| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9997382884061764, | |
| "eval_steps": 100, | |
| "global_step": 955, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0010468463752944255, | |
| "grad_norm": 5.5, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -0.3494967222213745, | |
| "logits/rejected": -0.3728627860546112, | |
| "logps/chosen": -285.8127136230469, | |
| "logps/rejected": -212.7957000732422, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.010468463752944255, | |
| "grad_norm": 6.0, | |
| "learning_rate": 5.208333333333334e-07, | |
| "logits/chosen": -0.5402188301086426, | |
| "logits/rejected": -0.5500054955482483, | |
| "logps/chosen": -315.3569030761719, | |
| "logps/rejected": -278.0135192871094, | |
| "loss": 0.6971, | |
| "rewards/accuracies": 0.4236111044883728, | |
| "rewards/chosen": -0.0025262832641601562, | |
| "rewards/margins": -0.0020421938970685005, | |
| "rewards/rejected": -0.0004840891051571816, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02093692750588851, | |
| "grad_norm": 5.46875, | |
| "learning_rate": 1.0416666666666667e-06, | |
| "logits/chosen": -0.5040005445480347, | |
| "logits/rejected": -0.5246113538742065, | |
| "logps/chosen": -306.70074462890625, | |
| "logps/rejected": -271.21923828125, | |
| "loss": 0.6942, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.007291222922503948, | |
| "rewards/margins": 0.007475622929632664, | |
| "rewards/rejected": -0.00018439842096995562, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.031405391258832765, | |
| "grad_norm": 5.9375, | |
| "learning_rate": 1.5625e-06, | |
| "logits/chosen": -0.5098769664764404, | |
| "logits/rejected": -0.5173646807670593, | |
| "logps/chosen": -291.0880432128906, | |
| "logps/rejected": -252.43124389648438, | |
| "loss": 0.6896, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.007684030570089817, | |
| "rewards/margins": 0.009358614683151245, | |
| "rewards/rejected": -0.001674586907029152, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04187385501177702, | |
| "grad_norm": 5.375, | |
| "learning_rate": 2.0833333333333334e-06, | |
| "logits/chosen": -0.4819061756134033, | |
| "logits/rejected": -0.5171386003494263, | |
| "logps/chosen": -306.039794921875, | |
| "logps/rejected": -244.5795135498047, | |
| "loss": 0.6816, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.026073966175317764, | |
| "rewards/margins": 0.02752821519970894, | |
| "rewards/rejected": -0.0014542475109919906, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.05234231876472128, | |
| "grad_norm": 4.5625, | |
| "learning_rate": 2.604166666666667e-06, | |
| "logits/chosen": -0.5450257658958435, | |
| "logits/rejected": -0.5735383033752441, | |
| "logps/chosen": -305.06109619140625, | |
| "logps/rejected": -282.95440673828125, | |
| "loss": 0.6762, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": 0.05157756805419922, | |
| "rewards/margins": 0.039979297667741776, | |
| "rewards/rejected": 0.011598266661167145, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06281078251766553, | |
| "grad_norm": 5.40625, | |
| "learning_rate": 3.125e-06, | |
| "logits/chosen": -0.5655590891838074, | |
| "logits/rejected": -0.5666534900665283, | |
| "logps/chosen": -290.9037780761719, | |
| "logps/rejected": -254.7635955810547, | |
| "loss": 0.6644, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": 0.08077142387628555, | |
| "rewards/margins": 0.037061307579278946, | |
| "rewards/rejected": 0.04371010884642601, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07327924627060979, | |
| "grad_norm": 5.0625, | |
| "learning_rate": 3.6458333333333333e-06, | |
| "logits/chosen": -0.5553646683692932, | |
| "logits/rejected": -0.5701892375946045, | |
| "logps/chosen": -286.6636962890625, | |
| "logps/rejected": -263.30523681640625, | |
| "loss": 0.6466, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.16360881924629211, | |
| "rewards/margins": 0.1136423796415329, | |
| "rewards/rejected": 0.04996642470359802, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.08374771002355404, | |
| "grad_norm": 4.78125, | |
| "learning_rate": 4.166666666666667e-06, | |
| "logits/chosen": -0.4944031834602356, | |
| "logits/rejected": -0.5216720700263977, | |
| "logps/chosen": -287.98211669921875, | |
| "logps/rejected": -273.87139892578125, | |
| "loss": 0.6293, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.25989845395088196, | |
| "rewards/margins": 0.19782108068466187, | |
| "rewards/rejected": 0.06207737326622009, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.0942161737764983, | |
| "grad_norm": 5.3125, | |
| "learning_rate": 4.6875000000000004e-06, | |
| "logits/chosen": -0.5244244337081909, | |
| "logits/rejected": -0.5471149682998657, | |
| "logps/chosen": -330.5142517089844, | |
| "logps/rejected": -295.2393798828125, | |
| "loss": 0.6177, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.2595275342464447, | |
| "rewards/margins": 0.25044435262680054, | |
| "rewards/rejected": 0.009083209559321404, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.10468463752944256, | |
| "grad_norm": 5.71875, | |
| "learning_rate": 4.9997324926814375e-06, | |
| "logits/chosen": -0.5673449039459229, | |
| "logits/rejected": -0.5620633959770203, | |
| "logps/chosen": -275.5569763183594, | |
| "logps/rejected": -289.50457763671875, | |
| "loss": 0.6275, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.3272210955619812, | |
| "rewards/margins": 0.25637245178222656, | |
| "rewards/rejected": 0.07084863632917404, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.10468463752944256, | |
| "eval_logits/chosen": -0.5618711709976196, | |
| "eval_logits/rejected": -0.5581737160682678, | |
| "eval_logps/chosen": -289.9532775878906, | |
| "eval_logps/rejected": -276.77532958984375, | |
| "eval_loss": 0.612218976020813, | |
| "eval_rewards/accuracies": 0.6919999718666077, | |
| "eval_rewards/chosen": 0.2593880295753479, | |
| "eval_rewards/margins": 0.2692645788192749, | |
| "eval_rewards/rejected": -0.00987652875483036, | |
| "eval_runtime": 351.8724, | |
| "eval_samples_per_second": 5.684, | |
| "eval_steps_per_second": 0.355, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11515310128238682, | |
| "grad_norm": 4.25, | |
| "learning_rate": 4.996723692767927e-06, | |
| "logits/chosen": -0.6067845225334167, | |
| "logits/rejected": -0.6330805420875549, | |
| "logps/chosen": -288.8458557128906, | |
| "logps/rejected": -276.23870849609375, | |
| "loss": 0.6078, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.27882444858551025, | |
| "rewards/margins": 0.34665778279304504, | |
| "rewards/rejected": -0.0678333044052124, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12562156503533106, | |
| "grad_norm": 5.0625, | |
| "learning_rate": 4.9903757462135984e-06, | |
| "logits/chosen": -0.5565638542175293, | |
| "logits/rejected": -0.5734590291976929, | |
| "logps/chosen": -260.9988098144531, | |
| "logps/rejected": -252.0515594482422, | |
| "loss": 0.591, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.15620097517967224, | |
| "rewards/margins": 0.29570046067237854, | |
| "rewards/rejected": -0.1394994556903839, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.1360900287882753, | |
| "grad_norm": 5.5, | |
| "learning_rate": 4.980697142834315e-06, | |
| "logits/chosen": -0.5258967876434326, | |
| "logits/rejected": -0.5430957078933716, | |
| "logps/chosen": -301.23260498046875, | |
| "logps/rejected": -336.9810485839844, | |
| "loss": 0.5936, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.17174401879310608, | |
| "rewards/margins": 0.31216299533843994, | |
| "rewards/rejected": -0.14041896164417267, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.14655849254121958, | |
| "grad_norm": 4.75, | |
| "learning_rate": 4.967700826904229e-06, | |
| "logits/chosen": -0.5994604825973511, | |
| "logits/rejected": -0.6112743616104126, | |
| "logps/chosen": -281.85223388671875, | |
| "logps/rejected": -274.6047058105469, | |
| "loss": 0.5781, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.12341972440481186, | |
| "rewards/margins": 0.4266698360443115, | |
| "rewards/rejected": -0.30325013399124146, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15702695629416383, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 4.951404179843963e-06, | |
| "logits/chosen": -0.5988300442695618, | |
| "logits/rejected": -0.5718582272529602, | |
| "logps/chosen": -308.1268310546875, | |
| "logps/rejected": -279.3246765136719, | |
| "loss": 0.5775, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": 0.22290559113025665, | |
| "rewards/margins": 0.45703545212745667, | |
| "rewards/rejected": -0.2341299057006836, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.16749542004710807, | |
| "grad_norm": 4.5, | |
| "learning_rate": 4.931828996974498e-06, | |
| "logits/chosen": -0.5415905117988586, | |
| "logits/rejected": -0.5388587713241577, | |
| "logps/chosen": -297.1036682128906, | |
| "logps/rejected": -271.18475341796875, | |
| "loss": 0.5459, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.1913338154554367, | |
| "rewards/margins": 0.5190017819404602, | |
| "rewards/rejected": -0.3276679515838623, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.17796388380005235, | |
| "grad_norm": 4.875, | |
| "learning_rate": 4.909001458367867e-06, | |
| "logits/chosen": -0.6261523365974426, | |
| "logits/rejected": -0.6151515245437622, | |
| "logps/chosen": -288.1965637207031, | |
| "logps/rejected": -275.8837585449219, | |
| "loss": 0.5809, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.018318424001336098, | |
| "rewards/margins": 0.4437525272369385, | |
| "rewards/rejected": -0.42543405294418335, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.1884323475529966, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 4.882952093833628e-06, | |
| "logits/chosen": -0.6297336220741272, | |
| "logits/rejected": -0.6054444313049316, | |
| "logps/chosen": -303.1052551269531, | |
| "logps/rejected": -265.9317932128906, | |
| "loss": 0.5599, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.07898494601249695, | |
| "rewards/margins": 0.4595085084438324, | |
| "rewards/rejected": -0.5384935140609741, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.19890081130594087, | |
| "grad_norm": 5.875, | |
| "learning_rate": 4.853715742087947e-06, | |
| "logits/chosen": -0.5653724074363708, | |
| "logits/rejected": -0.5484523177146912, | |
| "logps/chosen": -276.7878723144531, | |
| "logps/rejected": -283.9636535644531, | |
| "loss": 0.5658, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.021318156272172928, | |
| "rewards/margins": 0.4857168197631836, | |
| "rewards/rejected": -0.46439868211746216, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "grad_norm": 6.5, | |
| "learning_rate": 4.821331504159906e-06, | |
| "logits/chosen": -0.5703824162483215, | |
| "logits/rejected": -0.5851908922195435, | |
| "logps/chosen": -298.11962890625, | |
| "logps/rejected": -256.55633544921875, | |
| "loss": 0.5726, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.06406483054161072, | |
| "rewards/margins": 0.5046994090080261, | |
| "rewards/rejected": -0.5687642097473145, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "eval_logits/chosen": -0.5266345739364624, | |
| "eval_logits/rejected": -0.510294497013092, | |
| "eval_logps/chosen": -293.3344421386719, | |
| "eval_logps/rejected": -283.0293273925781, | |
| "eval_loss": 0.552874743938446, | |
| "eval_rewards/accuracies": 0.7039999961853027, | |
| "eval_rewards/chosen": -0.07872959226369858, | |
| "eval_rewards/margins": 0.5565416812896729, | |
| "eval_rewards/rejected": -0.6352713108062744, | |
| "eval_runtime": 349.2838, | |
| "eval_samples_per_second": 5.726, | |
| "eval_steps_per_second": 0.358, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21983773881182936, | |
| "grad_norm": 6.21875, | |
| "learning_rate": 4.7858426910973435e-06, | |
| "logits/chosen": -0.6147966980934143, | |
| "logits/rejected": -0.6168125867843628, | |
| "logps/chosen": -280.411865234375, | |
| "logps/rejected": -273.9303283691406, | |
| "loss": 0.5479, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.01772790588438511, | |
| "rewards/margins": 0.5174022912979126, | |
| "rewards/rejected": -0.5351301431655884, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.23030620256477363, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 4.747296766042161e-06, | |
| "logits/chosen": -0.5869513750076294, | |
| "logits/rejected": -0.5712444186210632, | |
| "logps/chosen": -319.57781982421875, | |
| "logps/rejected": -271.85662841796875, | |
| "loss": 0.5371, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.07761907577514648, | |
| "rewards/margins": 0.5722238421440125, | |
| "rewards/rejected": -0.49460476636886597, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.24077466631771788, | |
| "grad_norm": 5.84375, | |
| "learning_rate": 4.705745280752586e-06, | |
| "logits/chosen": -0.6202965974807739, | |
| "logits/rejected": -0.5799211263656616, | |
| "logps/chosen": -292.62506103515625, | |
| "logps/rejected": -289.6524658203125, | |
| "loss": 0.5593, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.0319809652864933, | |
| "rewards/margins": 0.5654061436653137, | |
| "rewards/rejected": -0.5973870754241943, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.2512431300706621, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 4.661243806657256e-06, | |
| "logits/chosen": -0.5910071134567261, | |
| "logits/rejected": -0.5533290505409241, | |
| "logps/chosen": -300.802734375, | |
| "logps/rejected": -264.6828918457031, | |
| "loss": 0.5557, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.033292002975940704, | |
| "rewards/margins": 0.5110405087471008, | |
| "rewards/rejected": -0.5443325042724609, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.26171159382360637, | |
| "grad_norm": 5.71875, | |
| "learning_rate": 4.613851860533367e-06, | |
| "logits/chosen": -0.6005420684814453, | |
| "logits/rejected": -0.5617684125900269, | |
| "logps/chosen": -294.0511169433594, | |
| "logps/rejected": -260.35443115234375, | |
| "loss": 0.5825, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.09464361518621445, | |
| "rewards/margins": 0.5028691291809082, | |
| "rewards/rejected": -0.40822547674179077, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.2721800575765506, | |
| "grad_norm": 5.6875, | |
| "learning_rate": 4.563632824908252e-06, | |
| "logits/chosen": -0.5696755647659302, | |
| "logits/rejected": -0.5352145433425903, | |
| "logps/chosen": -293.357666015625, | |
| "logps/rejected": -279.78173828125, | |
| "loss": 0.5358, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.2980680465698242, | |
| "rewards/margins": 0.81305992603302, | |
| "rewards/rejected": -0.5149918794631958, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.2826485213294949, | |
| "grad_norm": 5.59375, | |
| "learning_rate": 4.510653863290871e-06, | |
| "logits/chosen": -0.5663573741912842, | |
| "logits/rejected": -0.5443081259727478, | |
| "logps/chosen": -296.19744873046875, | |
| "logps/rejected": -303.76507568359375, | |
| "loss": 0.5498, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.06666271388530731, | |
| "rewards/margins": 0.5538051724433899, | |
| "rewards/rejected": -0.4871424734592438, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.29311698508243916, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 4.454985830346574e-06, | |
| "logits/chosen": -0.6322006583213806, | |
| "logits/rejected": -0.598243236541748, | |
| "logps/chosen": -301.9215087890625, | |
| "logps/rejected": -285.16412353515625, | |
| "loss": 0.5696, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.03948987275362015, | |
| "rewards/margins": 0.49146270751953125, | |
| "rewards/rejected": -0.4519728720188141, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.3035854488353834, | |
| "grad_norm": 4.65625, | |
| "learning_rate": 4.396703177135262e-06, | |
| "logits/chosen": -0.5908337831497192, | |
| "logits/rejected": -0.5626215934753418, | |
| "logps/chosen": -287.64849853515625, | |
| "logps/rejected": -258.74005126953125, | |
| "loss": 0.526, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.14304542541503906, | |
| "rewards/margins": 0.6406429409980774, | |
| "rewards/rejected": -0.49759751558303833, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.31405391258832765, | |
| "grad_norm": 5.75, | |
| "learning_rate": 4.335883851539693e-06, | |
| "logits/chosen": -0.6105285882949829, | |
| "logits/rejected": -0.5751320123672485, | |
| "logps/chosen": -297.8828125, | |
| "logps/rejected": -294.6984558105469, | |
| "loss": 0.5429, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.1768026053905487, | |
| "rewards/margins": 0.6462854743003845, | |
| "rewards/rejected": -0.8230881690979004, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.31405391258832765, | |
| "eval_logits/chosen": -0.4910103976726532, | |
| "eval_logits/rejected": -0.4688774645328522, | |
| "eval_logps/chosen": -294.2773132324219, | |
| "eval_logps/rejected": -285.1317138671875, | |
| "eval_loss": 0.5380240678787231, | |
| "eval_rewards/accuracies": 0.7260000109672546, | |
| "eval_rewards/chosen": -0.1730155199766159, | |
| "eval_rewards/margins": 0.6724972128868103, | |
| "eval_rewards/rejected": -0.8455128073692322, | |
| "eval_runtime": 349.3616, | |
| "eval_samples_per_second": 5.725, | |
| "eval_steps_per_second": 0.358, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3245223763412719, | |
| "grad_norm": 6.625, | |
| "learning_rate": 4.2726091940171055e-06, | |
| "logits/chosen": -0.5600444078445435, | |
| "logits/rejected": -0.5925548672676086, | |
| "logps/chosen": -296.61083984375, | |
| "logps/rejected": -342.74749755859375, | |
| "loss": 0.5077, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.023765895515680313, | |
| "rewards/margins": 0.7941875457763672, | |
| "rewards/rejected": -0.7704216241836548, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.33499084009421615, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 4.206963828813555e-06, | |
| "logits/chosen": -0.6118437647819519, | |
| "logits/rejected": -0.5762773752212524, | |
| "logps/chosen": -297.16278076171875, | |
| "logps/rejected": -280.4404602050781, | |
| "loss": 0.531, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.036954768002033234, | |
| "rewards/margins": 0.7416526079177856, | |
| "rewards/rejected": -0.7786073684692383, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.34545930384716045, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 4.139035550786495e-06, | |
| "logits/chosen": -0.6456430554389954, | |
| "logits/rejected": -0.5850483179092407, | |
| "logps/chosen": -290.33062744140625, | |
| "logps/rejected": -261.4778137207031, | |
| "loss": 0.5252, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.11032617092132568, | |
| "rewards/margins": 0.6117701530456543, | |
| "rewards/rejected": -0.7220962643623352, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.3559277676001047, | |
| "grad_norm": 6.1875, | |
| "learning_rate": 4.068915207986931e-06, | |
| "logits/chosen": -0.6017712354660034, | |
| "logits/rejected": -0.5405057072639465, | |
| "logps/chosen": -298.4487609863281, | |
| "logps/rejected": -258.83343505859375, | |
| "loss": 0.5244, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.27785810828208923, | |
| "rewards/margins": 0.7664127945899963, | |
| "rewards/rejected": -1.0442708730697632, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.36639623135304894, | |
| "grad_norm": 6.25, | |
| "learning_rate": 3.996696580158211e-06, | |
| "logits/chosen": -0.5467532873153687, | |
| "logits/rejected": -0.525019645690918, | |
| "logps/chosen": -337.39007568359375, | |
| "logps/rejected": -292.0143127441406, | |
| "loss": 0.5204, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.04514966905117035, | |
| "rewards/margins": 0.7053651809692383, | |
| "rewards/rejected": -0.7505148649215698, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.3768646951059932, | |
| "grad_norm": 5.03125, | |
| "learning_rate": 3.922476253313921e-06, | |
| "logits/chosen": -0.5269440412521362, | |
| "logits/rejected": -0.5347733497619629, | |
| "logps/chosen": -275.55755615234375, | |
| "logps/rejected": -298.5218811035156, | |
| "loss": 0.5549, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.23508167266845703, | |
| "rewards/margins": 0.6767452955245972, | |
| "rewards/rejected": -0.9118270874023438, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.38733315885893743, | |
| "grad_norm": 5.78125, | |
| "learning_rate": 3.846353490562664e-06, | |
| "logits/chosen": -0.5508509278297424, | |
| "logits/rejected": -0.5576980113983154, | |
| "logps/chosen": -290.8457946777344, | |
| "logps/rejected": -265.32647705078125, | |
| "loss": 0.508, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.06573277711868286, | |
| "rewards/margins": 0.8556879758834839, | |
| "rewards/rejected": -0.9214208722114563, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.39780162261188173, | |
| "grad_norm": 5.28125, | |
| "learning_rate": 3.768430099352445e-06, | |
| "logits/chosen": -0.5762262344360352, | |
| "logits/rejected": -0.5799992084503174, | |
| "logps/chosen": -306.72760009765625, | |
| "logps/rejected": -279.5985107421875, | |
| "loss": 0.5075, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.010313669219613075, | |
| "rewards/margins": 0.9469502568244934, | |
| "rewards/rejected": -0.9572639465332031, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.408270086364826, | |
| "grad_norm": 6.40625, | |
| "learning_rate": 3.6888102953122307e-06, | |
| "logits/chosen": -0.6278566122055054, | |
| "logits/rejected": -0.593431830406189, | |
| "logps/chosen": -262.35601806640625, | |
| "logps/rejected": -263.1413879394531, | |
| "loss": 0.5558, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.06627392768859863, | |
| "rewards/margins": 0.7061600089073181, | |
| "rewards/rejected": -0.639886200428009, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "grad_norm": 5.21875, | |
| "learning_rate": 3.607600562872785e-06, | |
| "logits/chosen": -0.5806395411491394, | |
| "logits/rejected": -0.5490812659263611, | |
| "logps/chosen": -284.69036865234375, | |
| "logps/rejected": -275.1679992675781, | |
| "loss": 0.5054, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.10525357723236084, | |
| "rewards/margins": 0.7453383803367615, | |
| "rewards/rejected": -0.6400848031044006, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "eval_logits/chosen": -0.45346179604530334, | |
| "eval_logits/rejected": -0.42607197165489197, | |
| "eval_logps/chosen": -293.4172668457031, | |
| "eval_logps/rejected": -285.14593505859375, | |
| "eval_loss": 0.5331673622131348, | |
| "eval_rewards/accuracies": 0.7239999771118164, | |
| "eval_rewards/chosen": -0.08701039105653763, | |
| "eval_rewards/margins": 0.7599251866340637, | |
| "eval_rewards/rejected": -0.8469354510307312, | |
| "eval_runtime": 349.2048, | |
| "eval_samples_per_second": 5.727, | |
| "eval_steps_per_second": 0.358, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.42920701387071447, | |
| "grad_norm": 5.1875, | |
| "learning_rate": 3.5249095128531863e-06, | |
| "logits/chosen": -0.6033052206039429, | |
| "logits/rejected": -0.5547356605529785, | |
| "logps/chosen": -279.104248046875, | |
| "logps/rejected": -276.4415588378906, | |
| "loss": 0.5154, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.09319617599248886, | |
| "rewards/margins": 0.780913770198822, | |
| "rewards/rejected": -0.8741100430488586, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.4396754776236587, | |
| "grad_norm": 5.9375, | |
| "learning_rate": 3.4408477372034743e-06, | |
| "logits/chosen": -0.5911046266555786, | |
| "logits/rejected": -0.5682761669158936, | |
| "logps/chosen": -309.4401550292969, | |
| "logps/rejected": -297.2247009277344, | |
| "loss": 0.5621, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.05225636810064316, | |
| "rewards/margins": 0.636865496635437, | |
| "rewards/rejected": -0.6891219615936279, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.45014394137660296, | |
| "grad_norm": 6.34375, | |
| "learning_rate": 3.355527661097728e-06, | |
| "logits/chosen": -0.5924620032310486, | |
| "logits/rejected": -0.595079243183136, | |
| "logps/chosen": -280.9834899902344, | |
| "logps/rejected": -281.77105712890625, | |
| "loss": 0.5226, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.19577598571777344, | |
| "rewards/margins": 0.608073353767395, | |
| "rewards/rejected": -0.8038493990898132, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.46061240512954726, | |
| "grad_norm": 5.5625, | |
| "learning_rate": 3.269063392575352e-06, | |
| "logits/chosen": -0.5576502680778503, | |
| "logits/rejected": -0.5509397983551025, | |
| "logps/chosen": -328.8077087402344, | |
| "logps/rejected": -306.9642333984375, | |
| "loss": 0.5141, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.0049751936458051205, | |
| "rewards/margins": 0.7285463213920593, | |
| "rewards/rejected": -0.733521580696106, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.4710808688824915, | |
| "grad_norm": 6.4375, | |
| "learning_rate": 3.181570569931697e-06, | |
| "logits/chosen": -0.5819513201713562, | |
| "logits/rejected": -0.5696187019348145, | |
| "logps/chosen": -287.6014709472656, | |
| "logps/rejected": -283.2913513183594, | |
| "loss": 0.5033, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.24654550850391388, | |
| "rewards/margins": 0.6834465265274048, | |
| "rewards/rejected": -0.9299919009208679, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.48154933263543576, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 3.09316620706208e-06, | |
| "logits/chosen": -0.5356208086013794, | |
| "logits/rejected": -0.5477502346038818, | |
| "logps/chosen": -308.5799560546875, | |
| "logps/rejected": -289.5464172363281, | |
| "loss": 0.493, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.07389375567436218, | |
| "rewards/margins": 0.9217998385429382, | |
| "rewards/rejected": -0.995693564414978, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.49201779638838, | |
| "grad_norm": 5.3125, | |
| "learning_rate": 3.0039685369660785e-06, | |
| "logits/chosen": -0.5380591154098511, | |
| "logits/rejected": -0.49667781591415405, | |
| "logps/chosen": -283.11297607421875, | |
| "logps/rejected": -267.4956359863281, | |
| "loss": 0.5366, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.035861968994140625, | |
| "rewards/margins": 0.8068122863769531, | |
| "rewards/rejected": -0.7709503173828125, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5024862601413242, | |
| "grad_norm": 5.375, | |
| "learning_rate": 2.91409685362137e-06, | |
| "logits/chosen": -0.5487444996833801, | |
| "logits/rejected": -0.5391730666160583, | |
| "logps/chosen": -280.1451110839844, | |
| "logps/rejected": -277.56793212890625, | |
| "loss": 0.5096, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.06115647405385971, | |
| "rewards/margins": 0.7818087339401245, | |
| "rewards/rejected": -0.8429651260375977, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.5129547238942685, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 2.8236713524386085e-06, | |
| "logits/chosen": -0.6103276014328003, | |
| "logits/rejected": -0.5744163393974304, | |
| "logps/chosen": -279.46942138671875, | |
| "logps/rejected": -257.06500244140625, | |
| "loss": 0.5051, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.14644916355609894, | |
| "rewards/margins": 0.8112751841545105, | |
| "rewards/rejected": -0.6648260354995728, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.5234231876472127, | |
| "grad_norm": 5.5625, | |
| "learning_rate": 2.7328129695107205e-06, | |
| "logits/chosen": -0.5192793607711792, | |
| "logits/rejected": -0.5326711535453796, | |
| "logps/chosen": -264.8558349609375, | |
| "logps/rejected": -274.16455078125, | |
| "loss": 0.5508, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.14748027920722961, | |
| "rewards/margins": 0.9845119714736938, | |
| "rewards/rejected": -0.8370317220687866, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5234231876472127, | |
| "eval_logits/chosen": -0.43347951769828796, | |
| "eval_logits/rejected": -0.4044938385486603, | |
| "eval_logps/chosen": -292.7540283203125, | |
| "eval_logps/rejected": -284.7646179199219, | |
| "eval_loss": 0.5267207622528076, | |
| "eval_rewards/accuracies": 0.7179999947547913, | |
| "eval_rewards/chosen": -0.02068748138844967, | |
| "eval_rewards/margins": 0.7881156206130981, | |
| "eval_rewards/rejected": -0.8088030219078064, | |
| "eval_runtime": 349.1516, | |
| "eval_samples_per_second": 5.728, | |
| "eval_steps_per_second": 0.358, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.533891651400157, | |
| "grad_norm": 4.65625, | |
| "learning_rate": 2.641643219871597e-06, | |
| "logits/chosen": -0.5523602962493896, | |
| "logits/rejected": -0.5135177969932556, | |
| "logps/chosen": -313.77227783203125, | |
| "logps/rejected": -297.81756591796875, | |
| "loss": 0.5053, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.082377590239048, | |
| "rewards/margins": 0.7289600968360901, | |
| "rewards/rejected": -0.8113376498222351, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.5443601151531012, | |
| "grad_norm": 6.84375, | |
| "learning_rate": 2.5502840349805074e-06, | |
| "logits/chosen": -0.5441479086875916, | |
| "logits/rejected": -0.5340668559074402, | |
| "logps/chosen": -309.962158203125, | |
| "logps/rejected": -297.2099304199219, | |
| "loss": 0.5476, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.005710450001060963, | |
| "rewards/margins": 0.8396749496459961, | |
| "rewards/rejected": -0.8339645266532898, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.5548285789060455, | |
| "grad_norm": 5.125, | |
| "learning_rate": 2.4588575996495797e-06, | |
| "logits/chosen": -0.4979037344455719, | |
| "logits/rejected": -0.5032162666320801, | |
| "logps/chosen": -272.52459716796875, | |
| "logps/rejected": -262.5303039550781, | |
| "loss": 0.5183, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.2069997489452362, | |
| "rewards/margins": 0.838523268699646, | |
| "rewards/rejected": -1.0455230474472046, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.5652970426589898, | |
| "grad_norm": 5.59375, | |
| "learning_rate": 2.367486188632446e-06, | |
| "logits/chosen": -0.5421893000602722, | |
| "logits/rejected": -0.539043128490448, | |
| "logps/chosen": -286.2185363769531, | |
| "logps/rejected": -325.55206298828125, | |
| "loss": 0.5085, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.05336964130401611, | |
| "rewards/margins": 0.8456710577011108, | |
| "rewards/rejected": -0.8990408182144165, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.575765506411934, | |
| "grad_norm": 4.34375, | |
| "learning_rate": 2.276292003092593e-06, | |
| "logits/chosen": -0.5617792010307312, | |
| "logits/rejected": -0.5426347851753235, | |
| "logps/chosen": -257.9263610839844, | |
| "logps/rejected": -265.53057861328125, | |
| "loss": 0.486, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.004675072617828846, | |
| "rewards/margins": 0.9126423597335815, | |
| "rewards/rejected": -0.9173175096511841, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5862339701648783, | |
| "grad_norm": 4.84375, | |
| "learning_rate": 2.1853970071701415e-06, | |
| "logits/chosen": -0.5527303218841553, | |
| "logits/rejected": -0.5213286876678467, | |
| "logps/chosen": -279.15924072265625, | |
| "logps/rejected": -280.146484375, | |
| "loss": 0.5179, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.01912788115441799, | |
| "rewards/margins": 0.8087967038154602, | |
| "rewards/rejected": -0.7896687388420105, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.5967024339178225, | |
| "grad_norm": 5.71875, | |
| "learning_rate": 2.0949227648656194e-06, | |
| "logits/chosen": -0.5781675577163696, | |
| "logits/rejected": -0.5574057102203369, | |
| "logps/chosen": -295.38726806640625, | |
| "logps/rejected": -262.20684814453125, | |
| "loss": 0.5412, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.023471500724554062, | |
| "rewards/margins": 0.8612861633300781, | |
| "rewards/rejected": -0.8847578167915344, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.6071708976707668, | |
| "grad_norm": 5.8125, | |
| "learning_rate": 2.00499027745888e-06, | |
| "logits/chosen": -0.5529965162277222, | |
| "logits/rejected": -0.5318363308906555, | |
| "logps/chosen": -299.06695556640625, | |
| "logps/rejected": -297.82098388671875, | |
| "loss": 0.5449, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.07610292732715607, | |
| "rewards/margins": 0.7717056274414062, | |
| "rewards/rejected": -0.6956026554107666, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.6176393614237111, | |
| "grad_norm": 5.625, | |
| "learning_rate": 1.915719821680624e-06, | |
| "logits/chosen": -0.5771445035934448, | |
| "logits/rejected": -0.5279287099838257, | |
| "logps/chosen": -286.1835021972656, | |
| "logps/rejected": -281.80828857421875, | |
| "loss": 0.5206, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": 0.37562283873558044, | |
| "rewards/margins": 0.8612510561943054, | |
| "rewards/rejected": -0.485628217458725, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "grad_norm": 5.75, | |
| "learning_rate": 1.8272307888529276e-06, | |
| "logits/chosen": -0.5050173997879028, | |
| "logits/rejected": -0.4642249643802643, | |
| "logps/chosen": -262.1390380859375, | |
| "logps/rejected": -278.6227111816406, | |
| "loss": 0.5338, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.2233264446258545, | |
| "rewards/margins": 0.8637221455574036, | |
| "rewards/rejected": -0.6403957009315491, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "eval_logits/chosen": -0.4303784966468811, | |
| "eval_logits/rejected": -0.40024831891059875, | |
| "eval_logps/chosen": -290.5658874511719, | |
| "eval_logps/rejected": -282.5771179199219, | |
| "eval_loss": 0.5262989401817322, | |
| "eval_rewards/accuracies": 0.7300000190734863, | |
| "eval_rewards/chosen": 0.19812585413455963, | |
| "eval_rewards/margins": 0.788177490234375, | |
| "eval_rewards/rejected": -0.5900516510009766, | |
| "eval_runtime": 349.0523, | |
| "eval_samples_per_second": 5.73, | |
| "eval_steps_per_second": 0.358, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6385762889295996, | |
| "grad_norm": 4.53125, | |
| "learning_rate": 1.739641525213929e-06, | |
| "logits/chosen": -0.5362275838851929, | |
| "logits/rejected": -0.5250522494316101, | |
| "logps/chosen": -264.38677978515625, | |
| "logps/rejected": -269.70172119140625, | |
| "loss": 0.4947, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.2012002021074295, | |
| "rewards/margins": 0.8771921396255493, | |
| "rewards/rejected": -0.675991952419281, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.6490447526825438, | |
| "grad_norm": 4.65625, | |
| "learning_rate": 1.6530691736402317e-06, | |
| "logits/chosen": -0.547565221786499, | |
| "logits/rejected": -0.5210016965866089, | |
| "logps/chosen": -292.267822265625, | |
| "logps/rejected": -281.25054931640625, | |
| "loss": 0.5076, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.06052354723215103, | |
| "rewards/margins": 0.8467457890510559, | |
| "rewards/rejected": -0.7862221598625183, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.6595132164354881, | |
| "grad_norm": 5.5625, | |
| "learning_rate": 1.5676295169786864e-06, | |
| "logits/chosen": -0.5603654980659485, | |
| "logits/rejected": -0.5221408605575562, | |
| "logps/chosen": -287.28192138671875, | |
| "logps/rejected": -272.3338317871094, | |
| "loss": 0.5266, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.08671960234642029, | |
| "rewards/margins": 0.8218274116516113, | |
| "rewards/rejected": -0.908547043800354, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.6699816801884323, | |
| "grad_norm": 5.40625, | |
| "learning_rate": 1.4834368231970922e-06, | |
| "logits/chosen": -0.5882564783096313, | |
| "logits/rejected": -0.5383986234664917, | |
| "logps/chosen": -288.31964111328125, | |
| "logps/rejected": -273.548583984375, | |
| "loss": 0.5013, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.17558792233467102, | |
| "rewards/margins": 0.7076612710952759, | |
| "rewards/rejected": -0.8832491040229797, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.6804501439413766, | |
| "grad_norm": 5.3125, | |
| "learning_rate": 1.4006036925609245e-06, | |
| "logits/chosen": -0.5553715825080872, | |
| "logits/rejected": -0.5215333104133606, | |
| "logps/chosen": -301.14434814453125, | |
| "logps/rejected": -250.81289672851562, | |
| "loss": 0.539, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.25257399678230286, | |
| "rewards/margins": 0.822566032409668, | |
| "rewards/rejected": -1.0751399993896484, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6909186076943209, | |
| "grad_norm": 5.1875, | |
| "learning_rate": 1.3192409070404582e-06, | |
| "logits/chosen": -0.5745340585708618, | |
| "logits/rejected": -0.551754355430603, | |
| "logps/chosen": -305.2149658203125, | |
| "logps/rejected": -307.12396240234375, | |
| "loss": 0.5097, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.0797838568687439, | |
| "rewards/margins": 0.7875758409500122, | |
| "rewards/rejected": -0.8673597574234009, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.7013870714472651, | |
| "grad_norm": 6.6875, | |
| "learning_rate": 1.2394572821496953e-06, | |
| "logits/chosen": -0.5750129818916321, | |
| "logits/rejected": -0.5432807207107544, | |
| "logps/chosen": -278.31634521484375, | |
| "logps/rejected": -259.3590087890625, | |
| "loss": 0.5169, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.12945926189422607, | |
| "rewards/margins": 0.7823888659477234, | |
| "rewards/rejected": -0.9118480682373047, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.7118555352002094, | |
| "grad_norm": 4.9375, | |
| "learning_rate": 1.1613595214152713e-06, | |
| "logits/chosen": -0.6039143800735474, | |
| "logits/rejected": -0.5614811182022095, | |
| "logps/chosen": -288.0231628417969, | |
| "logps/rejected": -276.5443420410156, | |
| "loss": 0.5014, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.20713257789611816, | |
| "rewards/margins": 0.7958000302314758, | |
| "rewards/rejected": -1.0029326677322388, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.7223239989531536, | |
| "grad_norm": 4.1875, | |
| "learning_rate": 1.0850520736699362e-06, | |
| "logits/chosen": -0.5435994863510132, | |
| "logits/rejected": -0.5199321508407593, | |
| "logps/chosen": -342.0343933105469, | |
| "logps/rejected": -316.83929443359375, | |
| "loss": 0.5367, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.19384647905826569, | |
| "rewards/margins": 0.8790246248245239, | |
| "rewards/rejected": -1.072871208190918, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.7327924627060979, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 1.0106369933615043e-06, | |
| "logits/chosen": -0.5901396870613098, | |
| "logits/rejected": -0.551986813545227, | |
| "logps/chosen": -316.48187255859375, | |
| "logps/rejected": -263.9020080566406, | |
| "loss": 0.5064, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.20547008514404297, | |
| "rewards/margins": 0.7208881378173828, | |
| "rewards/rejected": -0.9263582229614258, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7327924627060979, | |
| "eval_logits/chosen": -0.4080447256565094, | |
| "eval_logits/rejected": -0.37605100870132446, | |
| "eval_logps/chosen": -294.5545959472656, | |
| "eval_logps/rejected": -286.7521057128906, | |
| "eval_loss": 0.5174716114997864, | |
| "eval_rewards/accuracies": 0.7300000190734863, | |
| "eval_rewards/chosen": -0.20074398815631866, | |
| "eval_rewards/margins": 0.8068088889122009, | |
| "eval_rewards/rejected": -1.0075528621673584, | |
| "eval_runtime": 349.1622, | |
| "eval_samples_per_second": 5.728, | |
| "eval_steps_per_second": 0.358, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7432609264590422, | |
| "grad_norm": 6.03125, | |
| "learning_rate": 9.382138040640714e-07, | |
| "logits/chosen": -0.5988644361495972, | |
| "logits/rejected": -0.5565031170845032, | |
| "logps/chosen": -265.79888916015625, | |
| "logps/rejected": -279.4545593261719, | |
| "loss": 0.5573, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.14814189076423645, | |
| "rewards/margins": 0.8269087076187134, | |
| "rewards/rejected": -0.9750504493713379, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.7537293902119864, | |
| "grad_norm": 5.3125, | |
| "learning_rate": 8.678793653740633e-07, | |
| "logits/chosen": -0.5240329504013062, | |
| "logits/rejected": -0.522124171257019, | |
| "logps/chosen": -264.15338134765625, | |
| "logps/rejected": -263.60552978515625, | |
| "loss": 0.5183, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.0942729264497757, | |
| "rewards/margins": 0.8182271122932434, | |
| "rewards/rejected": -0.9125000834465027, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.7641978539649307, | |
| "grad_norm": 4.6875, | |
| "learning_rate": 7.997277433690984e-07, | |
| "logits/chosen": -0.5428770780563354, | |
| "logits/rejected": -0.4947339594364166, | |
| "logps/chosen": -302.0629577636719, | |
| "logps/rejected": -287.492431640625, | |
| "loss": 0.5134, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.12038590759038925, | |
| "rewards/margins": 0.7666383981704712, | |
| "rewards/rejected": -0.8870244026184082, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.7746663177178749, | |
| "grad_norm": 4.6875, | |
| "learning_rate": 7.338500848029603e-07, | |
| "logits/chosen": -0.49188748002052307, | |
| "logits/rejected": -0.5088191628456116, | |
| "logps/chosen": -292.02117919921875, | |
| "logps/rejected": -275.2965393066406, | |
| "loss": 0.4939, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.023575307801365852, | |
| "rewards/margins": 0.7873858213424683, | |
| "rewards/rejected": -0.7638105154037476, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.7851347814708192, | |
| "grad_norm": 5.0625, | |
| "learning_rate": 6.70334495204884e-07, | |
| "logits/chosen": -0.5374957919120789, | |
| "logits/rejected": -0.516038179397583, | |
| "logps/chosen": -324.4206848144531, | |
| "logps/rejected": -286.3302917480469, | |
| "loss": 0.5036, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.0536157488822937, | |
| "rewards/margins": 0.7997394800186157, | |
| "rewards/rejected": -0.746123731136322, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7956032452237635, | |
| "grad_norm": 5.34375, | |
| "learning_rate": 6.092659210462232e-07, | |
| "logits/chosen": -0.5625780820846558, | |
| "logits/rejected": -0.5423771142959595, | |
| "logps/chosen": -269.4684143066406, | |
| "logps/rejected": -268.89886474609375, | |
| "loss": 0.5496, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.04108331725001335, | |
| "rewards/margins": 0.7257132530212402, | |
| "rewards/rejected": -0.7667967081069946, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.8060717089767077, | |
| "grad_norm": 5.6875, | |
| "learning_rate": 5.507260361320738e-07, | |
| "logits/chosen": -0.5479773283004761, | |
| "logits/rejected": -0.5508543252944946, | |
| "logps/chosen": -285.4389953613281, | |
| "logps/rejected": -279.5918884277344, | |
| "loss": 0.5066, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.08319403976202011, | |
| "rewards/margins": 0.6797414422035217, | |
| "rewards/rejected": -0.5965473651885986, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.816540172729652, | |
| "grad_norm": 5.21875, | |
| "learning_rate": 4.947931323697983e-07, | |
| "logits/chosen": -0.5364648103713989, | |
| "logits/rejected": -0.5092785954475403, | |
| "logps/chosen": -286.8380432128906, | |
| "logps/rejected": -279.98956298828125, | |
| "loss": 0.522, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.04336369037628174, | |
| "rewards/margins": 0.6655139327049255, | |
| "rewards/rejected": -0.7088775634765625, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.8270086364825961, | |
| "grad_norm": 5.4375, | |
| "learning_rate": 4.4154201506053985e-07, | |
| "logits/chosen": -0.5611253976821899, | |
| "logits/rejected": -0.5496819019317627, | |
| "logps/chosen": -301.17742919921875, | |
| "logps/rejected": -264.5899658203125, | |
| "loss": 0.5154, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.043927911669015884, | |
| "rewards/margins": 0.7878991365432739, | |
| "rewards/rejected": -0.8318269848823547, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "grad_norm": 4.875, | |
| "learning_rate": 3.910439028537638e-07, | |
| "logits/chosen": -0.5692980289459229, | |
| "logits/rejected": -0.5227521657943726, | |
| "logps/chosen": -348.53631591796875, | |
| "logps/rejected": -302.92352294921875, | |
| "loss": 0.5349, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.09300284087657928, | |
| "rewards/margins": 0.6956671476364136, | |
| "rewards/rejected": -0.6026642322540283, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "eval_logits/chosen": -0.4160803556442261, | |
| "eval_logits/rejected": -0.3852657377719879, | |
| "eval_logps/chosen": -292.3984069824219, | |
| "eval_logps/rejected": -284.57269287109375, | |
| "eval_loss": 0.5197243690490723, | |
| "eval_rewards/accuracies": 0.7200000286102295, | |
| "eval_rewards/chosen": 0.014875907450914383, | |
| "eval_rewards/margins": 0.8044845461845398, | |
| "eval_rewards/rejected": -0.7896087169647217, | |
| "eval_runtime": 349.1752, | |
| "eval_samples_per_second": 5.728, | |
| "eval_steps_per_second": 0.358, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8479455639884846, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 3.4336633249862084e-07, | |
| "logits/chosen": -0.5947756767272949, | |
| "logits/rejected": -0.5246752500534058, | |
| "logps/chosen": -320.2247009277344, | |
| "logps/rejected": -289.26373291015625, | |
| "loss": 0.5012, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.009494239464402199, | |
| "rewards/margins": 0.8274568319320679, | |
| "rewards/rejected": -0.8179625272750854, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.8584140277414289, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 2.98573068519539e-07, | |
| "logits/chosen": -0.570965588092804, | |
| "logits/rejected": -0.5559743046760559, | |
| "logps/chosen": -308.2550048828125, | |
| "logps/rejected": -294.05462646484375, | |
| "loss": 0.5287, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.03936102241277695, | |
| "rewards/margins": 0.7610028982162476, | |
| "rewards/rejected": -0.7216418981552124, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.8688824914943732, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 2.5672401793681854e-07, | |
| "logits/chosen": -0.5910287499427795, | |
| "logits/rejected": -0.5737439393997192, | |
| "logps/chosen": -275.55743408203125, | |
| "logps/rejected": -270.0027770996094, | |
| "loss": 0.5101, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.0218903087079525, | |
| "rewards/margins": 0.816811203956604, | |
| "rewards/rejected": -0.8387015461921692, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.8793509552473174, | |
| "grad_norm": 4.9375, | |
| "learning_rate": 2.178751501463036e-07, | |
| "logits/chosen": -0.5559400916099548, | |
| "logits/rejected": -0.5404902696609497, | |
| "logps/chosen": -315.4130859375, | |
| "logps/rejected": -308.131591796875, | |
| "loss": 0.5091, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.04114828258752823, | |
| "rewards/margins": 0.7911642789840698, | |
| "rewards/rejected": -0.8323125839233398, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.8898194190002617, | |
| "grad_norm": 5.4375, | |
| "learning_rate": 1.820784220652766e-07, | |
| "logits/chosen": -0.6067087054252625, | |
| "logits/rejected": -0.5707641839981079, | |
| "logps/chosen": -347.02484130859375, | |
| "logps/rejected": -280.36297607421875, | |
| "loss": 0.5052, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.15417176485061646, | |
| "rewards/margins": 0.8600176572799683, | |
| "rewards/rejected": -0.7058458924293518, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9002878827532059, | |
| "grad_norm": 5.46875, | |
| "learning_rate": 1.4938170864468636e-07, | |
| "logits/chosen": -0.5292789936065674, | |
| "logits/rejected": -0.509742021560669, | |
| "logps/chosen": -290.58807373046875, | |
| "logps/rejected": -269.5320739746094, | |
| "loss": 0.4873, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.0017050117021426558, | |
| "rewards/margins": 0.8745778799057007, | |
| "rewards/rejected": -0.8728728294372559, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.9107563465061502, | |
| "grad_norm": 5.625, | |
| "learning_rate": 1.1982873884064466e-07, | |
| "logits/chosen": -0.49922800064086914, | |
| "logits/rejected": -0.4975685477256775, | |
| "logps/chosen": -288.2568054199219, | |
| "logps/rejected": -277.89642333984375, | |
| "loss": 0.5241, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.032976943999528885, | |
| "rewards/margins": 0.6596282720565796, | |
| "rewards/rejected": -0.6926053166389465, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.9212248102590945, | |
| "grad_norm": 5.78125, | |
| "learning_rate": 9.345903713082305e-08, | |
| "logits/chosen": -0.58247309923172, | |
| "logits/rejected": -0.574453592300415, | |
| "logps/chosen": -315.4455871582031, | |
| "logps/rejected": -281.5334167480469, | |
| "loss": 0.5391, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": 0.010342386551201344, | |
| "rewards/margins": 0.6320812702178955, | |
| "rewards/rejected": -0.621738851070404, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.9316932740120387, | |
| "grad_norm": 5.0625, | |
| "learning_rate": 7.030787065396866e-08, | |
| "logits/chosen": -0.5529004335403442, | |
| "logits/rejected": -0.5129944682121277, | |
| "logps/chosen": -319.9122314453125, | |
| "logps/rejected": -294.0019836425781, | |
| "loss": 0.5227, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.045671284198760986, | |
| "rewards/margins": 0.709865391254425, | |
| "rewards/rejected": -0.6641942262649536, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.942161737764983, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 5.0406202043228604e-08, | |
| "logits/chosen": -0.5645865201950073, | |
| "logits/rejected": -0.5604526400566101, | |
| "logps/chosen": -334.4140930175781, | |
| "logps/rejected": -275.6079406738281, | |
| "loss": 0.4775, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.1654743105173111, | |
| "rewards/margins": 0.9833230972290039, | |
| "rewards/rejected": -0.817848801612854, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.942161737764983, | |
| "eval_logits/chosen": -0.4151337742805481, | |
| "eval_logits/rejected": -0.3841632902622223, | |
| "eval_logps/chosen": -292.3968200683594, | |
| "eval_logps/rejected": -284.6649169921875, | |
| "eval_loss": 0.5181210041046143, | |
| "eval_rewards/accuracies": 0.7260000109672546, | |
| "eval_rewards/chosen": 0.01503300666809082, | |
| "eval_rewards/margins": 0.8138652443885803, | |
| "eval_rewards/rejected": -0.7988322973251343, | |
| "eval_runtime": 349.0965, | |
| "eval_samples_per_second": 5.729, | |
| "eval_steps_per_second": 0.358, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9526302015179272, | |
| "grad_norm": 5.0625, | |
| "learning_rate": 3.378064801637687e-08, | |
| "logits/chosen": -0.6055993437767029, | |
| "logits/rejected": -0.557610034942627, | |
| "logps/chosen": -315.5115966796875, | |
| "logps/rejected": -314.4410400390625, | |
| "loss": 0.5209, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.11909689754247665, | |
| "rewards/margins": 0.7703868746757507, | |
| "rewards/rejected": -0.6512900590896606, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.9630986652708715, | |
| "grad_norm": 4.6875, | |
| "learning_rate": 2.0453443778310766e-08, | |
| "logits/chosen": -0.5314048528671265, | |
| "logits/rejected": -0.49928751587867737, | |
| "logps/chosen": -330.26824951171875, | |
| "logps/rejected": -305.7253723144531, | |
| "loss": 0.533, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.07462009787559509, | |
| "rewards/margins": 0.8954026103019714, | |
| "rewards/rejected": -0.820782482624054, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.9735671290238157, | |
| "grad_norm": 4.78125, | |
| "learning_rate": 1.0442413283435759e-08, | |
| "logits/chosen": -0.5311998724937439, | |
| "logits/rejected": -0.49351367354393005, | |
| "logps/chosen": -318.3543701171875, | |
| "logps/rejected": -278.8912048339844, | |
| "loss": 0.514, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.13066162168979645, | |
| "rewards/margins": 1.1091994047164917, | |
| "rewards/rejected": -0.9785377383232117, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.98403559277676, | |
| "grad_norm": 4.28125, | |
| "learning_rate": 3.760945397705828e-09, | |
| "logits/chosen": -0.5610886812210083, | |
| "logits/rejected": -0.5083616971969604, | |
| "logps/chosen": -291.5441589355469, | |
| "logps/rejected": -261.2994079589844, | |
| "loss": 0.4851, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": 0.18183191120624542, | |
| "rewards/margins": 1.0001600980758667, | |
| "rewards/rejected": -0.8183282613754272, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.9945040565297043, | |
| "grad_norm": 4.875, | |
| "learning_rate": 4.1797599220405605e-10, | |
| "logits/chosen": -0.576126217842102, | |
| "logits/rejected": -0.5575291514396667, | |
| "logps/chosen": -295.9472961425781, | |
| "logps/rejected": -279.5827331542969, | |
| "loss": 0.5083, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.11896193027496338, | |
| "rewards/margins": 0.8570976257324219, | |
| "rewards/rejected": -0.7381356358528137, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.9997382884061764, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5431846207973221, | |
| "train_runtime": 19104.9346, | |
| "train_samples_per_second": 3.2, | |
| "train_steps_per_second": 0.05 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 955, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100000, | |
| "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 | |
| } | |