| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984, | |
| "eval_steps": 500, | |
| "global_step": 468, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.0638297872340426e-07, | |
| "logits/chosen": 0.18604117631912231, | |
| "logits/rejected": 0.34631967544555664, | |
| "logps/chosen": -460.0769348144531, | |
| "logps/rejected": -351.57135009765625, | |
| "loss": 0.3612, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.0638297872340427e-06, | |
| "logits/chosen": 0.05747946724295616, | |
| "logits/rejected": 0.16539901494979858, | |
| "logps/chosen": -330.9287414550781, | |
| "logps/rejected": -328.71575927734375, | |
| "loss": 0.3819, | |
| "rewards/accuracies": 0.3888888955116272, | |
| "rewards/chosen": 2.3913149561849423e-05, | |
| "rewards/margins": -1.7764228687155992e-05, | |
| "rewards/rejected": 4.167737643001601e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.1276595744680853e-06, | |
| "logits/chosen": 0.15373219549655914, | |
| "logits/rejected": 0.19493858516216278, | |
| "logps/chosen": -325.29803466796875, | |
| "logps/rejected": -315.1011047363281, | |
| "loss": 0.3718, | |
| "rewards/accuracies": 0.4312500059604645, | |
| "rewards/chosen": -1.974666338355746e-05, | |
| "rewards/margins": 1.2878153654583002e-07, | |
| "rewards/rejected": -1.9875456928275526e-05, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.191489361702128e-06, | |
| "logits/chosen": 0.11452794075012207, | |
| "logits/rejected": 0.1842522919178009, | |
| "logps/chosen": -371.60943603515625, | |
| "logps/rejected": -327.19366455078125, | |
| "loss": 0.3676, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -5.7096302043646574e-05, | |
| "rewards/margins": 3.7131747376406565e-05, | |
| "rewards/rejected": -9.422805305803195e-05, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 4.255319148936171e-06, | |
| "logits/chosen": 0.18893569707870483, | |
| "logits/rejected": 0.2084759920835495, | |
| "logps/chosen": -337.7607727050781, | |
| "logps/rejected": -353.72503662109375, | |
| "loss": 0.3678, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.00010834180284291506, | |
| "rewards/margins": 0.00026063303812406957, | |
| "rewards/rejected": -0.0001522912352811545, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 4.999373573764188e-06, | |
| "logits/chosen": 0.08207504451274872, | |
| "logits/rejected": 0.20786412060260773, | |
| "logps/chosen": -384.91656494140625, | |
| "logps/rejected": -347.6080627441406, | |
| "loss": 0.3741, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.0010863704374060035, | |
| "rewards/margins": 0.0014893051702529192, | |
| "rewards/rejected": -0.0004029346746392548, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.988245838331339e-06, | |
| "logits/chosen": 0.13469652831554413, | |
| "logits/rejected": 0.17153413593769073, | |
| "logps/chosen": -372.34124755859375, | |
| "logps/rejected": -329.98626708984375, | |
| "loss": 0.3646, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.0023965700529515743, | |
| "rewards/margins": 0.0026933744084089994, | |
| "rewards/rejected": -0.0002968042972497642, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.963268819535228e-06, | |
| "logits/chosen": 0.15689334273338318, | |
| "logits/rejected": 0.20978419482707977, | |
| "logps/chosen": -364.29888916015625, | |
| "logps/rejected": -377.21087646484375, | |
| "loss": 0.3705, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.0023818810004740953, | |
| "rewards/margins": 0.0072565278969705105, | |
| "rewards/rejected": -0.0048746466636657715, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.9245815365216115e-06, | |
| "logits/chosen": 0.11595060676336288, | |
| "logits/rejected": 0.18608702719211578, | |
| "logps/chosen": -328.6256103515625, | |
| "logps/rejected": -365.93896484375, | |
| "loss": 0.3663, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.002104334533214569, | |
| "rewards/margins": 0.011472588405013084, | |
| "rewards/rejected": -0.013576922006905079, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.872399318152594e-06, | |
| "logits/chosen": 0.08527339994907379, | |
| "logits/rejected": 0.11540959030389786, | |
| "logps/chosen": -352.0487060546875, | |
| "logps/rejected": -361.36578369140625, | |
| "loss": 0.3456, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.011672710999846458, | |
| "rewards/margins": 0.01901327446103096, | |
| "rewards/rejected": -0.03068598173558712, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.807012604511542e-06, | |
| "logits/chosen": 0.14592930674552917, | |
| "logits/rejected": 0.18039533495903015, | |
| "logps/chosen": -414.8818359375, | |
| "logps/rejected": -435.3639221191406, | |
| "loss": 0.34, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.04601982235908508, | |
| "rewards/margins": 0.034013133496046066, | |
| "rewards/rejected": -0.08003295958042145, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.728785330347771e-06, | |
| "logits/chosen": 0.13395074009895325, | |
| "logits/rejected": 0.2120208740234375, | |
| "logps/chosen": -432.98486328125, | |
| "logps/rejected": -553.6598510742188, | |
| "loss": 0.305, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.11783289909362793, | |
| "rewards/margins": 0.08183668553829193, | |
| "rewards/rejected": -0.19966959953308105, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.63815289945858e-06, | |
| "logits/chosen": 0.14807412028312683, | |
| "logits/rejected": 0.21360798180103302, | |
| "logps/chosen": -565.0419311523438, | |
| "logps/rejected": -595.3816528320312, | |
| "loss": 0.3434, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.18253007531166077, | |
| "rewards/margins": 0.060289014130830765, | |
| "rewards/rejected": -0.24281907081604004, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 4.535619761282989e-06, | |
| "logits/chosen": 0.10818658024072647, | |
| "logits/rejected": 0.18194182217121124, | |
| "logps/chosen": -462.3949279785156, | |
| "logps/rejected": -513.8490600585938, | |
| "loss": 0.3173, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.1326686441898346, | |
| "rewards/margins": 0.0804656594991684, | |
| "rewards/rejected": -0.2131342887878418, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.42175660319555e-06, | |
| "logits/chosen": 0.13493295013904572, | |
| "logits/rejected": 0.22737479209899902, | |
| "logps/chosen": -532.4095458984375, | |
| "logps/rejected": -585.3455200195312, | |
| "loss": 0.3196, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.16629299521446228, | |
| "rewards/margins": 0.08414360880851746, | |
| "rewards/rejected": -0.25043657422065735, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.297197174127619e-06, | |
| "logits/chosen": 0.17478415369987488, | |
| "logits/rejected": 0.24990789592266083, | |
| "logps/chosen": -479.77862548828125, | |
| "logps/rejected": -553.7377319335938, | |
| "loss": 0.3207, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.14474426209926605, | |
| "rewards/margins": 0.09155549854040146, | |
| "rewards/rejected": -0.2362997829914093, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.162634757195418e-06, | |
| "logits/chosen": 0.13402113318443298, | |
| "logits/rejected": 0.2551622688770294, | |
| "logps/chosen": -496.40081787109375, | |
| "logps/rejected": -558.84326171875, | |
| "loss": 0.3187, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.13809171319007874, | |
| "rewards/margins": 0.08086591213941574, | |
| "rewards/rejected": -0.21895763278007507, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.018818310967843e-06, | |
| "logits/chosen": 0.12252243608236313, | |
| "logits/rejected": 0.16481925547122955, | |
| "logps/chosen": -484.56353759765625, | |
| "logps/rejected": -554.2274780273438, | |
| "loss": 0.3255, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.14209917187690735, | |
| "rewards/margins": 0.09574152529239655, | |
| "rewards/rejected": -0.2378406971693039, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.866548300851254e-06, | |
| "logits/chosen": 0.08216498792171478, | |
| "logits/rejected": 0.17952165007591248, | |
| "logps/chosen": -482.576171875, | |
| "logps/rejected": -579.781982421875, | |
| "loss": 0.3047, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.14626096189022064, | |
| "rewards/margins": 0.08570893108844757, | |
| "rewards/rejected": -0.23196987807750702, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 3.706672243793271e-06, | |
| "logits/chosen": 0.07855963706970215, | |
| "logits/rejected": 0.07844971120357513, | |
| "logps/chosen": -468.69061279296875, | |
| "logps/rejected": -544.0549926757812, | |
| "loss": 0.2935, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.1354297697544098, | |
| "rewards/margins": 0.09144213050603867, | |
| "rewards/rejected": -0.22687189280986786, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.5400799911032357e-06, | |
| "logits/chosen": 0.10545216500759125, | |
| "logits/rejected": 0.1882828176021576, | |
| "logps/chosen": -501.1815490722656, | |
| "logps/rejected": -601.7250366210938, | |
| "loss": 0.3041, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.15607957541942596, | |
| "rewards/margins": 0.10436417162418365, | |
| "rewards/rejected": -0.2604437470436096, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.3676987756445894e-06, | |
| "logits/chosen": 0.10487540811300278, | |
| "logits/rejected": 0.11818546056747437, | |
| "logps/chosen": -470.6344299316406, | |
| "logps/rejected": -565.8145751953125, | |
| "loss": 0.3148, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.16682696342468262, | |
| "rewards/margins": 0.09452913701534271, | |
| "rewards/rejected": -0.26135605573654175, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.1904880509659397e-06, | |
| "logits/chosen": 0.13482534885406494, | |
| "logits/rejected": 0.20024776458740234, | |
| "logps/chosen": -510.10528564453125, | |
| "logps/rejected": -605.6468505859375, | |
| "loss": 0.3172, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.16119703650474548, | |
| "rewards/margins": 0.1006912812590599, | |
| "rewards/rejected": -0.261888325214386, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 3.0094341510955697e-06, | |
| "logits/chosen": 0.10005593299865723, | |
| "logits/rejected": 0.1791614145040512, | |
| "logps/chosen": -532.1925048828125, | |
| "logps/rejected": -624.0726318359375, | |
| "loss": 0.3106, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.17168700695037842, | |
| "rewards/margins": 0.09216924011707306, | |
| "rewards/rejected": -0.26385626196861267, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 2.825544800722376e-06, | |
| "logits/chosen": 0.10918021202087402, | |
| "logits/rejected": 0.18382051587104797, | |
| "logps/chosen": -508.88494873046875, | |
| "logps/rejected": -571.0011596679688, | |
| "loss": 0.309, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.15795882046222687, | |
| "rewards/margins": 0.09128745645284653, | |
| "rewards/rejected": -0.2492462694644928, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 2.639843506318899e-06, | |
| "logits/chosen": 0.12444597482681274, | |
| "logits/rejected": 0.16049379110336304, | |
| "logps/chosen": -509.65631103515625, | |
| "logps/rejected": -609.0398559570312, | |
| "loss": 0.2924, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.1560695320367813, | |
| "rewards/margins": 0.11202052980661392, | |
| "rewards/rejected": -0.26809003949165344, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2.4533638594248094e-06, | |
| "logits/chosen": 0.11506851017475128, | |
| "logits/rejected": 0.1052849292755127, | |
| "logps/chosen": -535.1851806640625, | |
| "logps/rejected": -609.4058837890625, | |
| "loss": 0.3139, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.18579542636871338, | |
| "rewards/margins": 0.10177616029977798, | |
| "rewards/rejected": -0.28757157921791077, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.2671437837980943e-06, | |
| "logits/chosen": 0.11827238649129868, | |
| "logits/rejected": 0.12189098447561264, | |
| "logps/chosen": -542.7433471679688, | |
| "logps/rejected": -633.5802612304688, | |
| "loss": 0.2867, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.19100908935070038, | |
| "rewards/margins": 0.09670811891555786, | |
| "rewards/rejected": -0.28771719336509705, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.082219758453629e-06, | |
| "logits/chosen": 0.0904841274023056, | |
| "logits/rejected": 0.20232203602790833, | |
| "logps/chosen": -575.5374755859375, | |
| "logps/rejected": -672.4290161132812, | |
| "loss": 0.3105, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.2171885222196579, | |
| "rewards/margins": 0.10769243538379669, | |
| "rewards/rejected": -0.3248809278011322, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.899621048743019e-06, | |
| "logits/chosen": 0.07801838964223862, | |
| "logits/rejected": 0.16570156812667847, | |
| "logps/chosen": -556.6237182617188, | |
| "logps/rejected": -643.4578857421875, | |
| "loss": 0.3145, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.20063337683677673, | |
| "rewards/margins": 0.1107235923409462, | |
| "rewards/rejected": -0.31135696172714233, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.7203639775848423e-06, | |
| "logits/chosen": 0.07458348572254181, | |
| "logits/rejected": 0.08251482248306274, | |
| "logps/chosen": -524.4200439453125, | |
| "logps/rejected": -666.1168823242188, | |
| "loss": 0.2934, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.18233875930309296, | |
| "rewards/margins": 0.13505366444587708, | |
| "rewards/rejected": -0.3173924386501312, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1.5454462687309445e-06, | |
| "logits/chosen": 0.06379405409097672, | |
| "logits/rejected": 0.14568018913269043, | |
| "logps/chosen": -504.7530822753906, | |
| "logps/rejected": -650.941650390625, | |
| "loss": 0.2814, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.17345736920833588, | |
| "rewards/margins": 0.1400579959154129, | |
| "rewards/rejected": -0.3135153353214264, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.3758414935535147e-06, | |
| "logits/chosen": 0.0773477703332901, | |
| "logits/rejected": 0.12101063877344131, | |
| "logps/chosen": -536.0679931640625, | |
| "logps/rejected": -659.2662353515625, | |
| "loss": 0.3014, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.18032298982143402, | |
| "rewards/margins": 0.11869337409734726, | |
| "rewards/rejected": -0.29901641607284546, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.2124936522614622e-06, | |
| "logits/chosen": 0.09862785786390305, | |
| "logits/rejected": 0.14109982550144196, | |
| "logps/chosen": -487.79071044921875, | |
| "logps/rejected": -610.7267456054688, | |
| "loss": 0.2935, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.18224193155765533, | |
| "rewards/margins": 0.10070188343524933, | |
| "rewards/rejected": -0.2829437851905823, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0563119197063934e-06, | |
| "logits/chosen": 0.09601452201604843, | |
| "logits/rejected": 0.13355228304862976, | |
| "logps/chosen": -478.78143310546875, | |
| "logps/rejected": -622.6370239257812, | |
| "loss": 0.31, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.17518463730812073, | |
| "rewards/margins": 0.11871640384197235, | |
| "rewards/rejected": -0.2939010262489319, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 9.081655850224449e-07, | |
| "logits/chosen": 0.14273716509342194, | |
| "logits/rejected": 0.12122112512588501, | |
| "logps/chosen": -535.0431518554688, | |
| "logps/rejected": -643.0104370117188, | |
| "loss": 0.3196, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.19643327593803406, | |
| "rewards/margins": 0.10987784713506699, | |
| "rewards/rejected": -0.30631113052368164, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.688792132653111e-07, | |
| "logits/chosen": 0.1597367525100708, | |
| "logits/rejected": 0.18071278929710388, | |
| "logps/chosen": -597.0608520507812, | |
| "logps/rejected": -692.2689208984375, | |
| "loss": 0.2965, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.20307877659797668, | |
| "rewards/margins": 0.12918424606323242, | |
| "rewards/rejected": -0.3322630524635315, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 6.392280559802341e-07, | |
| "logits/chosen": 0.1371072232723236, | |
| "logits/rejected": 0.16206106543540955, | |
| "logps/chosen": -573.9052124023438, | |
| "logps/rejected": -700.1799926757812, | |
| "loss": 0.2689, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.2099401205778122, | |
| "rewards/margins": 0.13856378197669983, | |
| "rewards/rejected": -0.34850388765335083, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 5.199337362431792e-07, | |
| "logits/chosen": 0.10267746448516846, | |
| "logits/rejected": 0.1483933925628662, | |
| "logps/chosen": -554.7241821289062, | |
| "logps/rejected": -637.0789184570312, | |
| "loss": 0.2999, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.19773730635643005, | |
| "rewards/margins": 0.11398313194513321, | |
| "rewards/rejected": -0.31172046065330505, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.1166023219176176e-07, | |
| "logits/chosen": 0.11247100681066513, | |
| "logits/rejected": 0.1309679001569748, | |
| "logps/chosen": -557.1490478515625, | |
| "logps/rejected": -678.8016357421875, | |
| "loss": 0.2945, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.1877584308385849, | |
| "rewards/margins": 0.12902548909187317, | |
| "rewards/rejected": -0.3167839050292969, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 3.150101814011136e-07, | |
| "logits/chosen": 0.16857033967971802, | |
| "logits/rejected": 0.18106935918331146, | |
| "logps/chosen": -571.0828857421875, | |
| "logps/rejected": -638.4654541015625, | |
| "loss": 0.3045, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.19992712140083313, | |
| "rewards/margins": 0.10413169860839844, | |
| "rewards/rejected": -0.30405884981155396, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 2.3052152667409289e-07, | |
| "logits/chosen": 0.09823437035083771, | |
| "logits/rejected": 0.23346427083015442, | |
| "logps/chosen": -537.1575317382812, | |
| "logps/rejected": -657.79931640625, | |
| "loss": 0.3101, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.19983352720737457, | |
| "rewards/margins": 0.11520209163427353, | |
| "rewards/rejected": -0.3150356113910675, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.5866452191498488e-07, | |
| "logits/chosen": 0.13151055574417114, | |
| "logits/rejected": 0.17384907603263855, | |
| "logps/chosen": -562.8436279296875, | |
| "logps/rejected": -688.2677612304688, | |
| "loss": 0.2892, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.21342253684997559, | |
| "rewards/margins": 0.1261134147644043, | |
| "rewards/rejected": -0.3395359516143799, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 9.983911475163727e-08, | |
| "logits/chosen": 0.11608059704303741, | |
| "logits/rejected": 0.13708294928073883, | |
| "logps/chosen": -503.0716857910156, | |
| "logps/rejected": -606.0911865234375, | |
| "loss": 0.2983, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.18081562221050262, | |
| "rewards/margins": 0.11758317053318024, | |
| "rewards/rejected": -0.29839879274368286, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 5.437272047405712e-08, | |
| "logits/chosen": 0.11100079119205475, | |
| "logits/rejected": 0.13695240020751953, | |
| "logps/chosen": -530.9528198242188, | |
| "logps/rejected": -668.3643798828125, | |
| "loss": 0.3095, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.2087739259004593, | |
| "rewards/margins": 0.13447535037994385, | |
| "rewards/rejected": -0.34324929118156433, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.251839967945535e-08, | |
| "logits/chosen": 0.022059569135308266, | |
| "logits/rejected": 0.12616530060768127, | |
| "logps/chosen": -492.80181884765625, | |
| "logps/rejected": -643.1315307617188, | |
| "loss": 0.2834, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.18623578548431396, | |
| "rewards/margins": 0.14532844722270966, | |
| "rewards/rejected": -0.3315642178058624, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 4.453449766758933e-09, | |
| "logits/chosen": 0.10801200568675995, | |
| "logits/rejected": 0.11458040773868561, | |
| "logps/chosen": -531.0709228515625, | |
| "logps/rejected": -645.5343627929688, | |
| "loss": 0.3043, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.1950865238904953, | |
| "rewards/margins": 0.11565764993429184, | |
| "rewards/rejected": -0.3107442259788513, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 468, | |
| "total_flos": 0.0, | |
| "train_loss": 0.2399272450015076, | |
| "train_runtime": 4417.3525, | |
| "train_samples_per_second": 3.396, | |
| "train_steps_per_second": 0.106 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 468, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |