| { | |
| "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": 1.1945625860018705, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -2.5192830562591553, | |
| "logits/rejected": -2.3547825813293457, | |
| "logps/chosen": -297.60443115234375, | |
| "logps/rejected": -252.4619903564453, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": -0.00020415784092620015, | |
| "rewards/margins": -0.0002505290030967444, | |
| "rewards/rejected": 4.637122037820518e-05, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.010468463752944255, | |
| "grad_norm": 1.1009278086693854, | |
| "learning_rate": 5.208333333333334e-07, | |
| "logits/chosen": -2.2455766201019287, | |
| "logits/rejected": -2.215245008468628, | |
| "logps/chosen": -275.6755065917969, | |
| "logps/rejected": -254.76722717285156, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.6111111044883728, | |
| "rewards/chosen": 0.004448441788554192, | |
| "rewards/margins": 0.0008290203404612839, | |
| "rewards/rejected": 0.0036194208078086376, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02093692750588851, | |
| "grad_norm": 1.165704750760885, | |
| "learning_rate": 1.0416666666666667e-06, | |
| "logits/chosen": -2.2313215732574463, | |
| "logits/rejected": -2.114736795425415, | |
| "logps/chosen": -277.5883483886719, | |
| "logps/rejected": -255.2056427001953, | |
| "loss": 0.6907, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": 0.026263630017638206, | |
| "rewards/margins": 0.005699009168893099, | |
| "rewards/rejected": 0.020564619451761246, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.031405391258832765, | |
| "grad_norm": 1.1872789709669673, | |
| "learning_rate": 1.5625e-06, | |
| "logits/chosen": -2.3138914108276367, | |
| "logits/rejected": -2.2109274864196777, | |
| "logps/chosen": -281.3846740722656, | |
| "logps/rejected": -262.41693115234375, | |
| "loss": 0.6858, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.04216086491942406, | |
| "rewards/margins": 0.014992751181125641, | |
| "rewards/rejected": 0.027168119326233864, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04187385501177702, | |
| "grad_norm": 1.1753743538686479, | |
| "learning_rate": 2.0833333333333334e-06, | |
| "logits/chosen": -2.307976245880127, | |
| "logits/rejected": -2.2140889167785645, | |
| "logps/chosen": -268.4030456542969, | |
| "logps/rejected": -255.44882202148438, | |
| "loss": 0.6814, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.05265005677938461, | |
| "rewards/margins": 0.026613134890794754, | |
| "rewards/rejected": 0.026036927476525307, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.05234231876472128, | |
| "grad_norm": 1.1656515626034387, | |
| "learning_rate": 2.604166666666667e-06, | |
| "logits/chosen": -2.2787346839904785, | |
| "logits/rejected": -2.175128936767578, | |
| "logps/chosen": -227.7914581298828, | |
| "logps/rejected": -206.5706024169922, | |
| "loss": 0.6764, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.05774398520588875, | |
| "rewards/margins": 0.042526550590991974, | |
| "rewards/rejected": 0.015217426232993603, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06281078251766553, | |
| "grad_norm": 1.324103463890143, | |
| "learning_rate": 3.125e-06, | |
| "logits/chosen": -2.2864975929260254, | |
| "logits/rejected": -2.185832977294922, | |
| "logps/chosen": -264.6636657714844, | |
| "logps/rejected": -228.9823455810547, | |
| "loss": 0.6687, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.04733141511678696, | |
| "rewards/margins": 0.05985499545931816, | |
| "rewards/rejected": -0.012523581273853779, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07327924627060979, | |
| "grad_norm": 1.4908224965531012, | |
| "learning_rate": 3.6458333333333333e-06, | |
| "logits/chosen": -2.128446102142334, | |
| "logits/rejected": -2.080828905105591, | |
| "logps/chosen": -256.41363525390625, | |
| "logps/rejected": -262.66949462890625, | |
| "loss": 0.6557, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.0010238643735647202, | |
| "rewards/margins": 0.10270833969116211, | |
| "rewards/rejected": -0.10373219102621078, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.08374771002355404, | |
| "grad_norm": 2.5274509004519445, | |
| "learning_rate": 4.166666666666667e-06, | |
| "logits/chosen": -2.2609763145446777, | |
| "logits/rejected": -2.1039209365844727, | |
| "logps/chosen": -263.5312194824219, | |
| "logps/rejected": -256.4105529785156, | |
| "loss": 0.6402, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.12527017295360565, | |
| "rewards/margins": 0.11860889196395874, | |
| "rewards/rejected": -0.2438790500164032, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.0942161737764983, | |
| "grad_norm": 2.5638356680875156, | |
| "learning_rate": 4.6875000000000004e-06, | |
| "logits/chosen": -2.1489036083221436, | |
| "logits/rejected": -2.081789493560791, | |
| "logps/chosen": -270.69091796875, | |
| "logps/rejected": -291.17962646484375, | |
| "loss": 0.6304, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.1973474770784378, | |
| "rewards/margins": 0.19165988266468048, | |
| "rewards/rejected": -0.3890073299407959, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.10468463752944256, | |
| "grad_norm": 2.7331412184768507, | |
| "learning_rate": 4.9997324926814375e-06, | |
| "logits/chosen": -2.193147659301758, | |
| "logits/rejected": -2.0836679935455322, | |
| "logps/chosen": -303.36590576171875, | |
| "logps/rejected": -320.6068420410156, | |
| "loss": 0.6212, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.31098657846450806, | |
| "rewards/margins": 0.2229224145412445, | |
| "rewards/rejected": -0.5339089632034302, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.10468463752944256, | |
| "eval_logits/chosen": -2.0932729244232178, | |
| "eval_logits/rejected": -2.008643627166748, | |
| "eval_logps/chosen": -298.0505676269531, | |
| "eval_logps/rejected": -299.1472473144531, | |
| "eval_loss": 0.6321468353271484, | |
| "eval_rewards/accuracies": 0.6944444179534912, | |
| "eval_rewards/chosen": -0.3312907814979553, | |
| "eval_rewards/margins": 0.213734969496727, | |
| "eval_rewards/rejected": -0.5450257658958435, | |
| "eval_runtime": 321.7711, | |
| "eval_samples_per_second": 6.216, | |
| "eval_steps_per_second": 0.196, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11515310128238682, | |
| "grad_norm": 2.3956891782186904, | |
| "learning_rate": 4.996723692767927e-06, | |
| "logits/chosen": -2.1808857917785645, | |
| "logits/rejected": -2.05126690864563, | |
| "logps/chosen": -281.98193359375, | |
| "logps/rejected": -265.5942687988281, | |
| "loss": 0.6236, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.3388732373714447, | |
| "rewards/margins": 0.2357875108718872, | |
| "rewards/rejected": -0.5746607184410095, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12562156503533106, | |
| "grad_norm": 4.49315717191337, | |
| "learning_rate": 4.9903757462135984e-06, | |
| "logits/chosen": -2.1513657569885254, | |
| "logits/rejected": -2.0812501907348633, | |
| "logps/chosen": -287.4888610839844, | |
| "logps/rejected": -328.67156982421875, | |
| "loss": 0.5888, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.41489553451538086, | |
| "rewards/margins": 0.31543681025505066, | |
| "rewards/rejected": -0.7303323149681091, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.1360900287882753, | |
| "grad_norm": 3.426776723546068, | |
| "learning_rate": 4.980697142834315e-06, | |
| "logits/chosen": -2.1264045238494873, | |
| "logits/rejected": -2.0155797004699707, | |
| "logps/chosen": -381.717041015625, | |
| "logps/rejected": -350.6758728027344, | |
| "loss": 0.5946, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.8191909790039062, | |
| "rewards/margins": 0.3090600371360779, | |
| "rewards/rejected": -1.1282509565353394, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.14655849254121958, | |
| "grad_norm": 3.004217560772131, | |
| "learning_rate": 4.967700826904229e-06, | |
| "logits/chosen": -2.0693180561065674, | |
| "logits/rejected": -2.010124683380127, | |
| "logps/chosen": -363.1732482910156, | |
| "logps/rejected": -400.6826171875, | |
| "loss": 0.5746, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.9419809579849243, | |
| "rewards/margins": 0.41463392972946167, | |
| "rewards/rejected": -1.3566150665283203, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15702695629416383, | |
| "grad_norm": 4.5725215394059004, | |
| "learning_rate": 4.951404179843963e-06, | |
| "logits/chosen": -2.154370069503784, | |
| "logits/rejected": -2.016098976135254, | |
| "logps/chosen": -362.11627197265625, | |
| "logps/rejected": -358.12823486328125, | |
| "loss": 0.5888, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.7362005710601807, | |
| "rewards/margins": 0.44253450632095337, | |
| "rewards/rejected": -1.1787351369857788, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.16749542004710807, | |
| "grad_norm": 7.788576778528557, | |
| "learning_rate": 4.931828996974498e-06, | |
| "logits/chosen": -2.099804162979126, | |
| "logits/rejected": -1.9289453029632568, | |
| "logps/chosen": -348.2688903808594, | |
| "logps/rejected": -349.8648986816406, | |
| "loss": 0.5524, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.6180304288864136, | |
| "rewards/margins": 0.5966934561729431, | |
| "rewards/rejected": -1.2147239446640015, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.17796388380005235, | |
| "grad_norm": 3.73369227909531, | |
| "learning_rate": 4.909001458367867e-06, | |
| "logits/chosen": -2.005589008331299, | |
| "logits/rejected": -1.8591816425323486, | |
| "logps/chosen": -357.9122619628906, | |
| "logps/rejected": -401.5499267578125, | |
| "loss": 0.5578, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.9827691912651062, | |
| "rewards/margins": 0.5925935506820679, | |
| "rewards/rejected": -1.5753626823425293, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.1884323475529966, | |
| "grad_norm": 5.142786695761387, | |
| "learning_rate": 4.882952093833628e-06, | |
| "logits/chosen": -1.906904935836792, | |
| "logits/rejected": -1.8502197265625, | |
| "logps/chosen": -384.1532287597656, | |
| "logps/rejected": -441.2289123535156, | |
| "loss": 0.5357, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.2132524251937866, | |
| "rewards/margins": 0.6486458778381348, | |
| "rewards/rejected": -1.861898422241211, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.19890081130594087, | |
| "grad_norm": 4.970082596077688, | |
| "learning_rate": 4.853715742087947e-06, | |
| "logits/chosen": -1.7953016757965088, | |
| "logits/rejected": -1.749053716659546, | |
| "logps/chosen": -331.3048400878906, | |
| "logps/rejected": -420.5511779785156, | |
| "loss": 0.5344, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.8520873188972473, | |
| "rewards/margins": 0.6949248313903809, | |
| "rewards/rejected": -1.5470120906829834, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "grad_norm": 4.380709251136357, | |
| "learning_rate": 4.821331504159906e-06, | |
| "logits/chosen": -1.931880235671997, | |
| "logits/rejected": -1.7673060894012451, | |
| "logps/chosen": -395.6458435058594, | |
| "logps/rejected": -393.07818603515625, | |
| "loss": 0.5618, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.8311824798583984, | |
| "rewards/margins": 0.598331868648529, | |
| "rewards/rejected": -1.4295144081115723, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "eval_logits/chosen": -1.7550567388534546, | |
| "eval_logits/rejected": -1.669370174407959, | |
| "eval_logps/chosen": -346.90643310546875, | |
| "eval_logps/rejected": -381.2445983886719, | |
| "eval_loss": 0.5600804686546326, | |
| "eval_rewards/accuracies": 0.7222222089767456, | |
| "eval_rewards/chosen": -0.8198498487472534, | |
| "eval_rewards/margins": 0.5461496114730835, | |
| "eval_rewards/rejected": -1.365999460220337, | |
| "eval_runtime": 319.6036, | |
| "eval_samples_per_second": 6.258, | |
| "eval_steps_per_second": 0.197, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21983773881182936, | |
| "grad_norm": 4.390548860079351, | |
| "learning_rate": 4.7858426910973435e-06, | |
| "logits/chosen": -1.82965886592865, | |
| "logits/rejected": -1.7543909549713135, | |
| "logps/chosen": -382.2540588378906, | |
| "logps/rejected": -415.6708984375, | |
| "loss": 0.5746, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -1.0466556549072266, | |
| "rewards/margins": 0.42583417892456055, | |
| "rewards/rejected": -1.4724897146224976, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.23030620256477363, | |
| "grad_norm": 5.154433364870237, | |
| "learning_rate": 4.747296766042161e-06, | |
| "logits/chosen": -1.791577696800232, | |
| "logits/rejected": -1.676790475845337, | |
| "logps/chosen": -430.439697265625, | |
| "logps/rejected": -446.4627380371094, | |
| "loss": 0.5594, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.3720312118530273, | |
| "rewards/margins": 0.6089810132980347, | |
| "rewards/rejected": -1.9810121059417725, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.24077466631771788, | |
| "grad_norm": 5.067287557540918, | |
| "learning_rate": 4.705745280752586e-06, | |
| "logits/chosen": -1.6005672216415405, | |
| "logits/rejected": -1.5293024778366089, | |
| "logps/chosen": -364.32769775390625, | |
| "logps/rejected": -393.8982849121094, | |
| "loss": 0.5585, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.1689074039459229, | |
| "rewards/margins": 0.4595082402229309, | |
| "rewards/rejected": -1.6284157037734985, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.2512431300706621, | |
| "grad_norm": 4.187057753178847, | |
| "learning_rate": 4.661243806657256e-06, | |
| "logits/chosen": -1.6649366617202759, | |
| "logits/rejected": -1.6055676937103271, | |
| "logps/chosen": -377.30517578125, | |
| "logps/rejected": -407.87017822265625, | |
| "loss": 0.5306, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.2085318565368652, | |
| "rewards/margins": 0.6119499802589417, | |
| "rewards/rejected": -1.8204820156097412, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.26171159382360637, | |
| "grad_norm": 4.31385003289797, | |
| "learning_rate": 4.613851860533367e-06, | |
| "logits/chosen": -1.6592628955841064, | |
| "logits/rejected": -1.5831575393676758, | |
| "logps/chosen": -392.666259765625, | |
| "logps/rejected": -415.00933837890625, | |
| "loss": 0.5341, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.2613078355789185, | |
| "rewards/margins": 0.6352987289428711, | |
| "rewards/rejected": -1.896606683731079, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.2721800575765506, | |
| "grad_norm": 6.08274430115293, | |
| "learning_rate": 4.563632824908252e-06, | |
| "logits/chosen": -1.5671354532241821, | |
| "logits/rejected": -1.4325566291809082, | |
| "logps/chosen": -446.5708923339844, | |
| "logps/rejected": -492.8990173339844, | |
| "loss": 0.5228, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.7604526281356812, | |
| "rewards/margins": 0.6336302161216736, | |
| "rewards/rejected": -2.394083261489868, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.2826485213294949, | |
| "grad_norm": 4.56724778147284, | |
| "learning_rate": 4.510653863290871e-06, | |
| "logits/chosen": -1.456993579864502, | |
| "logits/rejected": -1.3645284175872803, | |
| "logps/chosen": -401.00244140625, | |
| "logps/rejected": -461.67816162109375, | |
| "loss": 0.5277, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.2706773281097412, | |
| "rewards/margins": 0.8170326352119446, | |
| "rewards/rejected": -2.087709903717041, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.29311698508243916, | |
| "grad_norm": 3.947817632488332, | |
| "learning_rate": 4.454985830346574e-06, | |
| "logits/chosen": -1.5088273286819458, | |
| "logits/rejected": -1.3900407552719116, | |
| "logps/chosen": -385.47772216796875, | |
| "logps/rejected": -426.82049560546875, | |
| "loss": 0.5485, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.9744614362716675, | |
| "rewards/margins": 0.75079345703125, | |
| "rewards/rejected": -1.725255012512207, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.3035854488353834, | |
| "grad_norm": 4.730568661724723, | |
| "learning_rate": 4.396703177135262e-06, | |
| "logits/chosen": -1.517740249633789, | |
| "logits/rejected": -1.351360559463501, | |
| "logps/chosen": -388.90789794921875, | |
| "logps/rejected": -416.22052001953125, | |
| "loss": 0.526, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.0588579177856445, | |
| "rewards/margins": 0.6205722093582153, | |
| "rewards/rejected": -1.6794300079345703, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.31405391258832765, | |
| "grad_norm": 5.829501633867066, | |
| "learning_rate": 4.335883851539693e-06, | |
| "logits/chosen": -1.2372510433197021, | |
| "logits/rejected": -1.2240632772445679, | |
| "logps/chosen": -386.8778381347656, | |
| "logps/rejected": -448.6904296875, | |
| "loss": 0.54, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.441900610923767, | |
| "rewards/margins": 0.6276523470878601, | |
| "rewards/rejected": -2.0695528984069824, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.31405391258832765, | |
| "eval_logits/chosen": -1.1714633703231812, | |
| "eval_logits/rejected": -1.0704221725463867, | |
| "eval_logps/chosen": -417.12750244140625, | |
| "eval_logps/rejected": -478.07476806640625, | |
| "eval_loss": 0.5264545679092407, | |
| "eval_rewards/accuracies": 0.7460317611694336, | |
| "eval_rewards/chosen": -1.5220601558685303, | |
| "eval_rewards/margins": 0.8122406601905823, | |
| "eval_rewards/rejected": -2.334300994873047, | |
| "eval_runtime": 318.2186, | |
| "eval_samples_per_second": 6.285, | |
| "eval_steps_per_second": 0.198, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3245223763412719, | |
| "grad_norm": 4.0412766328088585, | |
| "learning_rate": 4.2726091940171055e-06, | |
| "logits/chosen": -1.3419673442840576, | |
| "logits/rejected": -1.1779625415802002, | |
| "logps/chosen": -421.1036682128906, | |
| "logps/rejected": -466.26275634765625, | |
| "loss": 0.5369, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.4869836568832397, | |
| "rewards/margins": 0.8415184020996094, | |
| "rewards/rejected": -2.3285021781921387, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.33499084009421615, | |
| "grad_norm": 7.228642390991299, | |
| "learning_rate": 4.206963828813555e-06, | |
| "logits/chosen": -1.3606576919555664, | |
| "logits/rejected": -1.266966700553894, | |
| "logps/chosen": -374.7814025878906, | |
| "logps/rejected": -450.57489013671875, | |
| "loss": 0.5399, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.3771132230758667, | |
| "rewards/margins": 0.7092502117156982, | |
| "rewards/rejected": -2.0863633155822754, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.34545930384716045, | |
| "grad_norm": 5.998674043872773, | |
| "learning_rate": 4.139035550786495e-06, | |
| "logits/chosen": -1.305854082107544, | |
| "logits/rejected": -1.2263256311416626, | |
| "logps/chosen": -429.99188232421875, | |
| "logps/rejected": -497.3521423339844, | |
| "loss": 0.5206, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.8139232397079468, | |
| "rewards/margins": 0.7226920127868652, | |
| "rewards/rejected": -2.5366153717041016, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.3559277676001047, | |
| "grad_norm": 6.244527338235992, | |
| "learning_rate": 4.068915207986931e-06, | |
| "logits/chosen": -1.0698177814483643, | |
| "logits/rejected": -0.9774864315986633, | |
| "logps/chosen": -438.493408203125, | |
| "logps/rejected": -487.9413146972656, | |
| "loss": 0.5553, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.8784738779067993, | |
| "rewards/margins": 0.7664145231246948, | |
| "rewards/rejected": -2.644888401031494, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.36639623135304894, | |
| "grad_norm": 6.05372732646654, | |
| "learning_rate": 3.996696580158211e-06, | |
| "logits/chosen": -1.419528603553772, | |
| "logits/rejected": -1.3026950359344482, | |
| "logps/chosen": -399.3412170410156, | |
| "logps/rejected": -459.59063720703125, | |
| "loss": 0.5144, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.4473296403884888, | |
| "rewards/margins": 0.7528419494628906, | |
| "rewards/rejected": -2.2001712322235107, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.3768646951059932, | |
| "grad_norm": 7.477190042107064, | |
| "learning_rate": 3.922476253313921e-06, | |
| "logits/chosen": -1.3192278146743774, | |
| "logits/rejected": -1.2596288919448853, | |
| "logps/chosen": -388.9505920410156, | |
| "logps/rejected": -450.23748779296875, | |
| "loss": 0.476, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.3493094444274902, | |
| "rewards/margins": 0.7934570908546448, | |
| "rewards/rejected": -2.1427664756774902, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.38733315885893743, | |
| "grad_norm": 5.411077344670459, | |
| "learning_rate": 3.846353490562664e-06, | |
| "logits/chosen": -1.3181861639022827, | |
| "logits/rejected": -1.2644492387771606, | |
| "logps/chosen": -406.28851318359375, | |
| "logps/rejected": -484.81646728515625, | |
| "loss": 0.5189, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.618009328842163, | |
| "rewards/margins": 0.912137508392334, | |
| "rewards/rejected": -2.530146837234497, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.39780162261188173, | |
| "grad_norm": 6.255769900471538, | |
| "learning_rate": 3.768430099352445e-06, | |
| "logits/chosen": -1.2576202154159546, | |
| "logits/rejected": -1.1985712051391602, | |
| "logps/chosen": -426.02978515625, | |
| "logps/rejected": -499.4483337402344, | |
| "loss": 0.5291, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.724285364151001, | |
| "rewards/margins": 0.7168464660644531, | |
| "rewards/rejected": -2.441131830215454, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.408270086364826, | |
| "grad_norm": 5.374401251386735, | |
| "learning_rate": 3.6888102953122307e-06, | |
| "logits/chosen": -1.3139002323150635, | |
| "logits/rejected": -1.2482521533966064, | |
| "logps/chosen": -366.7402648925781, | |
| "logps/rejected": -402.67144775390625, | |
| "loss": 0.5445, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.0828481912612915, | |
| "rewards/margins": 0.5530051589012146, | |
| "rewards/rejected": -1.6358531713485718, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "grad_norm": 5.700868192112112, | |
| "learning_rate": 3.607600562872785e-06, | |
| "logits/chosen": -1.2806731462478638, | |
| "logits/rejected": -1.2076570987701416, | |
| "logps/chosen": -397.6805725097656, | |
| "logps/rejected": -457.99578857421875, | |
| "loss": 0.5261, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.534794807434082, | |
| "rewards/margins": 0.6010990738868713, | |
| "rewards/rejected": -2.1358938217163086, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "eval_logits/chosen": -1.201329231262207, | |
| "eval_logits/rejected": -1.1013988256454468, | |
| "eval_logps/chosen": -430.45263671875, | |
| "eval_logps/rejected": -497.2759094238281, | |
| "eval_loss": 0.5082111954689026, | |
| "eval_rewards/accuracies": 0.7539682388305664, | |
| "eval_rewards/chosen": -1.6553115844726562, | |
| "eval_rewards/margins": 0.8710008859634399, | |
| "eval_rewards/rejected": -2.5263123512268066, | |
| "eval_runtime": 314.7811, | |
| "eval_samples_per_second": 6.354, | |
| "eval_steps_per_second": 0.2, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.42920701387071447, | |
| "grad_norm": 4.78368697455832, | |
| "learning_rate": 3.5249095128531863e-06, | |
| "logits/chosen": -1.1789991855621338, | |
| "logits/rejected": -1.0596911907196045, | |
| "logps/chosen": -449.33709716796875, | |
| "logps/rejected": -515.829833984375, | |
| "loss": 0.5275, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.8276984691619873, | |
| "rewards/margins": 0.9437017440795898, | |
| "rewards/rejected": -2.7714004516601562, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.4396754776236587, | |
| "grad_norm": 5.718886468906387, | |
| "learning_rate": 3.4408477372034743e-06, | |
| "logits/chosen": -1.3089560270309448, | |
| "logits/rejected": -1.111132025718689, | |
| "logps/chosen": -451.94354248046875, | |
| "logps/rejected": -472.0818786621094, | |
| "loss": 0.5212, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.6895692348480225, | |
| "rewards/margins": 0.77433842420578, | |
| "rewards/rejected": -2.4639077186584473, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.45014394137660296, | |
| "grad_norm": 5.4739752222845, | |
| "learning_rate": 3.355527661097728e-06, | |
| "logits/chosen": -1.2457327842712402, | |
| "logits/rejected": -1.1701027154922485, | |
| "logps/chosen": -408.38629150390625, | |
| "logps/rejected": -495.34906005859375, | |
| "loss": 0.5228, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.8043445348739624, | |
| "rewards/margins": 0.8327828645706177, | |
| "rewards/rejected": -2.63712739944458, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.46061240512954726, | |
| "grad_norm": 6.85581768096026, | |
| "learning_rate": 3.269063392575352e-06, | |
| "logits/chosen": -1.2453696727752686, | |
| "logits/rejected": -1.202413558959961, | |
| "logps/chosen": -399.9681091308594, | |
| "logps/rejected": -447.87744140625, | |
| "loss": 0.5256, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.5209221839904785, | |
| "rewards/margins": 0.6652868986129761, | |
| "rewards/rejected": -2.186208963394165, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.4710808688824915, | |
| "grad_norm": 5.776991001419176, | |
| "learning_rate": 3.181570569931697e-06, | |
| "logits/chosen": -1.3612130880355835, | |
| "logits/rejected": -1.2961633205413818, | |
| "logps/chosen": -399.5633239746094, | |
| "logps/rejected": -510.95513916015625, | |
| "loss": 0.5105, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.5496407747268677, | |
| "rewards/margins": 0.9577849507331848, | |
| "rewards/rejected": -2.507425308227539, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.48154933263543576, | |
| "grad_norm": 5.051312773178446, | |
| "learning_rate": 3.09316620706208e-06, | |
| "logits/chosen": -1.2932283878326416, | |
| "logits/rejected": -1.2328197956085205, | |
| "logps/chosen": -472.6172790527344, | |
| "logps/rejected": -538.5615234375, | |
| "loss": 0.5286, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.122169256210327, | |
| "rewards/margins": 0.7884066104888916, | |
| "rewards/rejected": -2.9105758666992188, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.49201779638838, | |
| "grad_norm": 4.905312405134056, | |
| "learning_rate": 3.0039685369660785e-06, | |
| "logits/chosen": -1.299459457397461, | |
| "logits/rejected": -1.164813756942749, | |
| "logps/chosen": -433.37554931640625, | |
| "logps/rejected": -470.384521484375, | |
| "loss": 0.4942, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.7430375814437866, | |
| "rewards/margins": 0.6865721940994263, | |
| "rewards/rejected": -2.429609775543213, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5024862601413242, | |
| "grad_norm": 6.146340380005326, | |
| "learning_rate": 2.91409685362137e-06, | |
| "logits/chosen": -1.2403475046157837, | |
| "logits/rejected": -1.0839545726776123, | |
| "logps/chosen": -513.665771484375, | |
| "logps/rejected": -583.8973388671875, | |
| "loss": 0.5046, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.4166836738586426, | |
| "rewards/margins": 1.000232219696045, | |
| "rewards/rejected": -3.4169158935546875, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.5129547238942685, | |
| "grad_norm": 5.112445065707855, | |
| "learning_rate": 2.8236713524386085e-06, | |
| "logits/chosen": -1.2714743614196777, | |
| "logits/rejected": -1.0941110849380493, | |
| "logps/chosen": -532.9962158203125, | |
| "logps/rejected": -588.236328125, | |
| "loss": 0.5025, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.491058826446533, | |
| "rewards/margins": 0.900057315826416, | |
| "rewards/rejected": -3.39111590385437, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.5234231876472127, | |
| "grad_norm": 5.578978194196055, | |
| "learning_rate": 2.7328129695107205e-06, | |
| "logits/chosen": -1.2111141681671143, | |
| "logits/rejected": -1.1049137115478516, | |
| "logps/chosen": -476.11932373046875, | |
| "logps/rejected": -544.0899658203125, | |
| "loss": 0.5107, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.1048712730407715, | |
| "rewards/margins": 0.9492176175117493, | |
| "rewards/rejected": -3.054089069366455, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5234231876472127, | |
| "eval_logits/chosen": -1.0955979824066162, | |
| "eval_logits/rejected": -0.9851866364479065, | |
| "eval_logps/chosen": -509.9847717285156, | |
| "eval_logps/rejected": -587.1475830078125, | |
| "eval_loss": 0.5058528184890747, | |
| "eval_rewards/accuracies": 0.75, | |
| "eval_rewards/chosen": -2.4506328105926514, | |
| "eval_rewards/margins": 0.9743964076042175, | |
| "eval_rewards/rejected": -3.4250295162200928, | |
| "eval_runtime": 306.1826, | |
| "eval_samples_per_second": 6.532, | |
| "eval_steps_per_second": 0.206, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.533891651400157, | |
| "grad_norm": 5.171650655838442, | |
| "learning_rate": 2.641643219871597e-06, | |
| "logits/chosen": -1.1813862323760986, | |
| "logits/rejected": -1.0965768098831177, | |
| "logps/chosen": -508.13092041015625, | |
| "logps/rejected": -570.53759765625, | |
| "loss": 0.4794, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.4878056049346924, | |
| "rewards/margins": 0.8628519773483276, | |
| "rewards/rejected": -3.3506579399108887, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.5443601151531012, | |
| "grad_norm": 6.785478355965952, | |
| "learning_rate": 2.5502840349805074e-06, | |
| "logits/chosen": -1.1492969989776611, | |
| "logits/rejected": -1.0285675525665283, | |
| "logps/chosen": -577.2451171875, | |
| "logps/rejected": -618.67236328125, | |
| "loss": 0.536, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -3.018739938735962, | |
| "rewards/margins": 0.8701656460762024, | |
| "rewards/rejected": -3.8889052867889404, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.5548285789060455, | |
| "grad_norm": 5.9061132192029975, | |
| "learning_rate": 2.4588575996495797e-06, | |
| "logits/chosen": -1.1598259210586548, | |
| "logits/rejected": -1.0490505695343018, | |
| "logps/chosen": -515.8734130859375, | |
| "logps/rejected": -585.2592163085938, | |
| "loss": 0.5089, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.4302353858947754, | |
| "rewards/margins": 0.9659484624862671, | |
| "rewards/rejected": -3.396183729171753, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.5652970426589898, | |
| "grad_norm": 4.6033180245939604, | |
| "learning_rate": 2.367486188632446e-06, | |
| "logits/chosen": -1.1840062141418457, | |
| "logits/rejected": -1.05405592918396, | |
| "logps/chosen": -469.9341735839844, | |
| "logps/rejected": -503.66156005859375, | |
| "loss": 0.5311, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.1539134979248047, | |
| "rewards/margins": 0.7668038606643677, | |
| "rewards/rejected": -2.920717239379883, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.575765506411934, | |
| "grad_norm": 4.876993529338895, | |
| "learning_rate": 2.276292003092593e-06, | |
| "logits/chosen": -1.1908996105194092, | |
| "logits/rejected": -1.03009033203125, | |
| "logps/chosen": -489.2470703125, | |
| "logps/rejected": -519.9906616210938, | |
| "loss": 0.5265, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -2.272523880004883, | |
| "rewards/margins": 0.6718829870223999, | |
| "rewards/rejected": -2.9444069862365723, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5862339701648783, | |
| "grad_norm": 6.4140589895092734, | |
| "learning_rate": 2.1853970071701415e-06, | |
| "logits/chosen": -1.0713722705841064, | |
| "logits/rejected": -0.9895181655883789, | |
| "logps/chosen": -517.431396484375, | |
| "logps/rejected": -573.0247802734375, | |
| "loss": 0.527, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.533311605453491, | |
| "rewards/margins": 0.8953849077224731, | |
| "rewards/rejected": -3.428696393966675, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.5967024339178225, | |
| "grad_norm": 6.376011574902463, | |
| "learning_rate": 2.0949227648656194e-06, | |
| "logits/chosen": -1.129098653793335, | |
| "logits/rejected": -1.0245158672332764, | |
| "logps/chosen": -559.3487548828125, | |
| "logps/rejected": -636.18310546875, | |
| "loss": 0.5053, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -2.986375331878662, | |
| "rewards/margins": 0.8951998949050903, | |
| "rewards/rejected": -3.8815758228302, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.6071708976707668, | |
| "grad_norm": 5.245103814265102, | |
| "learning_rate": 2.00499027745888e-06, | |
| "logits/chosen": -1.081526517868042, | |
| "logits/rejected": -0.9703742861747742, | |
| "logps/chosen": -521.400146484375, | |
| "logps/rejected": -581.4735107421875, | |
| "loss": 0.5167, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.8668112754821777, | |
| "rewards/margins": 0.8452316522598267, | |
| "rewards/rejected": -3.712043046951294, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.6176393614237111, | |
| "grad_norm": 5.7527775937649, | |
| "learning_rate": 1.915719821680624e-06, | |
| "logits/chosen": -1.287595510482788, | |
| "logits/rejected": -1.2315866947174072, | |
| "logps/chosen": -470.63995361328125, | |
| "logps/rejected": -567.0520629882812, | |
| "loss": 0.5018, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.246957778930664, | |
| "rewards/margins": 0.9269870519638062, | |
| "rewards/rejected": -3.1739444732666016, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "grad_norm": 4.552980884850473, | |
| "learning_rate": 1.8272307888529276e-06, | |
| "logits/chosen": -1.1632342338562012, | |
| "logits/rejected": -1.0386791229248047, | |
| "logps/chosen": -524.4891357421875, | |
| "logps/rejected": -568.3989868164062, | |
| "loss": 0.4851, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.2760519981384277, | |
| "rewards/margins": 0.7800448536872864, | |
| "rewards/rejected": -3.0560970306396484, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "eval_logits/chosen": -1.107803463935852, | |
| "eval_logits/rejected": -0.9969872832298279, | |
| "eval_logps/chosen": -492.1783447265625, | |
| "eval_logps/rejected": -567.8048706054688, | |
| "eval_loss": 0.5023476481437683, | |
| "eval_rewards/accuracies": 0.7678571343421936, | |
| "eval_rewards/chosen": -2.272569179534912, | |
| "eval_rewards/margins": 0.9590328931808472, | |
| "eval_rewards/rejected": -3.231602191925049, | |
| "eval_runtime": 309.4155, | |
| "eval_samples_per_second": 6.464, | |
| "eval_steps_per_second": 0.204, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6385762889295996, | |
| "grad_norm": 5.93434046945835, | |
| "learning_rate": 1.739641525213929e-06, | |
| "logits/chosen": -1.1899484395980835, | |
| "logits/rejected": -1.128395438194275, | |
| "logps/chosen": -505.1640625, | |
| "logps/rejected": -569.9411010742188, | |
| "loss": 0.5131, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.3301897048950195, | |
| "rewards/margins": 1.009691834449768, | |
| "rewards/rejected": -3.339881420135498, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.6490447526825438, | |
| "grad_norm": 6.7592991694860345, | |
| "learning_rate": 1.6530691736402317e-06, | |
| "logits/chosen": -1.3041086196899414, | |
| "logits/rejected": -1.1299916505813599, | |
| "logps/chosen": -520.1345825195312, | |
| "logps/rejected": -570.2188110351562, | |
| "loss": 0.4665, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -2.1966605186462402, | |
| "rewards/margins": 1.0572277307510376, | |
| "rewards/rejected": -3.253887891769409, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.6595132164354881, | |
| "grad_norm": 6.939788728853941, | |
| "learning_rate": 1.5676295169786864e-06, | |
| "logits/chosen": -1.2350585460662842, | |
| "logits/rejected": -1.1085925102233887, | |
| "logps/chosen": -506.601806640625, | |
| "logps/rejected": -558.4306640625, | |
| "loss": 0.4883, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.2913689613342285, | |
| "rewards/margins": 0.9455882906913757, | |
| "rewards/rejected": -3.23695707321167, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.6699816801884323, | |
| "grad_norm": 6.885905876648702, | |
| "learning_rate": 1.4834368231970922e-06, | |
| "logits/chosen": -1.1120684146881104, | |
| "logits/rejected": -0.9865466952323914, | |
| "logps/chosen": -484.1168518066406, | |
| "logps/rejected": -577.2811279296875, | |
| "loss": 0.4707, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -2.489487648010254, | |
| "rewards/margins": 1.0948190689086914, | |
| "rewards/rejected": -3.5843067169189453, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.6804501439413766, | |
| "grad_norm": 5.355709320338061, | |
| "learning_rate": 1.4006036925609245e-06, | |
| "logits/chosen": -1.2093435525894165, | |
| "logits/rejected": -1.1428296566009521, | |
| "logps/chosen": -484.78680419921875, | |
| "logps/rejected": -564.5701293945312, | |
| "loss": 0.5169, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.2595038414001465, | |
| "rewards/margins": 0.8043657541275024, | |
| "rewards/rejected": -3.0638692378997803, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6909186076943209, | |
| "grad_norm": 5.122013631042651, | |
| "learning_rate": 1.3192409070404582e-06, | |
| "logits/chosen": -1.2926921844482422, | |
| "logits/rejected": -1.1912837028503418, | |
| "logps/chosen": -462.49700927734375, | |
| "logps/rejected": -535.156005859375, | |
| "loss": 0.4902, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.105215311050415, | |
| "rewards/margins": 0.9164140820503235, | |
| "rewards/rejected": -3.0216293334960938, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.7013870714472651, | |
| "grad_norm": 7.049199430649973, | |
| "learning_rate": 1.2394572821496953e-06, | |
| "logits/chosen": -1.144698977470398, | |
| "logits/rejected": -0.9912746548652649, | |
| "logps/chosen": -518.4602661132812, | |
| "logps/rejected": -603.5948486328125, | |
| "loss": 0.4773, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.6358141899108887, | |
| "rewards/margins": 1.113600492477417, | |
| "rewards/rejected": -3.7494144439697266, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.7118555352002094, | |
| "grad_norm": 7.264843779332727, | |
| "learning_rate": 1.1613595214152713e-06, | |
| "logits/chosen": -1.1309657096862793, | |
| "logits/rejected": -1.0589998960494995, | |
| "logps/chosen": -539.7699584960938, | |
| "logps/rejected": -620.6878662109375, | |
| "loss": 0.5266, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.733842134475708, | |
| "rewards/margins": 0.9637987017631531, | |
| "rewards/rejected": -3.6976406574249268, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.7223239989531536, | |
| "grad_norm": 5.030210764921467, | |
| "learning_rate": 1.0850520736699362e-06, | |
| "logits/chosen": -1.288641333580017, | |
| "logits/rejected": -1.134615182876587, | |
| "logps/chosen": -510.186767578125, | |
| "logps/rejected": -550.937744140625, | |
| "loss": 0.4574, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.3800137042999268, | |
| "rewards/margins": 0.9033814668655396, | |
| "rewards/rejected": -3.2833950519561768, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.7327924627060979, | |
| "grad_norm": 5.9328337595528495, | |
| "learning_rate": 1.0106369933615043e-06, | |
| "logits/chosen": -1.1679879426956177, | |
| "logits/rejected": -1.1002038717269897, | |
| "logps/chosen": -487.10821533203125, | |
| "logps/rejected": -576.1431884765625, | |
| "loss": 0.4681, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.3741798400878906, | |
| "rewards/margins": 0.9901224970817566, | |
| "rewards/rejected": -3.364302158355713, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7327924627060979, | |
| "eval_logits/chosen": -1.119031548500061, | |
| "eval_logits/rejected": -1.006774663925171, | |
| "eval_logps/chosen": -496.6231994628906, | |
| "eval_logps/rejected": -581.5197143554688, | |
| "eval_loss": 0.49932044744491577, | |
| "eval_rewards/accuracies": 0.7678571343421936, | |
| "eval_rewards/chosen": -2.3170175552368164, | |
| "eval_rewards/margins": 1.0517328977584839, | |
| "eval_rewards/rejected": -3.3687500953674316, | |
| "eval_runtime": 280.6333, | |
| "eval_samples_per_second": 7.127, | |
| "eval_steps_per_second": 0.224, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7432609264590422, | |
| "grad_norm": 5.3428016079168215, | |
| "learning_rate": 9.382138040640714e-07, | |
| "logits/chosen": -1.1798994541168213, | |
| "logits/rejected": -1.021723985671997, | |
| "logps/chosen": -505.2616271972656, | |
| "logps/rejected": -571.1856689453125, | |
| "loss": 0.547, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.396669864654541, | |
| "rewards/margins": 0.9654728174209595, | |
| "rewards/rejected": -3.362142562866211, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.7537293902119864, | |
| "grad_norm": 5.0438516064442505, | |
| "learning_rate": 8.678793653740633e-07, | |
| "logits/chosen": -1.3297260999679565, | |
| "logits/rejected": -1.1736423969268799, | |
| "logps/chosen": -527.0916748046875, | |
| "logps/rejected": -579.3074340820312, | |
| "loss": 0.478, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.070934534072876, | |
| "rewards/margins": 1.0328184366226196, | |
| "rewards/rejected": -3.103752851486206, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.7641978539649307, | |
| "grad_norm": 5.464567536353577, | |
| "learning_rate": 7.997277433690984e-07, | |
| "logits/chosen": -1.2094228267669678, | |
| "logits/rejected": -1.076755404472351, | |
| "logps/chosen": -460.71563720703125, | |
| "logps/rejected": -538.8247680664062, | |
| "loss": 0.4966, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.2227933406829834, | |
| "rewards/margins": 0.9530885815620422, | |
| "rewards/rejected": -3.175881862640381, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.7746663177178749, | |
| "grad_norm": 5.377248875033102, | |
| "learning_rate": 7.338500848029603e-07, | |
| "logits/chosen": -1.1969387531280518, | |
| "logits/rejected": -1.0555990934371948, | |
| "logps/chosen": -529.8873291015625, | |
| "logps/rejected": -568.1295166015625, | |
| "loss": 0.5065, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.4915201663970947, | |
| "rewards/margins": 0.9319826364517212, | |
| "rewards/rejected": -3.4235024452209473, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.7851347814708192, | |
| "grad_norm": 5.342695362281337, | |
| "learning_rate": 6.70334495204884e-07, | |
| "logits/chosen": -1.0425455570220947, | |
| "logits/rejected": -0.9723536372184753, | |
| "logps/chosen": -487.095947265625, | |
| "logps/rejected": -599.2386474609375, | |
| "loss": 0.4857, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -2.589409112930298, | |
| "rewards/margins": 1.040583848953247, | |
| "rewards/rejected": -3.629992723464966, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7956032452237635, | |
| "grad_norm": 4.96165517698307, | |
| "learning_rate": 6.092659210462232e-07, | |
| "logits/chosen": -1.172639012336731, | |
| "logits/rejected": -1.0221275091171265, | |
| "logps/chosen": -510.41253662109375, | |
| "logps/rejected": -564.8956909179688, | |
| "loss": 0.5034, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.53062105178833, | |
| "rewards/margins": 0.8949772119522095, | |
| "rewards/rejected": -3.42559814453125, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.8060717089767077, | |
| "grad_norm": 4.707914305263311, | |
| "learning_rate": 5.507260361320738e-07, | |
| "logits/chosen": -1.2693157196044922, | |
| "logits/rejected": -1.2365710735321045, | |
| "logps/chosen": -522.3062744140625, | |
| "logps/rejected": -619.4640502929688, | |
| "loss": 0.4736, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.363852024078369, | |
| "rewards/margins": 0.9652940034866333, | |
| "rewards/rejected": -3.329145908355713, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.816540172729652, | |
| "grad_norm": 5.592342234404946, | |
| "learning_rate": 4.947931323697983e-07, | |
| "logits/chosen": -1.2886607646942139, | |
| "logits/rejected": -1.0684127807617188, | |
| "logps/chosen": -534.323486328125, | |
| "logps/rejected": -556.2030639648438, | |
| "loss": 0.5036, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.277695894241333, | |
| "rewards/margins": 0.8380621075630188, | |
| "rewards/rejected": -3.115757942199707, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.8270086364825961, | |
| "grad_norm": 6.076661367759154, | |
| "learning_rate": 4.4154201506053985e-07, | |
| "logits/chosen": -1.1675662994384766, | |
| "logits/rejected": -1.065953254699707, | |
| "logps/chosen": -489.9781188964844, | |
| "logps/rejected": -578.2179565429688, | |
| "loss": 0.5128, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.482016086578369, | |
| "rewards/margins": 0.9305012822151184, | |
| "rewards/rejected": -3.4125168323516846, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "grad_norm": 6.131830839970953, | |
| "learning_rate": 3.910439028537638e-07, | |
| "logits/chosen": -1.1610690355300903, | |
| "logits/rejected": -1.1072094440460205, | |
| "logps/chosen": -469.524658203125, | |
| "logps/rejected": -585.119140625, | |
| "loss": 0.4852, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.3968334197998047, | |
| "rewards/margins": 1.120810866355896, | |
| "rewards/rejected": -3.5176444053649902, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "eval_logits/chosen": -1.1353023052215576, | |
| "eval_logits/rejected": -1.0236940383911133, | |
| "eval_logps/chosen": -504.6183166503906, | |
| "eval_logps/rejected": -585.8155517578125, | |
| "eval_loss": 0.49497368931770325, | |
| "eval_rewards/accuracies": 0.773809552192688, | |
| "eval_rewards/chosen": -2.396967887878418, | |
| "eval_rewards/margins": 1.0147408246994019, | |
| "eval_rewards/rejected": -3.4117088317871094, | |
| "eval_runtime": 274.6399, | |
| "eval_samples_per_second": 7.282, | |
| "eval_steps_per_second": 0.229, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8479455639884846, | |
| "grad_norm": 4.892138077626307, | |
| "learning_rate": 3.4336633249862084e-07, | |
| "logits/chosen": -1.209530234336853, | |
| "logits/rejected": -1.0349524021148682, | |
| "logps/chosen": -532.2420043945312, | |
| "logps/rejected": -591.3436279296875, | |
| "loss": 0.5065, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.4228711128234863, | |
| "rewards/margins": 0.9818238019943237, | |
| "rewards/rejected": -3.4046947956085205, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.8584140277414289, | |
| "grad_norm": 6.133736149907814, | |
| "learning_rate": 2.98573068519539e-07, | |
| "logits/chosen": -1.2474277019500732, | |
| "logits/rejected": -1.1728675365447998, | |
| "logps/chosen": -516.3637084960938, | |
| "logps/rejected": -589.2227783203125, | |
| "loss": 0.5066, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.5452017784118652, | |
| "rewards/margins": 0.8416748046875, | |
| "rewards/rejected": -3.3868765830993652, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.8688824914943732, | |
| "grad_norm": 4.448941268784857, | |
| "learning_rate": 2.5672401793681854e-07, | |
| "logits/chosen": -1.1796165704727173, | |
| "logits/rejected": -1.1301778554916382, | |
| "logps/chosen": -493.81610107421875, | |
| "logps/rejected": -576.2391357421875, | |
| "loss": 0.5119, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -2.471508502960205, | |
| "rewards/margins": 0.8166677355766296, | |
| "rewards/rejected": -3.2881767749786377, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.8793509552473174, | |
| "grad_norm": 6.561312283813159, | |
| "learning_rate": 2.178751501463036e-07, | |
| "logits/chosen": -1.2200143337249756, | |
| "logits/rejected": -1.120086431503296, | |
| "logps/chosen": -525.2439575195312, | |
| "logps/rejected": -629.3662719726562, | |
| "loss": 0.4823, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.501779317855835, | |
| "rewards/margins": 1.0803827047348022, | |
| "rewards/rejected": -3.5821621417999268, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.8898194190002617, | |
| "grad_norm": 5.675008200582371, | |
| "learning_rate": 1.820784220652766e-07, | |
| "logits/chosen": -1.148503065109253, | |
| "logits/rejected": -0.9931659698486328, | |
| "logps/chosen": -511.488037109375, | |
| "logps/rejected": -576.7986450195312, | |
| "loss": 0.4672, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -2.456411361694336, | |
| "rewards/margins": 0.9856443405151367, | |
| "rewards/rejected": -3.4420554637908936, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9002878827532059, | |
| "grad_norm": 5.847430828911881, | |
| "learning_rate": 1.4938170864468636e-07, | |
| "logits/chosen": -1.2107369899749756, | |
| "logits/rejected": -1.0609266757965088, | |
| "logps/chosen": -531.3970947265625, | |
| "logps/rejected": -590.2992553710938, | |
| "loss": 0.4693, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.60589861869812, | |
| "rewards/margins": 0.8416641354560852, | |
| "rewards/rejected": -3.4475624561309814, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.9107563465061502, | |
| "grad_norm": 5.678184676582534, | |
| "learning_rate": 1.1982873884064466e-07, | |
| "logits/chosen": -1.24862539768219, | |
| "logits/rejected": -1.0584386587142944, | |
| "logps/chosen": -544.3020629882812, | |
| "logps/rejected": -598.3751220703125, | |
| "loss": 0.4947, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.632056713104248, | |
| "rewards/margins": 1.0834977626800537, | |
| "rewards/rejected": -3.7155539989471436, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.9212248102590945, | |
| "grad_norm": 5.3183871343635944, | |
| "learning_rate": 9.345903713082305e-08, | |
| "logits/chosen": -1.2423183917999268, | |
| "logits/rejected": -1.1057523488998413, | |
| "logps/chosen": -514.6922607421875, | |
| "logps/rejected": -574.3685302734375, | |
| "loss": 0.4801, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.4431517124176025, | |
| "rewards/margins": 0.9935353994369507, | |
| "rewards/rejected": -3.4366869926452637, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.9316932740120387, | |
| "grad_norm": 5.735081148749753, | |
| "learning_rate": 7.030787065396866e-08, | |
| "logits/chosen": -1.1413437128067017, | |
| "logits/rejected": -1.0328372716903687, | |
| "logps/chosen": -508.3773498535156, | |
| "logps/rejected": -619.8019409179688, | |
| "loss": 0.4938, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.5983500480651855, | |
| "rewards/margins": 1.2036244869232178, | |
| "rewards/rejected": -3.8019745349884033, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.942161737764983, | |
| "grad_norm": 6.618746821621782, | |
| "learning_rate": 5.0406202043228604e-08, | |
| "logits/chosen": -1.045906901359558, | |
| "logits/rejected": -0.9974561929702759, | |
| "logps/chosen": -517.8667602539062, | |
| "logps/rejected": -669.0172729492188, | |
| "loss": 0.4907, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.5165352821350098, | |
| "rewards/margins": 1.2452183961868286, | |
| "rewards/rejected": -3.761753559112549, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.942161737764983, | |
| "eval_logits/chosen": -1.1023893356323242, | |
| "eval_logits/rejected": -0.9901031255722046, | |
| "eval_logps/chosen": -521.706298828125, | |
| "eval_logps/rejected": -608.1346435546875, | |
| "eval_loss": 0.494513601064682, | |
| "eval_rewards/accuracies": 0.7777777910232544, | |
| "eval_rewards/chosen": -2.567847490310669, | |
| "eval_rewards/margins": 1.0670523643493652, | |
| "eval_rewards/rejected": -3.634899854660034, | |
| "eval_runtime": 302.7434, | |
| "eval_samples_per_second": 6.606, | |
| "eval_steps_per_second": 0.208, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9526302015179272, | |
| "grad_norm": 6.10762031606348, | |
| "learning_rate": 3.378064801637687e-08, | |
| "logits/chosen": -1.1716662645339966, | |
| "logits/rejected": -0.9894771575927734, | |
| "logps/chosen": -494.2669982910156, | |
| "logps/rejected": -552.8900146484375, | |
| "loss": 0.4935, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.474764585494995, | |
| "rewards/margins": 1.0439238548278809, | |
| "rewards/rejected": -3.518688201904297, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.9630986652708715, | |
| "grad_norm": 5.683369015946174, | |
| "learning_rate": 2.0453443778310766e-08, | |
| "logits/chosen": -1.1771481037139893, | |
| "logits/rejected": -1.0182334184646606, | |
| "logps/chosen": -545.4282836914062, | |
| "logps/rejected": -589.4589233398438, | |
| "loss": 0.5081, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.6610922813415527, | |
| "rewards/margins": 0.9007646441459656, | |
| "rewards/rejected": -3.561856746673584, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.9735671290238157, | |
| "grad_norm": 5.950595947719059, | |
| "learning_rate": 1.0442413283435759e-08, | |
| "logits/chosen": -1.1201808452606201, | |
| "logits/rejected": -0.9143557548522949, | |
| "logps/chosen": -556.3572998046875, | |
| "logps/rejected": -596.0765380859375, | |
| "loss": 0.493, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.5759315490722656, | |
| "rewards/margins": 1.0089161396026611, | |
| "rewards/rejected": -3.5848472118377686, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.98403559277676, | |
| "grad_norm": 7.5374296673981425, | |
| "learning_rate": 3.760945397705828e-09, | |
| "logits/chosen": -1.1323813199996948, | |
| "logits/rejected": -1.0745770931243896, | |
| "logps/chosen": -517.08056640625, | |
| "logps/rejected": -605.00146484375, | |
| "loss": 0.4829, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.5434603691101074, | |
| "rewards/margins": 0.8878037333488464, | |
| "rewards/rejected": -3.4312641620635986, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.9945040565297043, | |
| "grad_norm": 5.8745190839702515, | |
| "learning_rate": 4.1797599220405605e-10, | |
| "logits/chosen": -1.1730302572250366, | |
| "logits/rejected": -0.9989528656005859, | |
| "logps/chosen": -524.8547973632812, | |
| "logps/rejected": -583.33203125, | |
| "loss": 0.4861, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.5831165313720703, | |
| "rewards/margins": 0.908293604850769, | |
| "rewards/rejected": -3.49141001701355, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.9997382884061764, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5319095570379527, | |
| "train_runtime": 23762.0752, | |
| "train_samples_per_second": 2.573, | |
| "train_steps_per_second": 0.04 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 955, | |
| "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": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |