| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9983510322538092, | |
| "eval_steps": 400, | |
| "global_step": 473, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.002110678715124332, | |
| "grad_norm": 5.946747818020539, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": -0.680223822593689, | |
| "logits/rejected": -1.0431472063064575, | |
| "logps/chosen": -237.51611328125, | |
| "logps/rejected": -224.62709045410156, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.010553393575621661, | |
| "grad_norm": 4.788472074755126, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -0.605172336101532, | |
| "logits/rejected": -1.07023286819458, | |
| "logps/chosen": -302.76666259765625, | |
| "logps/rejected": -281.7729187011719, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.3515625, | |
| "rewards/chosen": 0.00022320327116176486, | |
| "rewards/margins": 8.367415284737945e-05, | |
| "rewards/rejected": 0.00013952911831438541, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.021106787151243322, | |
| "grad_norm": 4.238337049927753, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -0.6779512763023376, | |
| "logits/rejected": -1.0506097078323364, | |
| "logps/chosen": -270.19256591796875, | |
| "logps/rejected": -256.5780029296875, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": -7.003974315011874e-05, | |
| "rewards/margins": -0.0007576555944979191, | |
| "rewards/rejected": 0.0006876158877275884, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03166018072686498, | |
| "grad_norm": 5.613762220129112, | |
| "learning_rate": 1.5624999999999999e-07, | |
| "logits/chosen": -0.6236697435379028, | |
| "logits/rejected": -1.136690378189087, | |
| "logps/chosen": -294.53656005859375, | |
| "logps/rejected": -257.5579528808594, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.0019345388282090425, | |
| "rewards/margins": 0.0002388877037446946, | |
| "rewards/rejected": 0.0016956512117758393, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.042213574302486644, | |
| "grad_norm": 5.5506034027854625, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -0.5656238794326782, | |
| "logits/rejected": -1.1458613872528076, | |
| "logps/chosen": -277.18695068359375, | |
| "logps/rejected": -242.61972045898438, | |
| "loss": 0.6923, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.004845106042921543, | |
| "rewards/margins": 0.001295976573601365, | |
| "rewards/rejected": 0.0035491292364895344, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.052766967878108306, | |
| "grad_norm": 5.6701208184749, | |
| "learning_rate": 2.604166666666667e-07, | |
| "logits/chosen": -0.5958523750305176, | |
| "logits/rejected": -1.1272190809249878, | |
| "logps/chosen": -305.53717041015625, | |
| "logps/rejected": -268.25347900390625, | |
| "loss": 0.691, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.015507316216826439, | |
| "rewards/margins": 0.0047125560231506824, | |
| "rewards/rejected": 0.010794761590659618, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.06332036145372996, | |
| "grad_norm": 6.300590779192755, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -0.594957709312439, | |
| "logits/rejected": -1.086895227432251, | |
| "logps/chosen": -308.94012451171875, | |
| "logps/rejected": -276.35650634765625, | |
| "loss": 0.6879, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.03275737911462784, | |
| "rewards/margins": 0.01083751954138279, | |
| "rewards/rejected": 0.02191985584795475, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.07387375502935163, | |
| "grad_norm": 5.279197636228696, | |
| "learning_rate": 3.645833333333333e-07, | |
| "logits/chosen": -0.566003680229187, | |
| "logits/rejected": -1.1707605123519897, | |
| "logps/chosen": -305.409912109375, | |
| "logps/rejected": -279.29486083984375, | |
| "loss": 0.6838, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.05798542499542236, | |
| "rewards/margins": 0.023088786751031876, | |
| "rewards/rejected": 0.03489663824439049, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.08442714860497329, | |
| "grad_norm": 4.841427121898649, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -0.6702648401260376, | |
| "logits/rejected": -1.063316822052002, | |
| "logps/chosen": -295.71905517578125, | |
| "logps/rejected": -275.6202087402344, | |
| "loss": 0.681, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.06462486833333969, | |
| "rewards/margins": 0.02073287032544613, | |
| "rewards/rejected": 0.04389199614524841, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.09498054218059494, | |
| "grad_norm": 4.337699418611637, | |
| "learning_rate": 4.6874999999999996e-07, | |
| "logits/chosen": -0.7736238241195679, | |
| "logits/rejected": -1.1706254482269287, | |
| "logps/chosen": -272.80596923828125, | |
| "logps/rejected": -252.0467987060547, | |
| "loss": 0.6729, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.06801488995552063, | |
| "rewards/margins": 0.04081468656659126, | |
| "rewards/rejected": 0.027200212702155113, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.10553393575621661, | |
| "grad_norm": 5.035982549882043, | |
| "learning_rate": 4.999726797933858e-07, | |
| "logits/chosen": -0.8857172131538391, | |
| "logits/rejected": -1.2423771619796753, | |
| "logps/chosen": -284.9183044433594, | |
| "logps/rejected": -266.7237854003906, | |
| "loss": 0.6673, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.03259050473570824, | |
| "rewards/margins": 0.05623316764831543, | |
| "rewards/rejected": -0.023642662912607193, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.11608732933183827, | |
| "grad_norm": 5.2802868788136035, | |
| "learning_rate": 4.99665396039775e-07, | |
| "logits/chosen": -0.8473213315010071, | |
| "logits/rejected": -1.336387276649475, | |
| "logps/chosen": -276.6699523925781, | |
| "logps/rejected": -261.4330749511719, | |
| "loss": 0.6554, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.028915267437696457, | |
| "rewards/margins": 0.08701522648334503, | |
| "rewards/rejected": -0.11593051254749298, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.12664072290745992, | |
| "grad_norm": 9.944712985206676, | |
| "learning_rate": 4.99017099386437e-07, | |
| "logits/chosen": -0.911908745765686, | |
| "logits/rejected": -1.4581682682037354, | |
| "logps/chosen": -287.9392395019531, | |
| "logps/rejected": -270.0166931152344, | |
| "loss": 0.6552, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.10472609847784042, | |
| "rewards/margins": 0.09941879659891129, | |
| "rewards/rejected": -0.20414488017559052, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.13719411648308158, | |
| "grad_norm": 5.431017253590544, | |
| "learning_rate": 4.980286753286194e-07, | |
| "logits/chosen": -0.8732226490974426, | |
| "logits/rejected": -1.378962755203247, | |
| "logps/chosen": -294.68267822265625, | |
| "logps/rejected": -273.043701171875, | |
| "loss": 0.6509, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.15889093279838562, | |
| "rewards/margins": 0.0964367687702179, | |
| "rewards/rejected": -0.2553277015686035, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.14774751005870326, | |
| "grad_norm": 6.0993507114780305, | |
| "learning_rate": 4.967014739346915e-07, | |
| "logits/chosen": -0.9360530972480774, | |
| "logits/rejected": -1.3933957815170288, | |
| "logps/chosen": -306.38653564453125, | |
| "logps/rejected": -282.32769775390625, | |
| "loss": 0.6403, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.14856548607349396, | |
| "rewards/margins": 0.1044367104768753, | |
| "rewards/rejected": -0.25300222635269165, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.15830090363432492, | |
| "grad_norm": 6.805381088591199, | |
| "learning_rate": 4.950373080021136e-07, | |
| "logits/chosen": -0.9195225834846497, | |
| "logits/rejected": -1.3844249248504639, | |
| "logps/chosen": -316.6197509765625, | |
| "logps/rejected": -294.3293762207031, | |
| "loss": 0.6296, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.15873293578624725, | |
| "rewards/margins": 0.13153603672981262, | |
| "rewards/rejected": -0.2902689576148987, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.16885429720994657, | |
| "grad_norm": 11.435330521176741, | |
| "learning_rate": 4.930384505813737e-07, | |
| "logits/chosen": -1.0555646419525146, | |
| "logits/rejected": -1.482975721359253, | |
| "logps/chosen": -320.57818603515625, | |
| "logps/rejected": -304.3819885253906, | |
| "loss": 0.6345, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.2796502113342285, | |
| "rewards/margins": 0.15378037095069885, | |
| "rewards/rejected": -0.43343058228492737, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.17940769078556823, | |
| "grad_norm": 10.893293221473195, | |
| "learning_rate": 4.907076318712738e-07, | |
| "logits/chosen": -0.9902318120002747, | |
| "logits/rejected": -1.4965012073516846, | |
| "logps/chosen": -321.46295166015625, | |
| "logps/rejected": -311.6328125, | |
| "loss": 0.6289, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -0.3334459364414215, | |
| "rewards/margins": 0.1901421993970871, | |
| "rewards/rejected": -0.5235881209373474, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.18996108436118989, | |
| "grad_norm": 7.423510047109125, | |
| "learning_rate": 4.88048035489807e-07, | |
| "logits/chosen": -0.9904667139053345, | |
| "logits/rejected": -1.5588419437408447, | |
| "logps/chosen": -308.77227783203125, | |
| "logps/rejected": -290.52130126953125, | |
| "loss": 0.6187, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.33037030696868896, | |
| "rewards/margins": 0.15141813457012177, | |
| "rewards/rejected": -0.4817884564399719, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20051447793681157, | |
| "grad_norm": 10.503540684222791, | |
| "learning_rate": 4.85063294125718e-07, | |
| "logits/chosen": -1.0931203365325928, | |
| "logits/rejected": -1.59712815284729, | |
| "logps/chosen": -327.8336486816406, | |
| "logps/rejected": -313.16302490234375, | |
| "loss": 0.616, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.4475630819797516, | |
| "rewards/margins": 0.1690167635679245, | |
| "rewards/rejected": -0.6165798306465149, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.21106787151243322, | |
| "grad_norm": 7.941678777423352, | |
| "learning_rate": 4.817574845766874e-07, | |
| "logits/chosen": -1.1208244562149048, | |
| "logits/rejected": -1.5354535579681396, | |
| "logps/chosen": -351.09759521484375, | |
| "logps/rejected": -341.6310729980469, | |
| "loss": 0.6132, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.5545604825019836, | |
| "rewards/margins": 0.21660009026527405, | |
| "rewards/rejected": -0.7711606025695801, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.22162126508805488, | |
| "grad_norm": 7.975671452180151, | |
| "learning_rate": 4.781351221809166e-07, | |
| "logits/chosen": -1.0387537479400635, | |
| "logits/rejected": -1.6272369623184204, | |
| "logps/chosen": -360.13323974609375, | |
| "logps/rejected": -349.0940856933594, | |
| "loss": 0.6115, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.5049761533737183, | |
| "rewards/margins": 0.21938514709472656, | |
| "rewards/rejected": -0.7243613004684448, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.23217465866367654, | |
| "grad_norm": 8.034746512723116, | |
| "learning_rate": 4.742011546497182e-07, | |
| "logits/chosen": -1.1847660541534424, | |
| "logits/rejected": -1.6106970310211182, | |
| "logps/chosen": -334.15869140625, | |
| "logps/rejected": -328.2769470214844, | |
| "loss": 0.6055, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.4516691267490387, | |
| "rewards/margins": 0.2418258637189865, | |
| "rewards/rejected": -0.693494975566864, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.2427280522392982, | |
| "grad_norm": 10.287751316470317, | |
| "learning_rate": 4.6996095530953875e-07, | |
| "logits/chosen": -1.339599370956421, | |
| "logits/rejected": -1.7150366306304932, | |
| "logps/chosen": -331.191162109375, | |
| "logps/rejected": -330.8086853027344, | |
| "loss": 0.5971, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.5352717638015747, | |
| "rewards/margins": 0.21640977263450623, | |
| "rewards/rejected": -0.7516815662384033, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.25328144581491985, | |
| "grad_norm": 12.105502619722833, | |
| "learning_rate": 4.654203157626399e-07, | |
| "logits/chosen": -1.1615684032440186, | |
| "logits/rejected": -1.5826199054718018, | |
| "logps/chosen": -356.81915283203125, | |
| "logps/rejected": -361.2477722167969, | |
| "loss": 0.5989, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.7497292757034302, | |
| "rewards/margins": 0.2648738920688629, | |
| "rewards/rejected": -1.0146030187606812, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.26383483939054153, | |
| "grad_norm": 11.641374719855692, | |
| "learning_rate": 4.605854379764673e-07, | |
| "logits/chosen": -1.228808045387268, | |
| "logits/rejected": -1.633094072341919, | |
| "logps/chosen": -376.2483825683594, | |
| "logps/rejected": -372.6545104980469, | |
| "loss": 0.5901, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.8822706937789917, | |
| "rewards/margins": 0.2663251757621765, | |
| "rewards/rejected": -1.1485958099365234, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.27438823296616316, | |
| "grad_norm": 18.11761811423972, | |
| "learning_rate": 4.5546292581250857e-07, | |
| "logits/chosen": -1.2683994770050049, | |
| "logits/rejected": -1.7013158798217773, | |
| "logps/chosen": -393.85699462890625, | |
| "logps/rejected": -393.3715515136719, | |
| "loss": 0.5931, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.0844448804855347, | |
| "rewards/margins": 0.284206360578537, | |
| "rewards/rejected": -1.3686515092849731, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.28494162654178484, | |
| "grad_norm": 11.265967096929554, | |
| "learning_rate": 4.5005977600621275e-07, | |
| "logits/chosen": -1.2262264490127563, | |
| "logits/rejected": -1.720760703086853, | |
| "logps/chosen": -390.81500244140625, | |
| "logps/rejected": -386.7849426269531, | |
| "loss": 0.577, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.1187413930892944, | |
| "rewards/margins": 0.2867385149002075, | |
| "rewards/rejected": -1.4054800271987915, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.2954950201174065, | |
| "grad_norm": 13.458747198104215, | |
| "learning_rate": 4.443833686102919e-07, | |
| "logits/chosen": -1.2382609844207764, | |
| "logits/rejected": -1.6747297048568726, | |
| "logps/chosen": -434.0775451660156, | |
| "logps/rejected": -440.32598876953125, | |
| "loss": 0.5758, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.3505640029907227, | |
| "rewards/margins": 0.3462044298648834, | |
| "rewards/rejected": -1.6967684030532837, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.30604841369302815, | |
| "grad_norm": 11.083230607187067, | |
| "learning_rate": 4.384414569144561e-07, | |
| "logits/chosen": -1.2794346809387207, | |
| "logits/rejected": -1.7727094888687134, | |
| "logps/chosen": -410.8555603027344, | |
| "logps/rejected": -412.8260803222656, | |
| "loss": 0.5823, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.3353734016418457, | |
| "rewards/margins": 0.34854626655578613, | |
| "rewards/rejected": -1.6839195489883423, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.31660180726864984, | |
| "grad_norm": 9.628316005499197, | |
| "learning_rate": 4.3224215685535287e-07, | |
| "logits/chosen": -1.2036387920379639, | |
| "logits/rejected": -1.6848615407943726, | |
| "logps/chosen": -425.97796630859375, | |
| "logps/rejected": -431.17547607421875, | |
| "loss": 0.5789, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.3492869138717651, | |
| "rewards/margins": 0.35322755575180054, | |
| "rewards/rejected": -1.702514410018921, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.32715520084427147, | |
| "grad_norm": 13.446042568474843, | |
| "learning_rate": 4.2579393593117364e-07, | |
| "logits/chosen": -1.2631900310516357, | |
| "logits/rejected": -1.6935676336288452, | |
| "logps/chosen": -405.7359619140625, | |
| "logps/rejected": -407.66510009765625, | |
| "loss": 0.5651, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.19693124294281, | |
| "rewards/margins": 0.2678401470184326, | |
| "rewards/rejected": -1.4647715091705322, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.33770859441989315, | |
| "grad_norm": 11.045996690620928, | |
| "learning_rate": 4.191056016360699e-07, | |
| "logits/chosen": -1.3399848937988281, | |
| "logits/rejected": -1.648470163345337, | |
| "logps/chosen": -443.5994567871094, | |
| "logps/rejected": -469.8695373535156, | |
| "loss": 0.5645, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.471341848373413, | |
| "rewards/margins": 0.4636549949645996, | |
| "rewards/rejected": -1.9349968433380127, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.34826198799551483, | |
| "grad_norm": 13.778566050110996, | |
| "learning_rate": 4.121862894301754e-07, | |
| "logits/chosen": -1.3236744403839111, | |
| "logits/rejected": -1.8871498107910156, | |
| "logps/chosen": -420.1505432128906, | |
| "logps/rejected": -415.3837890625, | |
| "loss": 0.5593, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.1470205783843994, | |
| "rewards/margins": 0.3792787194252014, | |
| "rewards/rejected": -1.526299238204956, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.35881538157113646, | |
| "grad_norm": 17.339401498715958, | |
| "learning_rate": 4.050454502616667e-07, | |
| "logits/chosen": -1.3061871528625488, | |
| "logits/rejected": -1.6525911092758179, | |
| "logps/chosen": -458.53338623046875, | |
| "logps/rejected": -494.4827575683594, | |
| "loss": 0.5455, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.6396774053573608, | |
| "rewards/margins": 0.5506707429885864, | |
| "rewards/rejected": -2.1903481483459473, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.36936877514675814, | |
| "grad_norm": 14.142408652824205, | |
| "learning_rate": 3.976928376579047e-07, | |
| "logits/chosen": -1.2991268634796143, | |
| "logits/rejected": -1.7594095468521118, | |
| "logps/chosen": -505.6692810058594, | |
| "logps/rejected": -530.7013549804688, | |
| "loss": 0.5295, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.9093101024627686, | |
| "rewards/margins": 0.5646597743034363, | |
| "rewards/rejected": -2.4739699363708496, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.37992216872237977, | |
| "grad_norm": 12.91331776704347, | |
| "learning_rate": 3.9013849440328945e-07, | |
| "logits/chosen": -1.2874772548675537, | |
| "logits/rejected": -1.6723277568817139, | |
| "logps/chosen": -453.1542053222656, | |
| "logps/rejected": -469.8594665527344, | |
| "loss": 0.5487, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.7126022577285767, | |
| "rewards/margins": 0.4102846682071686, | |
| "rewards/rejected": -2.1228866577148438, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.39047556229800146, | |
| "grad_norm": 13.372482096217857, | |
| "learning_rate": 3.8239273882202473e-07, | |
| "logits/chosen": -1.3676090240478516, | |
| "logits/rejected": -1.727839708328247, | |
| "logps/chosen": -470.1155700683594, | |
| "logps/rejected": -490.20196533203125, | |
| "loss": 0.5615, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.8151922225952148, | |
| "rewards/margins": 0.47195178270339966, | |
| "rewards/rejected": -2.287144184112549, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 0.40102895587362314, | |
| "grad_norm": 19.85364607320994, | |
| "learning_rate": 3.7446615068452804e-07, | |
| "logits/chosen": -1.3825256824493408, | |
| "logits/rejected": -1.7227319478988647, | |
| "logps/chosen": -438.42333984375, | |
| "logps/rejected": -469.5264587402344, | |
| "loss": 0.5311, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.7287547588348389, | |
| "rewards/margins": 0.5413612723350525, | |
| "rewards/rejected": -2.2701163291931152, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.41158234944924477, | |
| "grad_norm": 14.668927735081262, | |
| "learning_rate": 3.6636955675673743e-07, | |
| "logits/chosen": -1.2241297960281372, | |
| "logits/rejected": -1.7972831726074219, | |
| "logps/chosen": -444.69830322265625, | |
| "logps/rejected": -449.8846740722656, | |
| "loss": 0.5395, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.570357084274292, | |
| "rewards/margins": 0.42953139543533325, | |
| "rewards/rejected": -1.9998886585235596, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.42213574302486645, | |
| "grad_norm": 12.855557980372362, | |
| "learning_rate": 3.5811401601205093e-07, | |
| "logits/chosen": -1.3073621988296509, | |
| "logits/rejected": -1.7404594421386719, | |
| "logps/chosen": -453.62286376953125, | |
| "logps/rejected": -476.1055603027344, | |
| "loss": 0.516, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -1.6229336261749268, | |
| "rewards/margins": 0.5728715062141418, | |
| "rewards/rejected": -2.195805072784424, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4326891366004881, | |
| "grad_norm": 14.723820955973457, | |
| "learning_rate": 3.497108045260995e-07, | |
| "logits/chosen": -1.3508846759796143, | |
| "logits/rejected": -1.8392765522003174, | |
| "logps/chosen": -464.30926513671875, | |
| "logps/rejected": -480.6266174316406, | |
| "loss": 0.5468, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.9108375310897827, | |
| "rewards/margins": 0.4916614592075348, | |
| "rewards/rejected": -2.402498960494995, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 0.44324253017610976, | |
| "grad_norm": 17.043565936987463, | |
| "learning_rate": 3.411714000749838e-07, | |
| "logits/chosen": -1.374245047569275, | |
| "logits/rejected": -1.8004543781280518, | |
| "logps/chosen": -487.89776611328125, | |
| "logps/rejected": -509.8401794433594, | |
| "loss": 0.5286, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.9728209972381592, | |
| "rewards/margins": 0.46407732367515564, | |
| "rewards/rejected": -2.4368982315063477, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.4537959237517314, | |
| "grad_norm": 24.760116135102702, | |
| "learning_rate": 3.3250746645801287e-07, | |
| "logits/chosen": -1.5120536088943481, | |
| "logits/rejected": -1.846258521080017, | |
| "logps/chosen": -495.29779052734375, | |
| "logps/rejected": -514.2366943359375, | |
| "loss": 0.5482, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.207698345184326, | |
| "rewards/margins": 0.46926626563072205, | |
| "rewards/rejected": -2.676964521408081, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 0.4643493173273531, | |
| "grad_norm": 16.143780206492696, | |
| "learning_rate": 3.237308375663571e-07, | |
| "logits/chosen": -1.3949377536773682, | |
| "logits/rejected": -1.7159026861190796, | |
| "logps/chosen": -467.3490295410156, | |
| "logps/rejected": -508.98480224609375, | |
| "loss": 0.5367, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.080124855041504, | |
| "rewards/margins": 0.5874677896499634, | |
| "rewards/rejected": -2.6675925254821777, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.47490271090297476, | |
| "grad_norm": 21.361838942730817, | |
| "learning_rate": 3.148535012193767e-07, | |
| "logits/chosen": -1.4045331478118896, | |
| "logits/rejected": -1.745840072631836, | |
| "logps/chosen": -509.15594482421875, | |
| "logps/rejected": -545.216064453125, | |
| "loss": 0.5216, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.0763394832611084, | |
| "rewards/margins": 0.5761786699295044, | |
| "rewards/rejected": -2.6525182723999023, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.4854561044785964, | |
| "grad_norm": 14.456644926293217, | |
| "learning_rate": 3.0588758279070183e-07, | |
| "logits/chosen": -1.3919737339019775, | |
| "logits/rejected": -1.8210327625274658, | |
| "logps/chosen": -463.69171142578125, | |
| "logps/rejected": -494.8233947753906, | |
| "loss": 0.5352, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.015812397003174, | |
| "rewards/margins": 0.515998363494873, | |
| "rewards/rejected": -2.531810760498047, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.49600949805421807, | |
| "grad_norm": 14.589143727916596, | |
| "learning_rate": 2.968453286464312e-07, | |
| "logits/chosen": -1.4191744327545166, | |
| "logits/rejected": -1.7561525106430054, | |
| "logps/chosen": -487.16912841796875, | |
| "logps/rejected": -515.9456787109375, | |
| "loss": 0.5387, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.961477279663086, | |
| "rewards/margins": 0.5488137602806091, | |
| "rewards/rejected": -2.51029109954834, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 0.5065628916298397, | |
| "grad_norm": 14.508599811690631, | |
| "learning_rate": 2.8773908941806877e-07, | |
| "logits/chosen": -1.4843761920928955, | |
| "logits/rejected": -1.8826162815093994, | |
| "logps/chosen": -506.026123046875, | |
| "logps/rejected": -531.3427124023438, | |
| "loss": 0.5289, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.0544593334198, | |
| "rewards/margins": 0.5065608024597168, | |
| "rewards/rejected": -2.5610203742980957, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5171162852054614, | |
| "grad_norm": 18.773519116584172, | |
| "learning_rate": 2.785813031330473e-07, | |
| "logits/chosen": -1.443350076675415, | |
| "logits/rejected": -1.7564016580581665, | |
| "logps/chosen": -486.8233947753906, | |
| "logps/rejected": -526.447021484375, | |
| "loss": 0.5172, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.9753189086914062, | |
| "rewards/margins": 0.6450583338737488, | |
| "rewards/rejected": -2.6203773021698, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 0.5276696787810831, | |
| "grad_norm": 18.389045572362814, | |
| "learning_rate": 2.693844782258779e-07, | |
| "logits/chosen": -1.510719656944275, | |
| "logits/rejected": -1.871638536453247, | |
| "logps/chosen": -502.90521240234375, | |
| "logps/rejected": -547.2249145507812, | |
| "loss": 0.4928, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.225248098373413, | |
| "rewards/margins": 0.747429370880127, | |
| "rewards/rejected": -2.972677707672119, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5382230723567047, | |
| "grad_norm": 16.8014567042791, | |
| "learning_rate": 2.601611764531342e-07, | |
| "logits/chosen": -1.4652379751205444, | |
| "logits/rejected": -1.7754548788070679, | |
| "logps/chosen": -507.2633361816406, | |
| "logps/rejected": -574.4364013671875, | |
| "loss": 0.5166, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.2433393001556396, | |
| "rewards/margins": 0.8110122680664062, | |
| "rewards/rejected": -3.054351329803467, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 0.5487764659323263, | |
| "grad_norm": 13.726764888156497, | |
| "learning_rate": 2.5092399573560323e-07, | |
| "logits/chosen": -1.4230600595474243, | |
| "logits/rejected": -1.8214447498321533, | |
| "logps/chosen": -478.06817626953125, | |
| "logps/rejected": -509.787109375, | |
| "loss": 0.5249, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.9923759698867798, | |
| "rewards/margins": 0.5705231428146362, | |
| "rewards/rejected": -2.562898874282837, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.559329859507948, | |
| "grad_norm": 23.65560305674686, | |
| "learning_rate": 2.4168555295104124e-07, | |
| "logits/chosen": -1.6187779903411865, | |
| "logits/rejected": -1.9773069620132446, | |
| "logps/chosen": -536.1935424804688, | |
| "logps/rejected": -572.4215698242188, | |
| "loss": 0.5145, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.2998838424682617, | |
| "rewards/margins": 0.6668075919151306, | |
| "rewards/rejected": -2.966691493988037, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 0.5698832530835697, | |
| "grad_norm": 16.778556885067378, | |
| "learning_rate": 2.3245846670103626e-07, | |
| "logits/chosen": -1.6787500381469727, | |
| "logits/rejected": -2.1560122966766357, | |
| "logps/chosen": -517.3394775390625, | |
| "logps/rejected": -558.0335693359375, | |
| "loss": 0.4913, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.354661464691162, | |
| "rewards/margins": 0.7168524265289307, | |
| "rewards/rejected": -3.0715138912200928, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5804366466591914, | |
| "grad_norm": 17.263807575864956, | |
| "learning_rate": 2.232553400755159e-07, | |
| "logits/chosen": -1.6897958517074585, | |
| "logits/rejected": -2.048074960708618, | |
| "logps/chosen": -483.93780517578125, | |
| "logps/rejected": -522.008056640625, | |
| "loss": 0.5194, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.1912150382995605, | |
| "rewards/margins": 0.6004900932312012, | |
| "rewards/rejected": -2.791705369949341, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.590990040234813, | |
| "grad_norm": 14.97406750721532, | |
| "learning_rate": 2.1408874343844294e-07, | |
| "logits/chosen": -1.5942695140838623, | |
| "logits/rejected": -1.9944686889648438, | |
| "logps/chosen": -510.0712890625, | |
| "logps/rejected": -560.1776123046875, | |
| "loss": 0.5099, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.1629281044006348, | |
| "rewards/margins": 0.6950211524963379, | |
| "rewards/rejected": -2.8579492568969727, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6015434338104346, | |
| "grad_norm": 15.808244586573563, | |
| "learning_rate": 2.049711972582101e-07, | |
| "logits/chosen": -1.5963830947875977, | |
| "logits/rejected": -1.9408931732177734, | |
| "logps/chosen": -526.6187744140625, | |
| "logps/rejected": -573.6658325195312, | |
| "loss": 0.5005, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -2.284257411956787, | |
| "rewards/margins": 0.7334116697311401, | |
| "rewards/rejected": -3.0176689624786377, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 0.6120968273860563, | |
| "grad_norm": 32.53721598643591, | |
| "learning_rate": 1.9591515500618588e-07, | |
| "logits/chosen": -1.6634728908538818, | |
| "logits/rejected": -2.111931324005127, | |
| "logps/chosen": -564.41748046875, | |
| "logps/rejected": -615.9747314453125, | |
| "loss": 0.527, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.686408519744873, | |
| "rewards/margins": 0.7001844644546509, | |
| "rewards/rejected": -3.3865933418273926, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.622650220961678, | |
| "grad_norm": 19.340424118071393, | |
| "learning_rate": 1.8693298614677112e-07, | |
| "logits/chosen": -1.658970594406128, | |
| "logits/rejected": -1.9972753524780273, | |
| "logps/chosen": -569.0438232421875, | |
| "logps/rejected": -618.6884765625, | |
| "loss": 0.4914, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.6297943592071533, | |
| "rewards/margins": 0.8370558619499207, | |
| "rewards/rejected": -3.4668502807617188, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 0.6332036145372997, | |
| "grad_norm": 17.931673600276827, | |
| "learning_rate": 1.7803695924219814e-07, | |
| "logits/chosen": -1.5948402881622314, | |
| "logits/rejected": -2.000213146209717, | |
| "logps/chosen": -551.8944091796875, | |
| "logps/rejected": -608.9307250976562, | |
| "loss": 0.4796, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -2.4895987510681152, | |
| "rewards/margins": 0.8136266469955444, | |
| "rewards/rejected": -3.3032257556915283, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6437570081129214, | |
| "grad_norm": 18.41740038504812, | |
| "learning_rate": 1.6923922519515067e-07, | |
| "logits/chosen": -1.5580310821533203, | |
| "logits/rejected": -2.058936595916748, | |
| "logps/chosen": -568.0121459960938, | |
| "logps/rejected": -599.84765625, | |
| "loss": 0.5062, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.7541260719299316, | |
| "rewards/margins": 0.6605782508850098, | |
| "rewards/rejected": -3.4147045612335205, | |
| "step": 305 | |
| }, | |
| { | |
| "epoch": 0.6543104016885429, | |
| "grad_norm": 18.94630227910009, | |
| "learning_rate": 1.605518006520924e-07, | |
| "logits/chosen": -1.5944321155548096, | |
| "logits/rejected": -1.9596939086914062, | |
| "logps/chosen": -544.1249389648438, | |
| "logps/rejected": -598.9169921875, | |
| "loss": 0.5156, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.5852460861206055, | |
| "rewards/margins": 0.7831751108169556, | |
| "rewards/rejected": -3.3684210777282715, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6648637952641646, | |
| "grad_norm": 17.846645357695152, | |
| "learning_rate": 1.519865515899731e-07, | |
| "logits/chosen": -1.6307493448257446, | |
| "logits/rejected": -2.0079474449157715, | |
| "logps/chosen": -536.2100830078125, | |
| "logps/rejected": -575.9131469726562, | |
| "loss": 0.5119, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.501380443572998, | |
| "rewards/margins": 0.6081604361534119, | |
| "rewards/rejected": -3.1095407009124756, | |
| "step": 315 | |
| }, | |
| { | |
| "epoch": 0.6754171888397863, | |
| "grad_norm": 17.811680184641038, | |
| "learning_rate": 1.4355517710873182e-07, | |
| "logits/chosen": -1.6296656131744385, | |
| "logits/rejected": -2.132188558578491, | |
| "logps/chosen": -553.8519287109375, | |
| "logps/rejected": -594.5834350585938, | |
| "loss": 0.5043, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.6355135440826416, | |
| "rewards/margins": 0.7011954188346863, | |
| "rewards/rejected": -3.3367092609405518, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.685970582415408, | |
| "grad_norm": 22.014031239903254, | |
| "learning_rate": 1.3526919345173318e-07, | |
| "logits/chosen": -1.7465054988861084, | |
| "logits/rejected": -2.223958969116211, | |
| "logps/chosen": -591.5750732421875, | |
| "logps/rejected": -645.5902709960938, | |
| "loss": 0.5127, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -3.071688413619995, | |
| "rewards/margins": 0.8220219612121582, | |
| "rewards/rejected": -3.8937106132507324, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.6965239759910297, | |
| "grad_norm": 18.950149468318013, | |
| "learning_rate": 1.2713991827596443e-07, | |
| "logits/chosen": -1.8067715167999268, | |
| "logits/rejected": -2.2385947704315186, | |
| "logps/chosen": -603.2128295898438, | |
| "logps/rejected": -631.5573120117188, | |
| "loss": 0.5004, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.8057212829589844, | |
| "rewards/margins": 0.7445579767227173, | |
| "rewards/rejected": -3.5502796173095703, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7070773695666512, | |
| "grad_norm": 19.529222069564966, | |
| "learning_rate": 1.191784551934773e-07, | |
| "logits/chosen": -1.6992321014404297, | |
| "logits/rejected": -2.138716697692871, | |
| "logps/chosen": -544.4804077148438, | |
| "logps/rejected": -600.5407104492188, | |
| "loss": 0.495, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -2.6194682121276855, | |
| "rewards/margins": 0.8088088035583496, | |
| "rewards/rejected": -3.428276777267456, | |
| "step": 335 | |
| }, | |
| { | |
| "epoch": 0.7176307631422729, | |
| "grad_norm": 18.18480447945621, | |
| "learning_rate": 1.1139567860518953e-07, | |
| "logits/chosen": -1.7028363943099976, | |
| "logits/rejected": -2.206139326095581, | |
| "logps/chosen": -540.541748046875, | |
| "logps/rejected": -595.2509155273438, | |
| "loss": 0.4766, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.7220633029937744, | |
| "rewards/margins": 0.8165687322616577, | |
| "rewards/rejected": -3.5386319160461426, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7281841567178946, | |
| "grad_norm": 24.785455969446467, | |
| "learning_rate": 1.0380221884776128e-07, | |
| "logits/chosen": -1.6632766723632812, | |
| "logits/rejected": -2.1659774780273438, | |
| "logps/chosen": -634.0614624023438, | |
| "logps/rejected": -691.1161499023438, | |
| "loss": 0.4602, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -3.2313294410705566, | |
| "rewards/margins": 0.9306994676589966, | |
| "rewards/rejected": -4.1620283126831055, | |
| "step": 345 | |
| }, | |
| { | |
| "epoch": 0.7387375502935163, | |
| "grad_norm": 21.96462703390364, | |
| "learning_rate": 9.640844767383405e-08, | |
| "logits/chosen": -1.7271251678466797, | |
| "logits/rejected": -2.200099468231201, | |
| "logps/chosen": -614.051025390625, | |
| "logps/rejected": -659.74853515625, | |
| "loss": 0.524, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -3.4554572105407715, | |
| "rewards/margins": 0.7755366563796997, | |
| "rewards/rejected": -4.23099422454834, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.749290943869138, | |
| "grad_norm": 19.92963560086107, | |
| "learning_rate": 8.922446408546378e-08, | |
| "logits/chosen": -1.659334421157837, | |
| "logits/rejected": -2.0364480018615723, | |
| "logps/chosen": -569.6084594726562, | |
| "logps/rejected": -638.3173828125, | |
| "loss": 0.4554, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -2.770272731781006, | |
| "rewards/margins": 0.9671792984008789, | |
| "rewards/rejected": -3.7374520301818848, | |
| "step": 355 | |
| }, | |
| { | |
| "epoch": 0.7598443374447595, | |
| "grad_norm": 17.97428748366273, | |
| "learning_rate": 8.22600805400994e-08, | |
| "logits/chosen": -1.5649785995483398, | |
| "logits/rejected": -2.075366735458374, | |
| "logps/chosen": -527.646484375, | |
| "logps/rejected": -579.1947021484375, | |
| "loss": 0.4778, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.405059576034546, | |
| "rewards/margins": 0.786846399307251, | |
| "rewards/rejected": -3.191905975341797, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7703977310203812, | |
| "grad_norm": 25.05638368769761, | |
| "learning_rate": 7.552480954794558e-08, | |
| "logits/chosen": -1.6564147472381592, | |
| "logits/rejected": -2.168598175048828, | |
| "logps/chosen": -607.8692626953125, | |
| "logps/rejected": -662.06494140625, | |
| "loss": 0.4862, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -3.0341908931732178, | |
| "rewards/margins": 0.822891116142273, | |
| "rewards/rejected": -3.857081890106201, | |
| "step": 365 | |
| }, | |
| { | |
| "epoch": 0.7809511245960029, | |
| "grad_norm": 22.12728668677665, | |
| "learning_rate": 6.902785067901854e-08, | |
| "logits/chosen": -1.7829809188842773, | |
| "logits/rejected": -2.062894105911255, | |
| "logps/chosen": -574.40673828125, | |
| "logps/rejected": -647.600830078125, | |
| "loss": 0.491, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.980964183807373, | |
| "rewards/margins": 0.8619121313095093, | |
| "rewards/rejected": -3.84287691116333, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7915045181716246, | |
| "grad_norm": 19.35197929971672, | |
| "learning_rate": 6.277807799763973e-08, | |
| "logits/chosen": -1.6468517780303955, | |
| "logits/rejected": -2.1115472316741943, | |
| "logps/chosen": -583.3654174804688, | |
| "logps/rejected": -647.2584228515625, | |
| "loss": 0.4817, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -2.734621524810791, | |
| "rewards/margins": 0.9541465044021606, | |
| "rewards/rejected": -3.688767910003662, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.8020579117472463, | |
| "grad_norm": 25.507938089288015, | |
| "learning_rate": 5.678402794153145e-08, | |
| "logits/chosen": -1.7440725564956665, | |
| "logits/rejected": -2.137272357940674, | |
| "logps/chosen": -546.3848876953125, | |
| "logps/rejected": -603.1316528320312, | |
| "loss": 0.4873, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.7099640369415283, | |
| "rewards/margins": 0.749721884727478, | |
| "rewards/rejected": -3.459686279296875, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8126113053228678, | |
| "grad_norm": 23.303144237592594, | |
| "learning_rate": 5.105388766206969e-08, | |
| "logits/chosen": -1.7952743768692017, | |
| "logits/rejected": -2.218257188796997, | |
| "logps/chosen": -593.2301635742188, | |
| "logps/rejected": -648.0328979492188, | |
| "loss": 0.4938, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -2.9787325859069824, | |
| "rewards/margins": 0.8751193284988403, | |
| "rewards/rejected": -3.853851318359375, | |
| "step": 385 | |
| }, | |
| { | |
| "epoch": 0.8231646988984895, | |
| "grad_norm": 18.27389690660161, | |
| "learning_rate": 4.5595483841620484e-08, | |
| "logits/chosen": -1.7571223974227905, | |
| "logits/rejected": -2.144810914993286, | |
| "logps/chosen": -578.3859252929688, | |
| "logps/rejected": -649.9483642578125, | |
| "loss": 0.5063, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -3.023244619369507, | |
| "rewards/margins": 0.9022436141967773, | |
| "rewards/rejected": -3.925487995147705, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8337180924741112, | |
| "grad_norm": 19.91862919187348, | |
| "learning_rate": 4.0416272003232526e-08, | |
| "logits/chosen": -1.6357412338256836, | |
| "logits/rejected": -2.0107734203338623, | |
| "logps/chosen": -543.9173583984375, | |
| "logps/rejected": -597.3966064453125, | |
| "loss": 0.4669, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.689850330352783, | |
| "rewards/margins": 0.8007882833480835, | |
| "rewards/rejected": -3.4906387329101562, | |
| "step": 395 | |
| }, | |
| { | |
| "epoch": 0.8442714860497329, | |
| "grad_norm": 22.868215506762866, | |
| "learning_rate": 3.552332632729041e-08, | |
| "logits/chosen": -1.7554508447647095, | |
| "logits/rejected": -2.100292444229126, | |
| "logps/chosen": -604.8919067382812, | |
| "logps/rejected": -656.6446533203125, | |
| "loss": 0.4967, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -3.00014591217041, | |
| "rewards/margins": 0.7955247163772583, | |
| "rewards/rejected": -3.795670986175537, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8442714860497329, | |
| "eval_logits/chosen": -2.6870288848876953, | |
| "eval_logits/rejected": -2.53877329826355, | |
| "eval_logps/chosen": -555.4608154296875, | |
| "eval_logps/rejected": -610.8331298828125, | |
| "eval_loss": 0.6258421540260315, | |
| "eval_rewards/accuracies": 0.6639676094055176, | |
| "eval_rewards/chosen": -2.919989824295044, | |
| "eval_rewards/margins": 0.4445248544216156, | |
| "eval_rewards/rejected": -3.3645148277282715, | |
| "eval_runtime": 305.2745, | |
| "eval_samples_per_second": 6.473, | |
| "eval_steps_per_second": 0.809, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8548248796253546, | |
| "grad_norm": 18.976507998784875, | |
| "learning_rate": 3.092332998903416e-08, | |
| "logits/chosen": -1.6837856769561768, | |
| "logits/rejected": -2.183306932449341, | |
| "logps/chosen": -543.5635375976562, | |
| "logps/rejected": -615.6289672851562, | |
| "loss": 0.4408, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -2.6628594398498535, | |
| "rewards/margins": 1.0260940790176392, | |
| "rewards/rejected": -3.688953399658203, | |
| "step": 405 | |
| }, | |
| { | |
| "epoch": 0.8653782732009762, | |
| "grad_norm": 20.75221237108568, | |
| "learning_rate": 2.6622566030146455e-08, | |
| "logits/chosen": -1.7799179553985596, | |
| "logits/rejected": -2.1815567016601562, | |
| "logps/chosen": -591.3876953125, | |
| "logps/rejected": -658.4329223632812, | |
| "loss": 0.4724, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -3.0632710456848145, | |
| "rewards/margins": 0.9336662292480469, | |
| "rewards/rejected": -3.9969372749328613, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8759316667765978, | |
| "grad_norm": 20.339013632703708, | |
| "learning_rate": 2.26269087768734e-08, | |
| "logits/chosen": -1.7808520793914795, | |
| "logits/rejected": -2.20698881149292, | |
| "logps/chosen": -610.84716796875, | |
| "logps/rejected": -696.8916015625, | |
| "loss": 0.4667, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -3.1746575832366943, | |
| "rewards/margins": 1.094146490097046, | |
| "rewards/rejected": -4.268804550170898, | |
| "step": 415 | |
| }, | |
| { | |
| "epoch": 0.8864850603522195, | |
| "grad_norm": 21.448619329065966, | |
| "learning_rate": 1.894181581640106e-08, | |
| "logits/chosen": -1.7119935750961304, | |
| "logits/rejected": -2.181185722351074, | |
| "logps/chosen": -640.0302734375, | |
| "logps/rejected": -725.7114868164062, | |
| "loss": 0.4371, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -3.318742036819458, | |
| "rewards/margins": 1.2371790409088135, | |
| "rewards/rejected": -4.555922031402588, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.8970384539278412, | |
| "grad_norm": 31.199200608117586, | |
| "learning_rate": 1.5572320542448143e-08, | |
| "logits/chosen": -1.8320446014404297, | |
| "logits/rejected": -2.2155635356903076, | |
| "logps/chosen": -637.1694946289062, | |
| "logps/rejected": -691.4876708984375, | |
| "loss": 0.4918, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -3.344149112701416, | |
| "rewards/margins": 0.7779822945594788, | |
| "rewards/rejected": -4.12213134765625, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.9075918475034628, | |
| "grad_norm": 26.35733359263408, | |
| "learning_rate": 1.2523025280255729e-08, | |
| "logits/chosen": -1.8720299005508423, | |
| "logits/rejected": -2.2721853256225586, | |
| "logps/chosen": -598.7555541992188, | |
| "logps/rejected": -676.6019287109375, | |
| "loss": 0.4668, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -3.2352309226989746, | |
| "rewards/margins": 1.1256295442581177, | |
| "rewards/rejected": -4.360860347747803, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9181452410790845, | |
| "grad_norm": 19.592377697948315, | |
| "learning_rate": 9.798095000364214e-09, | |
| "logits/chosen": -1.8192625045776367, | |
| "logits/rejected": -2.1697540283203125, | |
| "logps/chosen": -593.2144775390625, | |
| "logps/rejected": -647.7749633789062, | |
| "loss": 0.5005, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -3.3029427528381348, | |
| "rewards/margins": 0.7348226308822632, | |
| "rewards/rejected": -4.0377655029296875, | |
| "step": 435 | |
| }, | |
| { | |
| "epoch": 0.9286986346547061, | |
| "grad_norm": 18.25740720427561, | |
| "learning_rate": 7.401251629764876e-09, | |
| "logits/chosen": -1.8516120910644531, | |
| "logits/rejected": -2.2366607189178467, | |
| "logps/chosen": -612.3639526367188, | |
| "logps/rejected": -662.9229736328125, | |
| "loss": 0.4957, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -3.2708640098571777, | |
| "rewards/margins": 0.781753420829773, | |
| "rewards/rejected": -4.052617073059082, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9392520282303278, | |
| "grad_norm": 21.707645823041325, | |
| "learning_rate": 5.335768968195098e-09, | |
| "logits/chosen": -1.822199821472168, | |
| "logits/rejected": -2.2232327461242676, | |
| "logps/chosen": -590.3006591796875, | |
| "logps/rejected": -674.7203369140625, | |
| "loss": 0.4531, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -3.1192424297332764, | |
| "rewards/margins": 1.0382407903671265, | |
| "rewards/rejected": -4.1574835777282715, | |
| "step": 445 | |
| }, | |
| { | |
| "epoch": 0.9498054218059495, | |
| "grad_norm": 27.24335169269175, | |
| "learning_rate": 3.604468216521883e-09, | |
| "logits/chosen": -1.8075069189071655, | |
| "logits/rejected": -2.1066629886627197, | |
| "logps/chosen": -587.6980590820312, | |
| "logps/rejected": -656.5093994140625, | |
| "loss": 0.4569, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -3.0228023529052734, | |
| "rewards/margins": 0.8714979290962219, | |
| "rewards/rejected": -3.8943004608154297, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9603588153815711, | |
| "grad_norm": 30.5949031304229, | |
| "learning_rate": 2.2097141233206884e-09, | |
| "logits/chosen": -1.764469861984253, | |
| "logits/rejected": -2.0643749237060547, | |
| "logps/chosen": -625.1185302734375, | |
| "logps/rejected": -686.2703857421875, | |
| "loss": 0.4951, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -3.299448013305664, | |
| "rewards/margins": 0.8464337587356567, | |
| "rewards/rejected": -4.1458821296691895, | |
| "step": 455 | |
| }, | |
| { | |
| "epoch": 0.9709122089571928, | |
| "grad_norm": 16.854970901618884, | |
| "learning_rate": 1.1534117549133472e-09, | |
| "logits/chosen": -1.8324248790740967, | |
| "logits/rejected": -2.1692023277282715, | |
| "logps/chosen": -600.291015625, | |
| "logps/rejected": -676.2047119140625, | |
| "loss": 0.4721, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -3.200017213821411, | |
| "rewards/margins": 0.9616624712944031, | |
| "rewards/rejected": -4.161679267883301, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.9814656025328145, | |
| "grad_norm": 20.996442076100582, | |
| "learning_rate": 4.3700389327672173e-10, | |
| "logits/chosen": -1.6799993515014648, | |
| "logits/rejected": -2.161038398742676, | |
| "logps/chosen": -642.9143676757812, | |
| "logps/rejected": -710.483154296875, | |
| "loss": 0.4775, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -3.233483076095581, | |
| "rewards/margins": 1.045021414756775, | |
| "rewards/rejected": -4.278504371643066, | |
| "step": 465 | |
| }, | |
| { | |
| "epoch": 0.9920189961084361, | |
| "grad_norm": 31.480125798687133, | |
| "learning_rate": 6.146906537587982e-11, | |
| "logits/chosen": -1.7645127773284912, | |
| "logits/rejected": -2.224125385284424, | |
| "logps/chosen": -594.4464721679688, | |
| "logps/rejected": -659.8152465820312, | |
| "loss": 0.4866, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.988931179046631, | |
| "rewards/margins": 1.064064383506775, | |
| "rewards/rejected": -4.052995204925537, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.9983510322538092, | |
| "step": 473, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5460537698505796, | |
| "train_runtime": 22947.8791, | |
| "train_samples_per_second": 2.643, | |
| "train_steps_per_second": 0.021 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 473, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000000, | |
| "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 | |
| } | |