| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 100, | |
| "global_step": 478, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": -2.7515339851379395, | |
| "logits/rejected": -2.7739691734313965, | |
| "logps/chosen": -226.12843322753906, | |
| "logps/rejected": -195.60536193847656, | |
| "loss": 0.3749, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -2.7849607467651367, | |
| "logits/rejected": -2.7732653617858887, | |
| "logps/chosen": -296.5021667480469, | |
| "logps/rejected": -257.125732421875, | |
| "loss": 0.3841, | |
| "rewards/accuracies": 0.4166666567325592, | |
| "rewards/chosen": 9.646587386669125e-06, | |
| "rewards/margins": -4.828277451451868e-05, | |
| "rewards/rejected": 5.7929337344830856e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -2.7799129486083984, | |
| "logits/rejected": -2.7518584728240967, | |
| "logps/chosen": -267.0304870605469, | |
| "logps/rejected": -265.03131103515625, | |
| "loss": 0.3837, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -1.761986277415417e-05, | |
| "rewards/margins": 0.0019770762883126736, | |
| "rewards/rejected": -0.001994695980101824, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.7776148319244385, | |
| "logits/rejected": -2.7521443367004395, | |
| "logps/chosen": -246.96499633789062, | |
| "logps/rejected": -252.1887664794922, | |
| "loss": 0.3827, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.000771665945649147, | |
| "rewards/margins": 0.007972861640155315, | |
| "rewards/rejected": -0.008744527585804462, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.759718418121338, | |
| "logits/rejected": -2.7176804542541504, | |
| "logps/chosen": -276.6578674316406, | |
| "logps/rejected": -259.2043151855469, | |
| "loss": 0.3725, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.01379473228007555, | |
| "rewards/margins": 0.010001211427152157, | |
| "rewards/rejected": -0.023795943707227707, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.999733114418725e-07, | |
| "logits/chosen": -2.740699291229248, | |
| "logits/rejected": -2.732423782348633, | |
| "logps/chosen": -310.40411376953125, | |
| "logps/rejected": -309.11676025390625, | |
| "loss": 0.3717, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.020774774253368378, | |
| "rewards/margins": 0.06383714824914932, | |
| "rewards/rejected": -0.0846119225025177, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.990398100856366e-07, | |
| "logits/chosen": -2.707352638244629, | |
| "logits/rejected": -2.6780617237091064, | |
| "logps/chosen": -292.3264465332031, | |
| "logps/rejected": -283.150634765625, | |
| "loss": 0.3518, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.09367800503969193, | |
| "rewards/margins": 0.06513573229312897, | |
| "rewards/rejected": -0.1588137447834015, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.967775735898179e-07, | |
| "logits/chosen": -2.7030491828918457, | |
| "logits/rejected": -2.6767783164978027, | |
| "logps/chosen": -272.22491455078125, | |
| "logps/rejected": -275.2875671386719, | |
| "loss": 0.331, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.18802210688591003, | |
| "rewards/margins": 0.11484960466623306, | |
| "rewards/rejected": -0.3028717637062073, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.931986719649298e-07, | |
| "logits/chosen": -2.7163443565368652, | |
| "logits/rejected": -2.7092785835266113, | |
| "logps/chosen": -290.34381103515625, | |
| "logps/rejected": -307.81549072265625, | |
| "loss": 0.2975, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.30055341124534607, | |
| "rewards/margins": 0.24591656029224396, | |
| "rewards/rejected": -0.5464699864387512, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.883222001996351e-07, | |
| "logits/chosen": -2.7387337684631348, | |
| "logits/rejected": -2.747392416000366, | |
| "logps/chosen": -279.77520751953125, | |
| "logps/rejected": -291.8474426269531, | |
| "loss": 0.2752, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.40700283646583557, | |
| "rewards/margins": 0.30104246735572815, | |
| "rewards/rejected": -0.7080453634262085, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.821741763807186e-07, | |
| "logits/chosen": -2.7377610206604004, | |
| "logits/rejected": -2.7128090858459473, | |
| "logps/chosen": -286.3442687988281, | |
| "logps/rejected": -326.53643798828125, | |
| "loss": 0.2667, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.4444156587123871, | |
| "rewards/margins": 0.26415666937828064, | |
| "rewards/rejected": -0.708572268486023, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": -2.7741379737854004, | |
| "eval_logits/rejected": -2.7592568397521973, | |
| "eval_logps/chosen": -302.3351745605469, | |
| "eval_logps/rejected": -336.5635986328125, | |
| "eval_loss": 0.26699498295783997, | |
| "eval_rewards/accuracies": 0.6796875, | |
| "eval_rewards/chosen": -0.4529554545879364, | |
| "eval_rewards/margins": 0.3391478657722473, | |
| "eval_rewards/rejected": -0.7921034097671509, | |
| "eval_runtime": 53.381, | |
| "eval_samples_per_second": 37.466, | |
| "eval_steps_per_second": 0.599, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.747874028753375e-07, | |
| "logits/chosen": -2.7634060382843018, | |
| "logits/rejected": -2.7215521335601807, | |
| "logps/chosen": -326.7304382324219, | |
| "logps/rejected": -301.48797607421875, | |
| "loss": 0.2584, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.5723937749862671, | |
| "rewards/margins": 0.20699183642864227, | |
| "rewards/rejected": -0.7793856263160706, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.662012913161997e-07, | |
| "logits/chosen": -2.7727789878845215, | |
| "logits/rejected": -2.747377872467041, | |
| "logps/chosen": -332.1087951660156, | |
| "logps/rejected": -324.8017883300781, | |
| "loss": 0.2269, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.5284228920936584, | |
| "rewards/margins": 0.3542647063732147, | |
| "rewards/rejected": -0.8826876878738403, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.5646165232345103e-07, | |
| "logits/chosen": -2.7477874755859375, | |
| "logits/rejected": -2.763334274291992, | |
| "logps/chosen": -318.76171875, | |
| "logps/rejected": -344.4560852050781, | |
| "loss": 0.2668, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.4845491051673889, | |
| "rewards/margins": 0.34464871883392334, | |
| "rewards/rejected": -0.829197883605957, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.456204510851956e-07, | |
| "logits/chosen": -2.707592248916626, | |
| "logits/rejected": -2.695403575897217, | |
| "logps/chosen": -314.1436462402344, | |
| "logps/rejected": -335.11773681640625, | |
| "loss": 0.2446, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.5327329039573669, | |
| "rewards/margins": 0.3184421956539154, | |
| "rewards/rejected": -0.8511751294136047, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.337355301007335e-07, | |
| "logits/chosen": -2.7098329067230225, | |
| "logits/rejected": -2.701035976409912, | |
| "logps/chosen": -322.45220947265625, | |
| "logps/rejected": -355.2679748535156, | |
| "loss": 0.227, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.6261399984359741, | |
| "rewards/margins": 0.47124552726745605, | |
| "rewards/rejected": -1.0973855257034302, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.2087030056579986e-07, | |
| "logits/chosen": -2.603577136993408, | |
| "logits/rejected": -2.5530428886413574, | |
| "logps/chosen": -349.7845764160156, | |
| "logps/rejected": -361.4437561035156, | |
| "loss": 0.2254, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.7117608189582825, | |
| "rewards/margins": 0.37124234437942505, | |
| "rewards/rejected": -1.083003282546997, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.070934040463998e-07, | |
| "logits/chosen": -2.6051487922668457, | |
| "logits/rejected": -2.566930055618286, | |
| "logps/chosen": -350.85601806640625, | |
| "logps/rejected": -397.63140869140625, | |
| "loss": 0.2194, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.7997169494628906, | |
| "rewards/margins": 0.4845522940158844, | |
| "rewards/rejected": -1.2842693328857422, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.9247834624635404e-07, | |
| "logits/chosen": -2.6019983291625977, | |
| "logits/rejected": -2.587660312652588, | |
| "logps/chosen": -366.35992431640625, | |
| "logps/rejected": -348.6190185546875, | |
| "loss": 0.2364, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.6343470215797424, | |
| "rewards/margins": 0.3612142503261566, | |
| "rewards/rejected": -0.9955613017082214, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.7710310482256523e-07, | |
| "logits/chosen": -2.6051242351531982, | |
| "logits/rejected": -2.577193021774292, | |
| "logps/chosen": -329.8570556640625, | |
| "logps/rejected": -313.30438232421875, | |
| "loss": 0.2243, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.6197746396064758, | |
| "rewards/margins": 0.35030925273895264, | |
| "rewards/rejected": -0.9700838327407837, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.610497133404795e-07, | |
| "logits/chosen": -2.5546486377716064, | |
| "logits/rejected": -2.511190176010132, | |
| "logps/chosen": -349.89056396484375, | |
| "logps/rejected": -380.19647216796875, | |
| "loss": 0.2068, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.8633190393447876, | |
| "rewards/margins": 0.445193350315094, | |
| "rewards/rejected": -1.3085120916366577, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": -2.5673255920410156, | |
| "eval_logits/rejected": -2.5512256622314453, | |
| "eval_logps/chosen": -340.4659729003906, | |
| "eval_logps/rejected": -394.05877685546875, | |
| "eval_loss": 0.2086561620235443, | |
| "eval_rewards/accuracies": 0.68359375, | |
| "eval_rewards/chosen": -0.8342636823654175, | |
| "eval_rewards/margins": 0.532791256904602, | |
| "eval_rewards/rejected": -1.3670549392700195, | |
| "eval_runtime": 53.3257, | |
| "eval_samples_per_second": 37.505, | |
| "eval_steps_per_second": 0.6, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.4440382358952115e-07, | |
| "logits/chosen": -2.49550199508667, | |
| "logits/rejected": -2.4403936862945557, | |
| "logps/chosen": -356.9299011230469, | |
| "logps/rejected": -390.6685791015625, | |
| "loss": 0.1834, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.8836824297904968, | |
| "rewards/margins": 0.6037726998329163, | |
| "rewards/rejected": -1.487455129623413, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.272542485937368e-07, | |
| "logits/chosen": -2.481822967529297, | |
| "logits/rejected": -2.4627151489257812, | |
| "logps/chosen": -325.0804138183594, | |
| "logps/rejected": -375.494873046875, | |
| "loss": 0.1975, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.8965877294540405, | |
| "rewards/margins": 0.5446338653564453, | |
| "rewards/rejected": -1.4412214756011963, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.096924887558854e-07, | |
| "logits/chosen": -2.5111708641052246, | |
| "logits/rejected": -2.5036470890045166, | |
| "logps/chosen": -354.4523010253906, | |
| "logps/rejected": -382.4290466308594, | |
| "loss": 0.216, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.7817985415458679, | |
| "rewards/margins": 0.5005447268486023, | |
| "rewards/rejected": -1.2823432683944702, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.9181224366319943e-07, | |
| "logits/chosen": -2.5602169036865234, | |
| "logits/rejected": -2.5355300903320312, | |
| "logps/chosen": -356.17913818359375, | |
| "logps/rejected": -396.00250244140625, | |
| "loss": 0.1939, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.9149402379989624, | |
| "rewards/margins": 0.42837992310523987, | |
| "rewards/rejected": -1.3433201313018799, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.7370891215954565e-07, | |
| "logits/chosen": -2.5077805519104004, | |
| "logits/rejected": -2.4881882667541504, | |
| "logps/chosen": -377.3251953125, | |
| "logps/rejected": -386.83270263671875, | |
| "loss": 0.209, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.9626044034957886, | |
| "rewards/margins": 0.3591587543487549, | |
| "rewards/rejected": -1.321763277053833, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.55479083351317e-07, | |
| "logits/chosen": -2.534468173980713, | |
| "logits/rejected": -2.5206878185272217, | |
| "logps/chosen": -350.94195556640625, | |
| "logps/rejected": -361.4455261230469, | |
| "loss": 0.2209, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.7781239748001099, | |
| "rewards/margins": 0.5087354779243469, | |
| "rewards/rejected": -1.2868595123291016, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.3722002126275822e-07, | |
| "logits/chosen": -2.4776508808135986, | |
| "logits/rejected": -2.455139398574829, | |
| "logps/chosen": -372.4471435546875, | |
| "logps/rejected": -390.30316162109375, | |
| "loss": 0.2018, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.8379491567611694, | |
| "rewards/margins": 0.4740463197231293, | |
| "rewards/rejected": -1.311995506286621, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.19029145890313e-07, | |
| "logits/chosen": -2.4426798820495605, | |
| "logits/rejected": -2.421225070953369, | |
| "logps/chosen": -383.3365478515625, | |
| "logps/rejected": -408.34246826171875, | |
| "loss": 0.1931, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.0771123170852661, | |
| "rewards/margins": 0.5110739469528198, | |
| "rewards/rejected": -1.5881863832473755, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 2.0100351342479216e-07, | |
| "logits/chosen": -2.4331817626953125, | |
| "logits/rejected": -2.4107987880706787, | |
| "logps/chosen": -416.7959899902344, | |
| "logps/rejected": -442.81951904296875, | |
| "loss": 0.1776, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.1279302835464478, | |
| "rewards/margins": 0.5917536020278931, | |
| "rewards/rejected": -1.7196838855743408, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.8323929841460178e-07, | |
| "logits/chosen": -2.44761323928833, | |
| "logits/rejected": -2.425476312637329, | |
| "logps/chosen": -379.3946838378906, | |
| "logps/rejected": -423.1104431152344, | |
| "loss": 0.2095, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.083051085472107, | |
| "rewards/margins": 0.5393560528755188, | |
| "rewards/rejected": -1.6224071979522705, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": -2.479083776473999, | |
| "eval_logits/rejected": -2.464531421661377, | |
| "eval_logps/chosen": -340.87713623046875, | |
| "eval_logps/rejected": -401.1194152832031, | |
| "eval_loss": 0.22332347929477692, | |
| "eval_rewards/accuracies": 0.7109375, | |
| "eval_rewards/chosen": -0.8383752107620239, | |
| "eval_rewards/margins": 0.5992860794067383, | |
| "eval_rewards/rejected": -1.4376611709594727, | |
| "eval_runtime": 53.3117, | |
| "eval_samples_per_second": 37.515, | |
| "eval_steps_per_second": 0.6, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.6583128063291573e-07, | |
| "logits/chosen": -2.459578037261963, | |
| "logits/rejected": -2.4456562995910645, | |
| "logps/chosen": -363.7841796875, | |
| "logps/rejected": -435.59478759765625, | |
| "loss": 0.2183, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8949993848800659, | |
| "rewards/margins": 0.6970807313919067, | |
| "rewards/rejected": -1.592079997062683, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.488723393865766e-07, | |
| "logits/chosen": -2.4861679077148438, | |
| "logits/rejected": -2.437126874923706, | |
| "logps/chosen": -382.5135498046875, | |
| "logps/rejected": -415.387939453125, | |
| "loss": 0.2205, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.8941608667373657, | |
| "rewards/margins": 0.5314512252807617, | |
| "rewards/rejected": -1.4256120920181274, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3245295796480788e-07, | |
| "logits/chosen": -2.408893346786499, | |
| "logits/rejected": -2.405259609222412, | |
| "logps/chosen": -369.69049072265625, | |
| "logps/rejected": -420.62384033203125, | |
| "loss": 0.211, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.9662488102912903, | |
| "rewards/margins": 0.5407388210296631, | |
| "rewards/rejected": -1.5069875717163086, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.1666074087171627e-07, | |
| "logits/chosen": -2.400749444961548, | |
| "logits/rejected": -2.397270917892456, | |
| "logps/chosen": -420.1078186035156, | |
| "logps/rejected": -454.15069580078125, | |
| "loss": 0.2107, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.1604535579681396, | |
| "rewards/margins": 0.46516379714012146, | |
| "rewards/rejected": -1.625617265701294, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0157994641835734e-07, | |
| "logits/chosen": -2.448854684829712, | |
| "logits/rejected": -2.404040813446045, | |
| "logps/chosen": -332.3216247558594, | |
| "logps/rejected": -371.5892028808594, | |
| "loss": 0.208, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.8533551096916199, | |
| "rewards/margins": 0.6172915697097778, | |
| "rewards/rejected": -1.470646619796753, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.729103716819111e-08, | |
| "logits/chosen": -2.3937747478485107, | |
| "logits/rejected": -2.3905582427978516, | |
| "logps/chosen": -368.26849365234375, | |
| "logps/rejected": -392.41937255859375, | |
| "loss": 0.2182, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.8935198783874512, | |
| "rewards/margins": 0.5953238010406494, | |
| "rewards/rejected": -1.4888436794281006, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.387025063449081e-08, | |
| "logits/chosen": -2.410921096801758, | |
| "logits/rejected": -2.3979973793029785, | |
| "logps/chosen": -351.77972412109375, | |
| "logps/rejected": -412.9110412597656, | |
| "loss": 0.2042, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.9797185063362122, | |
| "rewards/margins": 0.6103909611701965, | |
| "rewards/rejected": -1.5901095867156982, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 6.138919252022435e-08, | |
| "logits/chosen": -2.4055371284484863, | |
| "logits/rejected": -2.365029811859131, | |
| "logps/chosen": -429.22540283203125, | |
| "logps/rejected": -424.9950256347656, | |
| "loss": 0.211, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.043166160583496, | |
| "rewards/margins": 0.5213008522987366, | |
| "rewards/rejected": -1.5644668340682983, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.991445467064689e-08, | |
| "logits/chosen": -2.3971471786499023, | |
| "logits/rejected": -2.3788516521453857, | |
| "logps/chosen": -353.85052490234375, | |
| "logps/rejected": -394.6015319824219, | |
| "loss": 0.2059, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.9903608560562134, | |
| "rewards/margins": 0.528742253780365, | |
| "rewards/rejected": -1.519102931022644, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.9507259776993954e-08, | |
| "logits/chosen": -2.4183950424194336, | |
| "logits/rejected": -2.3998031616210938, | |
| "logps/chosen": -349.13397216796875, | |
| "logps/rejected": -394.29034423828125, | |
| "loss": 0.204, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.842970073223114, | |
| "rewards/margins": 0.6520211696624756, | |
| "rewards/rejected": -1.4949913024902344, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": -2.4425089359283447, | |
| "eval_logits/rejected": -2.4274942874908447, | |
| "eval_logps/chosen": -346.5982666015625, | |
| "eval_logps/rejected": -409.02783203125, | |
| "eval_loss": 0.21434982120990753, | |
| "eval_rewards/accuracies": 0.703125, | |
| "eval_rewards/chosen": -0.8955863118171692, | |
| "eval_rewards/margins": 0.6211590766906738, | |
| "eval_rewards/rejected": -1.5167453289031982, | |
| "eval_runtime": 53.3158, | |
| "eval_samples_per_second": 37.512, | |
| "eval_steps_per_second": 0.6, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.022313472693447e-08, | |
| "logits/chosen": -2.346083164215088, | |
| "logits/rejected": -2.3530008792877197, | |
| "logps/chosen": -359.29718017578125, | |
| "logps/rejected": -380.91094970703125, | |
| "loss": 0.2113, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.9364844560623169, | |
| "rewards/margins": 0.4689885973930359, | |
| "rewards/rejected": -1.4054731130599976, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.2111614344599684e-08, | |
| "logits/chosen": -2.3877267837524414, | |
| "logits/rejected": -2.3703396320343018, | |
| "logps/chosen": -347.065673828125, | |
| "logps/rejected": -389.3710021972656, | |
| "loss": 0.2072, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.8826552629470825, | |
| "rewards/margins": 0.5836518406867981, | |
| "rewards/rejected": -1.4663069248199463, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.521597710086439e-08, | |
| "logits/chosen": -2.427062511444092, | |
| "logits/rejected": -2.3968257904052734, | |
| "logps/chosen": -384.0503845214844, | |
| "logps/rejected": -394.9156799316406, | |
| "loss": 0.2142, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.0325697660446167, | |
| "rewards/margins": 0.47820836305618286, | |
| "rewards/rejected": -1.5107781887054443, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 9.57301420397924e-09, | |
| "logits/chosen": -2.4799857139587402, | |
| "logits/rejected": -2.432332992553711, | |
| "logps/chosen": -368.3861083984375, | |
| "logps/rejected": -357.58380126953125, | |
| "loss": 0.2142, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.9250670671463013, | |
| "rewards/margins": 0.450716108083725, | |
| "rewards/rejected": -1.375783085823059, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 5.212833302556258e-09, | |
| "logits/chosen": -2.4059414863586426, | |
| "logits/rejected": -2.3920960426330566, | |
| "logps/chosen": -371.8204650878906, | |
| "logps/rejected": -386.82269287109375, | |
| "loss": 0.215, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.8487032651901245, | |
| "rewards/margins": 0.4881046414375305, | |
| "rewards/rejected": -1.3368079662322998, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.158697848236607e-09, | |
| "logits/chosen": -2.456326961517334, | |
| "logits/rejected": -2.4320716857910156, | |
| "logps/chosen": -348.0980529785156, | |
| "logps/rejected": -381.2003173828125, | |
| "loss": 0.2155, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.8624986410140991, | |
| "rewards/margins": 0.5545080900192261, | |
| "rewards/rejected": -1.4170067310333252, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 4.269029751107489e-10, | |
| "logits/chosen": -2.4110119342803955, | |
| "logits/rejected": -2.383718967437744, | |
| "logps/chosen": -323.4446716308594, | |
| "logps/rejected": -392.60064697265625, | |
| "loss": 0.2178, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.8438596725463867, | |
| "rewards/margins": 0.6573068499565125, | |
| "rewards/rejected": -1.501166582107544, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 478, | |
| "total_flos": 0.0, | |
| "train_loss": 0.24105885531363627, | |
| "train_runtime": 3941.0619, | |
| "train_samples_per_second": 15.512, | |
| "train_steps_per_second": 0.121 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 478, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |