{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.4984423676012461, "eval_steps": 500, "global_step": 40, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 8193.375, "completions/mean_length": 3989.6012573242188, "completions/min_length": 431.625, "epoch": 0.012461059190031152, "grad_norm": 0.17126550499342053, "kl": 0.0, "learning_rate": 1e-06, "loss": -0.12229467928409576, "memory(GiB)": 167.35, "reward": 0.34970238618552685, "reward_std": 0.24919933266937733, "rewards/AnswerTagAccuracyORM/mean": 0.34970238618552685, "rewards/AnswerTagAccuracyORM/std": 0.44819867983460426, "step": 1, "train_speed(iter/s)": 0.000657 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.20833333333333331, "completions/max_length": 8193.125, "completions/mean_length": 3849.834930419922, "completions/min_length": 370.75, "epoch": 0.024922118380062305, "grad_norm": 0.1997916996720489, "kl": 0.0, "learning_rate": 9.996046986136508e-07, "loss": -0.09795808792114258, "memory(GiB)": 167.35, "reward": 0.3898809542879462, "reward_std": 0.3207090422511101, "rewards/AnswerTagAccuracyORM/mean": 0.3898809542879462, "rewards/AnswerTagAccuracyORM/std": 0.45823679864406586, "step": 2, "train_speed(iter/s)": 0.000679 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.29910714285714285, "completions/max_length": 8193.25, "completions/mean_length": 4399.15478515625, "completions/min_length": 428.75, "epoch": 0.037383177570093455, "grad_norm": 0.17121898013903475, "kl": 0.000213623046875, "learning_rate": 9.984194195073478e-07, "loss": -0.1223849356174469, "memory(GiB)": 174.45, "reward": 0.37648809468373656, "reward_std": 0.2712590955197811, "rewards/AnswerTagAccuracyORM/mean": 0.37648809468373656, "rewards/AnswerTagAccuracyORM/std": 0.4078557249158621, "step": 3, "train_speed(iter/s)": 0.000679 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.2157738095238095, "completions/max_length": 8193.75, "completions/mean_length": 3801.852783203125, "completions/min_length": 397.75, "epoch": 0.04984423676012461, "grad_norm": 0.18133669602016783, "kl": 0.00020253658294677734, "learning_rate": 9.964460368509865e-07, "loss": -0.1141134724020958, "memory(GiB)": 174.45, "reward": 0.47916668094694614, "reward_std": 0.303694985806942, "rewards/AnswerTagAccuracyORM/mean": 0.47916668094694614, "rewards/AnswerTagAccuracyORM/std": 0.44507846981287, "step": 4, "train_speed(iter/s)": 0.000686 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.1919642857142857, "completions/max_length": 8193.75, "completions/mean_length": 3705.5669860839844, "completions/min_length": 372.375, "epoch": 0.06230529595015576, "grad_norm": 0.2246349455331878, "kl": 0.0002810955047607422, "learning_rate": 9.936876709681666e-07, "loss": -0.1321631819009781, "memory(GiB)": 174.45, "reward": 0.43154763616621494, "reward_std": 0.35136963427066803, "rewards/AnswerTagAccuracyORM/mean": 0.43154763616621494, "rewards/AnswerTagAccuracyORM/std": 0.4658842086791992, "step": 5, "train_speed(iter/s)": 0.00069 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.20238095238095238, "completions/max_length": 8194.875, "completions/mean_length": 3702.392852783203, "completions/min_length": 364.625, "epoch": 0.07476635514018691, "grad_norm": 0.17656783163918816, "kl": 0.0002797842025756836, "learning_rate": 9.901486834023181e-07, "loss": -0.08321866393089294, "memory(GiB)": 174.45, "reward": 0.3764880960807204, "reward_std": 0.2821493409574032, "rewards/AnswerTagAccuracyORM/mean": 0.3764880960807204, "rewards/AnswerTagAccuracyORM/std": 0.4614675045013428, "step": 6, "train_speed(iter/s)": 0.000693 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.28720238095238093, "completions/max_length": 8193.625, "completions/mean_length": 4263.456878662109, "completions/min_length": 327.625, "epoch": 0.08722741433021806, "grad_norm": 0.2524499795692924, "kl": 0.0003044605255126953, "learning_rate": 9.858346700202048e-07, "loss": -0.19307070970535278, "memory(GiB)": 174.45, "reward": 0.377976194024086, "reward_std": 0.3151053488254547, "rewards/AnswerTagAccuracyORM/mean": 0.377976194024086, "rewards/AnswerTagAccuracyORM/std": 0.4634386822581291, "step": 7, "train_speed(iter/s)": 0.000691 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.20833333333333331, "completions/max_length": 8193.25, "completions/mean_length": 3797.3007202148438, "completions/min_length": 272.5, "epoch": 0.09968847352024922, "grad_norm": 0.20136930786075843, "kl": 0.00032448768615722656, "learning_rate": 9.807524521637102e-07, "loss": -0.10552932322025299, "memory(GiB)": 174.45, "reward": 0.4092261902987957, "reward_std": 0.3334960360080004, "rewards/AnswerTagAccuracyORM/mean": 0.4092261902987957, "rewards/AnswerTagAccuracyORM/std": 0.47217297181487083, "step": 8, "train_speed(iter/s)": 0.000693 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.2663690476190476, "completions/max_length": 8193.25, "completions/mean_length": 4201.8289794921875, "completions/min_length": 466.0, "epoch": 0.11214953271028037, "grad_norm": 0.19988066240856855, "kl": 0.0003037452697753906, "learning_rate": 9.749100658638914e-07, "loss": -0.12433922290802002, "memory(GiB)": 174.52, "reward": 0.3482142873108387, "reward_std": 0.2973194606602192, "rewards/AnswerTagAccuracyORM/mean": 0.3482142873108387, "rewards/AnswerTagAccuracyORM/std": 0.4571182616055012, "step": 9, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.23660714285714288, "completions/max_length": 8193.5, "completions/mean_length": 4137.879486083984, "completions/min_length": 402.25, "epoch": 0.12461059190031153, "grad_norm": 0.23579962953472336, "kl": 0.00039386749267578125, "learning_rate": 9.68316749134364e-07, "loss": -0.15604804456233978, "memory(GiB)": 175.38, "reward": 0.37648810259997845, "reward_std": 0.33304327353835106, "rewards/AnswerTagAccuracyORM/mean": 0.37648810259997845, "rewards/AnswerTagAccuracyORM/std": 0.46259133517742157, "step": 10, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.1949404761904762, "completions/max_length": 8193.0, "completions/mean_length": 3524.0000915527344, "completions/min_length": 352.5, "epoch": 0.13707165109034267, "grad_norm": 0.18587876448756557, "kl": 0.0004878044128417969, "learning_rate": 9.609829273641032e-07, "loss": -0.09066524356603622, "memory(GiB)": 175.38, "reward": 0.3913690634071827, "reward_std": 0.2906598076224327, "rewards/AnswerTagAccuracyORM/mean": 0.3913690634071827, "rewards/AnswerTagAccuracyORM/std": 0.48316407203674316, "step": 11, "train_speed(iter/s)": 0.000694 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.33035714285714285, "completions/max_length": 8193.0, "completions/mean_length": 4620.495635986328, "completions/min_length": 343.25, "epoch": 0.14953271028037382, "grad_norm": 0.21227366684003468, "kl": 0.00048542022705078125, "learning_rate": 9.529201968327616e-07, "loss": -0.179769366979599, "memory(GiB)": 175.38, "reward": 0.3452381007373333, "reward_std": 0.24404709972441196, "rewards/AnswerTagAccuracyORM/mean": 0.3452381007373333, "rewards/AnswerTagAccuracyORM/std": 0.4610680975019932, "step": 12, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.19345238095238096, "completions/max_length": 8193.0, "completions/mean_length": 3493.071502685547, "completions/min_length": 343.625, "epoch": 0.16199376947040497, "grad_norm": 0.19459450978177167, "kl": 0.000698089599609375, "learning_rate": 9.441413063745659e-07, "loss": -0.12030751258134842, "memory(GiB)": 175.38, "reward": 0.45238096080720425, "reward_std": 0.3026049826294184, "rewards/AnswerTagAccuracyORM/mean": 0.45238096080720425, "rewards/AnswerTagAccuracyORM/std": 0.4734005853533745, "step": 13, "train_speed(iter/s)": 0.000694 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.24404761904761904, "completions/max_length": 8193.125, "completions/mean_length": 3945.2025146484375, "completions/min_length": 317.5, "epoch": 0.17445482866043613, "grad_norm": 0.2160426116072706, "kl": 0.0008077621459960938, "learning_rate": 9.346601372197913e-07, "loss": -0.17013423144817352, "memory(GiB)": 175.38, "reward": 0.4226190559566021, "reward_std": 0.31002783961594105, "rewards/AnswerTagAccuracyORM/mean": 0.4226190559566021, "rewards/AnswerTagAccuracyORM/std": 0.48812782764434814, "step": 14, "train_speed(iter/s)": 0.000695 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.2693452380952381, "completions/max_length": 8193.25, "completions/mean_length": 3999.154815673828, "completions/min_length": 381.0, "epoch": 0.18691588785046728, "grad_norm": 0.2338145519921423, "kl": 0.0007433891296386719, "learning_rate": 9.24491681045682e-07, "loss": -0.14010122418403625, "memory(GiB)": 175.38, "reward": 0.3779761968180537, "reward_std": 0.2741878144443035, "rewards/AnswerTagAccuracyORM/mean": 0.3779761968180537, "rewards/AnswerTagAccuracyORM/std": 0.46018435433506966, "step": 15, "train_speed(iter/s)": 0.000694 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.24255952380952378, "completions/max_length": 8193.25, "completions/mean_length": 3811.431610107422, "completions/min_length": 405.25, "epoch": 0.19937694704049844, "grad_norm": 0.24163120591345413, "kl": 0.0007557868957519531, "learning_rate": 9.136520162715286e-07, "loss": -0.15715962648391724, "memory(GiB)": 175.88, "reward": 0.5014881025999784, "reward_std": 0.2836833242326975, "rewards/AnswerTagAccuracyORM/mean": 0.5014881025999784, "rewards/AnswerTagAccuracyORM/std": 0.4741004407405853, "step": 16, "train_speed(iter/s)": 0.000695 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.30952380952380953, "completions/max_length": 8193.375, "completions/mean_length": 4581.294769287109, "completions/min_length": 462.75, "epoch": 0.2118380062305296, "grad_norm": 0.20943523953763443, "kl": 0.0010576248168945312, "learning_rate": 9.021582826353824e-07, "loss": -0.17735408246517181, "memory(GiB)": 175.88, "reward": 0.345238103531301, "reward_std": 0.2590687647461891, "rewards/AnswerTagAccuracyORM/mean": 0.345238103531301, "rewards/AnswerTagAccuracyORM/std": 0.4563480280339718, "step": 17, "train_speed(iter/s)": 0.000694 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.27232142857142855, "completions/max_length": 8193.5, "completions/mean_length": 4101.574493408203, "completions/min_length": 369.125, "epoch": 0.22429906542056074, "grad_norm": 0.20605735960275373, "kl": 0.001071929931640625, "learning_rate": 8.90028654092606e-07, "loss": -0.08183983713388443, "memory(GiB)": 175.88, "reward": 0.39583334792405367, "reward_std": 0.24742238968610764, "rewards/AnswerTagAccuracyORM/mean": 0.39583334792405367, "rewards/AnswerTagAccuracyORM/std": 0.4427494630217552, "step": 18, "train_speed(iter/s)": 0.000694 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.2723214285714286, "completions/max_length": 8196.125, "completions/mean_length": 3904.400390625, "completions/min_length": 349.125, "epoch": 0.2367601246105919, "grad_norm": 0.22278161024164303, "kl": 0.0016794204711914062, "learning_rate": 8.77282310079115e-07, "loss": -0.20427300035953522, "memory(GiB)": 175.88, "reward": 0.4389881044626236, "reward_std": 0.2799454350024462, "rewards/AnswerTagAccuracyORM/mean": 0.4389881044626236, "rewards/AnswerTagAccuracyORM/std": 0.4796442426741123, "step": 19, "train_speed(iter/s)": 0.000694 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.32589285714285715, "completions/max_length": 8198.5, "completions/mean_length": 4417.791748046875, "completions/min_length": 493.75, "epoch": 0.24922118380062305, "grad_norm": 0.17009334695071343, "kl": 0.001983642578125, "learning_rate": 8.639394051847471e-07, "loss": -0.07000445574522018, "memory(GiB)": 175.93, "reward": 0.31250000186264515, "reward_std": 0.21009514946490526, "rewards/AnswerTagAccuracyORM/mean": 0.31250000186264515, "rewards/AnswerTagAccuracyORM/std": 0.4253981076180935, "step": 20, "train_speed(iter/s)": 0.000693 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.261904761904762, "completions/max_length": 8193.0, "completions/mean_length": 4160.090850830078, "completions/min_length": 358.75, "epoch": 0.2616822429906542, "grad_norm": 0.18862458424536294, "kl": 0.002643585205078125, "learning_rate": 8.500210372847126e-07, "loss": -0.14565826952457428, "memory(GiB)": 175.93, "reward": 0.42857143841683865, "reward_std": 0.27233397029340267, "rewards/AnswerTagAccuracyORM/mean": 0.42857143841683865, "rewards/AnswerTagAccuracyORM/std": 0.4581611305475235, "step": 21, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 8193.125, "completions/mean_length": 4088.5298767089844, "completions/min_length": 376.375, "epoch": 0.27414330218068533, "grad_norm": 0.19231284564227827, "kl": 0.0030651092529296875, "learning_rate": 8.355492141795184e-07, "loss": -0.17381058633327484, "memory(GiB)": 175.93, "reward": 0.3854166716337204, "reward_std": 0.26877069287002087, "rewards/AnswerTagAccuracyORM/mean": 0.3854166716337204, "rewards/AnswerTagAccuracyORM/std": 0.4839600883424282, "step": 22, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.28720238095238093, "completions/max_length": 8193.125, "completions/mean_length": 4086.6087036132812, "completions/min_length": 368.875, "epoch": 0.2866043613707165, "grad_norm": 0.2125176134875177, "kl": 0.00290679931640625, "learning_rate": 8.2054681879611e-07, "loss": -0.22260838747024536, "memory(GiB)": 175.93, "reward": 0.4613095372915268, "reward_std": 0.25964014790952206, "rewards/AnswerTagAccuracyORM/mean": 0.4613095372915268, "rewards/AnswerTagAccuracyORM/std": 0.4591619521379471, "step": 23, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.244047619047619, "completions/max_length": 8193.375, "completions/mean_length": 3719.543182373047, "completions/min_length": 375.5, "epoch": 0.29906542056074764, "grad_norm": 0.21050284745364475, "kl": 0.00313568115234375, "learning_rate": 8.050375730052621e-07, "loss": -0.130848228931427, "memory(GiB)": 175.93, "reward": 0.404761909507215, "reward_std": 0.28319618850946426, "rewards/AnswerTagAccuracyORM/mean": 0.404761909507215, "rewards/AnswerTagAccuracyORM/std": 0.46318022534251213, "step": 24, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 8193.375, "completions/mean_length": 4491.202484130859, "completions/min_length": 415.25, "epoch": 0.3115264797507788, "grad_norm": 0.1899503253060254, "kl": 0.0037708282470703125, "learning_rate": 7.890460001124241e-07, "loss": -0.14017701148986816, "memory(GiB)": 175.93, "reward": 0.3750000046566129, "reward_std": 0.2724082116037607, "rewards/AnswerTagAccuracyORM/mean": 0.3750000046566129, "rewards/AnswerTagAccuracyORM/std": 0.4636641815304756, "step": 25, "train_speed(iter/s)": 0.000691 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.23809523809523808, "completions/max_length": 8193.375, "completions/mean_length": 3894.7633666992188, "completions/min_length": 407.25, "epoch": 0.32398753894080995, "grad_norm": 0.21607441644369615, "kl": 0.004718780517578125, "learning_rate": 7.725973860813338e-07, "loss": -0.13692191243171692, "memory(GiB)": 175.93, "reward": 0.46130953915417194, "reward_std": 0.2574587743729353, "rewards/AnswerTagAccuracyORM/mean": 0.46130953915417194, "rewards/AnswerTagAccuracyORM/std": 0.46890895441174507, "step": 26, "train_speed(iter/s)": 0.000691 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.23214285714285715, "completions/max_length": 8193.25, "completions/mean_length": 3892.3154907226562, "completions/min_length": 490.5, "epoch": 0.3364485981308411, "grad_norm": 0.2714960413830388, "kl": 0.004787445068359375, "learning_rate": 7.557177395517111e-07, "loss": -0.12591125071048737, "memory(GiB)": 176.07, "reward": 0.41369048738852143, "reward_std": 0.2556753307580948, "rewards/AnswerTagAccuracyORM/mean": 0.41369048738852143, "rewards/AnswerTagAccuracyORM/std": 0.46406359039247036, "step": 27, "train_speed(iter/s)": 0.000691 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.21726190476190477, "completions/max_length": 8193.125, "completions/mean_length": 3599.0833740234375, "completions/min_length": 376.75, "epoch": 0.34890965732087226, "grad_norm": 0.19644845815079176, "kl": 0.003849029541015625, "learning_rate": 7.38433750714253e-07, "loss": -0.1492602676153183, "memory(GiB)": 176.07, "reward": 0.4702381044626236, "reward_std": 0.2548898998647928, "rewards/AnswerTagAccuracyORM/mean": 0.4702381044626236, "rewards/AnswerTagAccuracyORM/std": 0.4880012050271034, "step": 28, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.24255952380952378, "completions/max_length": 8193.125, "completions/mean_length": 4016.561065673828, "completions/min_length": 468.875, "epoch": 0.3613707165109034, "grad_norm": 0.17926535607961544, "kl": 0.0053863525390625, "learning_rate": 7.207727491079559e-07, "loss": -0.09940721839666367, "memory(GiB)": 176.07, "reward": 0.4523809663951397, "reward_std": 0.2573994603008032, "rewards/AnswerTagAccuracyORM/mean": 0.4523809663951397, "rewards/AnswerTagAccuracyORM/std": 0.48338139802217484, "step": 29, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.2619047619047619, "completions/max_length": 8193.375, "completions/mean_length": 4025.3720703125, "completions/min_length": 408.625, "epoch": 0.37383177570093457, "grad_norm": 0.1833169510245964, "kl": 0.00562286376953125, "learning_rate": 7.027626604064969e-07, "loss": -0.10118662565946579, "memory(GiB)": 176.07, "reward": 0.3854166753590107, "reward_std": 0.23373469151556492, "rewards/AnswerTagAccuracyORM/mean": 0.3854166753590107, "rewards/AnswerTagAccuracyORM/std": 0.46597452089190483, "step": 30, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.2663690476190476, "completions/max_length": 8193.25, "completions/mean_length": 3996.0833740234375, "completions/min_length": 357.875, "epoch": 0.3862928348909657, "grad_norm": 0.21555033230693768, "kl": 0.00638580322265625, "learning_rate": 6.844319622620039e-07, "loss": -0.17128993570804596, "memory(GiB)": 177.2, "reward": 0.4315476343035698, "reward_std": 0.287859283387661, "rewards/AnswerTagAccuracyORM/mean": 0.4315476343035698, "rewards/AnswerTagAccuracyORM/std": 0.4801550768315792, "step": 31, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.2604166666666667, "completions/max_length": 8193.125, "completions/mean_length": 3904.589385986328, "completions/min_length": 405.5, "epoch": 0.3987538940809969, "grad_norm": 0.19674394061545156, "kl": 0.006427764892578125, "learning_rate": 6.658096392760339e-07, "loss": -0.1568790227174759, "memory(GiB)": 177.2, "reward": 0.4315476268529892, "reward_std": 0.2220794102177024, "rewards/AnswerTagAccuracyORM/mean": 0.4315476268529892, "rewards/AnswerTagAccuracyORM/std": 0.48653851449489594, "step": 32, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.22767857142857145, "completions/max_length": 8193.875, "completions/mean_length": 3937.4747924804688, "completions/min_length": 423.875, "epoch": 0.411214953271028, "grad_norm": 0.19013545416962976, "kl": 0.007045745849609375, "learning_rate": 6.469251371689604e-07, "loss": -0.15400366485118866, "memory(GiB)": 177.2, "reward": 0.5223214328289032, "reward_std": 0.30748542957007885, "rewards/AnswerTagAccuracyORM/mean": 0.5223214328289032, "rewards/AnswerTagAccuracyORM/std": 0.48715949431061745, "step": 33, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.17261904761904764, "completions/max_length": 8193.375, "completions/mean_length": 3657.462890625, "completions/min_length": 391.75, "epoch": 0.4236760124610592, "grad_norm": 0.149800501823744, "kl": 0.00738525390625, "learning_rate": 6.278083162202373e-07, "loss": -0.054694823920726776, "memory(GiB)": 177.2, "reward": 0.4613095410168171, "reward_std": 0.21404170151799917, "rewards/AnswerTagAccuracyORM/mean": 0.4613095410168171, "rewards/AnswerTagAccuracyORM/std": 0.49066925421357155, "step": 34, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.244047619047619, "completions/max_length": 8193.125, "completions/mean_length": 3713.2679138183594, "completions/min_length": 407.125, "epoch": 0.43613707165109034, "grad_norm": 0.2141010331036125, "kl": 0.008380889892578125, "learning_rate": 6.084894040531589e-07, "loss": -0.14978839457035065, "memory(GiB)": 177.2, "reward": 0.5000000176951289, "reward_std": 0.2632578443735838, "rewards/AnswerTagAccuracyORM/mean": 0.5000000176951289, "rewards/AnswerTagAccuracyORM/std": 0.43221472948789597, "step": 35, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.25744047619047616, "completions/max_length": 8194.125, "completions/mean_length": 4263.843811035156, "completions/min_length": 441.125, "epoch": 0.4485981308411215, "grad_norm": 0.17438120553123407, "kl": 0.006961822509765625, "learning_rate": 5.889989478387752e-07, "loss": -0.0636586844921112, "memory(GiB)": 177.2, "reward": 0.41220239736139774, "reward_std": 0.2634734772145748, "rewards/AnswerTagAccuracyORM/mean": 0.41220239736139774, "rewards/AnswerTagAccuracyORM/std": 0.47433316707611084, "step": 36, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.2946428571428571, "completions/max_length": 8193.5, "completions/mean_length": 4341.491119384766, "completions/min_length": 374.625, "epoch": 0.46105919003115264, "grad_norm": 0.21429795338500487, "kl": 0.00689697265625, "learning_rate": 5.693677659945342e-07, "loss": -0.15549130737781525, "memory(GiB)": 177.2, "reward": 0.4330357238650322, "reward_std": 0.30585045740008354, "rewards/AnswerTagAccuracyORM/mean": 0.4330357238650322, "rewards/AnswerTagAccuracyORM/std": 0.4805232435464859, "step": 37, "train_speed(iter/s)": 0.000691 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.255952380952381, "completions/max_length": 8153.5, "completions/mean_length": 3888.888381958008, "completions/min_length": 513.875, "epoch": 0.4735202492211838, "grad_norm": 0.23492644678534505, "kl": 0.008266448974609375, "learning_rate": 5.496268994540309e-07, "loss": -0.13862904906272888, "memory(GiB)": 177.2, "reward": 0.49107144121080637, "reward_std": 0.2540656551718712, "rewards/AnswerTagAccuracyORM/mean": 0.49107144121080637, "rewards/AnswerTagAccuracyORM/std": 0.450797863304615, "step": 38, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.3169642857142857, "completions/max_length": 8193.125, "completions/mean_length": 4171.233703613281, "completions/min_length": 372.125, "epoch": 0.48598130841121495, "grad_norm": 0.21053247802635136, "kl": 0.006114959716796875, "learning_rate": 5.298075625849099e-07, "loss": -0.1665022075176239, "memory(GiB)": 177.2, "reward": 0.4077381081879139, "reward_std": 0.2519750725477934, "rewards/AnswerTagAccuracyORM/mean": 0.4077381081879139, "rewards/AnswerTagAccuracyORM/std": 0.4829714484512806, "step": 39, "train_speed(iter/s)": 0.000692 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.2693452380952381, "completions/max_length": 8193.625, "completions/mean_length": 4133.266448974609, "completions/min_length": 360.75, "epoch": 0.4984423676012461, "grad_norm": 0.2650285808547177, "kl": 0.00934600830078125, "learning_rate": 5.09941093832535e-07, "loss": -0.19298426806926727, "memory(GiB)": 177.2, "reward": 0.4479166716337204, "reward_std": 0.28238569758832455, "rewards/AnswerTagAccuracyORM/mean": 0.4479166716337204, "rewards/AnswerTagAccuracyORM/std": 0.46365100890398026, "step": 40, "train_speed(iter/s)": 0.000692 } ], "logging_steps": 1, "max_steps": 80, "num_input_tokens_seen": 0, "num_train_epochs": 1, "save_steps": 20, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 4, "trial_name": null, "trial_params": null }