076421800069511, | |
| "rewards/rejected": 0.00040355639066547155, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03183446080382014, | |
| "grad_norm": 1.2288534081464715, | |
| "learning_rate": 1.5873015873015872e-07, | |
| "logits/chosen": -1.9721990823745728, | |
| "logits/rejected": -1.980015516281128, | |
| "logps/chosen": -197.92391967773438, | |
| "logps/rejected": -207.22412109375, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 8.275737855001353e-06, | |
| "rewards/margins": -0.0001316081325057894, | |
| "rewards/rejected": 0.00013988386490382254, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.0477516912057302, | |
| "grad_norm": 0.9705773937715004, | |
| "learning_rate": 2.3809523809523806e-07, | |
| "logits/chosen": -1.8532848358154297, | |
| "logits/rejected": -1.8753681182861328, | |
| "logps/chosen": -192.031982421875, | |
| "logps/rejected": -195.5606231689453, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": 0.0007524025277234614, | |
| "rewards/margins": 0.000668005901388824, | |
| "rewards/rejected": 8.439658995484933e-05, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.06366892160764027, | |
| "grad_norm": 1.3879715707434512, | |
| "learning_rate": 3.1746031746031743e-07, | |
| "logits/chosen": -1.9114805459976196, | |
| "logits/rejected": -1.9526256322860718, | |
| "logps/chosen": -198.54586791992188, | |
| "logps/rejected": -190.65158081054688, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.515625, | |
| "rewards/chosen": 0.0004263836017344147, | |
| "rewards/margins": 0.00022551305301021785, | |
| "rewards/rejected": 0.0002008705196203664, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.07958615200955034, | |
| "grad_norm": 0.9223665834941889, | |
| "learning_rate": 3.968253968253968e-07, | |
| "logits/chosen": -1.9061291217803955, | |
| "logits/rejected": -1.963629126548767, | |
| "logps/chosen": -190.56007385253906, | |
| "logps/rejected": -193.8275146484375, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.0014020069502294064, | |
| "rewards/margins": 0.00029031833400949836, | |
| "rewards/rejected": 0.001111688674427569, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.0955033824114604, | |
| "grad_norm": 1.6880319723067283, | |
| "learning_rate": 4.761904761904761e-07, | |
| "logits/chosen": -1.9793567657470703, | |
| "logits/rejected": -2.001311779022217, | |
| "logps/chosen": -196.06846618652344, | |
| "logps/rejected": -188.52914428710938, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.528124988079071, | |
| "rewards/chosen": 0.00374850956723094, | |
| "rewards/margins": 0.0002347224799450487, | |
| "rewards/rejected": 0.003513787407428026, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.11142061281337047, | |
| "grad_norm": 1.2950341728538266, | |
| "learning_rate": 4.998106548810311e-07, | |
| "logits/chosen": -1.9034782648086548, | |
| "logits/rejected": -1.9826198816299438, | |
| "logps/chosen": -196.86940002441406, | |
| "logps/rejected": -177.41644287109375, | |
| "loss": 0.6926, | |
| "rewards/accuracies": 0.5718749761581421, | |
| "rewards/chosen": 0.005933411885052919, | |
| "rewards/margins": 0.001384641625918448, | |
| "rewards/rejected": 0.004548769909888506, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.12733784321528055, | |
| "grad_norm": 2.34883059858935, | |
| "learning_rate": 4.988839406031596e-07, | |
| "logits/chosen": -1.9306480884552002, | |
| "logits/rejected": -1.9404945373535156, | |
| "logps/chosen": -176.39120483398438, | |
| "logps/rejected": -193.53347778320312, | |
| "loss": 0.6922, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.010212738066911697, | |
| "rewards/margins": 0.0014647066127508879, | |
| "rewards/rejected": 0.008748031221330166, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.14325507361719061, | |
| "grad_norm": 1.8215899265671025, | |
| "learning_rate": 4.971879403278432e-07, | |
| "logits/chosen": -1.8801124095916748, | |
| "logits/rejected": -1.9323303699493408, | |
| "logps/chosen": -189.4450225830078, | |
| "logps/rejected": -190.84414672851562, | |
| "loss": 0.6922, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 0.011725414544343948, | |
| "rewards/margins": 0.0013827414950355887, | |
| "rewards/rejected": 0.01034267246723175, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.15917230401910068, | |
| "grad_norm": 1.208023660425821, | |
| "learning_rate": 4.947278962947386e-07, | |
| "logits/chosen": -1.8471660614013672, | |
| "logits/rejected": -1.8932346105575562, | |
| "logps/chosen": -193.34718322753906, | |
| "logps/rejected": -196.87457275390625, | |
| "loss": 0.6915, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": 0.015522779896855354, | |
| "rewards/margins": 0.002660988597199321, | |
| "rewards/rejected": 0.01286179106682539, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.17508953442101075, | |
| "grad_norm": 1.1987469741900727, | |
| "learning_rate": 4.915114123589732e-07, | |
| "logits/chosen": -1.8427928686141968, | |
| "logits/rejected": -1.9131923913955688, | |
| "logps/chosen": -177.29983520507812, | |
| "logps/rejected": -178.62013244628906, | |
| "loss": 0.6911, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.019433852285146713, | |
| "rewards/margins": 0.003971750847995281, | |
| "rewards/rejected": 0.015462102368474007, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.1910067648229208, | |
| "grad_norm": 1.3513440295821657, | |
| "learning_rate": 4.875484304880629e-07, | |
| "logits/chosen": -1.852002501487732, | |
| "logits/rejected": -1.8977711200714111, | |
| "logps/chosen": -196.73301696777344, | |
| "logps/rejected": -198.54873657226562, | |
| "loss": 0.6905, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.02165614813566208, | |
| "rewards/margins": 0.006273286882787943, | |
| "rewards/rejected": 0.015382861718535423, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.20692399522483088, | |
| "grad_norm": 0.919308465194322, | |
| "learning_rate": 4.828512000318616e-07, | |
| "logits/chosen": -1.8836482763290405, | |
| "logits/rejected": -1.9497747421264648, | |
| "logps/chosen": -197.17617797851562, | |
| "logps/rejected": -199.19927978515625, | |
| "loss": 0.6895, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.026965487748384476, | |
| "rewards/margins": 0.008914025500416756, | |
| "rewards/rejected": 0.01805146224796772, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.22284122562674094, | |
| "grad_norm": 1.3141295336399639, | |
| "learning_rate": 4.774342398605221e-07, | |
| "logits/chosen": -1.9415794610977173, | |
| "logits/rejected": -1.9936542510986328, | |
| "logps/chosen": -191.3373565673828, | |
| "logps/rejected": -182.86744689941406, | |
| "loss": 0.6894, | |
| "rewards/accuracies": 0.565625011920929, | |
| "rewards/chosen": 0.02900194190442562, | |
| "rewards/margins": 0.007812701165676117, | |
| "rewards/rejected": 0.021189238876104355, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.238758456028651, | |
| "grad_norm": 1.0426885097254153, | |
| "learning_rate": 4.713142934875005e-07, | |
| "logits/chosen": -1.9385217428207397, | |
| "logits/rejected": -2.025717258453369, | |
| "logps/chosen": -201.0162811279297, | |
| "logps/rejected": -189.43167114257812, | |
| "loss": 0.6882, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": 0.030758211389183998, | |
| "rewards/margins": 0.014848137274384499, | |
| "rewards/rejected": 0.01591007225215435, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.2546756864305611, | |
| "grad_norm": 0.9993599317482367, | |
| "learning_rate": 4.64510277316316e-07, | |
| "logits/chosen": -1.883013367652893, | |
| "logits/rejected": -1.9087369441986084, | |
| "logps/chosen": -177.09432983398438, | |
| "logps/rejected": -181.0247802734375, | |
| "loss": 0.6875, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": 0.019768565893173218, | |
| "rewards/margins": 0.01374894194304943, | |
| "rewards/rejected": 0.006019626744091511, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.27059291683247116, | |
| "grad_norm": 6.045830822367488, | |
| "learning_rate": 4.570432221710314e-07, | |
| "logits/chosen": -1.8615553379058838, | |
| "logits/rejected": -1.9155362844467163, | |
| "logps/chosen": -191.3380126953125, | |
| "logps/rejected": -196.19644165039062, | |
| "loss": 0.6866, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.01975223794579506, | |
| "rewards/margins": 0.01370079256594181, | |
| "rewards/rejected": 0.006051443982869387, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.28651014723438123, | |
| "grad_norm": 3.6863233560878217, | |
| "learning_rate": 4.4893620829118124e-07, | |
| "logits/chosen": -1.8656437397003174, | |
| "logits/rejected": -1.9271419048309326, | |
| "logps/chosen": -197.45355224609375, | |
| "logps/rejected": -187.43008422851562, | |
| "loss": 0.6861, | |
| "rewards/accuracies": 0.621874988079071, | |
| "rewards/chosen": 0.011210992000997066, | |
| "rewards/margins": 0.020882535725831985, | |
| "rewards/rejected": -0.009671543724834919, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3024273776362913, | |
| "grad_norm": 1.0448028144351018, | |
| "learning_rate": 4.40214293992074e-07, | |
| "logits/chosen": -1.9013402462005615, | |
| "logits/rejected": -1.9314658641815186, | |
| "logps/chosen": -189.4186553955078, | |
| "logps/rejected": -181.21237182617188, | |
| "loss": 0.685, | |
| "rewards/accuracies": 0.6031249761581421, | |
| "rewards/chosen": -0.005181929562240839, | |
| "rewards/margins": 0.016439586877822876, | |
| "rewards/rejected": -0.021621517837047577, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.31834460803820136, | |
| "grad_norm": 2.39968390976015, | |
| "learning_rate": 4.3090443821097566e-07, | |
| "logits/chosen": -1.7716095447540283, | |
| "logits/rejected": -1.8057435750961304, | |
| "logps/chosen": -191.8214111328125, | |
| "logps/rejected": -202.6089630126953, | |
| "loss": 0.6834, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.008436818607151508, | |
| "rewards/margins": 0.02256094664335251, | |
| "rewards/rejected": -0.030997764319181442, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.3342618384401114, | |
| "grad_norm": 1.2956793625076484, | |
| "learning_rate": 4.210354171785795e-07, | |
| "logits/chosen": -1.7898409366607666, | |
| "logits/rejected": -1.8095554113388062, | |
| "logps/chosen": -187.91336059570312, | |
| "logps/rejected": -191.72274780273438, | |
| "loss": 0.6822, | |
| "rewards/accuracies": 0.6031249761581421, | |
| "rewards/chosen": -0.006862046662718058, | |
| "rewards/margins": 0.026659756898880005, | |
| "rewards/rejected": -0.033521804958581924, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.3501790688420215, | |
| "grad_norm": 2.0041098363033623, | |
| "learning_rate": 4.1063773547332584e-07, | |
| "logits/chosen": -1.7102569341659546, | |
| "logits/rejected": -1.7631139755249023, | |
| "logps/chosen": -192.54238891601562, | |
| "logps/rejected": -196.4750213623047, | |
| "loss": 0.6811, | |
| "rewards/accuracies": 0.515625, | |
| "rewards/chosen": -0.02755718305706978, | |
| "rewards/margins": 0.021966466680169106, | |
| "rewards/rejected": -0.04952365159988403, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.36609629924393156, | |
| "grad_norm": 1.9332381756111658, | |
| "learning_rate": 3.997435317334988e-07, | |
| "logits/chosen": -1.7369827032089233, | |
| "logits/rejected": -1.7770426273345947, | |
| "logps/chosen": -191.87298583984375, | |
| "logps/rejected": -193.9591522216797, | |
| "loss": 0.6817, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.043014999479055405, | |
| "rewards/margins": 0.020200857892632484, | |
| "rewards/rejected": -0.06321585923433304, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.3820135296458416, | |
| "grad_norm": 1.8903913985972478, | |
| "learning_rate": 3.8838647931853684e-07, | |
| "logits/chosen": -1.7956278324127197, | |
| "logits/rejected": -1.7797701358795166, | |
| "logps/chosen": -191.063720703125, | |
| "logps/rejected": -194.3412322998047, | |
| "loss": 0.6807, | |
| "rewards/accuracies": 0.596875011920929, | |
| "rewards/chosen": -0.04468019679188728, | |
| "rewards/margins": 0.024412112310528755, | |
| "rewards/rejected": -0.06909231096506119, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.3979307600477517, | |
| "grad_norm": 2.209986978975659, | |
| "learning_rate": 3.7660168222660824e-07, | |
| "logits/chosen": -1.7686593532562256, | |
| "logits/rejected": -1.784973382949829, | |
| "logps/chosen": -210.4496612548828, | |
| "logps/rejected": -210.5813751220703, | |
| "loss": 0.6764, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.06448944658041, | |
| "rewards/margins": 0.04242020100355148, | |
| "rewards/rejected": -0.10690965503454208, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.41384799044966175, | |
| "grad_norm": 2.7492197743164577, | |
| "learning_rate": 3.6442556659016475e-07, | |
| "logits/chosen": -1.6692039966583252, | |
| "logits/rejected": -1.7762839794158936, | |
| "logps/chosen": -204.4394073486328, | |
| "logps/rejected": -207.07705688476562, | |
| "loss": 0.6753, | |
| "rewards/accuracies": 0.590624988079071, | |
| "rewards/chosen": -0.07524862885475159, | |
| "rewards/margins": 0.02850642427802086, | |
| "rewards/rejected": -0.10375505685806274, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.4297652208515718, | |
| "grad_norm": 2.266511648349345, | |
| "learning_rate": 3.5189576808485404e-07, | |
| "logits/chosen": -1.7136598825454712, | |
| "logits/rejected": -1.710269570350647, | |
| "logps/chosen": -208.0347137451172, | |
| "logps/rejected": -209.95751953125, | |
| "loss": 0.6737, | |
| "rewards/accuracies": 0.578125, | |
| "rewards/chosen": -0.08503381162881851, | |
| "rewards/margins": 0.0269694272428751, | |
| "rewards/rejected": -0.11200324445962906, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.4456824512534819, | |
| "grad_norm": 2.5291958520679354, | |
| "learning_rate": 3.390510155998023e-07, | |
| "logits/chosen": -1.628941297531128, | |
| "logits/rejected": -1.6508426666259766, | |
| "logps/chosen": -214.1870574951172, | |
| "logps/rejected": -212.94284057617188, | |
| "loss": 0.6707, | |
| "rewards/accuracies": 0.6468750238418579, | |
| "rewards/chosen": -0.12327191978693008, | |
| "rewards/margins": 0.07120023667812347, | |
| "rewards/rejected": -0.19447214901447296, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.46159968165539195, | |
| "grad_norm": 2.574405785518012, | |
| "learning_rate": 3.2593101152883795e-07, | |
| "logits/chosen": -1.6216236352920532, | |
| "logits/rejected": -1.6766021251678467, | |
| "logps/chosen": -216.7034149169922, | |
| "logps/rejected": -219.9063262939453, | |
| "loss": 0.6732, | |
| "rewards/accuracies": 0.621874988079071, | |
| "rewards/chosen": -0.16276732087135315, | |
| "rewards/margins": 0.05662418156862259, | |
| "rewards/rejected": -0.21939150989055634, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.477516912057302, | |
| "grad_norm": 2.6453554421569514, | |
| "learning_rate": 3.125763090526674e-07, | |
| "logits/chosen": -1.5658096075057983, | |
| "logits/rejected": -1.609926462173462, | |
| "logps/chosen": -212.8961944580078, | |
| "logps/rejected": -213.0232696533203, | |
| "loss": 0.6673, | |
| "rewards/accuracies": 0.596875011920929, | |
| "rewards/chosen": -0.14756593108177185, | |
| "rewards/margins": 0.06441614031791687, | |
| "rewards/rejected": -0.21198205649852753, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.4934341424592121, | |
| "grad_norm": 4.170174585407687, | |
| "learning_rate": 2.9902818679131775e-07, | |
| "logits/chosen": -1.5784204006195068, | |
| "logits/rejected": -1.5990747213363647, | |
| "logps/chosen": -216.91799926757812, | |
| "logps/rejected": -221.1397705078125, | |
| "loss": 0.665, | |
| "rewards/accuracies": 0.590624988079071, | |
| "rewards/chosen": -0.20333552360534668, | |
| "rewards/margins": 0.05764692276716232, | |
| "rewards/rejected": -0.2609824538230896, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.5093513728611222, | |
| "grad_norm": 2.5979901565172714, | |
| "learning_rate": 2.8532852121428733e-07, | |
| "logits/chosen": -1.5708476305007935, | |
| "logits/rejected": -1.610568642616272, | |
| "logps/chosen": -212.8919219970703, | |
| "logps/rejected": -209.1988983154297, | |
| "loss": 0.6602, | |
| "rewards/accuracies": 0.6031249761581421, | |
| "rewards/chosen": -0.21002519130706787, | |
| "rewards/margins": 0.07976146787405014, | |
| "rewards/rejected": -0.2897866368293762, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.5252686032630323, | |
| "grad_norm": 3.0268583107868676, | |
| "learning_rate": 2.715196572027789e-07, | |
| "logits/chosen": -1.5595386028289795, | |
| "logits/rejected": -1.579420804977417, | |
| "logps/chosen": -208.55126953125, | |
| "logps/rejected": -228.11337280273438, | |
| "loss": 0.6561, | |
| "rewards/accuracies": 0.6343749761581421, | |
| "rewards/chosen": -0.19552794098854065, | |
| "rewards/margins": 0.11572189629077911, | |
| "rewards/rejected": -0.31124982237815857, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.5411858336649423, | |
| "grad_norm": 2.608251205411858, | |
| "learning_rate": 2.5764427716409815e-07, | |
| "logits/chosen": -1.5596858263015747, | |
| "logits/rejected": -1.5745502710342407, | |
| "logps/chosen": -222.6412811279297, | |
| "logps/rejected": -227.70651245117188, | |
| "loss": 0.6544, | |
| "rewards/accuracies": 0.640625, | |
| "rewards/chosen": -0.21032941341400146, | |
| "rewards/margins": 0.09801909327507019, | |
| "rewards/rejected": -0.30834850668907166, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.5571030640668524, | |
| "grad_norm": 2.3596476657059666, | |
| "learning_rate": 2.4374526910277886e-07, | |
| "logits/chosen": -1.4910805225372314, | |
| "logits/rejected": -1.4995691776275635, | |
| "logps/chosen": -219.46536254882812, | |
| "logps/rejected": -235.89620971679688, | |
| "loss": 0.6613, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.2539146840572357, | |
| "rewards/margins": 0.09661410748958588, | |
| "rewards/rejected": -0.3505287766456604, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.5730202944687625, | |
| "grad_norm": 2.721512353744307, | |
| "learning_rate": 2.2986559405621886e-07, | |
| "logits/chosen": -1.5120307207107544, | |
| "logits/rejected": -1.5819119215011597, | |
| "logps/chosen": -236.832763671875, | |
| "logps/rejected": -235.5454559326172, | |
| "loss": 0.6668, | |
| "rewards/accuracies": 0.565625011920929, | |
| "rewards/chosen": -0.3000979721546173, | |
| "rewards/margins": 0.07621388137340546, | |
| "rewards/rejected": -0.37631186842918396, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.5889375248706725, | |
| "grad_norm": 3.020330440823815, | |
| "learning_rate": 2.160481533045751e-07, | |
| "logits/chosen": -1.4668768644332886, | |
| "logits/rejected": -1.5363471508026123, | |
| "logps/chosen": -218.35946655273438, | |
| "logps/rejected": -224.10208129882812, | |
| "loss": 0.6492, | |
| "rewards/accuracies": 0.6156250238418579, | |
| "rewards/chosen": -0.2719998359680176, | |
| "rewards/margins": 0.12549051642417908, | |
| "rewards/rejected": -0.39749038219451904, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.6048547552725826, | |
| "grad_norm": 2.9446946400840974, | |
| "learning_rate": 2.0233565576536564e-07, | |
| "logits/chosen": -1.485190749168396, | |
| "logits/rejected": -1.5274041891098022, | |
| "logps/chosen": -222.81765747070312, | |
| "logps/rejected": -222.657470703125, | |
| "loss": 0.6538, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.26167741417884827, | |
| "rewards/margins": 0.08558806031942368, | |
| "rewards/rejected": -0.34726545214653015, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.6207719856744927, | |
| "grad_norm": 3.1883509652873223, | |
| "learning_rate": 1.887704859826528e-07, | |
| "logits/chosen": -1.4792619943618774, | |
| "logits/rejected": -1.5055263042449951, | |
| "logps/chosen": -221.5879364013672, | |
| "logps/rejected": -239.82571411132812, | |
| "loss": 0.6495, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.2849220633506775, | |
| "rewards/margins": 0.13855159282684326, | |
| "rewards/rejected": -0.42347365617752075, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.6366892160764027, | |
| "grad_norm": 2.8954609378442657, | |
| "learning_rate": 1.7539457311884675e-07, | |
| "logits/chosen": -1.4238282442092896, | |
| "logits/rejected": -1.423081398010254, | |
| "logps/chosen": -229.7063446044922, | |
| "logps/rejected": -235.14236450195312, | |
| "loss": 0.6522, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.2554658353328705, | |
| "rewards/margins": 0.14818701148033142, | |
| "rewards/rejected": -0.4036528468132019, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.6526064464783128, | |
| "grad_norm": 3.203997856966496, | |
| "learning_rate": 1.6224926135406693e-07, | |
| "logits/chosen": -1.460460901260376, | |
| "logits/rejected": -1.4777477979660034, | |
| "logps/chosen": -236.0177764892578, | |
| "logps/rejected": -240.2278289794922, | |
| "loss": 0.6525, | |
| "rewards/accuracies": 0.596875011920929, | |
| "rewards/chosen": -0.32773512601852417, | |
| "rewards/margins": 0.08220280706882477, | |
| "rewards/rejected": -0.4099380075931549, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.6685236768802229, | |
| "grad_norm": 2.8800206835573094, | |
| "learning_rate": 1.4937518209365108e-07, | |
| "logits/chosen": -1.5116090774536133, | |
| "logits/rejected": -1.5359185934066772, | |
| "logps/chosen": -248.60263061523438, | |
| "logps/rejected": -245.75967407226562, | |
| "loss": 0.6527, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.3213188052177429, | |
| "rewards/margins": 0.10633780807256699, | |
| "rewards/rejected": -0.4276565909385681, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.6844409072821329, | |
| "grad_norm": 3.29734083048723, | |
| "learning_rate": 1.3681212837880977e-07, | |
| "logits/chosen": -1.4253356456756592, | |
| "logits/rejected": -1.444923758506775, | |
| "logps/chosen": -220.8928985595703, | |
| "logps/rejected": -238.1382293701172, | |
| "loss": 0.6502, | |
| "rewards/accuracies": 0.621874988079071, | |
| "rewards/chosen": -0.3292413353919983, | |
| "rewards/margins": 0.10124459117650986, | |
| "rewards/rejected": -0.43048587441444397, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.700358137684043, | |
| "grad_norm": 3.2357462634957908, | |
| "learning_rate": 1.2459893188861613e-07, | |
| "logits/chosen": -1.3886950016021729, | |
| "logits/rejected": -1.4186146259307861, | |
| "logps/chosen": -229.12069702148438, | |
| "logps/rejected": -241.5669708251953, | |
| "loss": 0.65, | |
| "rewards/accuracies": 0.6656249761581421, | |
| "rewards/chosen": -0.31468960642814636, | |
| "rewards/margins": 0.14469651877880096, | |
| "rewards/rejected": -0.4593861699104309, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.716275368085953, | |
| "grad_norm": 3.664386493120931, | |
| "learning_rate": 1.1277334291351145e-07, | |
| "logits/chosen": -1.4633034467697144, | |
| "logits/rejected": -1.4953584671020508, | |
| "logps/chosen": -229.84976196289062, | |
| "logps/rejected": -245.46762084960938, | |
| "loss": 0.6451, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.35235971212387085, | |
| "rewards/margins": 0.1600349247455597, | |
| "rewards/rejected": -0.5123946666717529, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.7321925984878631, | |
| "grad_norm": 3.054029038543547, | |
| "learning_rate": 1.0137191367132078e-07, | |
| "logits/chosen": -1.466225504875183, | |
| "logits/rejected": -1.5026957988739014, | |
| "logps/chosen": -227.568359375, | |
| "logps/rejected": -251.8888397216797, | |
| "loss": 0.6449, | |
| "rewards/accuracies": 0.6468750238418579, | |
| "rewards/chosen": -0.2990425229072571, | |
| "rewards/margins": 0.16379448771476746, | |
| "rewards/rejected": -0.46283698081970215, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.7481098288897732, | |
| "grad_norm": 3.604867452453406, | |
| "learning_rate": 9.042988532644249e-08, | |
| "logits/chosen": -1.447538137435913, | |
| "logits/rejected": -1.5032737255096436, | |
| "logps/chosen": -240.51644897460938, | |
| "logps/rejected": -250.5240020751953, | |
| "loss": 0.6527, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.36844930052757263, | |
| "rewards/margins": 0.1114746555685997, | |
| "rewards/rejected": -0.47992390394210815, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.7640270592916832, | |
| "grad_norm": 3.100922028441707, | |
| "learning_rate": 7.998107906142839e-08, | |
| "logits/chosen": -1.4335956573486328, | |
| "logits/rejected": -1.468783974647522, | |
| "logps/chosen": -220.85147094726562, | |
| "logps/rejected": -234.76596069335938, | |
| "loss": 0.6462, | |
| "rewards/accuracies": 0.640625, | |
| "rewards/chosen": -0.3198398947715759, | |
| "rewards/margins": 0.13547857105731964, | |
| "rewards/rejected": -0.45531851053237915, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.7799442896935933, | |
| "grad_norm": 3.0897060153169447, | |
| "learning_rate": 7.005779153764682e-08, | |
| "logits/chosen": -1.460636854171753, | |
| "logits/rejected": -1.481178641319275, | |
| "logps/chosen": -217.1626434326172, | |
| "logps/rejected": -241.839111328125, | |
| "loss": 0.6465, | |
| "rewards/accuracies": 0.628125011920929, | |
| "rewards/chosen": -0.33777886629104614, | |
| "rewards/margins": 0.1417325884103775, | |
| "rewards/rejected": -0.47951143980026245, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.7958615200955034, | |
| "grad_norm": 3.515426969892236, | |
| "learning_rate": 6.069069506815325e-08, | |
| "logits/chosen": -1.423098087310791, | |
| "logits/rejected": -1.4111202955245972, | |
| "logps/chosen": -228.85360717773438, | |
| "logps/rejected": -247.3131561279297, | |
| "loss": 0.6459, | |
| "rewards/accuracies": 0.6468750238418579, | |
| "rewards/chosen": -0.3209570348262787, | |
| "rewards/margins": 0.18556544184684753, | |
| "rewards/rejected": -0.506522536277771, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.8117787504974134, | |
| "grad_norm": 3.168259746643005, | |
| "learning_rate": 5.190874281132851e-08, | |
| "logits/chosen": -1.3531607389450073, | |
| "logits/rejected": -1.4315484762191772, | |
| "logps/chosen": -228.17135620117188, | |
| "logps/rejected": -237.56796264648438, | |
| "loss": 0.6446, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.3634713888168335, | |
| "rewards/margins": 0.11672137677669525, | |
| "rewards/rejected": -0.48019275069236755, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.8276959808993235, | |
| "grad_norm": 3.547783875081257, | |
| "learning_rate": 4.373907927832513e-08, | |
| "logits/chosen": -1.4724462032318115, | |
| "logits/rejected": -1.4711410999298096, | |
| "logps/chosen": -216.9486541748047, | |
| "logps/rejected": -229.16531372070312, | |
| "loss": 0.6497, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.34610506892204285, | |
| "rewards/margins": 0.10790134966373444, | |
| "rewards/rejected": -0.4540063738822937, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.8436132113012336, | |
| "grad_norm": 3.0690519001773744, | |
| "learning_rate": 3.620695643093924e-08, | |
| "logits/chosen": -1.436292290687561, | |
| "logits/rejected": -1.4620921611785889, | |
| "logps/chosen": -220.2271728515625, | |
| "logps/rejected": -242.4230499267578, | |
| "loss": 0.6444, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -0.2857658565044403, | |
| "rewards/margins": 0.16418084502220154, | |
| "rewards/rejected": -0.44994670152664185, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.8595304417031436, | |
| "grad_norm": 3.5137530293056676, | |
| "learning_rate": 2.9335655629243645e-08, | |
| "logits/chosen": -1.4339349269866943, | |
| "logits/rejected": -1.4756128787994385, | |
| "logps/chosen": -225.1971893310547, | |
| "logps/rejected": -240.7228546142578, | |
| "loss": 0.6474, | |
| "rewards/accuracies": 0.628125011920929, | |
| "rewards/chosen": -0.32935741543769836, | |
| "rewards/margins": 0.12228095531463623, | |
| "rewards/rejected": -0.4516383707523346, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.8754476721050537, | |
| "grad_norm": 3.303113230393075, | |
| "learning_rate": 2.31464156702382e-08, | |
| "logits/chosen": -1.4354156255722046, | |
| "logits/rejected": -1.4688793420791626, | |
| "logps/chosen": -228.87588500976562, | |
| "logps/rejected": -227.47872924804688, | |
| "loss": 0.6503, | |
| "rewards/accuracies": 0.609375, | |
| "rewards/chosen": -0.31748342514038086, | |
| "rewards/margins": 0.12047611176967621, | |
| "rewards/rejected": -0.4379595220088959, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.8913649025069638, | |
| "grad_norm": 3.6187583437530138, | |
| "learning_rate": 1.7658367139945228e-08, | |
| "logits/chosen": -1.484116554260254, | |
| "logits/rejected": -1.4669333696365356, | |
| "logps/chosen": -220.8059844970703, | |
| "logps/rejected": -235.5103759765625, | |
| "loss": 0.6422, | |
| "rewards/accuracies": 0.578125, | |
| "rewards/chosen": -0.34964197874069214, | |
| "rewards/margins": 0.1103706806898117, | |
| "rewards/rejected": -0.46001267433166504, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.9072821329088738, | |
| "grad_norm": 2.946862169915585, | |
| "learning_rate": 1.2888473281864597e-08, | |
| "logits/chosen": -1.3717347383499146, | |
| "logits/rejected": -1.4307266473770142, | |
| "logps/chosen": -230.9296112060547, | |
| "logps/rejected": -248.7758026123047, | |
| "loss": 0.6373, | |
| "rewards/accuracies": 0.6468750238418579, | |
| "rewards/chosen": -0.3366944193840027, | |
| "rewards/margins": 0.13169880211353302, | |
| "rewards/rejected": -0.4683932363986969, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.9231993633107839, | |
| "grad_norm": 3.986364311332544, | |
| "learning_rate": 8.851477564560061e-09, | |
| "logits/chosen": -1.4669787883758545, | |
| "logits/rejected": -1.5305414199829102, | |
| "logps/chosen": -227.6433563232422, | |
| "logps/rejected": -243.227294921875, | |
| "loss": 0.6434, | |
| "rewards/accuracies": 0.6468750238418579, | |
| "rewards/chosen": -0.32894089818000793, | |
| "rewards/margins": 0.15033873915672302, | |
| "rewards/rejected": -0.47927966713905334, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.939116593712694, | |
| "grad_norm": 3.4745032456527394, | |
| "learning_rate": 5.559858110443016e-09, | |
| "logits/chosen": -1.418752908706665, | |
| "logits/rejected": -1.463607668876648, | |
| "logps/chosen": -237.24252319335938, | |
| "logps/rejected": -252.40145874023438, | |
| "loss": 0.6384, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.3376900553703308, | |
| "rewards/margins": 0.15840700268745422, | |
| "rewards/rejected": -0.4960970878601074, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.955033824114604, | |
| "grad_norm": 6.105907684391068, | |
| "learning_rate": 3.023789126611137e-09, | |
| "logits/chosen": -1.4318501949310303, | |
| "logits/rejected": -1.4236361980438232, | |
| "logps/chosen": -224.8287811279297, | |
| "logps/rejected": -240.1597442626953, | |
| "loss": 0.6398, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.32297495007514954, | |
| "rewards/margins": 0.1879177987575531, | |
| "rewards/rejected": -0.5108927488327026, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.9709510545165141, | |
| "grad_norm": 7.921761427791219, | |
| "learning_rate": 1.2511094569571668e-09, | |
| "logits/chosen": -1.4697223901748657, | |
| "logits/rejected": -1.5091297626495361, | |
| "logps/chosen": -227.2742156982422, | |
| "logps/rejected": -238.57974243164062, | |
| "loss": 0.6414, | |
| "rewards/accuracies": 0.596875011920929, | |
| "rewards/chosen": -0.3365384042263031, | |
| "rewards/margins": 0.1458342969417572, | |
| "rewards/rejected": -0.4823727011680603, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.9868682849184242, | |
| "grad_norm": 3.1578814752663407, | |
| "learning_rate": 2.4729835275189016e-10, | |
| "logits/chosen": -1.4319922924041748, | |
| "logits/rejected": -1.4167181253433228, | |
| "logps/chosen": -224.8016357421875, | |
| "logps/rejected": -247.70339965820312, | |
| "loss": 0.6404, | |
| "rewards/accuracies": 0.6343749761581421, | |
| "rewards/chosen": -0.36974093317985535, | |
| "rewards/margins": 0.14402151107788086, | |
| "rewards/rejected": -0.5137624740600586, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.9996020692399522, | |
| "step": 628, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6662861807331159, | |
| "train_runtime": 15158.4407, | |
| "train_samples_per_second": 10.608, | |
| "train_steps_per_second": 0.041 | |
| } | |
| ], | |
| "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 | |
| } | |