| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.9936, | |
| "eval_steps": 100, | |
| "global_step": 1248, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 4e-08, | |
| "logits/chosen": 0.7221256494522095, | |
| "logits/rejected": 0.8745549917221069, | |
| "logps/chosen": -277.6833801269531, | |
| "logps/rejected": -189.9869384765625, | |
| "loss": 2500.0, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 4.0000000000000003e-07, | |
| "logits/chosen": 0.6776503324508667, | |
| "logits/rejected": 0.7872164845466614, | |
| "logps/chosen": -273.4142761230469, | |
| "logps/rejected": -216.78836059570312, | |
| "loss": 2505.8785, | |
| "rewards/accuracies": 0.3888888955116272, | |
| "rewards/chosen": 0.00015669333515688777, | |
| "rewards/margins": -0.0005680065951310098, | |
| "rewards/rejected": 0.0007246998138725758, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 8.000000000000001e-07, | |
| "logits/chosen": 0.6071363687515259, | |
| "logits/rejected": 0.8961409330368042, | |
| "logps/chosen": -252.6411590576172, | |
| "logps/rejected": -188.6440887451172, | |
| "loss": 2498.5875, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.0007678332040086389, | |
| "rewards/margins": 0.00017280881002079695, | |
| "rewards/rejected": 0.0005950243212282658, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 1.2000000000000002e-06, | |
| "logits/chosen": 0.5971012711524963, | |
| "logits/rejected": 0.7454315423965454, | |
| "logps/chosen": -272.0901184082031, | |
| "logps/rejected": -205.0573272705078, | |
| "loss": 2504.0305, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.00030031849746592343, | |
| "rewards/margins": -0.0003763613640330732, | |
| "rewards/rejected": 7.604288111906499e-05, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 1.6000000000000001e-06, | |
| "logits/chosen": 0.5639680027961731, | |
| "logits/rejected": 0.8156954646110535, | |
| "logps/chosen": -254.08987426757812, | |
| "logps/rejected": -199.48912048339844, | |
| "loss": 2499.0203, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": -0.00016066078387666494, | |
| "rewards/margins": 0.00013786301133222878, | |
| "rewards/rejected": -0.00029852380976080894, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "logits/chosen": 0.6404772996902466, | |
| "logits/rejected": 0.8446690440177917, | |
| "logps/chosen": -264.58636474609375, | |
| "logps/rejected": -211.51986694335938, | |
| "loss": 2499.9363, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 0.0005515815573744476, | |
| "rewards/margins": 3.0463817893178202e-05, | |
| "rewards/rejected": 0.0005211178213357925, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 2.4000000000000003e-06, | |
| "logits/chosen": 0.693733274936676, | |
| "logits/rejected": 0.8246955871582031, | |
| "logps/chosen": -266.50872802734375, | |
| "logps/rejected": -209.1096649169922, | |
| "loss": 2498.2361, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": 0.0005181363085284829, | |
| "rewards/margins": 0.00020206482440698892, | |
| "rewards/rejected": 0.0003160713822580874, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 2.8000000000000003e-06, | |
| "logits/chosen": 0.6724094152450562, | |
| "logits/rejected": 0.8244352340698242, | |
| "logps/chosen": -256.1933898925781, | |
| "logps/rejected": -198.73048400878906, | |
| "loss": 2492.3055, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": 0.00043690926395356655, | |
| "rewards/margins": 0.0007959330687299371, | |
| "rewards/rejected": -0.0003590236883610487, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 3.2000000000000003e-06, | |
| "logits/chosen": 0.6599764227867126, | |
| "logits/rejected": 0.8170899152755737, | |
| "logps/chosen": -273.2989807128906, | |
| "logps/rejected": -216.25827026367188, | |
| "loss": 2488.535, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": 0.0022243678104132414, | |
| "rewards/margins": 0.0011816672049462795, | |
| "rewards/rejected": 0.0010427006054669619, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 3.6000000000000003e-06, | |
| "logits/chosen": 0.6013220548629761, | |
| "logits/rejected": 0.8273889422416687, | |
| "logps/chosen": -268.2584533691406, | |
| "logps/rejected": -213.9058380126953, | |
| "loss": 2482.6732, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.003236656542867422, | |
| "rewards/margins": 0.0017823975067585707, | |
| "rewards/rejected": 0.001454259268939495, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.000000000000001e-06, | |
| "logits/chosen": 0.6121121644973755, | |
| "logits/rejected": 0.8102381825447083, | |
| "logps/chosen": -258.34576416015625, | |
| "logps/rejected": -212.81478881835938, | |
| "loss": 2477.3281, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 0.0035489716101437807, | |
| "rewards/margins": 0.0023237697314471006, | |
| "rewards/rejected": 0.001225201180204749, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "eval_logits/chosen": 0.8220780491828918, | |
| "eval_logits/rejected": 0.8795672655105591, | |
| "eval_logps/chosen": -256.7978210449219, | |
| "eval_logps/rejected": -233.720703125, | |
| "eval_loss": 2500.715576171875, | |
| "eval_rewards/accuracies": 0.49300000071525574, | |
| "eval_rewards/chosen": -0.0018364518182352185, | |
| "eval_rewards/margins": -1.2206258361402433e-05, | |
| "eval_rewards/rejected": -0.0018242454389110208, | |
| "eval_runtime": 442.8545, | |
| "eval_samples_per_second": 4.516, | |
| "eval_steps_per_second": 1.129, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.4e-06, | |
| "logits/chosen": 0.6344770193099976, | |
| "logits/rejected": 0.8470233082771301, | |
| "logps/chosen": -257.6260986328125, | |
| "logps/rejected": -205.66622924804688, | |
| "loss": 2466.1889, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.005115890875458717, | |
| "rewards/margins": 0.003451352473348379, | |
| "rewards/rejected": 0.0016645386349409819, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 4.800000000000001e-06, | |
| "logits/chosen": 0.6225894689559937, | |
| "logits/rejected": 0.7519047856330872, | |
| "logps/chosen": -273.39776611328125, | |
| "logps/rejected": -211.7797088623047, | |
| "loss": 2446.4322, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": 0.007990755140781403, | |
| "rewards/margins": 0.005494705401360989, | |
| "rewards/rejected": 0.0024960506707429886, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 4.999755441268144e-06, | |
| "logits/chosen": 0.5374079942703247, | |
| "logits/rejected": 0.8167151212692261, | |
| "logps/chosen": -283.385498046875, | |
| "logps/rejected": -228.09716796875, | |
| "loss": 2434.1764, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": 0.010817321017384529, | |
| "rewards/margins": 0.006822638213634491, | |
| "rewards/rejected": 0.003994682338088751, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 4.997799258487003e-06, | |
| "logits/chosen": 0.6246897578239441, | |
| "logits/rejected": 0.8092406988143921, | |
| "logps/chosen": -273.5729064941406, | |
| "logps/rejected": -220.0849609375, | |
| "loss": 2406.7561, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": 0.013732816092669964, | |
| "rewards/margins": 0.009700324386358261, | |
| "rewards/rejected": 0.004032492637634277, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 4.993888423734898e-06, | |
| "logits/chosen": 0.6373761892318726, | |
| "logits/rejected": 0.7924760580062866, | |
| "logps/chosen": -259.81890869140625, | |
| "logps/rejected": -198.0768585205078, | |
| "loss": 2378.9869, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.017864007502794266, | |
| "rewards/margins": 0.01282494980841875, | |
| "rewards/rejected": 0.00503905676305294, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 4.988025997434253e-06, | |
| "logits/chosen": 0.6211603283882141, | |
| "logits/rejected": 0.7996471524238586, | |
| "logps/chosen": -262.85723876953125, | |
| "logps/rejected": -209.2670135498047, | |
| "loss": 2347.3615, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.02218160405755043, | |
| "rewards/margins": 0.016304031014442444, | |
| "rewards/rejected": 0.005877572111785412, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 4.980216567224801e-06, | |
| "logits/chosen": 0.5590766668319702, | |
| "logits/rejected": 0.8358405232429504, | |
| "logps/chosen": -268.93450927734375, | |
| "logps/rejected": -217.07223510742188, | |
| "loss": 2314.8727, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.027453165501356125, | |
| "rewards/margins": 0.02034853585064411, | |
| "rewards/rejected": 0.0071046315133571625, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 4.970466244373527e-06, | |
| "logits/chosen": 0.5857344269752502, | |
| "logits/rejected": 0.8025603294372559, | |
| "logps/chosen": -269.49420166015625, | |
| "logps/rejected": -224.668701171875, | |
| "loss": 2323.1209, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.026119906455278397, | |
| "rewards/margins": 0.019484227523207664, | |
| "rewards/rejected": 0.006635676138103008, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 4.958782658992307e-06, | |
| "logits/chosen": 0.6378930807113647, | |
| "logits/rejected": 0.7625473737716675, | |
| "logps/chosen": -245.82064819335938, | |
| "logps/rejected": -212.6953582763672, | |
| "loss": 2353.0727, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.02194877900183201, | |
| "rewards/margins": 0.01612677052617073, | |
| "rewards/rejected": 0.0058220066130161285, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 4.945174954066957e-06, | |
| "logits/chosen": 0.5794962048530579, | |
| "logits/rejected": 0.7216407656669617, | |
| "logps/chosen": -262.74481201171875, | |
| "logps/rejected": -205.56753540039062, | |
| "loss": 2224.3488, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.040737103670835495, | |
| "rewards/margins": 0.03179619461297989, | |
| "rewards/rejected": 0.008940907195210457, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_logits/chosen": 0.7462335824966431, | |
| "eval_logits/rejected": 0.8050766587257385, | |
| "eval_logps/chosen": -258.5672912597656, | |
| "eval_logps/rejected": -235.52040100097656, | |
| "eval_loss": 2499.890380859375, | |
| "eval_rewards/accuracies": 0.5015000104904175, | |
| "eval_rewards/chosen": -0.019531190395355225, | |
| "eval_rewards/margins": 0.0002903530257754028, | |
| "eval_rewards/rejected": -0.019821541383862495, | |
| "eval_runtime": 412.3797, | |
| "eval_samples_per_second": 4.85, | |
| "eval_steps_per_second": 1.212, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 4.929653778302397e-06, | |
| "logits/chosen": 0.6170880198478699, | |
| "logits/rejected": 0.7665790915489197, | |
| "logps/chosen": -260.6783447265625, | |
| "logps/rejected": -205.8716278076172, | |
| "loss": 2165.3709, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.04661153629422188, | |
| "rewards/margins": 0.039808522909879684, | |
| "rewards/rejected": 0.006803011987358332, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 4.912231277789509e-06, | |
| "logits/chosen": 0.5444914102554321, | |
| "logits/rejected": 0.7560594081878662, | |
| "logps/chosen": -277.1501770019531, | |
| "logps/rejected": -227.8493194580078, | |
| "loss": 2199.4646, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": 0.05624958872795105, | |
| "rewards/margins": 0.037255048751831055, | |
| "rewards/rejected": 0.018994538113474846, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 4.892921086500219e-06, | |
| "logits/chosen": 0.5499123930931091, | |
| "logits/rejected": 0.6719511151313782, | |
| "logps/chosen": -243.00390625, | |
| "logps/rejected": -205.0264434814453, | |
| "loss": 2218.1998, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": 0.04481152817606926, | |
| "rewards/margins": 0.03403898701071739, | |
| "rewards/rejected": 0.010772541165351868, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 4.871738315618236e-06, | |
| "logits/chosen": 0.5359088182449341, | |
| "logits/rejected": 0.7013985514640808, | |
| "logps/chosen": -263.36322021484375, | |
| "logps/rejected": -217.5428924560547, | |
| "loss": 2121.8713, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.06463086605072021, | |
| "rewards/margins": 0.04686326906085014, | |
| "rewards/rejected": 0.017767589539289474, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 4.848699541713801e-06, | |
| "logits/chosen": 0.5078493356704712, | |
| "logits/rejected": 0.6534050703048706, | |
| "logps/chosen": -255.99612426757812, | |
| "logps/rejected": -209.0691680908203, | |
| "loss": 2179.4311, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.05505412817001343, | |
| "rewards/margins": 0.04023104906082153, | |
| "rewards/rejected": 0.014823079109191895, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.823822793771696e-06, | |
| "logits/chosen": 0.5251200795173645, | |
| "logits/rejected": 0.6907744407653809, | |
| "logps/chosen": -244.97616577148438, | |
| "logps/rejected": -193.0167694091797, | |
| "loss": 2039.2832, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.06581679731607437, | |
| "rewards/margins": 0.05999577045440674, | |
| "rewards/rejected": 0.005821021273732185, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 4.797127539082669e-06, | |
| "logits/chosen": 0.5102941393852234, | |
| "logits/rejected": 0.6884879469871521, | |
| "logps/chosen": -267.79193115234375, | |
| "logps/rejected": -216.37808227539062, | |
| "loss": 2084.2711, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": 0.06373900920152664, | |
| "rewards/margins": 0.05440465360879898, | |
| "rewards/rejected": 0.009334356524050236, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 4.7686346680093135e-06, | |
| "logits/chosen": 0.46306952834129333, | |
| "logits/rejected": 0.6997040510177612, | |
| "logps/chosen": -248.92501831054688, | |
| "logps/rejected": -192.95266723632812, | |
| "loss": 1956.5348, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.08462236821651459, | |
| "rewards/margins": 0.07454703748226166, | |
| "rewards/rejected": 0.01007532887160778, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 4.738366477638319e-06, | |
| "logits/chosen": 0.49755996465682983, | |
| "logits/rejected": 0.6823209524154663, | |
| "logps/chosen": -252.31265258789062, | |
| "logps/rejected": -203.94357299804688, | |
| "loss": 2082.6836, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": 0.060582734644412994, | |
| "rewards/margins": 0.0560123547911644, | |
| "rewards/rejected": 0.004570374730974436, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 4.7063466543318965e-06, | |
| "logits/chosen": 0.40442362427711487, | |
| "logits/rejected": 0.643377423286438, | |
| "logps/chosen": -253.7049102783203, | |
| "logps/rejected": -207.00146484375, | |
| "loss": 1898.0719, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.09280969202518463, | |
| "rewards/margins": 0.08169680833816528, | |
| "rewards/rejected": 0.011112888343632221, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "eval_logits/chosen": 0.6232779622077942, | |
| "eval_logits/rejected": 0.6844168305397034, | |
| "eval_logps/chosen": -262.24908447265625, | |
| "eval_logps/rejected": -239.25296020507812, | |
| "eval_loss": 2505.691162109375, | |
| "eval_rewards/accuracies": 0.5139999985694885, | |
| "eval_rewards/chosen": -0.05634931102395058, | |
| "eval_rewards/margins": 0.0007975505432114005, | |
| "eval_rewards/rejected": -0.057146865874528885, | |
| "eval_runtime": 411.7351, | |
| "eval_samples_per_second": 4.857, | |
| "eval_steps_per_second": 1.214, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 4.672600255192022e-06, | |
| "logits/chosen": 0.4478415846824646, | |
| "logits/rejected": 0.6094005703926086, | |
| "logps/chosen": -280.0593566894531, | |
| "logps/rejected": -230.06820678710938, | |
| "loss": 1902.6814, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.08015590906143188, | |
| "rewards/margins": 0.08636869490146637, | |
| "rewards/rejected": -0.006212792359292507, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 4.6371536884520115e-06, | |
| "logits/chosen": 0.4466761648654938, | |
| "logits/rejected": 0.5766214728355408, | |
| "logps/chosen": -260.2146911621094, | |
| "logps/rejected": -202.6782989501953, | |
| "loss": 1965.6586, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.08317091315984726, | |
| "rewards/margins": 0.07663293927907944, | |
| "rewards/rejected": 0.006537970155477524, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "learning_rate": 4.600034692810764e-06, | |
| "logits/chosen": 0.4064536690711975, | |
| "logits/rejected": 0.6527116894721985, | |
| "logps/chosen": -285.71063232421875, | |
| "logps/rejected": -232.9168243408203, | |
| "loss": 1807.0604, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.1063646450638771, | |
| "rewards/margins": 0.10833654552698135, | |
| "rewards/rejected": -0.001971897669136524, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 4.561272315725852e-06, | |
| "logits/chosen": 0.39116889238357544, | |
| "logits/rejected": 0.5679996609687805, | |
| "logps/chosen": -269.14727783203125, | |
| "logps/rejected": -226.08987426757812, | |
| "loss": 1983.8736, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.09996851533651352, | |
| "rewards/margins": 0.08038529753684998, | |
| "rewards/rejected": 0.019583214074373245, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 4.520896890682449e-06, | |
| "logits/chosen": 0.3980609178543091, | |
| "logits/rejected": 0.5835639834403992, | |
| "logps/chosen": -261.222412109375, | |
| "logps/rejected": -211.2742156982422, | |
| "loss": 1903.2615, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.11096767336130142, | |
| "rewards/margins": 0.10369744151830673, | |
| "rewards/rejected": 0.007270221598446369, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "learning_rate": 4.478940013455864e-06, | |
| "logits/chosen": 0.32668036222457886, | |
| "logits/rejected": 0.642693042755127, | |
| "logps/chosen": -248.02194213867188, | |
| "logps/rejected": -195.903564453125, | |
| "loss": 1819.0176, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.11030755937099457, | |
| "rewards/margins": 0.10598069429397583, | |
| "rewards/rejected": 0.004326865542680025, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "learning_rate": 4.435434517386281e-06, | |
| "logits/chosen": 0.40880435705184937, | |
| "logits/rejected": 0.6121729612350464, | |
| "logps/chosen": -256.9469909667969, | |
| "logps/rejected": -210.01284790039062, | |
| "loss": 1875.7564, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.10248272120952606, | |
| "rewards/margins": 0.09737477451562881, | |
| "rewards/rejected": 0.005107959732413292, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 4.39041444768504e-06, | |
| "logits/chosen": 0.35375842452049255, | |
| "logits/rejected": 0.5443102121353149, | |
| "logps/chosen": -271.12091064453125, | |
| "logps/rejected": -223.96395874023438, | |
| "loss": 1840.3432, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.10606244951486588, | |
| "rewards/margins": 0.11150339990854263, | |
| "rewards/rejected": -0.005440945271402597, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 4.343915034792569e-06, | |
| "logits/chosen": 0.33456122875213623, | |
| "logits/rejected": 0.6237267255783081, | |
| "logps/chosen": -248.20458984375, | |
| "logps/rejected": -200.05455017089844, | |
| "loss": 1773.2205, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.11527778208255768, | |
| "rewards/margins": 0.11671481281518936, | |
| "rewards/rejected": -0.0014370165299624205, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 4.295972666808811e-06, | |
| "logits/chosen": 0.3166826665401459, | |
| "logits/rejected": 0.5174378156661987, | |
| "logps/chosen": -285.5495910644531, | |
| "logps/rejected": -236.46630859375, | |
| "loss": 1879.8852, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.10617715120315552, | |
| "rewards/margins": 0.1061965599656105, | |
| "rewards/rejected": -1.941286063811276e-05, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "eval_logits/chosen": 0.5215258598327637, | |
| "eval_logits/rejected": 0.5839402079582214, | |
| "eval_logps/chosen": -266.0533447265625, | |
| "eval_logps/rejected": -243.10528564453125, | |
| "eval_loss": 2516.08349609375, | |
| "eval_rewards/accuracies": 0.5199999809265137, | |
| "eval_rewards/chosen": -0.09439140558242798, | |
| "eval_rewards/margins": 0.001278651412576437, | |
| "eval_rewards/rejected": -0.09567005187273026, | |
| "eval_runtime": 412.4154, | |
| "eval_samples_per_second": 4.849, | |
| "eval_steps_per_second": 1.212, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "learning_rate": 4.246624861017732e-06, | |
| "logits/chosen": 0.31735214591026306, | |
| "logits/rejected": 0.45617708563804626, | |
| "logps/chosen": -242.1129608154297, | |
| "logps/rejected": -200.98440551757812, | |
| "loss": 1814.2686, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.13390564918518066, | |
| "rewards/margins": 0.11606212705373764, | |
| "rewards/rejected": 0.01784351095557213, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "learning_rate": 4.195910234528186e-06, | |
| "logits/chosen": 0.34283334016799927, | |
| "logits/rejected": 0.532746434211731, | |
| "logps/chosen": -250.3861541748047, | |
| "logps/rejected": -209.1632843017578, | |
| "loss": 1950.3166, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.08673225343227386, | |
| "rewards/margins": 0.09020966291427612, | |
| "rewards/rejected": -0.003477415069937706, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "learning_rate": 4.143868474054098e-06, | |
| "logits/chosen": 0.2972554564476013, | |
| "logits/rejected": 0.4938802123069763, | |
| "logps/chosen": -245.431640625, | |
| "logps/rejected": -195.8451385498047, | |
| "loss": 1783.8299, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.1252884864807129, | |
| "rewards/margins": 0.11878180503845215, | |
| "rewards/rejected": 0.006506689824163914, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "learning_rate": 4.0905403048576545e-06, | |
| "logits/chosen": 0.3040269911289215, | |
| "logits/rejected": 0.5599544048309326, | |
| "logps/chosen": -254.08749389648438, | |
| "logps/rejected": -213.42941284179688, | |
| "loss": 1878.9361, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.1011713370680809, | |
| "rewards/margins": 0.10365436226129532, | |
| "rewards/rejected": -0.002483018906787038, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "learning_rate": 4.035967458879751e-06, | |
| "logits/chosen": 0.29073578119277954, | |
| "logits/rejected": 0.4535519480705261, | |
| "logps/chosen": -226.4980926513672, | |
| "logps/rejected": -191.19338989257812, | |
| "loss": 1882.7074, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.11003535985946655, | |
| "rewards/margins": 0.10973072052001953, | |
| "rewards/rejected": 0.000304625544231385, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "learning_rate": 3.980192642082682e-06, | |
| "logits/chosen": 0.27826622128486633, | |
| "logits/rejected": 0.4509051740169525, | |
| "logps/chosen": -246.6437225341797, | |
| "logps/rejected": -199.22531127929688, | |
| "loss": 1963.9541, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.10869649797677994, | |
| "rewards/margins": 0.10766948759555817, | |
| "rewards/rejected": 0.0010270171333104372, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 3.923259501030604e-06, | |
| "logits/chosen": 0.24556589126586914, | |
| "logits/rejected": 0.4926172196865082, | |
| "logps/chosen": -247.1698455810547, | |
| "logps/rejected": -208.8766326904297, | |
| "loss": 1971.4818, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.0876002386212349, | |
| "rewards/margins": 0.08931747823953629, | |
| "rewards/rejected": -0.0017172384541481733, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "learning_rate": 3.865212588733927e-06, | |
| "logits/chosen": 0.3068960905075073, | |
| "logits/rejected": 0.515642523765564, | |
| "logps/chosen": -253.49880981445312, | |
| "logps/rejected": -201.22650146484375, | |
| "loss": 1692.4139, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.12761953473091125, | |
| "rewards/margins": 0.15515872836112976, | |
| "rewards/rejected": -0.027539223432540894, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "learning_rate": 3.8060973297843773e-06, | |
| "logits/chosen": 0.2675209045410156, | |
| "logits/rejected": 0.4657977521419525, | |
| "logps/chosen": -260.93353271484375, | |
| "logps/rejected": -221.60214233398438, | |
| "loss": 1822.4361, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.10896619409322739, | |
| "rewards/margins": 0.10793409496545792, | |
| "rewards/rejected": 0.0010320901637896895, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "learning_rate": 3.7459599848079965e-06, | |
| "logits/chosen": 0.32330405712127686, | |
| "logits/rejected": 0.49660053849220276, | |
| "logps/chosen": -270.75567626953125, | |
| "logps/rejected": -223.2246856689453, | |
| "loss": 1917.2811, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": 0.1301083266735077, | |
| "rewards/margins": 0.10672901570796967, | |
| "rewards/rejected": 0.023379310965538025, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_logits/chosen": 0.4610958397388458, | |
| "eval_logits/rejected": 0.5243986248970032, | |
| "eval_logps/chosen": -268.1746826171875, | |
| "eval_logps/rejected": -245.2342987060547, | |
| "eval_loss": 2527.199462890625, | |
| "eval_rewards/accuracies": 0.5115000009536743, | |
| "eval_rewards/chosen": -0.11560481786727905, | |
| "eval_rewards/margins": 0.0013554621255025268, | |
| "eval_rewards/rejected": -0.11696028709411621, | |
| "eval_runtime": 411.7771, | |
| "eval_samples_per_second": 4.857, | |
| "eval_steps_per_second": 1.214, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "learning_rate": 3.684847614263898e-06, | |
| "logits/chosen": 0.2653389871120453, | |
| "logits/rejected": 0.5690582990646362, | |
| "logps/chosen": -266.63861083984375, | |
| "logps/rejected": -214.3343963623047, | |
| "loss": 1769.2539, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.1135379895567894, | |
| "rewards/margins": 0.13335202634334564, | |
| "rewards/rejected": -0.01981404609978199, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "learning_rate": 3.622808041617133e-06, | |
| "logits/chosen": 0.23678083717823029, | |
| "logits/rejected": 0.3923517167568207, | |
| "logps/chosen": -251.62295532226562, | |
| "logps/rejected": -213.0308074951172, | |
| "loss": 1775.6238, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.12413735687732697, | |
| "rewards/margins": 0.11967913061380386, | |
| "rewards/rejected": 0.0044582299888134, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "learning_rate": 3.559889815914441e-06, | |
| "logits/chosen": 0.28451576828956604, | |
| "logits/rejected": 0.5491675138473511, | |
| "logps/chosen": -242.1103057861328, | |
| "logps/rejected": -196.47967529296875, | |
| "loss": 1708.515, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.13650254905223846, | |
| "rewards/margins": 0.1416713297367096, | |
| "rewards/rejected": -0.005168789066374302, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "learning_rate": 3.496142173792219e-06, | |
| "logits/chosen": 0.23486502468585968, | |
| "logits/rejected": 0.4870077073574066, | |
| "logps/chosen": -244.1999969482422, | |
| "logps/rejected": -209.3772430419922, | |
| "loss": 1659.4633, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.13094991445541382, | |
| "rewards/margins": 0.14943310618400574, | |
| "rewards/rejected": -0.018483208492398262, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "learning_rate": 3.4316150009464023e-06, | |
| "logits/chosen": 0.2991487979888916, | |
| "logits/rejected": 0.5229448080062866, | |
| "logps/chosen": -264.29534912109375, | |
| "logps/rejected": -222.90609741210938, | |
| "loss": 1999.6244, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": 0.1066334992647171, | |
| "rewards/margins": 0.0911371111869812, | |
| "rewards/rejected": 0.015496388077735901, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "learning_rate": 3.366358793094433e-06, | |
| "logits/chosen": 0.18731743097305298, | |
| "logits/rejected": 0.40507373213768005, | |
| "logps/chosen": -245.1287078857422, | |
| "logps/rejected": -194.5263214111328, | |
| "loss": 1803.6146, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.15052266418933868, | |
| "rewards/margins": 0.13785871863365173, | |
| "rewards/rejected": 0.012663939967751503, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 3.3004246164598535e-06, | |
| "logits/chosen": 0.20718684792518616, | |
| "logits/rejected": 0.46155864000320435, | |
| "logps/chosen": -270.30792236328125, | |
| "logps/rejected": -233.5759735107422, | |
| "loss": 1710.7586, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.14311200380325317, | |
| "rewards/margins": 0.14155760407447815, | |
| "rewards/rejected": 0.0015544015914201736, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 3.233864067810446e-06, | |
| "logits/chosen": 0.3184027075767517, | |
| "logits/rejected": 0.4623110890388489, | |
| "logps/chosen": -252.27920532226562, | |
| "logps/rejected": -208.2698974609375, | |
| "loss": 1681.7773, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.137790247797966, | |
| "rewards/margins": 0.15522876381874084, | |
| "rewards/rejected": -0.017438489943742752, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "learning_rate": 3.1667292340812077e-06, | |
| "logits/chosen": 0.1437700092792511, | |
| "logits/rejected": 0.3409765958786011, | |
| "logps/chosen": -247.97341918945312, | |
| "logps/rejected": -218.58163452148438, | |
| "loss": 1856.1729, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.13051538169384003, | |
| "rewards/margins": 0.12071319669485092, | |
| "rewards/rejected": 0.009802192449569702, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 3.099072651613728e-06, | |
| "logits/chosen": 0.24014055728912354, | |
| "logits/rejected": 0.3376144766807556, | |
| "logps/chosen": -236.4192352294922, | |
| "logps/rejected": -202.04861450195312, | |
| "loss": 1799.3824, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.11552337557077408, | |
| "rewards/margins": 0.12494631111621857, | |
| "rewards/rejected": -0.009422937408089638, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "eval_logits/chosen": 0.40748730301856995, | |
| "eval_logits/rejected": 0.4713863432407379, | |
| "eval_logps/chosen": -270.2481994628906, | |
| "eval_logps/rejected": -247.35043334960938, | |
| "eval_loss": 2534.42919921875, | |
| "eval_rewards/accuracies": 0.5210000276565552, | |
| "eval_rewards/chosen": -0.13634006679058075, | |
| "eval_rewards/margins": 0.001781497965566814, | |
| "eval_rewards/rejected": -0.13812156021595, | |
| "eval_runtime": 412.3385, | |
| "eval_samples_per_second": 4.85, | |
| "eval_steps_per_second": 1.213, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "learning_rate": 3.0309472650438982e-06, | |
| "logits/chosen": 0.2844320833683014, | |
| "logits/rejected": 0.44563204050064087, | |
| "logps/chosen": -256.5829772949219, | |
| "logps/rejected": -221.15725708007812, | |
| "loss": 1905.6564, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.10678652673959732, | |
| "rewards/margins": 0.10849614441394806, | |
| "rewards/rejected": -0.001709613250568509, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "learning_rate": 2.9624063858701006e-06, | |
| "logits/chosen": 0.2729710638523102, | |
| "logits/rejected": 0.4474371075630188, | |
| "logps/chosen": -252.75082397460938, | |
| "logps/rejected": -210.4240264892578, | |
| "loss": 1794.8119, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.11081516742706299, | |
| "rewards/margins": 0.1245477944612503, | |
| "rewards/rejected": -0.013732627034187317, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "learning_rate": 2.8935036507343185e-06, | |
| "logits/chosen": 0.20590758323669434, | |
| "logits/rejected": 0.4346593916416168, | |
| "logps/chosen": -262.78900146484375, | |
| "logps/rejected": -220.1951141357422, | |
| "loss": 1844.0162, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": 0.11512579768896103, | |
| "rewards/margins": 0.12753179669380188, | |
| "rewards/rejected": -0.012406004592776299, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "learning_rate": 2.8242929794487926e-06, | |
| "logits/chosen": 0.18648579716682434, | |
| "logits/rejected": 0.4544282555580139, | |
| "logps/chosen": -268.1901550292969, | |
| "logps/rejected": -227.70193481445312, | |
| "loss": 1771.6072, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.1294766366481781, | |
| "rewards/margins": 0.14083310961723328, | |
| "rewards/rejected": -0.011356466449797153, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 2.7548285328010984e-06, | |
| "logits/chosen": 0.26485538482666016, | |
| "logits/rejected": 0.36643001437187195, | |
| "logps/chosen": -250.9102783203125, | |
| "logps/rejected": -210.93991088867188, | |
| "loss": 1854.1963, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.12583324313163757, | |
| "rewards/margins": 0.11704309284687042, | |
| "rewards/rejected": 0.008790150284767151, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "learning_rate": 2.6851646701706306e-06, | |
| "logits/chosen": 0.1529732346534729, | |
| "logits/rejected": 0.33808109164237976, | |
| "logps/chosen": -280.79803466796875, | |
| "logps/rejected": -242.85238647460938, | |
| "loss": 1942.4141, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.11561117321252823, | |
| "rewards/margins": 0.10545969009399414, | |
| "rewards/rejected": 0.010151493363082409, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "learning_rate": 2.6153559069897007e-06, | |
| "logits/chosen": 0.24622070789337158, | |
| "logits/rejected": 0.3663380444049835, | |
| "logps/chosen": -240.98788452148438, | |
| "logps/rejected": -203.41390991210938, | |
| "loss": 1846.866, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.10691628605127335, | |
| "rewards/margins": 0.11943434178829193, | |
| "rewards/rejected": -0.012518051080405712, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "learning_rate": 2.5454568720824937e-06, | |
| "logits/chosen": 0.14400359988212585, | |
| "logits/rejected": 0.4130190908908844, | |
| "logps/chosen": -238.1641082763672, | |
| "logps/rejected": -194.319580078125, | |
| "loss": 1685.1135, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.1472087949514389, | |
| "rewards/margins": 0.141302227973938, | |
| "rewards/rejected": 0.005906577687710524, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "learning_rate": 2.4755222649153014e-06, | |
| "logits/chosen": 0.19388817250728607, | |
| "logits/rejected": 0.38251611590385437, | |
| "logps/chosen": -235.1546173095703, | |
| "logps/rejected": -187.75552368164062, | |
| "loss": 1779.3379, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.1266268938779831, | |
| "rewards/margins": 0.13590212166309357, | |
| "rewards/rejected": -0.009275219403207302, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "learning_rate": 2.4056068127914803e-06, | |
| "logits/chosen": 0.21440652012825012, | |
| "logits/rejected": 0.46250271797180176, | |
| "logps/chosen": -270.1151123046875, | |
| "logps/rejected": -226.77603149414062, | |
| "loss": 1751.5762, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.12278898805379868, | |
| "rewards/margins": 0.15427103638648987, | |
| "rewards/rejected": -0.03148204833269119, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "eval_logits/chosen": 0.39058271050453186, | |
| "eval_logits/rejected": 0.45453789830207825, | |
| "eval_logps/chosen": -271.09881591796875, | |
| "eval_logps/rejected": -248.2779541015625, | |
| "eval_loss": 2531.35498046875, | |
| "eval_rewards/accuracies": 0.5180000066757202, | |
| "eval_rewards/chosen": -0.14484626054763794, | |
| "eval_rewards/margins": 0.0025505598168820143, | |
| "eval_rewards/rejected": -0.14739681780338287, | |
| "eval_runtime": 411.7503, | |
| "eval_samples_per_second": 4.857, | |
| "eval_steps_per_second": 1.214, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "learning_rate": 2.3357652280246125e-06, | |
| "logits/chosen": 0.22305497527122498, | |
| "logits/rejected": 0.4246363043785095, | |
| "logps/chosen": -244.64987182617188, | |
| "logps/rejected": -204.69613647460938, | |
| "loss": 1997.1527, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.10293789952993393, | |
| "rewards/margins": 0.09320969879627228, | |
| "rewards/rejected": 0.009728200733661652, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "learning_rate": 2.2660521651234036e-06, | |
| "logits/chosen": 0.25884026288986206, | |
| "logits/rejected": 0.44062429666519165, | |
| "logps/chosen": -264.1467590332031, | |
| "logps/rejected": -212.3171844482422, | |
| "loss": 1703.6529, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.14418098330497742, | |
| "rewards/margins": 0.1545061618089676, | |
| "rewards/rejected": -0.0103251738473773, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 2.1965221780218173e-06, | |
| "logits/chosen": 0.16370469331741333, | |
| "logits/rejected": 0.34310778975486755, | |
| "logps/chosen": -260.98602294921875, | |
| "logps/rejected": -223.19790649414062, | |
| "loss": 1666.0893, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.14241552352905273, | |
| "rewards/margins": 0.14847032725811005, | |
| "rewards/rejected": -0.006054792553186417, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "learning_rate": 2.1272296773879107e-06, | |
| "logits/chosen": 0.14960137009620667, | |
| "logits/rejected": 0.40142565965652466, | |
| "logps/chosen": -237.9215087890625, | |
| "logps/rejected": -199.8916015625, | |
| "loss": 1738.3658, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.13602004945278168, | |
| "rewards/margins": 0.14607930183410645, | |
| "rewards/rejected": -0.010059243068099022, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "learning_rate": 2.058228888044788e-06, | |
| "logits/chosen": 0.2741914689540863, | |
| "logits/rejected": 0.3872790038585663, | |
| "logps/chosen": -223.29269409179688, | |
| "logps/rejected": -191.80650329589844, | |
| "loss": 1910.0223, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": 0.10628987848758698, | |
| "rewards/margins": 0.11196577548980713, | |
| "rewards/rejected": -0.00567590119317174, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "learning_rate": 1.989573806536978e-06, | |
| "logits/chosen": 0.18802039325237274, | |
| "logits/rejected": 0.4464651048183441, | |
| "logps/chosen": -256.7940979003906, | |
| "logps/rejected": -208.06430053710938, | |
| "loss": 1710.393, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.15197846293449402, | |
| "rewards/margins": 0.1547323763370514, | |
| "rewards/rejected": -0.002753905486315489, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "learning_rate": 1.921318158875459e-06, | |
| "logits/chosen": 0.2419842779636383, | |
| "logits/rejected": 0.4348203241825104, | |
| "logps/chosen": -249.1728973388672, | |
| "logps/rejected": -207.82229614257812, | |
| "loss": 1884.0729, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.12094120681285858, | |
| "rewards/margins": 0.12061796337366104, | |
| "rewards/rejected": 0.0003232499584555626, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 1.8535153584943915e-06, | |
| "logits/chosen": 0.2110445499420166, | |
| "logits/rejected": 0.4180677533149719, | |
| "logps/chosen": -257.97967529296875, | |
| "logps/rejected": -223.7866668701172, | |
| "loss": 1903.2082, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": 0.12379207462072372, | |
| "rewards/margins": 0.1165727972984314, | |
| "rewards/rejected": 0.007219274528324604, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "learning_rate": 1.7862184644524422e-06, | |
| "logits/chosen": 0.2049965113401413, | |
| "logits/rejected": 0.4525831639766693, | |
| "logps/chosen": -256.8290710449219, | |
| "logps/rejected": -224.7571563720703, | |
| "loss": 1728.5514, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.14060300588607788, | |
| "rewards/margins": 0.1399126946926117, | |
| "rewards/rejected": 0.0006903231260366738, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "learning_rate": 1.7194801399114471e-06, | |
| "logits/chosen": 0.1295299082994461, | |
| "logits/rejected": 0.38342705368995667, | |
| "logps/chosen": -265.95697021484375, | |
| "logps/rejected": -224.35922241210938, | |
| "loss": 1711.1711, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.1655883491039276, | |
| "rewards/margins": 0.14990000426769257, | |
| "rewards/rejected": 0.015688356012105942, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "eval_logits/chosen": 0.37592813372612, | |
| "eval_logits/rejected": 0.44024261832237244, | |
| "eval_logps/chosen": -271.4833984375, | |
| "eval_logps/rejected": -248.64402770996094, | |
| "eval_loss": 2536.2451171875, | |
| "eval_rewards/accuracies": 0.5145000219345093, | |
| "eval_rewards/chosen": -0.14869214594364166, | |
| "eval_rewards/margins": 0.0023655793629586697, | |
| "eval_rewards/rejected": -0.15105770528316498, | |
| "eval_runtime": 412.1594, | |
| "eval_samples_per_second": 4.852, | |
| "eval_steps_per_second": 1.213, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "learning_rate": 1.6533526109248632e-06, | |
| "logits/chosen": 0.22814953327178955, | |
| "logits/rejected": 0.44856709241867065, | |
| "logps/chosen": -264.6068115234375, | |
| "logps/rejected": -218.6339874267578, | |
| "loss": 1694.8418, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.1343715339899063, | |
| "rewards/margins": 0.15012042224407196, | |
| "rewards/rejected": -0.015748897567391396, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "learning_rate": 1.5878876255682951e-06, | |
| "logits/chosen": 0.25933316349983215, | |
| "logits/rejected": 0.428168922662735, | |
| "logps/chosen": -271.8929748535156, | |
| "logps/rejected": -222.19677734375, | |
| "loss": 1799.3105, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.13050588965415955, | |
| "rewards/margins": 0.12607057392597198, | |
| "rewards/rejected": 0.00443531759083271, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "learning_rate": 1.5231364134440485e-06, | |
| "logits/chosen": 0.16305533051490784, | |
| "logits/rejected": 0.37009209394454956, | |
| "logps/chosen": -239.7517547607422, | |
| "logps/rejected": -201.59359741210938, | |
| "loss": 1877.6279, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.13841071724891663, | |
| "rewards/margins": 0.12821929156780243, | |
| "rewards/rejected": 0.0101914182305336, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "learning_rate": 1.4591496455914292e-06, | |
| "logits/chosen": 0.23846232891082764, | |
| "logits/rejected": 0.44133177399635315, | |
| "logps/chosen": -269.0438232421875, | |
| "logps/rejected": -235.73202514648438, | |
| "loss": 1826.1201, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.12339715659618378, | |
| "rewards/margins": 0.12225867807865143, | |
| "rewards/rejected": 0.0011384893441572785, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "learning_rate": 1.395977394834132e-06, | |
| "logits/chosen": 0.2057628184556961, | |
| "logits/rejected": 0.457078218460083, | |
| "logps/chosen": -222.45474243164062, | |
| "logps/rejected": -178.09011840820312, | |
| "loss": 1674.3545, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.14612336456775665, | |
| "rewards/margins": 0.1515127718448639, | |
| "rewards/rejected": -0.005389401223510504, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 1.3336690965957733e-06, | |
| "logits/chosen": 0.23303177952766418, | |
| "logits/rejected": 0.38948389887809753, | |
| "logps/chosen": -276.028564453125, | |
| "logps/rejected": -242.33120727539062, | |
| "loss": 1734.6137, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.13224345445632935, | |
| "rewards/margins": 0.1367109715938568, | |
| "rewards/rejected": -0.004467502236366272, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 1.2722735102142192e-06, | |
| "logits/chosen": 0.18385744094848633, | |
| "logits/rejected": 0.483973890542984, | |
| "logps/chosen": -234.6072540283203, | |
| "logps/rejected": -188.79147338867188, | |
| "loss": 1524.1409, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": 0.1483391523361206, | |
| "rewards/margins": 0.16734978556632996, | |
| "rewards/rejected": -0.019010629504919052, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "learning_rate": 1.2118386807849733e-06, | |
| "logits/chosen": 0.23927922546863556, | |
| "logits/rejected": 0.39777225255966187, | |
| "logps/chosen": -247.7241668701172, | |
| "logps/rejected": -209.2562255859375, | |
| "loss": 1804.3422, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": 0.11453523486852646, | |
| "rewards/margins": 0.13239163160324097, | |
| "rewards/rejected": -0.0178564190864563, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "learning_rate": 1.1524119015635116e-06, | |
| "logits/chosen": 0.14219129085540771, | |
| "logits/rejected": 0.3456658124923706, | |
| "logps/chosen": -244.8931121826172, | |
| "logps/rejected": -194.36337280273438, | |
| "loss": 1655.0236, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": 0.17296788096427917, | |
| "rewards/margins": 0.17738190293312073, | |
| "rewards/rejected": -0.004414013121277094, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "learning_rate": 1.0940396769559584e-06, | |
| "logits/chosen": 0.23592603206634521, | |
| "logits/rejected": 0.3958457410335541, | |
| "logps/chosen": -242.33761596679688, | |
| "logps/rejected": -191.34255981445312, | |
| "loss": 1894.4447, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.11816737800836563, | |
| "rewards/margins": 0.1154218316078186, | |
| "rewards/rejected": 0.002745547564700246, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "eval_logits/chosen": 0.3617529273033142, | |
| "eval_logits/rejected": 0.42615193128585815, | |
| "eval_logps/chosen": -272.0999755859375, | |
| "eval_logps/rejected": -249.2417449951172, | |
| "eval_loss": 2542.6298828125, | |
| "eval_rewards/accuracies": 0.5235000252723694, | |
| "eval_rewards/chosen": -0.15485772490501404, | |
| "eval_rewards/margins": 0.002177263842895627, | |
| "eval_rewards/rejected": -0.1570349782705307, | |
| "eval_runtime": 411.8184, | |
| "eval_samples_per_second": 4.857, | |
| "eval_steps_per_second": 1.214, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "learning_rate": 1.036767686127079e-06, | |
| "logits/chosen": 0.26092058420181274, | |
| "logits/rejected": 0.4202706217765808, | |
| "logps/chosen": -271.7019958496094, | |
| "logps/rejected": -229.44772338867188, | |
| "loss": 1612.1519, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.1469217836856842, | |
| "rewards/margins": 0.1562378704547882, | |
| "rewards/rejected": -0.009316088631749153, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "learning_rate": 9.806407472540644e-07, | |
| "logits/chosen": 0.16799505054950714, | |
| "logits/rejected": 0.41259631514549255, | |
| "logps/chosen": -245.1134033203125, | |
| "logps/rejected": -193.49716186523438, | |
| "loss": 1752.974, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.1420082449913025, | |
| "rewards/margins": 0.16994443535804749, | |
| "rewards/rejected": -0.027936194092035294, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "learning_rate": 9.257027824540823e-07, | |
| "logits/chosen": 0.15648850798606873, | |
| "logits/rejected": 0.3334013819694519, | |
| "logps/chosen": -237.5020294189453, | |
| "logps/rejected": -190.00733947753906, | |
| "loss": 1678.6684, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.1168346032500267, | |
| "rewards/margins": 0.14195187389850616, | |
| "rewards/rejected": -0.02511727437376976, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.01, | |
| "learning_rate": 8.719967834130385e-07, | |
| "logits/chosen": 0.17559798061847687, | |
| "logits/rejected": 0.38583293557167053, | |
| "logps/chosen": -226.3419647216797, | |
| "logps/rejected": -192.8260498046875, | |
| "loss": 1782.9469, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.13087774813175201, | |
| "rewards/margins": 0.14290814101696014, | |
| "rewards/rejected": -0.012030376121401787, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 3.04, | |
| "learning_rate": 8.195647777424479e-07, | |
| "logits/chosen": 0.19569836556911469, | |
| "logits/rejected": 0.420011043548584, | |
| "logps/chosen": -243.2490692138672, | |
| "logps/rejected": -200.2803497314453, | |
| "loss": 1728.6527, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.1271076500415802, | |
| "rewards/margins": 0.14936277270317078, | |
| "rewards/rejected": -0.02225511521100998, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 3.07, | |
| "learning_rate": 7.684477960907422e-07, | |
| "logits/chosen": 0.2223655879497528, | |
| "logits/rejected": 0.36084768176078796, | |
| "logps/chosen": -259.83001708984375, | |
| "logps/rejected": -217.5656280517578, | |
| "loss": 1787.5324, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.13919194042682648, | |
| "rewards/margins": 0.12921075522899628, | |
| "rewards/rejected": 0.009981190785765648, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 3.1, | |
| "learning_rate": 7.186858400347455e-07, | |
| "logits/chosen": 0.15693505108356476, | |
| "logits/rejected": 0.35678160190582275, | |
| "logps/chosen": -266.93768310546875, | |
| "logps/rejected": -233.291259765625, | |
| "loss": 1748.1148, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.12252594530582428, | |
| "rewards/margins": 0.13253001868724823, | |
| "rewards/rejected": -0.010004105977714062, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 3.14, | |
| "learning_rate": 6.703178507764618e-07, | |
| "logits/chosen": 0.21335263550281525, | |
| "logits/rejected": 0.3721942901611328, | |
| "logps/chosen": -250.0565948486328, | |
| "logps/rejected": -213.24624633789062, | |
| "loss": 1792.84, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.1454544961452484, | |
| "rewards/margins": 0.12651817500591278, | |
| "rewards/rejected": 0.018936317414045334, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "learning_rate": 6.233816786696414e-07, | |
| "logits/chosen": 0.17327558994293213, | |
| "logits/rejected": 0.41259893774986267, | |
| "logps/chosen": -242.4098358154297, | |
| "logps/rejected": -191.53460693359375, | |
| "loss": 1636.6745, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.16031518578529358, | |
| "rewards/margins": 0.1668807417154312, | |
| "rewards/rejected": -0.006565576884895563, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 3.2, | |
| "learning_rate": 5.77914053600005e-07, | |
| "logits/chosen": 0.15771836042404175, | |
| "logits/rejected": 0.384365975856781, | |
| "logps/chosen": -247.66213989257812, | |
| "logps/rejected": -207.638427734375, | |
| "loss": 1798.5389, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.13780926167964935, | |
| "rewards/margins": 0.1279788464307785, | |
| "rewards/rejected": 0.00983042549341917, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.2, | |
| "eval_logits/chosen": 0.3559388518333435, | |
| "eval_logits/rejected": 0.42022594809532166, | |
| "eval_logps/chosen": -272.4200439453125, | |
| "eval_logps/rejected": -249.57797241210938, | |
| "eval_loss": 2542.728759765625, | |
| "eval_rewards/accuracies": 0.5205000042915344, | |
| "eval_rewards/chosen": -0.15805859863758087, | |
| "eval_rewards/margins": 0.0023383116349577904, | |
| "eval_rewards/rejected": -0.16039690375328064, | |
| "eval_runtime": 411.7713, | |
| "eval_samples_per_second": 4.857, | |
| "eval_steps_per_second": 1.214, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "learning_rate": 5.339505562422851e-07, | |
| "logits/chosen": 0.18551324307918549, | |
| "logits/rejected": 0.3874700367450714, | |
| "logps/chosen": -258.8070983886719, | |
| "logps/rejected": -208.31655883789062, | |
| "loss": 1684.7502, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.13547667860984802, | |
| "rewards/margins": 0.14734682440757751, | |
| "rewards/rejected": -0.011870155110955238, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 3.26, | |
| "learning_rate": 4.915255902165734e-07, | |
| "logits/chosen": 0.16962969303131104, | |
| "logits/rejected": 0.3440442383289337, | |
| "logps/chosen": -257.40386962890625, | |
| "logps/rejected": -213.15682983398438, | |
| "loss": 1727.4756, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.1502576470375061, | |
| "rewards/margins": 0.15001319348812103, | |
| "rewards/rejected": 0.000244450056925416, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 3.3, | |
| "learning_rate": 4.506723551657879e-07, | |
| "logits/chosen": 0.08985213190317154, | |
| "logits/rejected": 0.3018781542778015, | |
| "logps/chosen": -238.0452117919922, | |
| "logps/rejected": -200.73129272460938, | |
| "loss": 1795.9609, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.1125258356332779, | |
| "rewards/margins": 0.11527595669031143, | |
| "rewards/rejected": -0.0027501187287271023, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "learning_rate": 4.11422820775299e-07, | |
| "logits/chosen": 0.18122617900371552, | |
| "logits/rejected": 0.3776589035987854, | |
| "logps/chosen": -273.4202575683594, | |
| "logps/rejected": -214.76803588867188, | |
| "loss": 1575.2917, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.14816126227378845, | |
| "rewards/margins": 0.1801377683877945, | |
| "rewards/rejected": -0.03197649493813515, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 3.36, | |
| "learning_rate": 3.7380770175506397e-07, | |
| "logits/chosen": 0.18917641043663025, | |
| "logits/rejected": 0.3701549172401428, | |
| "logps/chosen": -236.4595184326172, | |
| "logps/rejected": -196.47378540039062, | |
| "loss": 1757.2316, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.13911323249340057, | |
| "rewards/margins": 0.14288835227489471, | |
| "rewards/rejected": -0.0037751286290585995, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 3.39, | |
| "learning_rate": 3.3785643380384063e-07, | |
| "logits/chosen": 0.22518393397331238, | |
| "logits/rejected": 0.4300554394721985, | |
| "logps/chosen": -264.11102294921875, | |
| "logps/rejected": -221.0089569091797, | |
| "loss": 1780.259, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.12094493955373764, | |
| "rewards/margins": 0.13927613198757172, | |
| "rewards/rejected": -0.018331199884414673, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "learning_rate": 3.0359715057429186e-07, | |
| "logits/chosen": 0.15539857745170593, | |
| "logits/rejected": 0.330243319272995, | |
| "logps/chosen": -249.2795867919922, | |
| "logps/rejected": -213.9594268798828, | |
| "loss": 1857.3367, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.11874745041131973, | |
| "rewards/margins": 0.1254003345966339, | |
| "rewards/rejected": -0.0066528706811368465, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 3.46, | |
| "learning_rate": 2.710566616570048e-07, | |
| "logits/chosen": 0.22669236361980438, | |
| "logits/rejected": 0.3851965069770813, | |
| "logps/chosen": -243.06094360351562, | |
| "logps/rejected": -207.26522827148438, | |
| "loss": 1802.6373, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.11891194432973862, | |
| "rewards/margins": 0.1296943724155426, | |
| "rewards/rejected": -0.010782415047287941, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 3.49, | |
| "learning_rate": 2.40260431600654e-07, | |
| "logits/chosen": 0.11191525310277939, | |
| "logits/rejected": 0.3566216826438904, | |
| "logps/chosen": -258.7408142089844, | |
| "logps/rejected": -233.2893524169922, | |
| "loss": 1846.4627, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.13531699776649475, | |
| "rewards/margins": 0.1306256651878357, | |
| "rewards/rejected": 0.004691324662417173, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 3.52, | |
| "learning_rate": 2.1123255998472952e-07, | |
| "logits/chosen": 0.13580968976020813, | |
| "logits/rejected": 0.3724610209465027, | |
| "logps/chosen": -238.85391235351562, | |
| "logps/rejected": -202.35397338867188, | |
| "loss": 1834.9711, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.1293712556362152, | |
| "rewards/margins": 0.12213204801082611, | |
| "rewards/rejected": 0.007239216007292271, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.52, | |
| "eval_logits/chosen": 0.3531816899776459, | |
| "eval_logits/rejected": 0.417733371257782, | |
| "eval_logps/chosen": -272.4703369140625, | |
| "eval_logps/rejected": -249.63449096679688, | |
| "eval_loss": 2542.2373046875, | |
| "eval_rewards/accuracies": 0.5205000042915344, | |
| "eval_rewards/chosen": -0.15856170654296875, | |
| "eval_rewards/margins": 0.0024005102459341288, | |
| "eval_rewards/rejected": -0.16096222400665283, | |
| "eval_runtime": 411.8208, | |
| "eval_samples_per_second": 4.856, | |
| "eval_steps_per_second": 1.214, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.55, | |
| "learning_rate": 1.8399576256041525e-07, | |
| "logits/chosen": 0.27907487750053406, | |
| "logits/rejected": 0.39952999353408813, | |
| "logps/chosen": -293.59210205078125, | |
| "logps/rejected": -251.298583984375, | |
| "loss": 1979.5523, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.12343426048755646, | |
| "rewards/margins": 0.1068718284368515, | |
| "rewards/rejected": 0.016562417149543762, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 3.58, | |
| "learning_rate": 1.58571353474391e-07, | |
| "logits/chosen": 0.22743897140026093, | |
| "logits/rejected": 0.40824824571609497, | |
| "logps/chosen": -248.27365112304688, | |
| "logps/rejected": -217.0345916748047, | |
| "loss": 1971.9447, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.10497580468654633, | |
| "rewards/margins": 0.10696852207183838, | |
| "rewards/rejected": -0.0019927166868001223, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 3.62, | |
| "learning_rate": 1.3497922858944857e-07, | |
| "logits/chosen": 0.21788544952869415, | |
| "logits/rejected": 0.3408287763595581, | |
| "logps/chosen": -261.25634765625, | |
| "logps/rejected": -226.1531982421875, | |
| "loss": 1813.0029, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.12613138556480408, | |
| "rewards/margins": 0.13940462470054626, | |
| "rewards/rejected": -0.013273234479129314, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 3.65, | |
| "learning_rate": 1.1323784991499471e-07, | |
| "logits/chosen": 0.17835107445716858, | |
| "logits/rejected": 0.41546598076820374, | |
| "logps/chosen": -254.427978515625, | |
| "logps/rejected": -208.0858917236328, | |
| "loss": 1688.9838, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.13813026249408722, | |
| "rewards/margins": 0.1528724581003189, | |
| "rewards/rejected": -0.014742200262844563, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 3.68, | |
| "learning_rate": 9.336423115961002e-08, | |
| "logits/chosen": 0.13116911053657532, | |
| "logits/rejected": 0.34874704480171204, | |
| "logps/chosen": -251.4066162109375, | |
| "logps/rejected": -205.0511016845703, | |
| "loss": 1693.0176, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.14480547606945038, | |
| "rewards/margins": 0.1530844271183014, | |
| "rewards/rejected": -0.008278947323560715, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "learning_rate": 7.537392441697793e-08, | |
| "logits/chosen": 0.1982724368572235, | |
| "logits/rejected": 0.3957718014717102, | |
| "logps/chosen": -265.66033935546875, | |
| "logps/rejected": -223.394775390625, | |
| "loss": 1694.5328, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.12068979442119598, | |
| "rewards/margins": 0.15094828605651855, | |
| "rewards/rejected": -0.030258500948548317, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 3.74, | |
| "learning_rate": 5.928100799559938e-08, | |
| "logits/chosen": 0.2445194274187088, | |
| "logits/rejected": 0.34973251819610596, | |
| "logps/chosen": -266.8522033691406, | |
| "logps/rejected": -214.7125701904297, | |
| "loss": 1810.5473, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.12299786508083344, | |
| "rewards/margins": 0.13123488426208496, | |
| "rewards/rejected": -0.008237012661993504, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 3.78, | |
| "learning_rate": 4.5098075401815435e-08, | |
| "logits/chosen": 0.17370513081550598, | |
| "logits/rejected": 0.44302305579185486, | |
| "logps/chosen": -251.4030303955078, | |
| "logps/rejected": -217.8542938232422, | |
| "loss": 1799.0785, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.14102791249752045, | |
| "rewards/margins": 0.14266334474086761, | |
| "rewards/rejected": -0.0016354346880689263, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 3.81, | |
| "learning_rate": 3.283622548476445e-08, | |
| "logits/chosen": 0.2375222146511078, | |
| "logits/rejected": 0.45388326048851013, | |
| "logps/chosen": -243.04116821289062, | |
| "logps/rejected": -203.55050659179688, | |
| "loss": 1673.1176, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": 0.13139726221561432, | |
| "rewards/margins": 0.14136573672294617, | |
| "rewards/rejected": -0.009968474507331848, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 3.84, | |
| "learning_rate": 2.250505375098161e-08, | |
| "logits/chosen": 0.28876617550849915, | |
| "logits/rejected": 0.42698416113853455, | |
| "logps/chosen": -248.3870086669922, | |
| "logps/rejected": -211.6997833251953, | |
| "loss": 1765.5148, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.1342637687921524, | |
| "rewards/margins": 0.1372515708208084, | |
| "rewards/rejected": -0.002987801330164075, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.84, | |
| "eval_logits/chosen": 0.35145023465156555, | |
| "eval_logits/rejected": 0.41600194573402405, | |
| "eval_logps/chosen": -272.50103759765625, | |
| "eval_logps/rejected": -249.63572692871094, | |
| "eval_loss": 2546.17138671875, | |
| "eval_rewards/accuracies": 0.5220000147819519, | |
| "eval_rewards/chosen": -0.15886859595775604, | |
| "eval_rewards/margins": 0.00210593082010746, | |
| "eval_rewards/rejected": -0.16097451746463776, | |
| "eval_runtime": 411.6701, | |
| "eval_samples_per_second": 4.858, | |
| "eval_steps_per_second": 1.215, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "learning_rate": 1.4112644855438228e-08, | |
| "logits/chosen": 0.18522273004055023, | |
| "logits/rejected": 0.3528333306312561, | |
| "logps/chosen": -252.374755859375, | |
| "logps/rejected": -213.50112915039062, | |
| "loss": 1695.0207, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.1518661379814148, | |
| "rewards/margins": 0.1503283679485321, | |
| "rewards/rejected": 0.001537766307592392, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "learning_rate": 7.665566274897007e-09, | |
| "logits/chosen": 0.14011794328689575, | |
| "logits/rejected": 0.45459121465682983, | |
| "logps/chosen": -269.1119079589844, | |
| "logps/rejected": -213.98684692382812, | |
| "loss": 1599.6027, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.1741984784603119, | |
| "rewards/margins": 0.17417994141578674, | |
| "rewards/rejected": 1.8526614439906552e-05, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 3.94, | |
| "learning_rate": 3.1688631685364292e-09, | |
| "logits/chosen": 0.11093461513519287, | |
| "logits/rejected": 0.3354471027851105, | |
| "logps/chosen": -240.45419311523438, | |
| "logps/rejected": -200.0811004638672, | |
| "loss": 1762.0594, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.14106154441833496, | |
| "rewards/margins": 0.14005112648010254, | |
| "rewards/rejected": 0.0010104203829541802, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "learning_rate": 6.260544298619664e-10, | |
| "logits/chosen": 0.1383497416973114, | |
| "logits/rejected": 0.3753938674926758, | |
| "logps/chosen": -235.23385620117188, | |
| "logps/rejected": -191.31582641601562, | |
| "loss": 1685.3523, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.15916982293128967, | |
| "rewards/margins": 0.1575343906879425, | |
| "rewards/rejected": 0.0016354434192180634, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 3.99, | |
| "step": 1248, | |
| "total_flos": 0.0, | |
| "train_loss": 1918.4646684695513, | |
| "train_runtime": 14252.7073, | |
| "train_samples_per_second": 1.403, | |
| "train_steps_per_second": 0.088 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1248, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |