{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.6925373134328359,
  "eval_steps": 500,
  "global_step": 140,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 51.0,
      "completions/mean_length": 16.09375,
      "completions/min_length": 2.0,
      "epoch": 0.011940298507462687,
      "grad_norm": 3.1591623828063176,
      "kl": 0.0,
      "learning_rate": 2.2222222222222223e-05,
      "loss": -3.725290298461914e-09,
      "memory(GiB)": 77.11,
      "reward": 0.1484375,
      "reward_std": 0.7272911071777344,
      "rewards/Response_no_think_reward_1/mean": 0.1484375,
      "rewards/Response_no_think_reward_1/std": 0.9646495580673218,
      "step": 1,
      "train_speed(iter/s)": 0.001668
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.023880597014925373,
      "grad_norm": 3.150729002521691,
      "kl": 0.0,
      "learning_rate": 4.4444444444444447e-05,
      "loss": -3.725290298461914e-09,
      "memory(GiB)": 77.11,
      "step": 2,
      "train_speed(iter/s)": 0.002544
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0028647197177633643,
      "clip_ratio/low_min": 0.0028647197177633643,
      "clip_ratio/region_mean": 0.0028647197177633643,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 51.0,
      "completions/mean_length": 14.59375,
      "completions/min_length": 2.0,
      "epoch": 0.03582089552238806,
      "grad_norm": 2.067277005959865,
      "kl": 0.0022258755707298405,
      "learning_rate": 6.666666666666667e-05,
      "loss": -0.0004270445788279176,
      "memory(GiB)": 77.3,
      "reward": -0.3828125,
      "reward_std": 0.707726001739502,
      "rewards/Response_no_think_reward_1/mean": -0.3828125,
      "rewards/Response_no_think_reward_1/std": 0.923201322555542,
      "step": 3,
      "train_speed(iter/s)": 0.002313
    },
    {
      "clip_ratio/high_max": 0.003551179339410737,
      "clip_ratio/high_mean": 0.003551179339410737,
      "clip_ratio/low_mean": 0.025148934859316796,
      "clip_ratio/low_min": 0.025148934859316796,
      "clip_ratio/region_mean": 0.028700114460662007,
      "epoch": 0.04776119402985075,
      "grad_norm": 1.6109973009110337,
      "kl": 0.007886729676101822,
      "learning_rate": 8.888888888888889e-05,
      "loss": -0.010280434973537922,
      "memory(GiB)": 77.3,
      "step": 4,
      "train_speed(iter/s)": 0.002696
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0019339247373864055,
      "clip_ratio/low_min": 0.0019339247373864055,
      "clip_ratio/region_mean": 0.0019339247373864055,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 51.0,
      "completions/mean_length": 13.5390625,
      "completions/min_length": 3.0,
      "epoch": 0.05970149253731343,
      "grad_norm": 2.893722550507481,
      "kl": 0.010198775078606559,
      "learning_rate": 0.00011111111111111112,
      "loss": 0.0004082494997419417,
      "memory(GiB)": 77.52,
      "reward": 0.125,
      "reward_std": 0.493943989276886,
      "rewards/Response_no_think_reward_1/mean": 0.125,
      "rewards/Response_no_think_reward_1/std": 0.9881184101104736,
      "step": 5,
      "train_speed(iter/s)": 0.002517
    },
    {
      "clip_ratio/high_max": 0.028469400480389595,
      "clip_ratio/high_mean": 0.028469400480389595,
      "clip_ratio/low_mean": 0.04948427155613899,
      "clip_ratio/low_min": 0.04948427155613899,
      "clip_ratio/region_mean": 0.07795367110520601,
      "epoch": 0.07164179104477612,
      "grad_norm": 1.5585579234945925,
      "kl": 0.6935355372261256,
      "learning_rate": 0.00013333333333333334,
      "loss": -0.005084950476884842,
      "memory(GiB)": 79.48,
      "step": 6,
      "train_speed(iter/s)": 0.002759
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0010160960373468697,
      "clip_ratio/low_min": 0.0010160960373468697,
      "clip_ratio/region_mean": 0.0010160960373468697,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 51.0,
      "completions/mean_length": 11.9296875,
      "completions/min_length": 3.0,
      "epoch": 0.08358208955223881,
      "grad_norm": 1.7579177577957044,
      "kl": 0.07463712914614007,
      "learning_rate": 0.00015555555555555556,
      "loss": 0.0008710746187716722,
      "memory(GiB)": 79.48,
      "reward": -0.0390625,
      "reward_std": 0.6340709924697876,
      "rewards/Response_no_think_reward_1/mean": -0.0390625,
      "rewards/Response_no_think_reward_1/std": 0.9992307424545288,
      "step": 7,
      "train_speed(iter/s)": 0.002622
    },
    {
      "clip_ratio/high_max": 0.02714990673121065,
      "clip_ratio/high_mean": 0.02714990673121065,
      "clip_ratio/low_mean": 0.03809016165905632,
      "clip_ratio/low_min": 0.03809016165905632,
      "clip_ratio/region_mean": 0.06524006789550185,
      "epoch": 0.0955223880597015,
      "grad_norm": 1.2730547516178623,
      "kl": 0.21840606682235375,
      "learning_rate": 0.00017777777777777779,
      "loss": -0.013987341895699501,
      "memory(GiB)": 79.48,
      "step": 8,
      "train_speed(iter/s)": 0.002802
    },
    {
      "clip_ratio/high_max": 0.0009191176504828036,
      "clip_ratio/high_mean": 0.0009191176504828036,
      "clip_ratio/low_mean": 0.0018382353009656072,
      "clip_ratio/low_min": 0.0018382353009656072,
      "clip_ratio/region_mean": 0.0027573530096560717,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 51.0,
      "completions/mean_length": 10.3828125,
      "completions/min_length": 3.0,
      "epoch": 0.10746268656716418,
      "grad_norm": 3.3295900087180645,
      "kl": 0.35752939723897725,
      "learning_rate": 0.0002,
      "loss": 0.003791232593357563,
      "memory(GiB)": 79.48,
      "reward": 0.328125,
      "reward_std": 0.5329009890556335,
      "rewards/Response_no_think_reward_1/mean": 0.328125,
      "rewards/Response_no_think_reward_1/std": 0.9483460783958435,
      "step": 9,
      "train_speed(iter/s)": 0.002657
    },
    {
      "clip_ratio/high_max": 0.03471234819153324,
      "clip_ratio/high_mean": 0.03471234819153324,
      "clip_ratio/low_mean": 0.045793990429956466,
      "clip_ratio/low_min": 0.045793990429956466,
      "clip_ratio/region_mean": 0.08050633978564292,
      "epoch": 0.11940298507462686,
      "grad_norm": 3.037041883710677,
      "kl": 1.3611202164320275,
      "learning_rate": 0.00019997998037428526,
      "loss": -0.0012767184525728226,
      "memory(GiB)": 79.48,
      "step": 10,
      "train_speed(iter/s)": 0.002798
    },
    {
      "clip_ratio/high_max": 0.0022767678019590676,
      "clip_ratio/high_mean": 0.0022767678019590676,
      "clip_ratio/low_mean": 0.004576534847728908,
      "clip_ratio/low_min": 0.004576534847728908,
      "clip_ratio/region_mean": 0.006853302649687976,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 51.0,
      "completions/mean_length": 11.3828125,
      "completions/min_length": 3.0,
      "epoch": 0.13134328358208955,
      "grad_norm": 1.1988213220999966,
      "kl": 0.176690819360374,
      "learning_rate": 0.00019991992951284932,
      "loss": 0.0002530772762838751,
      "memory(GiB)": 79.48,
      "reward": 0.2890625,
      "reward_std": 0.36220550537109375,
      "rewards/Response_no_think_reward_1/mean": 0.2890625,
      "rewards/Response_no_think_reward_1/std": 0.9487027525901794,
      "step": 11,
      "train_speed(iter/s)": 0.002687
    },
    {
      "clip_ratio/high_max": 0.010013980441726744,
      "clip_ratio/high_mean": 0.010013980441726744,
      "clip_ratio/low_mean": 0.0431372388266027,
      "clip_ratio/low_min": 0.0431372388266027,
      "clip_ratio/region_mean": 0.053151219501160085,
      "epoch": 0.14328358208955225,
      "grad_norm": 4.4251813995519464,
      "kl": 3.680493631065474,
      "learning_rate": 0.00019981987145960755,
      "loss": 0.022724991664290428,
      "memory(GiB)": 79.48,
      "step": 12,
      "train_speed(iter/s)": 0.002804
    },
    {
      "clip_ratio/high_max": 0.000480769231216982,
      "clip_ratio/high_mean": 0.000480769231216982,
      "clip_ratio/low_mean": 0.0022044407669454813,
      "clip_ratio/low_min": 0.0022044407669454813,
      "clip_ratio/region_mean": 0.0026852099981624633,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 51.0,
      "completions/mean_length": 12.0703125,
      "completions/min_length": 3.0,
      "epoch": 0.15522388059701492,
      "grad_norm": 2.0136556426936285,
      "kl": 0.14468206313904375,
      "learning_rate": 0.00019967984627705548,
      "loss": -0.00039299592026509345,
      "memory(GiB)": 79.48,
      "reward": 0.0625,
      "reward_std": 0.6480826735496521,
      "rewards/Response_no_think_reward_1/mean": 0.0625,
      "rewards/Response_no_think_reward_1/std": 0.9861242175102234,
      "step": 13,
      "train_speed(iter/s)": 0.00269
    },
    {
      "clip_ratio/high_max": 0.016389093652833253,
      "clip_ratio/high_mean": 0.016389093652833253,
      "clip_ratio/low_mean": 0.03954102098941803,
      "clip_ratio/low_min": 0.03954102098941803,
      "clip_ratio/region_mean": 0.05593011493328959,
      "epoch": 0.16716417910447762,
      "grad_norm": 0.7268697285990734,
      "kl": 0.15212981263175607,
      "learning_rate": 0.00019949991003022808,
      "loss": -0.01738908141851425,
      "memory(GiB)": 79.48,
      "step": 14,
      "train_speed(iter/s)": 0.002789
    },
    {
      "clip_ratio/high_max": 0.0005787037080153823,
      "clip_ratio/high_mean": 0.0005787037080153823,
      "clip_ratio/low_mean": 0.004868808202445507,
      "clip_ratio/low_min": 0.004868808202445507,
      "clip_ratio/region_mean": 0.005447512026876211,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 51.0,
      "completions/mean_length": 9.8671875,
      "completions/min_length": 3.0,
      "epoch": 0.1791044776119403,
      "grad_norm": 1.881956635614757,
      "kl": 0.24015408230479807,
      "learning_rate": 0.0001992801347642515,
      "loss": 0.0037227380089461803,
      "memory(GiB)": 79.48,
      "reward": 0.0625,
      "reward_std": 0.15197435021400452,
      "rewards/Response_no_think_reward_1/mean": 0.0625,
      "rewards/Response_no_think_reward_1/std": 0.9700231552124023,
      "step": 15,
      "train_speed(iter/s)": 0.002716
    },
    {
      "clip_ratio/high_max": 0.005905566504225135,
      "clip_ratio/high_mean": 0.005905566504225135,
      "clip_ratio/low_mean": 0.03072124859318137,
      "clip_ratio/low_min": 0.03072124859318137,
      "clip_ratio/region_mean": 0.036626815563067794,
      "epoch": 0.191044776119403,
      "grad_norm": 9.037445474392891,
      "kl": 4.632834411575459,
      "learning_rate": 0.00019902060847549718,
      "loss": 0.06257862597703934,
      "memory(GiB)": 79.48,
      "step": 16,
      "train_speed(iter/s)": 0.002802
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.004252503567840904,
      "clip_ratio/low_min": 0.004252503567840904,
      "clip_ratio/region_mean": 0.004252503567840904,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 46.0,
      "completions/mean_length": 7.0,
      "completions/min_length": 3.0,
      "epoch": 0.20298507462686566,
      "grad_norm": 5.483351751991195,
      "kl": 3.0633848210127326,
      "learning_rate": 0.0001987214350763483,
      "loss": 0.026951301842927933,
      "memory(GiB)": 79.48,
      "reward": 0.1640625,
      "reward_std": 0.45976677536964417,
      "rewards/Response_no_think_reward_1/mean": 0.1640625,
      "rewards/Response_no_think_reward_1/std": 0.9116684198379517,
      "step": 17,
      "train_speed(iter/s)": 0.002741
    },
    {
      "clip_ratio/high_max": 0.02420679770875722,
      "clip_ratio/high_mean": 0.02420679770875722,
      "clip_ratio/low_mean": 0.029018129454925656,
      "clip_ratio/low_min": 0.029018129454925656,
      "clip_ratio/region_mean": 0.053224927629344165,
      "epoch": 0.21492537313432836,
      "grad_norm": 11.084807824277876,
      "kl": 4.480629495417816,
      "learning_rate": 0.00019838273435359444,
      "loss": 0.03358945623040199,
      "memory(GiB)": 79.48,
      "step": 18,
      "train_speed(iter/s)": 0.002818
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.009987752127926797,
      "clip_ratio/low_min": 0.009987752127926797,
      "clip_ratio/region_mean": 0.009987752127926797,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 51.0,
      "completions/mean_length": 8.6875,
      "completions/min_length": 3.0,
      "epoch": 0.22686567164179106,
      "grad_norm": 2.774235490038128,
      "kl": 1.3207979609724134,
      "learning_rate": 0.00019800464192046955,
      "loss": 0.012463792227208614,
      "memory(GiB)": 79.48,
      "reward": -0.0078125,
      "reward_std": 0.3263596296310425,
      "rewards/Response_no_think_reward_1/mean": -0.0078125,
      "rewards/Response_no_think_reward_1/std": 0.9515514969825745,
      "step": 19,
      "train_speed(iter/s)": 0.002746
    },
    {
      "clip_ratio/high_max": 0.013878540194127709,
      "clip_ratio/high_mean": 0.013878540194127709,
      "clip_ratio/low_mean": 0.053771398146636784,
      "clip_ratio/low_min": 0.053771398146636784,
      "clip_ratio/region_mean": 0.06764993839897215,
      "epoch": 0.23880597014925373,
      "grad_norm": 18.987604867751966,
      "kl": 9.466534710245469,
      "learning_rate": 0.00019758730916235356,
      "loss": 0.16672615706920624,
      "memory(GiB)": 79.48,
      "step": 20,
      "train_speed(iter/s)": 0.002807
    },
    {
      "clip_ratio/high_max": 0.005019778618589044,
      "clip_ratio/high_mean": 0.005019778618589044,
      "clip_ratio/low_mean": 0.00046641789958812296,
      "clip_ratio/low_min": 0.00046641789958812296,
      "clip_ratio/region_mean": 0.005486196518177167,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 51.0,
      "completions/mean_length": 8.234375,
      "completions/min_length": 3.0,
      "epoch": 0.2507462686567164,
      "grad_norm": 3.1541164754847397,
      "kl": 0.4539797595352866,
      "learning_rate": 0.00019713090317615876,
      "loss": 0.0023644084576517344,
      "memory(GiB)": 79.48,
      "reward": -0.1015625,
      "reward_std": 0.3404619097709656,
      "rewards/Response_no_think_reward_1/mean": -0.1015625,
      "rewards/Response_no_think_reward_1/std": 0.9788298010826111,
      "step": 21,
      "train_speed(iter/s)": 0.002748
    },
    {
      "clip_ratio/high_max": 0.02849817555397749,
      "clip_ratio/high_mean": 0.02849817555397749,
      "clip_ratio/low_mean": 0.02262113688630052,
      "clip_ratio/low_min": 0.02262113688630052,
      "clip_ratio/region_mean": 0.051119312876835465,
      "epoch": 0.2626865671641791,
      "grad_norm": 3.225175633228704,
      "kl": 0.27581185393501073,
      "learning_rate": 0.00019663560670342558,
      "loss": 0.0023057509679347277,
      "memory(GiB)": 79.48,
      "step": 22,
      "train_speed(iter/s)": 0.002811
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0015432098880410194,
      "clip_ratio/low_min": 0.0015432098880410194,
      "clip_ratio/region_mean": 0.0015432098880410194,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 51.0,
      "completions/mean_length": 9.1015625,
      "completions/min_length": 3.0,
      "epoch": 0.2746268656716418,
      "grad_norm": 1.602042628934705,
      "kl": 0.223259057267569,
      "learning_rate": 0.00019610161805715397,
      "loss": 0.0012852977961301804,
      "memory(GiB)": 79.48,
      "reward": 0.25,
      "reward_std": 0.45234403014183044,
      "rewards/Response_no_think_reward_1/mean": 0.25,
      "rewards/Response_no_think_reward_1/std": 0.9474374651908875,
      "step": 23,
      "train_speed(iter/s)": 0.002763
    },
    {
      "clip_ratio/high_max": 0.02699369314359501,
      "clip_ratio/high_mean": 0.02699369314359501,
      "clip_ratio/low_mean": 0.024013921502046287,
      "clip_ratio/low_min": 0.024013921502046287,
      "clip_ratio/region_mean": 0.05100761540234089,
      "epoch": 0.2865671641791045,
      "grad_norm": 1.2163889618122583,
      "kl": 0.48499861752497964,
      "learning_rate": 0.00019552915104240065,
      "loss": -0.007453325670212507,
      "memory(GiB)": 79.48,
      "step": 24,
      "train_speed(iter/s)": 0.002821
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.001253594527952373,
      "clip_ratio/low_min": 0.001253594527952373,
      "clip_ratio/region_mean": 0.001253594527952373,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 51.0,
      "completions/mean_length": 9.9375,
      "completions/min_length": 3.0,
      "epoch": 0.29850746268656714,
      "grad_norm": 1.3884331155836789,
      "kl": 0.5089308844180778,
      "learning_rate": 0.00019491843487067306,
      "loss": 0.00490172766149044,
      "memory(GiB)": 79.48,
      "reward": 0.2265625,
      "reward_std": 0.46621453762054443,
      "rewards/Response_no_think_reward_1/mean": 0.2265625,
      "rewards/Response_no_think_reward_1/std": 0.9324832558631897,
      "step": 25,
      "train_speed(iter/s)": 0.002775
    },
    {
      "clip_ratio/high_max": 0.01939821511041373,
      "clip_ratio/high_mean": 0.01939821511041373,
      "clip_ratio/low_mean": 0.046715788485016674,
      "clip_ratio/low_min": 0.046715788485016674,
      "clip_ratio/region_mean": 0.06611400278052315,
      "epoch": 0.31044776119402984,
      "grad_norm": 7.610505518747904,
      "kl": 2.1003341920441017,
      "learning_rate": 0.00019426971406815463,
      "loss": 0.024694038555026054,
      "memory(GiB)": 79.48,
      "step": 26,
      "train_speed(iter/s)": 0.002827
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0030820727697573602,
      "clip_ratio/low_min": 0.0030820727697573602,
      "clip_ratio/region_mean": 0.0030820727697573602,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 51.0,
      "completions/mean_length": 9.2265625,
      "completions/min_length": 3.0,
      "epoch": 0.32238805970149254,
      "grad_norm": 2.593834753837433,
      "kl": 0.6449481542804278,
      "learning_rate": 0.00019358324837779863,
      "loss": 0.007649811916053295,
      "memory(GiB)": 79.48,
      "reward": 0.2890625,
      "reward_std": 0.6288036704063416,
      "rewards/Response_no_think_reward_1/mean": 0.2890625,
      "rewards/Response_no_think_reward_1/std": 0.870806872844696,
      "step": 27,
      "train_speed(iter/s)": 0.00279
    },
    {
      "clip_ratio/high_max": 0.02224817470414564,
      "clip_ratio/high_mean": 0.02224817470414564,
      "clip_ratio/low_mean": 0.03963817795738578,
      "clip_ratio/low_min": 0.03963817795738578,
      "clip_ratio/region_mean": 0.061886353767476976,
      "epoch": 0.33432835820895523,
      "grad_norm": 4.001134871741585,
      "kl": 2.057775880966801,
      "learning_rate": 0.00019285931265532871,
      "loss": 0.024306027218699455,
      "memory(GiB)": 79.48,
      "step": 28,
      "train_speed(iter/s)": 0.002839
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0025941890198737383,
      "clip_ratio/low_min": 0.0025941890198737383,
      "clip_ratio/region_mean": 0.0025941890198737383,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 51.0,
      "completions/mean_length": 7.6640625,
      "completions/min_length": 3.0,
      "epoch": 0.34626865671641793,
      "grad_norm": 2.4322431330202017,
      "kl": 0.9759200520347804,
      "learning_rate": 0.0001920981967591891,
      "loss": 0.009954690001904964,
      "memory(GiB)": 79.48,
      "reward": 0.234375,
      "reward_std": 0.5239416360855103,
      "rewards/Response_no_think_reward_1/mean": 0.234375,
      "rewards/Response_no_think_reward_1/std": 0.8179518580436707,
      "step": 29,
      "train_speed(iter/s)": 0.002811
    },
    {
      "clip_ratio/high_max": 0.05181985488161445,
      "clip_ratio/high_mean": 0.05181985488161445,
      "clip_ratio/low_mean": 0.03574432339519262,
      "clip_ratio/low_min": 0.03574432339519262,
      "clip_ratio/region_mean": 0.0875641773454845,
      "epoch": 0.3582089552238806,
      "grad_norm": 2.469531358214721,
      "kl": 0.3759418617701158,
      "learning_rate": 0.00019130020543448704,
      "loss": 0.030206725001335144,
      "memory(GiB)": 79.48,
      "step": 30,
      "train_speed(iter/s)": 0.002857
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0007102272938936949,
      "clip_ratio/low_min": 0.0007102272938936949,
      "clip_ratio/region_mean": 0.0007102272938936949,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 51.0,
      "completions/mean_length": 7.4921875,
      "completions/min_length": 3.0,
      "epoch": 0.3701492537313433,
      "grad_norm": 25.502961039749618,
      "kl": 0.3221544102343614,
      "learning_rate": 0.00019046565819097545,
      "loss": 0.0061463117599487305,
      "memory(GiB)": 79.48,
      "reward": 0.265625,
      "reward_std": 0.3810158967971802,
      "rewards/Response_no_think_reward_1/mean": 0.265625,
      "rewards/Response_no_think_reward_1/std": 0.8082680702209473,
      "step": 31,
      "train_speed(iter/s)": 0.002819
    },
    {
      "clip_ratio/high_max": 0.0004340277810115367,
      "clip_ratio/high_mean": 0.0004340277810115367,
      "clip_ratio/low_mean": 0.015644782513845712,
      "clip_ratio/low_min": 0.015644782513845712,
      "clip_ratio/region_mean": 0.01607881032396108,
      "epoch": 0.382089552238806,
      "grad_norm": 2.386529018247983,
      "kl": 0.4242442559116171,
      "learning_rate": 0.0001895948891751234,
      "loss": 0.005282086320221424,
      "memory(GiB)": 79.48,
      "step": 32,
      "train_speed(iter/s)": 0.002862
    },
    {
      "clip_ratio/high_max": 0.0008561643771827221,
      "clip_ratio/high_mean": 0.0008561643771827221,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0008561643771827221,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 36.0,
      "completions/mean_length": 6.8984375,
      "completions/min_length": 3.0,
      "epoch": 0.3940298507462687,
      "grad_norm": 2.313143797352374,
      "kl": 1.4197996056172997,
      "learning_rate": 0.00018868824703632657,
      "loss": 0.011696252971887589,
      "memory(GiB)": 79.48,
      "reward": 0.046875,
      "reward_std": 0.6391555070877075,
      "rewards/Response_no_think_reward_1/mean": 0.046875,
      "rewards/Response_no_think_reward_1/std": 0.8215538859367371,
      "step": 33,
      "train_speed(iter/s)": 0.002823
    },
    {
      "clip_ratio/high_max": 0.03243769891560078,
      "clip_ratio/high_mean": 0.03243769891560078,
      "clip_ratio/low_mean": 0.057673962553963065,
      "clip_ratio/low_min": 0.057673962553963065,
      "clip_ratio/region_mean": 0.09011166216805577,
      "epoch": 0.4059701492537313,
      "grad_norm": 1.7539402550269323,
      "kl": 1.9242291511036456,
      "learning_rate": 0.00018774609478731046,
      "loss": -0.004083915147930384,
      "memory(GiB)": 79.48,
      "step": 34,
      "train_speed(iter/s)": 0.002863
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0010245901066809893,
      "clip_ratio/low_min": 0.0010245901066809893,
      "clip_ratio/region_mean": 0.0010245901066809893,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 51.0,
      "completions/mean_length": 7.7109375,
      "completions/min_length": 3.0,
      "epoch": 0.417910447761194,
      "grad_norm": 2.4226244872635485,
      "kl": 1.010538819245994,
      "learning_rate": 0.00018676880965878291,
      "loss": 0.015125149860978127,
      "memory(GiB)": 79.48,
      "reward": 0.0546875,
      "reward_std": 0.5867211818695068,
      "rewards/Response_no_think_reward_1/mean": 0.0546875,
      "rewards/Response_no_think_reward_1/std": 0.8539726138114929,
      "step": 35,
      "train_speed(iter/s)": 0.00283
    },
    {
      "clip_ratio/high_max": 0.030840664519928396,
      "clip_ratio/high_mean": 0.030840664519928396,
      "clip_ratio/low_mean": 0.06221208241186105,
      "clip_ratio/low_min": 0.06221208241186105,
      "clip_ratio/region_mean": 0.09305274731013924,
      "epoch": 0.4298507462686567,
      "grad_norm": 2.284561806665497,
      "kl": 2.4607611764222383,
      "learning_rate": 0.00018575678294839373,
      "loss": 0.010896118357777596,
      "memory(GiB)": 79.48,
      "step": 36,
      "train_speed(iter/s)": 0.002867
    },
    {
      "clip_ratio/high_max": 0.001179245300590992,
      "clip_ratio/high_mean": 0.001179245300590992,
      "clip_ratio/low_mean": 0.00924376118928194,
      "clip_ratio/low_min": 0.00924376118928194,
      "clip_ratio/region_mean": 0.010423006489872932,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 27.0,
      "completions/mean_length": 5.328125,
      "completions/min_length": 2.0,
      "epoch": 0.4417910447761194,
      "grad_norm": 3.3426494147901047,
      "kl": 0.3604982797114644,
      "learning_rate": 0.000184710419864062,
      "loss": 0.0029986475128680468,
      "memory(GiB)": 79.48,
      "reward": 0.328125,
      "reward_std": 0.34082794189453125,
      "rewards/Response_no_think_reward_1/mean": 0.328125,
      "rewards/Response_no_think_reward_1/std": 0.8794179558753967,
      "step": 37,
      "train_speed(iter/s)": 0.002821
    },
    {
      "clip_ratio/high_max": 0.011844758177176118,
      "clip_ratio/high_mean": 0.011844758177176118,
      "clip_ratio/low_mean": 0.05284481216222048,
      "clip_ratio/low_min": 0.05284481216222048,
      "clip_ratio/region_mean": 0.06468957080505788,
      "epoch": 0.4537313432835821,
      "grad_norm": 0.9022263920887504,
      "kl": 0.5205309305310948,
      "learning_rate": 0.00018363013936173393,
      "loss": -0.011295529082417488,
      "memory(GiB)": 79.48,
      "step": 38,
      "train_speed(iter/s)": 0.002857
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 51.0,
      "completions/mean_length": 5.328125,
      "completions/min_length": 3.0,
      "epoch": 0.46567164179104475,
      "grad_norm": 1.5271115661229067,
      "kl": 0.27687166831583454,
      "learning_rate": 0.00018251637397763597,
      "loss": 0.0014096113154664636,
      "memory(GiB)": 79.48,
      "reward": 0.046875,
      "reward_std": 0.3280481696128845,
      "rewards/Response_no_think_reward_1/mean": 0.046875,
      "rewards/Response_no_think_reward_1/std": 0.9379100203514099,
      "step": 39,
      "train_speed(iter/s)": 0.002785
    },
    {
      "clip_ratio/high_max": 0.04636762477457523,
      "clip_ratio/high_mean": 0.04636762477457523,
      "clip_ratio/low_mean": 0.02858343638945371,
      "clip_ratio/low_min": 0.02858343638945371,
      "clip_ratio/region_mean": 0.07495106221176684,
      "epoch": 0.47761194029850745,
      "grad_norm": 0.6061436422525854,
      "kl": 0.7858691609289963,
      "learning_rate": 0.00018136956965509064,
      "loss": -0.008444712497293949,
      "memory(GiB)": 79.48,
      "step": 40,
      "train_speed(iter/s)": 0.002819
    },
    {
      "clip_ratio/high_max": 0.001953125,
      "clip_ratio/high_mean": 0.001953125,
      "clip_ratio/low_mean": 0.0004006410308647901,
      "clip_ratio/low_min": 0.0004006410308647901,
      "clip_ratio/region_mean": 0.00235376603086479,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 51.0,
      "completions/mean_length": 8.296875,
      "completions/min_length": 3.0,
      "epoch": 0.48955223880597015,
      "grad_norm": 2.15651488822414,
      "kl": 0.5289879650626972,
      "learning_rate": 0.000180190185565964,
      "loss": 0.006278940010815859,
      "memory(GiB)": 79.48,
      "reward": 0.140625,
      "reward_std": 0.3606899678707123,
      "rewards/Response_no_think_reward_1/mean": 0.140625,
      "rewards/Response_no_think_reward_1/std": 0.9284173250198364,
      "step": 41,
      "train_speed(iter/s)": 0.002732
    },
    {
      "clip_ratio/high_max": 0.019255130202509463,
      "clip_ratio/high_mean": 0.019255130202509463,
      "clip_ratio/low_mean": 0.049764272407628596,
      "clip_ratio/low_min": 0.049764272407628596,
      "clip_ratio/region_mean": 0.06901940272655338,
      "epoch": 0.5014925373134328,
      "grad_norm": 7.678211490755348,
      "kl": 4.704732097394299,
      "learning_rate": 0.00017897869392681685,
      "loss": 0.036949772387742996,
      "memory(GiB)": 79.48,
      "step": 42,
      "train_speed(iter/s)": 0.002764
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0018887363257817924,
      "clip_ratio/low_min": 0.0018887363257817924,
      "clip_ratio/region_mean": 0.0018887363257817924,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 51.0,
      "completions/mean_length": 10.0390625,
      "completions/min_length": 3.0,
      "epoch": 0.5134328358208955,
      "grad_norm": 1.5758982248595066,
      "kl": 0.34190677378865075,
      "learning_rate": 0.00017773557980983262,
      "loss": 0.004829109646379948,
      "memory(GiB)": 79.48,
      "reward": 0.3671875,
      "reward_std": 0.355153352022171,
      "rewards/Response_no_think_reward_1/mean": 0.3671875,
      "rewards/Response_no_think_reward_1/std": 0.8406137228012085,
      "step": 43,
      "train_speed(iter/s)": 0.002649
    },
    {
      "clip_ratio/high_max": 0.010433136369101703,
      "clip_ratio/high_mean": 0.010433136369101703,
      "clip_ratio/low_mean": 0.0495308703975752,
      "clip_ratio/low_min": 0.0495308703975752,
      "clip_ratio/region_mean": 0.05996400595176965,
      "epoch": 0.5253731343283582,
      "grad_norm": 0.9699999569587556,
      "kl": 0.5136687369667925,
      "learning_rate": 0.00017646134094859815,
      "loss": -0.008213551715016365,
      "memory(GiB)": 79.48,
      "step": 44,
      "train_speed(iter/s)": 0.00268
    },
    {
      "clip_ratio/high_max": 0.0009920635493472219,
      "clip_ratio/high_mean": 0.0009920635493472219,
      "clip_ratio/low_mean": 0.0006793478387407959,
      "clip_ratio/low_min": 0.0006793478387407959,
      "clip_ratio/region_mean": 0.0016714113880880177,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 51.0,
      "completions/mean_length": 6.53125,
      "completions/min_length": 3.0,
      "epoch": 0.5373134328358209,
      "grad_norm": 2.6187879748810294,
      "kl": 0.4996573789976537,
      "learning_rate": 0.00017515648753881492,
      "loss": 0.00461646867915988,
      "memory(GiB)": 79.48,
      "reward": 0.390625,
      "reward_std": 0.4600672125816345,
      "rewards/Response_no_think_reward_1/mean": 0.390625,
      "rewards/Response_no_think_reward_1/std": 0.8251401782035828,
      "step": 45,
      "train_speed(iter/s)": 0.002583
    },
    {
      "clip_ratio/high_max": 0.023812146624550223,
      "clip_ratio/high_mean": 0.023812146624550223,
      "clip_ratio/low_mean": 0.030458416498731822,
      "clip_ratio/low_min": 0.030458416498731822,
      "clip_ratio/region_mean": 0.054270562424790114,
      "epoch": 0.5492537313432836,
      "grad_norm": 1.046619303634399,
      "kl": 1.1794932310585864,
      "learning_rate": 0.00017382154203402048,
      "loss": -0.007490159012377262,
      "memory(GiB)": 79.48,
      "step": 46,
      "train_speed(iter/s)": 0.002612
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0014244693447835743,
      "clip_ratio/low_min": 0.0014244693447835743,
      "clip_ratio/region_mean": 0.0014244693447835743,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 51.0,
      "completions/mean_length": 11.59375,
      "completions/min_length": 3.0,
      "epoch": 0.5611940298507463,
      "grad_norm": 1.8613572071063658,
      "kl": 0.6177221334655769,
      "learning_rate": 0.0001724570389364019,
      "loss": 0.004859148990362883,
      "memory(GiB)": 79.48,
      "reward": 0.0625,
      "reward_std": 0.47469794750213623,
      "rewards/Response_no_think_reward_1/mean": 0.0625,
      "rewards/Response_no_think_reward_1/std": 0.8761943578720093,
      "step": 47,
      "train_speed(iter/s)": 0.002518
    },
    {
      "clip_ratio/high_max": 0.009966709709260613,
      "clip_ratio/high_mean": 0.009966709709260613,
      "clip_ratio/low_mean": 0.11994545813649893,
      "clip_ratio/low_min": 0.11994545813649893,
      "clip_ratio/region_mean": 0.12991216825321317,
      "epoch": 0.573134328358209,
      "grad_norm": 0.9444759818819449,
      "kl": 0.9089975492097437,
      "learning_rate": 0.00017106352458278522,
      "loss": -0.009144289419054985,
      "memory(GiB)": 79.48,
      "step": 48,
      "train_speed(iter/s)": 0.002547
    },
    {
      "clip_ratio/high_max": 0.00021043770539108664,
      "clip_ratio/high_mean": 0.00021043770539108664,
      "clip_ratio/low_mean": 0.002260449342429638,
      "clip_ratio/low_min": 0.002260449342429638,
      "clip_ratio/region_mean": 0.0024708870478207245,
      "completions/clipped_ratio": 0.1328125,
      "completions/max_length": 51.0,
      "completions/mean_length": 12.921875,
      "completions/min_length": 3.0,
      "epoch": 0.5850746268656717,
      "grad_norm": 1.7364401500647022,
      "kl": 0.6272030578693375,
      "learning_rate": 0.00016964155692588616,
      "loss": 0.004394113551825285,
      "memory(GiB)": 79.48,
      "reward": 0.1484375,
      "reward_std": 0.424231618642807,
      "rewards/Response_no_think_reward_1/mean": 0.1484375,
      "rewards/Response_no_think_reward_1/std": 0.914363443851471,
      "step": 49,
      "train_speed(iter/s)": 0.002518
    },
    {
      "clip_ratio/high_max": 0.03189032070804387,
      "clip_ratio/high_mean": 0.03189032070804387,
      "clip_ratio/low_mean": 0.04090093384729698,
      "clip_ratio/low_min": 0.04090093384729698,
      "clip_ratio/region_mean": 0.07279125548666343,
      "epoch": 0.5970149253731343,
      "grad_norm": 4.461755913109612,
      "kl": 1.2612105239531957,
      "learning_rate": 0.00016819170531091017,
      "loss": 0.0014673060504719615,
      "memory(GiB)": 79.48,
      "step": 50,
      "train_speed(iter/s)": 0.002545
    },
    {
      "clip_ratio/high_max": 0.0015625000232830644,
      "clip_ratio/high_mean": 0.0015625000232830644,
      "clip_ratio/low_mean": 0.0016522249497938901,
      "clip_ratio/low_min": 0.0016522249497938901,
      "clip_ratio/region_mean": 0.0032147249730769545,
      "completions/clipped_ratio": 0.0859375,
      "completions/max_length": 51.0,
      "completions/mean_length": 10.46875,
      "completions/min_length": 3.0,
      "epoch": 0.608955223880597,
      "grad_norm": 2.519038758282151,
      "kl": 0.549895275675226,
      "learning_rate": 0.00016671455024759067,
      "loss": 0.0033532448578625917,
      "memory(GiB)": 79.48,
      "reward": 0.1796875,
      "reward_std": 0.39886242151260376,
      "rewards/Response_no_think_reward_1/mean": 0.1796875,
      "rewards/Response_no_think_reward_1/std": 0.9258628487586975,
      "step": 51,
      "train_speed(iter/s)": 0.002473
    },
    {
      "clip_ratio/high_max": 0.03367150644771755,
      "clip_ratio/high_mean": 0.03367150644771755,
      "clip_ratio/low_mean": 0.0576377569232136,
      "clip_ratio/low_min": 0.0576377569232136,
      "clip_ratio/region_mean": 0.09130926383659244,
      "epoch": 0.6208955223880597,
      "grad_norm": 0.9238461637378925,
      "kl": 0.672293990442995,
      "learning_rate": 0.00016521068317775753,
      "loss": -0.019059473648667336,
      "memory(GiB)": 79.48,
      "step": 52,
      "train_speed(iter/s)": 0.002499
    },
    {
      "clip_ratio/high_max": 0.0013558201026171446,
      "clip_ratio/high_mean": 0.0013558201026171446,
      "clip_ratio/low_mean": 0.010322635032935068,
      "clip_ratio/low_min": 0.010322635032935068,
      "clip_ratio/region_mean": 0.011678455251967534,
      "completions/clipped_ratio": 0.1796875,
      "completions/max_length": 51.0,
      "completions/mean_length": 14.0390625,
      "completions/min_length": 3.0,
      "epoch": 0.6328358208955224,
      "grad_norm": 6.547271710761981,
      "kl": 1.0316203383699758,
      "learning_rate": 0.00016368070623852792,
      "loss": 0.006902866996824741,
      "memory(GiB)": 79.48,
      "reward": 0.1328125,
      "reward_std": 0.561459481716156,
      "rewards/Response_no_think_reward_1/mean": 0.1328125,
      "rewards/Response_no_think_reward_1/std": 0.9081528782844543,
      "step": 53,
      "train_speed(iter/s)": 0.002484
    },
    {
      "clip_ratio/high_max": 0.0014829304564045742,
      "clip_ratio/high_mean": 0.0014829304564045742,
      "clip_ratio/low_mean": 0.1896184734068811,
      "clip_ratio/low_min": 0.1896184734068811,
      "clip_ratio/region_mean": 0.19110140530392528,
      "epoch": 0.6447761194029851,
      "grad_norm": 3.3472874723788295,
      "kl": 2.011909826847841,
      "learning_rate": 0.00016212523202121544,
      "loss": 0.00577022647485137,
      "memory(GiB)": 79.48,
      "step": 54,
      "train_speed(iter/s)": 0.002509
    },
    {
      "clip_ratio/high_max": 0.0017361111240461469,
      "clip_ratio/high_mean": 0.0017361111240461469,
      "clip_ratio/low_mean": 0.004226236138492823,
      "clip_ratio/low_min": 0.004226236138492823,
      "clip_ratio/region_mean": 0.0059623472625389695,
      "completions/clipped_ratio": 0.0859375,
      "completions/max_length": 51.0,
      "completions/mean_length": 10.125,
      "completions/min_length": 3.0,
      "epoch": 0.6567164179104478,
      "grad_norm": 2.1721724253324193,
      "kl": 1.4456337625160813,
      "learning_rate": 0.00016054488332605283,
      "loss": 0.007076151203364134,
      "memory(GiB)": 79.48,
      "reward": 0.3046875,
      "reward_std": 0.4615631103515625,
      "rewards/Response_no_think_reward_1/mean": 0.3046875,
      "rewards/Response_no_think_reward_1/std": 0.8654214143753052,
      "step": 55,
      "train_speed(iter/s)": 0.002495
    },
    {
      "clip_ratio/high_max": 0.07161095441551879,
      "clip_ratio/high_mean": 0.07161095441551879,
      "clip_ratio/low_mean": 0.0391811344306916,
      "clip_ratio/low_min": 0.0391811344306916,
      "clip_ratio/region_mean": 0.11079208552837372,
      "epoch": 0.6686567164179105,
      "grad_norm": 1.6216131312943962,
      "kl": 0.9301650315364896,
      "learning_rate": 0.00015894029291282758,
      "loss": -0.018307652324438095,
      "memory(GiB)": 79.48,
      "step": 56,
      "train_speed(iter/s)": 0.002519
    },
    {
      "clip_ratio/high_max": 0.0028409091755747795,
      "clip_ratio/high_mean": 0.0028409091755747795,
      "clip_ratio/low_mean": 0.007847597065847367,
      "clip_ratio/low_min": 0.007847597065847367,
      "clip_ratio/region_mean": 0.010688506241422147,
      "completions/clipped_ratio": 0.0703125,
      "completions/max_length": 51.0,
      "completions/mean_length": 8.2109375,
      "completions/min_length": 3.0,
      "epoch": 0.6805970149253732,
      "grad_norm": 4.528309911623671,
      "kl": 1.2233935340000244,
      "learning_rate": 0.00015731210324752972,
      "loss": 0.015045193955302238,
      "memory(GiB)": 79.48,
      "reward": 0.171875,
      "reward_std": 0.5126261711120605,
      "rewards/Response_no_think_reward_1/mean": 0.171875,
      "rewards/Response_no_think_reward_1/std": 0.9315922856330872,
      "step": 57,
      "train_speed(iter/s)": 0.002481
    },
    {
      "clip_ratio/high_max": 0.022757820785045624,
      "clip_ratio/high_mean": 0.022757820785045624,
      "clip_ratio/low_mean": 0.14223288698121905,
      "clip_ratio/low_min": 0.14223288698121905,
      "clip_ratio/region_mean": 0.1649907068349421,
      "epoch": 0.6925373134328359,
      "grad_norm": 75.05881389891198,
      "kl": 71.87919758663793,
      "learning_rate": 0.00015566096624511307,
      "loss": 0.9493598341941833,
      "memory(GiB)": 79.48,
      "step": 58,
      "train_speed(iter/s)": 0.002504
    },
    {
      "clip_ratio/high_max": 0.0005230125389061868,
      "clip_ratio/high_mean": 0.0005230125389061868,
      "clip_ratio/low_mean": 0.012553135456983,
      "clip_ratio/low_min": 0.012553135456983,
      "clip_ratio/region_mean": 0.013076147937681526,
      "completions/clipped_ratio": 0.1171875,
      "completions/max_length": 51.0,
      "completions/mean_length": 9.3828125,
      "completions/min_length": 3.0,
      "epoch": 0.7044776119402985,
      "grad_norm": 2.061647426051912,
      "kl": 3.4461456341377925,
      "learning_rate": 0.00015398754300847343,
      "loss": 0.024864066392183304,
      "memory(GiB)": 79.48,
      "reward": 0.3046875,
      "reward_std": 0.35084033012390137,
      "rewards/Response_no_think_reward_1/mean": 0.3046875,
      "rewards/Response_no_think_reward_1/std": 0.9010806679725647,
      "step": 59,
      "train_speed(iter/s)": 0.002474
    },
    {
      "clip_ratio/high_max": 0.010577498760540038,
      "clip_ratio/high_mean": 0.010577498760540038,
      "clip_ratio/low_mean": 0.15099563542753458,
      "clip_ratio/low_min": 0.15099563542753458,
      "clip_ratio/region_mean": 0.16157313529402018,
      "epoch": 0.7164179104477612,
      "grad_norm": 6.318543519105419,
      "kl": 9.211844031116925,
      "learning_rate": 0.00015229250356374804,
      "loss": 0.058415383100509644,
      "memory(GiB)": 79.48,
      "step": 60,
      "train_speed(iter/s)": 0.002497
    },
    {
      "clip_ratio/high_max": 0.0023436461633536965,
      "clip_ratio/high_mean": 0.0023436461633536965,
      "clip_ratio/low_mean": 0.01256567076779902,
      "clip_ratio/low_min": 0.01256567076779902,
      "clip_ratio/region_mean": 0.014909316785633564,
      "completions/clipped_ratio": 0.1640625,
      "completions/max_length": 51.0,
      "completions/mean_length": 9.0546875,
      "completions/min_length": 3.0,
      "epoch": 0.7283582089552239,
      "grad_norm": 8.650077130485371,
      "kl": 12.875115153714432,
      "learning_rate": 0.00015057652659204197,
      "loss": 0.11098186671733856,
      "memory(GiB)": 79.48,
      "reward": -0.2109375,
      "reward_std": 0.3248431980609894,
      "rewards/Response_no_think_reward_1/mean": -0.2109375,
      "rewards/Response_no_think_reward_1/std": 0.7060185074806213,
      "step": 61,
      "train_speed(iter/s)": 0.002485
    },
    {
      "clip_ratio/high_max": 0.04812772118020803,
      "clip_ratio/high_mean": 0.04812772118020803,
      "clip_ratio/low_mean": 0.016272668377496302,
      "clip_ratio/low_min": 0.016272668377496302,
      "clip_ratio/region_mean": 0.06440038897562772,
      "epoch": 0.7402985074626866,
      "grad_norm": 6.330882663211948,
      "kl": 4.831557432677073,
      "learning_rate": 0.00014884029915768944,
      "loss": 0.08532620966434479,
      "memory(GiB)": 79.48,
      "step": 62,
      "train_speed(iter/s)": 0.002507
    },
    {
      "clip_ratio/high_max": 0.0009259259095415473,
      "clip_ratio/high_mean": 0.0009259259095415473,
      "clip_ratio/low_mean": 0.005223244777880609,
      "clip_ratio/low_min": 0.005223244777880609,
      "clip_ratio/region_mean": 0.006149170687422156,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 49.0,
      "completions/mean_length": 6.8203125,
      "completions/min_length": 3.0,
      "epoch": 0.7522388059701492,
      "grad_norm": 0.8455863083310526,
      "kl": 3.406710424387711,
      "learning_rate": 0.00014708451643315827,
      "loss": 0.015459287911653519,
      "memory(GiB)": 79.48,
      "reward": 0.2890625,
      "reward_std": 0.09522313624620438,
      "rewards/Response_no_think_reward_1/mean": 0.2890625,
      "rewards/Response_no_think_reward_1/std": 0.9319553971290588,
      "step": 63,
      "train_speed(iter/s)": 0.002478
    },
    {
      "clip_ratio/high_max": 0.00038343557389453053,
      "clip_ratio/high_mean": 0.00038343557389453053,
      "clip_ratio/low_mean": 0.03692773682996631,
      "clip_ratio/low_min": 0.03692773682996631,
      "clip_ratio/region_mean": 0.037311173509806395,
      "epoch": 0.764179104477612,
      "grad_norm": 1.460493225059439,
      "kl": 3.925420918938471,
      "learning_rate": 0.00014530988142070803,
      "loss": 0.015160211361944675,
      "memory(GiB)": 79.48,
      "step": 64,
      "train_speed(iter/s)": 0.002499
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.02216397225856781,
      "clip_ratio/low_min": 0.02216397225856781,
      "clip_ratio/region_mean": 0.02216397225856781,
      "completions/clipped_ratio": 0.15625,
      "completions/max_length": 51.0,
      "completions/mean_length": 8.1484375,
      "completions/min_length": 3.0,
      "epoch": 0.7761194029850746,
      "grad_norm": 16.54465377888419,
      "kl": 13.583033993607387,
      "learning_rate": 0.00014351710467091336,
      "loss": 0.09176836162805557,
      "memory(GiB)": 79.48,
      "reward": 0.0,
      "reward_std": 0.44421693682670593,
      "rewards/Response_no_think_reward_1/mean": 0.0,
      "rewards/Response_no_think_reward_1/std": 0.8418202996253967,
      "step": 65,
      "train_speed(iter/s)": 0.00249
    },
    {
      "clip_ratio/high_max": 0.016292952001094818,
      "clip_ratio/high_mean": 0.016292952001094818,
      "clip_ratio/low_mean": 0.02969469508389011,
      "clip_ratio/low_min": 0.02969469508389011,
      "clip_ratio/region_mean": 0.04598764737602323,
      "epoch": 0.7880597014925373,
      "grad_norm": 7.478581616879351,
      "kl": 9.063416050747037,
      "learning_rate": 0.0001417069039981647,
      "loss": 0.08187955617904663,
      "memory(GiB)": 79.48,
      "step": 66,
      "train_speed(iter/s)": 0.00251
    },
    {
      "clip_ratio/high_max": 0.0003571428533177823,
      "clip_ratio/high_mean": 0.0003571428533177823,
      "clip_ratio/low_mean": 0.011069363099522889,
      "clip_ratio/low_min": 0.011069363099522889,
      "clip_ratio/region_mean": 0.011426505981944501,
      "completions/clipped_ratio": 0.1171875,
      "completions/max_length": 48.0,
      "completions/mean_length": 6.75,
      "completions/min_length": 3.0,
      "epoch": 0.8,
      "grad_norm": 2.789323653853436,
      "kl": 6.344451879562257,
      "learning_rate": 0.00013988000419326072,
      "loss": 0.04125535860657692,
      "memory(GiB)": 79.48,
      "reward": 0.21875,
      "reward_std": 0.40496253967285156,
      "rewards/Response_no_think_reward_1/mean": 0.21875,
      "rewards/Response_no_think_reward_1/std": 0.8778777122497559,
      "step": 67,
      "train_speed(iter/s)": 0.0025
    },
    {
      "clip_ratio/high_max": 0.036289573181420565,
      "clip_ratio/high_mean": 0.036289573181420565,
      "clip_ratio/low_mean": 0.07330909674055874,
      "clip_ratio/low_min": 0.07330909674055874,
      "clip_ratio/region_mean": 0.10959866992197931,
      "epoch": 0.8119402985074626,
      "grad_norm": 3.351094606732551,
      "kl": 7.85955839138478,
      "learning_rate": 0.00013803713673320772,
      "loss": 0.0294882133603096,
      "memory(GiB)": 79.48,
      "step": 68,
      "train_speed(iter/s)": 0.00252
    },
    {
      "clip_ratio/high_max": 0.0034414873807691038,
      "clip_ratio/high_mean": 0.0034414873807691038,
      "clip_ratio/low_mean": 0.017763152020052075,
      "clip_ratio/low_min": 0.017763152020052075,
      "clip_ratio/region_mean": 0.021204639226198196,
      "completions/clipped_ratio": 0.1953125,
      "completions/max_length": 50.0,
      "completions/mean_length": 9.0390625,
      "completions/min_length": 3.0,
      "epoch": 0.8238805970149253,
      "grad_norm": 2.3936208496329043,
      "kl": 8.319812071829801,
      "learning_rate": 0.00013617903948834155,
      "loss": 0.0558692142367363,
      "memory(GiB)": 79.48,
      "reward": 0.1484375,
      "reward_std": 0.5162962079048157,
      "rewards/Response_no_think_reward_1/mean": 0.1484375,
      "rewards/Response_no_think_reward_1/std": 0.7642591595649719,
      "step": 69,
      "train_speed(iter/s)": 0.002511
    },
    {
      "clip_ratio/high_max": 0.05922704888507724,
      "clip_ratio/high_mean": 0.05922704888507724,
      "clip_ratio/low_mean": 0.03970964252948761,
      "clip_ratio/low_min": 0.03970964252948761,
      "clip_ratio/region_mean": 0.09893669094890356,
      "epoch": 0.835820895522388,
      "grad_norm": 6.476273530366032,
      "kl": 5.813117837678874,
      "learning_rate": 0.00013430645642688988,
      "loss": 0.04987351968884468,
      "memory(GiB)": 79.48,
      "step": 70,
      "train_speed(iter/s)": 0.00253
    },
    {
      "clip_ratio/high_max": 0.002159281895728782,
      "clip_ratio/high_mean": 0.002159281895728782,
      "clip_ratio/low_mean": 0.010578428045846522,
      "clip_ratio/low_min": 0.010578428045846522,
      "clip_ratio/region_mean": 0.012737710087094456,
      "completions/clipped_ratio": 0.1171875,
      "completions/max_length": 44.0,
      "completions/mean_length": 9.2421875,
      "completions/min_length": 3.0,
      "epoch": 0.8477611940298507,
      "grad_norm": 3.9784798974103217,
      "kl": 7.147874560090713,
      "learning_rate": 0.00013242013731709287,
      "loss": 0.06330341100692749,
      "memory(GiB)": 79.48,
      "reward": 0.015625,
      "reward_std": 0.39743539690971375,
      "rewards/Response_no_think_reward_1/mean": 0.015625,
      "rewards/Response_no_think_reward_1/std": 0.6985291242599487,
      "step": 71,
      "train_speed(iter/s)": 0.002526
    },
    {
      "clip_ratio/high_max": 0.014909500489011407,
      "clip_ratio/high_mean": 0.014909500489011407,
      "clip_ratio/low_mean": 0.05899865587707609,
      "clip_ratio/low_min": 0.05899865587707609,
      "clip_ratio/region_mean": 0.07390815508551896,
      "epoch": 0.8597014925373134,
      "grad_norm": 107.14584369632281,
      "kl": 5.481248813099228,
      "learning_rate": 0.00013052083742700173,
      "loss": 0.5873211026191711,
      "memory(GiB)": 79.48,
      "step": 72,
      "train_speed(iter/s)": 0.002544
    },
    {
      "clip_ratio/high_max": 0.005898644914850593,
      "clip_ratio/high_mean": 0.005898644914850593,
      "clip_ratio/low_mean": 0.013190844911150634,
      "clip_ratio/low_min": 0.013190844911150634,
      "clip_ratio/region_mean": 0.01908948994241655,
      "completions/clipped_ratio": 0.1484375,
      "completions/max_length": 51.0,
      "completions/mean_length": 7.625,
      "completions/min_length": 3.0,
      "epoch": 0.8716417910447761,
      "grad_norm": 5.671656969858763,
      "kl": 5.887194114271551,
      "learning_rate": 0.0001286093172220748,
      "loss": 0.05823807045817375,
      "memory(GiB)": 79.48,
      "reward": 0.0078125,
      "reward_std": 0.47193682193756104,
      "rewards/Response_no_think_reward_1/mean": 0.0078125,
      "rewards/Response_no_think_reward_1/std": 0.8648526072502136,
      "step": 73,
      "train_speed(iter/s)": 0.002537
    },
    {
      "clip_ratio/high_max": 0.05940199224278331,
      "clip_ratio/high_mean": 0.05940199224278331,
      "clip_ratio/low_mean": 0.07782801624853164,
      "clip_ratio/low_min": 0.07782801624853164,
      "clip_ratio/region_mean": 0.1372300088405609,
      "epoch": 0.8835820895522388,
      "grad_norm": 5.588550612212405,
      "kl": 9.119772452046163,
      "learning_rate": 0.00012668634206069304,
      "loss": 0.05784386023879051,
      "memory(GiB)": 79.48,
      "step": 74,
      "train_speed(iter/s)": 0.002555
    },
    {
      "clip_ratio/high_max": 0.0005630630766972899,
      "clip_ratio/high_mean": 0.0005630630766972899,
      "clip_ratio/low_mean": 0.009406007709912956,
      "clip_ratio/low_min": 0.009406007709912956,
      "clip_ratio/region_mean": 0.009969070786610246,
      "completions/clipped_ratio": 0.0546875,
      "completions/max_length": 43.0,
      "completions/mean_length": 7.3671875,
      "completions/min_length": 3.0,
      "epoch": 0.8955223880597015,
      "grad_norm": 40.351102862526105,
      "kl": 37.61209568884806,
      "learning_rate": 0.00012475268188771627,
      "loss": 0.25722047686576843,
      "memory(GiB)": 79.48,
      "reward": 0.515625,
      "reward_std": 0.3077147901058197,
      "rewards/Response_no_think_reward_1/mean": 0.515625,
      "rewards/Response_no_think_reward_1/std": 0.5883966684341431,
      "step": 75,
      "train_speed(iter/s)": 0.00255
    },
    {
      "clip_ratio/high_max": 0.03623928187880665,
      "clip_ratio/high_mean": 0.03623928187880665,
      "clip_ratio/low_mean": 0.00604772218503058,
      "clip_ratio/low_min": 0.00604772218503058,
      "clip_ratio/region_mean": 0.04228700394742191,
      "epoch": 0.9074626865671642,
      "grad_norm": 17.208461552664357,
      "kl": 6.087301138206385,
      "learning_rate": 0.00012280911092620297,
      "loss": 0.16766606271266937,
      "memory(GiB)": 79.48,
      "step": 76,
      "train_speed(iter/s)": 0.002568
    },
    {
      "clip_ratio/high_max": 0.0016891892300918698,
      "clip_ratio/high_mean": 0.0016891892300918698,
      "clip_ratio/low_mean": 0.012179655604995787,
      "clip_ratio/low_min": 0.012179655604995787,
      "clip_ratio/region_mean": 0.013868844718672335,
      "completions/clipped_ratio": 0.125,
      "completions/max_length": 39.0,
      "completions/mean_length": 7.21875,
      "completions/min_length": 3.0,
      "epoch": 0.9194029850746268,
      "grad_norm": 4.3979207862679734,
      "kl": 7.126492358453106,
      "learning_rate": 0.00012085640736741708,
      "loss": 0.06262560933828354,
      "memory(GiB)": 79.48,
      "reward": 0.0546875,
      "reward_std": 0.39322251081466675,
      "rewards/Response_no_think_reward_1/mean": 0.0546875,
      "rewards/Response_no_think_reward_1/std": 0.7561672329902649,
      "step": 77,
      "train_speed(iter/s)": 0.002562
    },
    {
      "clip_ratio/high_max": 0.026138584362342954,
      "clip_ratio/high_mean": 0.026138584362342954,
      "clip_ratio/low_mean": 0.04480157361831516,
      "clip_ratio/low_min": 0.04480157361831516,
      "clip_ratio/region_mean": 0.07094015786424279,
      "epoch": 0.9313432835820895,
      "grad_norm": 2.916485581854171,
      "kl": 4.029124836612027,
      "learning_rate": 0.00011889535305924618,
      "loss": 0.04432527720928192,
      "memory(GiB)": 79.48,
      "step": 78,
      "train_speed(iter/s)": 0.002579
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0125213154242374,
      "clip_ratio/low_min": 0.0125213154242374,
      "clip_ratio/region_mean": 0.0125213154242374,
      "completions/clipped_ratio": 0.1015625,
      "completions/max_length": 41.0,
      "completions/mean_length": 7.5703125,
      "completions/min_length": 3.0,
      "epoch": 0.9432835820895522,
      "grad_norm": 5.401724081079944,
      "kl": 2.715451994285104,
      "learning_rate": 0.00011692673319315541,
      "loss": 0.025846321135759354,
      "memory(GiB)": 79.48,
      "reward": 0.15625,
      "reward_std": 0.32662156224250793,
      "rewards/Response_no_think_reward_1/mean": 0.15625,
      "rewards/Response_no_think_reward_1/std": 0.778019905090332,
      "step": 79,
      "train_speed(iter/s)": 0.002572
    },
    {
      "clip_ratio/high_max": 0.02864583395421505,
      "clip_ratio/high_mean": 0.02864583395421505,
      "clip_ratio/low_mean": 0.06598890712484717,
      "clip_ratio/low_min": 0.06598890712484717,
      "clip_ratio/region_mean": 0.09463474014773965,
      "epoch": 0.9552238805970149,
      "grad_norm": 1.7607357402294221,
      "kl": 3.6259496022976236,
      "learning_rate": 0.00011495133598980263,
      "loss": 0.011118962429463863,
      "memory(GiB)": 79.48,
      "step": 80,
      "train_speed(iter/s)": 0.002589
    },
    {
      "clip_ratio/high_max": 0.0010000000474974513,
      "clip_ratio/high_mean": 0.0010000000474974513,
      "clip_ratio/low_mean": 0.009746301162522286,
      "clip_ratio/low_min": 0.009746301162522286,
      "clip_ratio/region_mean": 0.010746301151812077,
      "completions/clipped_ratio": 0.1015625,
      "completions/max_length": 45.0,
      "completions/mean_length": 7.421875,
      "completions/min_length": 3.0,
      "epoch": 0.9671641791044776,
      "grad_norm": 6.226807855082435,
      "kl": 3.5353202793048695,
      "learning_rate": 0.00011296995238344084,
      "loss": 0.0407019704580307,
      "memory(GiB)": 79.48,
      "reward": 0.2109375,
      "reward_std": 0.3596799373626709,
      "rewards/Response_no_think_reward_1/mean": 0.2109375,
      "rewards/Response_no_think_reward_1/std": 0.78019118309021,
      "step": 81,
      "train_speed(iter/s)": 0.002582
    },
    {
      "clip_ratio/high_max": 0.025790343526750803,
      "clip_ratio/high_mean": 0.025790343526750803,
      "clip_ratio/low_mean": 0.027053400757722557,
      "clip_ratio/low_min": 0.027053400757722557,
      "clip_ratio/region_mean": 0.052843744633719325,
      "epoch": 0.9791044776119403,
      "grad_norm": 43.50825312796573,
      "kl": 2.1143818981945515,
      "learning_rate": 0.00011098337570523396,
      "loss": 0.1263183057308197,
      "memory(GiB)": 79.48,
      "step": 82,
      "train_speed(iter/s)": 0.002599
    },
    {
      "clip_ratio/high_max": 0.0045458003878593445,
      "clip_ratio/high_mean": 0.0045458003878593445,
      "clip_ratio/low_mean": 0.004786531790159643,
      "clip_ratio/low_min": 0.004786531790159643,
      "clip_ratio/region_mean": 0.009332332178018987,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 23.0,
      "completions/mean_length": 5.6796875,
      "completions/min_length": 3.0,
      "epoch": 1.0119402985074626,
      "grad_norm": 5.294876949688284,
      "kl": 1.5498897642164593,
      "learning_rate": 0.00010899240136561299,
      "loss": 0.011598482728004456,
      "memory(GiB)": 79.48,
      "reward": 0.4453125,
      "reward_std": 0.1938612163066864,
      "rewards/Response_no_think_reward_1/mean": 0.4453125,
      "rewards/Response_no_think_reward_1/std": 0.7187981605529785,
      "step": 83,
      "train_speed(iter/s)": 0.002589
    },
    {
      "clip_ratio/high_max": 0.025593542493879795,
      "clip_ratio/high_mean": 0.025593542493879795,
      "clip_ratio/low_mean": 0.025054577039554715,
      "clip_ratio/low_min": 0.025054577039554715,
      "clip_ratio/region_mean": 0.05064811953343451,
      "epoch": 1.0238805970149254,
      "grad_norm": 2.6737027093734143,
      "kl": 1.398324867649535,
      "learning_rate": 0.00010699782653579973,
      "loss": -0.00029344583163037896,
      "memory(GiB)": 79.48,
      "step": 84,
      "train_speed(iter/s)": 0.002605
    },
    {
      "clip_ratio/high_max": 0.0017993993242271245,
      "clip_ratio/high_mean": 0.0017993993242271245,
      "clip_ratio/low_mean": 0.01037665200419724,
      "clip_ratio/low_min": 0.01037665200419724,
      "clip_ratio/region_mean": 0.012176051270216703,
      "completions/clipped_ratio": 0.171875,
      "completions/max_length": 42.0,
      "completions/mean_length": 6.125,
      "completions/min_length": 3.0,
      "epoch": 1.035820895522388,
      "grad_norm": 2.5145149964606297,
      "kl": 3.1435449587497715,
      "learning_rate": 0.00010500044982862519,
      "loss": 0.024344690144062042,
      "memory(GiB)": 79.48,
      "reward": 0.4765625,
      "reward_std": 0.2772369980812073,
      "rewards/Response_no_think_reward_1/mean": 0.4765625,
      "rewards/Response_no_think_reward_1/std": 0.8414914011955261,
      "step": 85,
      "train_speed(iter/s)": 0.002595
    },
    {
      "clip_ratio/high_max": 0.021866617840714753,
      "clip_ratio/high_mean": 0.021866617840714753,
      "clip_ratio/low_mean": 0.07049822807312012,
      "clip_ratio/low_min": 0.07049822807312012,
      "clip_ratio/region_mean": 0.09236484486609697,
      "epoch": 1.0477611940298508,
      "grad_norm": 2.607332219212116,
      "kl": 4.641115245953188,
      "learning_rate": 0.00010300107097877114,
      "loss": 0.010013583116233349,
      "memory(GiB)": 79.48,
      "step": 86,
      "train_speed(iter/s)": 0.00261
    },
    {
      "clip_ratio/high_max": 0.0018382353009656072,
      "clip_ratio/high_mean": 0.0018382353009656072,
      "clip_ratio/low_mean": 0.0032051282469183207,
      "clip_ratio/low_min": 0.0032051282469183207,
      "clip_ratio/region_mean": 0.005043363547883928,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 20.0,
      "completions/mean_length": 5.4375,
      "completions/min_length": 3.0,
      "epoch": 1.0597014925373134,
      "grad_norm": 8.56914849705318,
      "kl": 0.7865612751156732,
      "learning_rate": 0.00010100049052256235,
      "loss": 0.013117787428200245,
      "memory(GiB)": 79.48,
      "reward": 0.53125,
      "reward_std": 0.1860944628715515,
      "rewards/Response_no_think_reward_1/mean": 0.53125,
      "rewards/Response_no_think_reward_1/std": 0.7202088236808777,
      "step": 87,
      "train_speed(iter/s)": 0.002597
    },
    {
      "clip_ratio/high_max": 0.009521116502583027,
      "clip_ratio/high_mean": 0.009521116502583027,
      "clip_ratio/low_mean": 0.0203368536895141,
      "clip_ratio/low_min": 0.0203368536895141,
      "clip_ratio/region_mean": 0.029857969493605196,
      "epoch": 1.0716417910447762,
      "grad_norm": 3.3791848142288035,
      "kl": 0.6961851019823371,
      "learning_rate": 9.899950947743767e-05,
      "loss": 0.00659452797845006,
      "memory(GiB)": 79.48,
      "step": 88,
      "train_speed(iter/s)": 0.002612
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.004996612435206771,
      "clip_ratio/low_min": 0.004996612435206771,
      "clip_ratio/region_mean": 0.004996612435206771,
      "completions/clipped_ratio": 0.1171875,
      "completions/max_length": 31.0,
      "completions/mean_length": 5.2265625,
      "completions/min_length": 3.0,
      "epoch": 1.0835820895522388,
      "grad_norm": 3.6519414582158323,
      "kl": 4.849585925694555,
      "learning_rate": 9.699892902122886e-05,
      "loss": 0.06144208833575249,
      "memory(GiB)": 79.48,
      "reward": 0.1640625,
      "reward_std": 0.3352247476577759,
      "rewards/Response_no_think_reward_1/mean": 0.1640625,
      "rewards/Response_no_think_reward_1/std": 0.8303053379058838,
      "step": 89,
      "train_speed(iter/s)": 0.002606
    },
    {
      "clip_ratio/high_max": 0.017538156593218446,
      "clip_ratio/high_mean": 0.017538156593218446,
      "clip_ratio/low_mean": 0.03697483101859689,
      "clip_ratio/low_min": 0.03697483101859689,
      "clip_ratio/region_mean": 0.054512988310307264,
      "epoch": 1.0955223880597016,
      "grad_norm": 13.104523239395307,
      "kl": 2.37760816863738,
      "learning_rate": 9.499955017137484e-05,
      "loss": 0.10894370079040527,
      "memory(GiB)": 79.48,
      "step": 90,
      "train_speed(iter/s)": 0.002621
    },
    {
      "clip_ratio/high_max": 0.004422514699399471,
      "clip_ratio/high_mean": 0.004422514699399471,
      "clip_ratio/low_mean": 0.007080804323777556,
      "clip_ratio/low_min": 0.007080804323777556,
      "clip_ratio/region_mean": 0.01150331913959235,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 7.0,
      "completions/mean_length": 4.6640625,
      "completions/min_length": 3.0,
      "epoch": 1.1074626865671642,
      "grad_norm": 3.292956583617785,
      "kl": 3.1563512758002616,
      "learning_rate": 9.30021734642003e-05,
      "loss": 0.03713127598166466,
      "memory(GiB)": 79.48,
      "reward": 0.421875,
      "reward_std": 0.2630079686641693,
      "rewards/Response_no_think_reward_1/mean": 0.421875,
      "rewards/Response_no_think_reward_1/std": 0.7382611632347107,
      "step": 91,
      "train_speed(iter/s)": 0.002603
    },
    {
      "clip_ratio/high_max": 0.012390351388603449,
      "clip_ratio/high_mean": 0.012390351388603449,
      "clip_ratio/low_mean": 0.01934311306104064,
      "clip_ratio/low_min": 0.01934311306104064,
      "clip_ratio/region_mean": 0.03173346375115216,
      "epoch": 1.1194029850746268,
      "grad_norm": 6.523254504794387,
      "kl": 2.134530827694107,
      "learning_rate": 9.100759863438702e-05,
      "loss": 0.04548652470111847,
      "memory(GiB)": 79.48,
      "step": 92,
      "train_speed(iter/s)": 0.002618
    },
    {
      "clip_ratio/high_max": 0.0018115942366421223,
      "clip_ratio/high_mean": 0.0018115942366421223,
      "clip_ratio/low_mean": 0.010487846680916846,
      "clip_ratio/low_min": 0.010487846680916846,
      "clip_ratio/region_mean": 0.012299440917558968,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 17.0,
      "completions/mean_length": 4.7109375,
      "completions/min_length": 3.0,
      "epoch": 1.1313432835820896,
      "grad_norm": 2.8653500050121345,
      "kl": 3.8929527901345864,
      "learning_rate": 8.901662429476607e-05,
      "loss": 0.04490555822849274,
      "memory(GiB)": 79.48,
      "reward": 0.1796875,
      "reward_std": 0.25065141916275024,
      "rewards/Response_no_think_reward_1/mean": 0.1796875,
      "rewards/Response_no_think_reward_1/std": 0.9173188805580139,
      "step": 93,
      "train_speed(iter/s)": 0.00261
    },
    {
      "clip_ratio/high_max": 0.019497282337397337,
      "clip_ratio/high_mean": 0.019497282337397337,
      "clip_ratio/low_mean": 0.05975891789421439,
      "clip_ratio/low_min": 0.05975891789421439,
      "clip_ratio/region_mean": 0.07925620023161173,
      "epoch": 1.1432835820895522,
      "grad_norm": 8.612775013595648,
      "kl": 4.982872966560535,
      "learning_rate": 8.703004761655917e-05,
      "loss": 0.07919233292341232,
      "memory(GiB)": 79.48,
      "step": 94,
      "train_speed(iter/s)": 0.002625
    },
    {
      "clip_ratio/high_max": 0.0027173913549631834,
      "clip_ratio/high_mean": 0.0027173913549631834,
      "clip_ratio/low_mean": 0.005094834603369236,
      "clip_ratio/low_min": 0.005094834603369236,
      "clip_ratio/region_mean": 0.007812225958332419,
      "completions/clipped_ratio": 0.0859375,
      "completions/max_length": 10.0,
      "completions/mean_length": 4.9375,
      "completions/min_length": 3.0,
      "epoch": 1.155223880597015,
      "grad_norm": 10.776677331970285,
      "kl": 9.849615207003808,
      "learning_rate": 8.504866401019737e-05,
      "loss": 0.1374131143093109,
      "memory(GiB)": 79.48,
      "reward": 0.3203125,
      "reward_std": 0.14966703951358795,
      "rewards/Response_no_think_reward_1/mean": 0.3203125,
      "rewards/Response_no_think_reward_1/std": 0.6266219019889832,
      "step": 95,
      "train_speed(iter/s)": 0.002622
    },
    {
      "clip_ratio/high_max": 0.019021738320589066,
      "clip_ratio/high_mean": 0.019021738320589066,
      "clip_ratio/low_mean": 0.005626177066005766,
      "clip_ratio/low_min": 0.005626177066005766,
      "clip_ratio/region_mean": 0.024647915386594832,
      "epoch": 1.1671641791044776,
      "grad_norm": 3.0583039624376402,
      "kl": 4.468168576029711,
      "learning_rate": 8.307326680684461e-05,
      "loss": 0.07699939608573914,
      "memory(GiB)": 79.48,
      "step": 96,
      "train_speed(iter/s)": 0.002636
    },
    {
      "clip_ratio/high_max": 0.007378174108453095,
      "clip_ratio/high_mean": 0.007378174108453095,
      "clip_ratio/low_mean": 0.00924376118928194,
      "clip_ratio/low_min": 0.00924376118928194,
      "clip_ratio/region_mean": 0.016621935297735035,
      "completions/clipped_ratio": 0.0546875,
      "completions/max_length": 12.0,
      "completions/mean_length": 4.9453125,
      "completions/min_length": 3.0,
      "epoch": 1.1791044776119404,
      "grad_norm": 9.444776640622303,
      "kl": 3.863095655280631,
      "learning_rate": 8.110464694075383e-05,
      "loss": 0.02421477437019348,
      "memory(GiB)": 79.48,
      "reward": 0.4375,
      "reward_std": 0.25460314750671387,
      "rewards/Response_no_think_reward_1/mean": 0.4375,
      "rewards/Response_no_think_reward_1/std": 0.7290377616882324,
      "step": 97,
      "train_speed(iter/s)": 0.002626
    },
    {
      "clip_ratio/high_max": 0.02187895681709051,
      "clip_ratio/high_mean": 0.02187895681709051,
      "clip_ratio/low_mean": 0.005211503244936466,
      "clip_ratio/low_min": 0.005211503244936466,
      "clip_ratio/region_mean": 0.027090460062026978,
      "epoch": 1.191044776119403,
      "grad_norm": 4.753171281552747,
      "kl": 3.498804785136599,
      "learning_rate": 7.914359263258295e-05,
      "loss": 0.02348589338362217,
      "memory(GiB)": 79.48,
      "step": 98,
      "train_speed(iter/s)": 0.00264
    },
    {
      "clip_ratio/high_max": 0.0014204545877873898,
      "clip_ratio/high_mean": 0.0014204545877873898,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014204545877873898,
      "completions/clipped_ratio": 0.0546875,
      "completions/max_length": 11.0,
      "completions/mean_length": 5.5859375,
      "completions/min_length": 3.0,
      "epoch": 1.2029850746268655,
      "grad_norm": 4.853951435141314,
      "kl": 7.24955918840169,
      "learning_rate": 7.719088907379706e-05,
      "loss": 0.08564330637454987,
      "memory(GiB)": 79.48,
      "reward": 0.359375,
      "reward_std": 0.1706565022468567,
      "rewards/Response_no_think_reward_1/mean": 0.359375,
      "rewards/Response_no_think_reward_1/std": 0.6608795523643494,
      "step": 99,
      "train_speed(iter/s)": 0.002636
    },
    {
      "clip_ratio/high_max": 0.01420454541221261,
      "clip_ratio/high_mean": 0.01420454541221261,
      "clip_ratio/low_mean": 0.0028409091755747795,
      "clip_ratio/low_min": 0.0028409091755747795,
      "clip_ratio/region_mean": 0.017045455053448677,
      "epoch": 1.2149253731343284,
      "grad_norm": 2.2986744180745324,
      "kl": 4.886410776793127,
      "learning_rate": 7.524731811228374e-05,
      "loss": 0.06466365605592728,
      "memory(GiB)": 79.48,
      "step": 100,
      "train_speed(iter/s)": 0.00265
    },
    {
      "clip_ratio/high_max": 0.00620507646817714,
      "clip_ratio/high_mean": 0.00620507646817714,
      "clip_ratio/low_mean": 0.008002021699212492,
      "clip_ratio/low_min": 0.008002021699212492,
      "clip_ratio/region_mean": 0.014207098283804953,
      "completions/clipped_ratio": 0.0859375,
      "completions/max_length": 19.0,
      "completions/mean_length": 6.2890625,
      "completions/min_length": 3.0,
      "epoch": 1.2268656716417912,
      "grad_norm": 35.84498974880178,
      "kl": 13.474940237792907,
      "learning_rate": 7.331365793930698e-05,
      "loss": 0.23400571942329407,
      "memory(GiB)": 79.48,
      "reward": 0.3671875,
      "reward_std": 0.1965562105178833,
      "rewards/Response_no_think_reward_1/mean": 0.3671875,
      "rewards/Response_no_think_reward_1/std": 0.7410472631454468,
      "step": 101,
      "train_speed(iter/s)": 0.002642
    },
    {
      "clip_ratio/high_max": 0.011945601785555482,
      "clip_ratio/high_mean": 0.011945601785555482,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.011945601785555482,
      "epoch": 1.2388059701492538,
      "grad_norm": 27.884607230951485,
      "kl": 1.4154490869450456,
      "learning_rate": 7.139068277792523e-05,
      "loss": 0.2148125171661377,
      "memory(GiB)": 79.48,
      "step": 102,
      "train_speed(iter/s)": 0.002655
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.007009345572441816,
      "clip_ratio/low_min": 0.007009345572441816,
      "clip_ratio/region_mean": 0.007009345572441816,
      "completions/clipped_ratio": 0.0546875,
      "completions/max_length": 40.0,
      "completions/mean_length": 5.640625,
      "completions/min_length": 3.0,
      "epoch": 1.2507462686567163,
      "grad_norm": 2.3687718583378876,
      "kl": 1.394652240909636,
      "learning_rate": 6.94791625729983e-05,
      "loss": 0.018386229872703552,
      "memory(GiB)": 79.48,
      "reward": 0.6328125,
      "reward_std": 0.20033246278762817,
      "rewards/Response_no_think_reward_1/mean": 0.6328125,
      "rewards/Response_no_think_reward_1/std": 0.6258360743522644,
      "step": 103,
      "train_speed(iter/s)": 0.002645
    },
    {
      "clip_ratio/high_max": 0.031393789453431964,
      "clip_ratio/high_mean": 0.031393789453431964,
      "clip_ratio/low_mean": 0.01482532313093543,
      "clip_ratio/low_min": 0.01482532313093543,
      "clip_ratio/region_mean": 0.04621911211870611,
      "epoch": 1.2626865671641792,
      "grad_norm": 1.3586707954485446,
      "kl": 1.0225394079461694,
      "learning_rate": 6.757986268290712e-05,
      "loss": 0.011677918955683708,
      "memory(GiB)": 79.48,
      "step": 104,
      "train_speed(iter/s)": 0.002658
    },
    {
      "clip_ratio/high_max": 0.00795477326028049,
      "clip_ratio/high_mean": 0.00795477326028049,
      "clip_ratio/low_mean": 0.007858287775889039,
      "clip_ratio/low_min": 0.007858287775889039,
      "clip_ratio/region_mean": 0.015813061269000173,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 32.0,
      "completions/mean_length": 5.828125,
      "completions/min_length": 3.0,
      "epoch": 1.2746268656716417,
      "grad_norm": 16.43559312212207,
      "kl": 1.8490474935271664,
      "learning_rate": 6.569354357311014e-05,
      "loss": 0.03543264418840408,
      "memory(GiB)": 79.48,
      "reward": 0.3671875,
      "reward_std": 0.31494253873825073,
      "rewards/Response_no_think_reward_1/mean": 0.3671875,
      "rewards/Response_no_think_reward_1/std": 0.7923958897590637,
      "step": 105,
      "train_speed(iter/s)": 0.002652
    },
    {
      "clip_ratio/high_max": 0.011914517963305116,
      "clip_ratio/high_mean": 0.011914517963305116,
      "clip_ratio/low_mean": 0.03227056178729981,
      "clip_ratio/low_min": 0.03227056178729981,
      "clip_ratio/region_mean": 0.04418508114758879,
      "epoch": 1.2865671641791045,
      "grad_norm": 2.5408508655554636,
      "kl": 1.8903868702007571,
      "learning_rate": 6.382096051165847e-05,
      "loss": 0.014127654023468494,
      "memory(GiB)": 79.48,
      "step": 106,
      "train_speed(iter/s)": 0.002665
    },
    {
      "clip_ratio/high_max": 0.0030637255404144526,
      "clip_ratio/high_mean": 0.0030637255404144526,
      "clip_ratio/low_mean": 0.015822806861251593,
      "clip_ratio/low_min": 0.015822806861251593,
      "clip_ratio/region_mean": 0.018886532401666045,
      "completions/clipped_ratio": 0.109375,
      "completions/max_length": 47.0,
      "completions/mean_length": 6.7109375,
      "completions/min_length": 3.0,
      "epoch": 1.2985074626865671,
      "grad_norm": 5.1535076282650945,
      "kl": 2.8769574181642383,
      "learning_rate": 6.19628632667923e-05,
      "loss": 0.038440439850091934,
      "memory(GiB)": 79.48,
      "reward": -0.0625,
      "reward_std": 0.4184814691543579,
      "rewards/Response_no_think_reward_1/mean": -0.0625,
      "rewards/Response_no_think_reward_1/std": 0.8394786715507507,
      "step": 107,
      "train_speed(iter/s)": 0.002655
    },
    {
      "clip_ratio/high_max": 0.011060049437219277,
      "clip_ratio/high_mean": 0.011060049437219277,
      "clip_ratio/low_mean": 0.05394765589153394,
      "clip_ratio/low_min": 0.05394765589153394,
      "clip_ratio/region_mean": 0.06500770541606471,
      "epoch": 1.31044776119403,
      "grad_norm": 38.3461665329109,
      "kl": 2.019717270624824,
      "learning_rate": 6.011999580673931e-05,
      "loss": 0.16980835795402527,
      "memory(GiB)": 79.48,
      "step": 108,
      "train_speed(iter/s)": 0.002668
    },
    {
      "clip_ratio/high_max": 0.0016025641234591603,
      "clip_ratio/high_mean": 0.0016025641234591603,
      "clip_ratio/low_mean": 0.0032051282469183207,
      "clip_ratio/low_min": 0.0032051282469183207,
      "clip_ratio/region_mean": 0.004807692486792803,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 25.0,
      "completions/mean_length": 5.703125,
      "completions/min_length": 3.0,
      "epoch": 1.3223880597014925,
      "grad_norm": 1.9241770555383604,
      "kl": 0.4095292093356875,
      "learning_rate": 5.829309600183536e-05,
      "loss": 0.003778851358219981,
      "memory(GiB)": 79.48,
      "reward": 0.453125,
      "reward_std": 0.24464011192321777,
      "rewards/Response_no_think_reward_1/mean": 0.453125,
      "rewards/Response_no_think_reward_1/std": 0.8405039310455322,
      "step": 109,
      "train_speed(iter/s)": 0.00266
    },
    {
      "clip_ratio/high_max": 0.016826923470944166,
      "clip_ratio/high_mean": 0.016826923470944166,
      "clip_ratio/low_mean": 0.025240384973585606,
      "clip_ratio/low_min": 0.025240384973585606,
      "clip_ratio/region_mean": 0.042067307978868484,
      "epoch": 1.3343283582089551,
      "grad_norm": 4.00720293405374,
      "kl": 0.3944609015534297,
      "learning_rate": 5.648289532908666e-05,
      "loss": 0.0013700753916054964,
      "memory(GiB)": 79.48,
      "step": 110,
      "train_speed(iter/s)": 0.002673
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.012351020704954863,
      "clip_ratio/low_min": 0.012351020704954863,
      "clip_ratio/region_mean": 0.012351020704954863,
      "completions/clipped_ratio": 0.2265625,
      "completions/max_length": 23.0,
      "completions/mean_length": 6.03125,
      "completions/min_length": 3.0,
      "epoch": 1.346268656716418,
      "grad_norm": 5.342919033145562,
      "kl": 6.366508552979212,
      "learning_rate": 5.4690118579292015e-05,
      "loss": 0.08252155035734177,
      "memory(GiB)": 79.48,
      "reward": -0.0703125,
      "reward_std": 0.5281283259391785,
      "rewards/Response_no_think_reward_1/mean": -0.0703125,
      "rewards/Response_no_think_reward_1/std": 0.7954951524734497,
      "step": 111,
      "train_speed(iter/s)": 0.002667
    },
    {
      "clip_ratio/high_max": 0.0056201552506536245,
      "clip_ratio/high_mean": 0.0056201552506536245,
      "clip_ratio/low_mean": 0.07672008802182972,
      "clip_ratio/low_min": 0.07672008802182972,
      "clip_ratio/region_mean": 0.08234024350531399,
      "epoch": 1.3582089552238805,
      "grad_norm": 4.841852348859136,
      "kl": 7.482738017570227,
      "learning_rate": 5.291548356684177e-05,
      "loss": 0.06593235582113266,
      "memory(GiB)": 79.48,
      "step": 112,
      "train_speed(iter/s)": 0.00268
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.00446532026398927,
      "clip_ratio/low_min": 0.00446532026398927,
      "clip_ratio/region_mean": 0.00446532026398927,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 10.0,
      "completions/mean_length": 5.328125,
      "completions/min_length": 3.0,
      "epoch": 1.3701492537313433,
      "grad_norm": 2.4231207405852158,
      "kl": 2.7402262951763987,
      "learning_rate": 5.115970084231059e-05,
      "loss": 0.021689381450414658,
      "memory(GiB)": 79.48,
      "reward": 0.4453125,
      "reward_std": 0.33350807428359985,
      "rewards/Response_no_think_reward_1/mean": 0.4453125,
      "rewards/Response_no_think_reward_1/std": 0.7817665934562683,
      "step": 113,
      "train_speed(iter/s)": 0.002673
    },
    {
      "clip_ratio/high_max": 0.02434239676222205,
      "clip_ratio/high_mean": 0.02434239676222205,
      "clip_ratio/low_mean": 0.0223974745022133,
      "clip_ratio/low_min": 0.0223974745022133,
      "clip_ratio/region_mean": 0.046739870798774064,
      "epoch": 1.382089552238806,
      "grad_norm": 1.0551090300064008,
      "kl": 2.37436256357978,
      "learning_rate": 4.9423473407958035e-05,
      "loss": 0.008349667303264141,
      "memory(GiB)": 79.48,
      "step": 114,
      "train_speed(iter/s)": 0.002685
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.011147994082421064,
      "clip_ratio/low_min": 0.011147994082421064,
      "clip_ratio/region_mean": 0.011147994082421064,
      "completions/clipped_ratio": 0.15625,
      "completions/max_length": 11.0,
      "completions/mean_length": 5.3515625,
      "completions/min_length": 3.0,
      "epoch": 1.3940298507462687,
      "grad_norm": 35.484376517766954,
      "kl": 24.603704601235222,
      "learning_rate": 4.7707496436252e-05,
      "loss": 0.3364083468914032,
      "memory(GiB)": 79.48,
      "reward": 0.1328125,
      "reward_std": 0.28117600083351135,
      "rewards/Response_no_think_reward_1/mean": 0.1328125,
      "rewards/Response_no_think_reward_1/std": 0.8264437913894653,
      "step": 115,
      "train_speed(iter/s)": 0.002678
    },
    {
      "clip_ratio/high_max": 0.002314814832061529,
      "clip_ratio/high_mean": 0.002314814832061529,
      "clip_ratio/low_mean": 0.0054197743302211165,
      "clip_ratio/low_min": 0.0054197743302211165,
      "clip_ratio/region_mean": 0.007734589162282646,
      "epoch": 1.4059701492537313,
      "grad_norm": 7.954623939964381,
      "kl": 12.087648045126116,
      "learning_rate": 4.601245699152659e-05,
      "loss": 0.2005119025707245,
      "memory(GiB)": 79.48,
      "step": 116,
      "train_speed(iter/s)": 0.00269
    },
    {
      "clip_ratio/high_max": 0.0005434782360680401,
      "clip_ratio/high_mean": 0.0005434782360680401,
      "clip_ratio/low_mean": 0.004596828715875745,
      "clip_ratio/low_min": 0.004596828715875745,
      "clip_ratio/region_mean": 0.005140306951943785,
      "completions/clipped_ratio": 0.0703125,
      "completions/max_length": 40.0,
      "completions/mean_length": 5.8671875,
      "completions/min_length": 3.0,
      "epoch": 1.417910447761194,
      "grad_norm": 3.80570877135741,
      "kl": 3.121465804113541,
      "learning_rate": 4.433903375488697e-05,
      "loss": 0.044249728322029114,
      "memory(GiB)": 79.48,
      "reward": 0.3046875,
      "reward_std": 0.38101160526275635,
      "rewards/Response_no_think_reward_1/mean": 0.3046875,
      "rewards/Response_no_think_reward_1/std": 0.8562746047973633,
      "step": 117,
      "train_speed(iter/s)": 0.002684
    },
    {
      "clip_ratio/high_max": 0.014024957199580967,
      "clip_ratio/high_mean": 0.014024957199580967,
      "clip_ratio/low_mean": 0.0131244600052014,
      "clip_ratio/low_min": 0.0131244600052014,
      "clip_ratio/region_mean": 0.02714941732119769,
      "epoch": 1.4298507462686567,
      "grad_norm": 3.857626589553114,
      "kl": 2.146273965074215,
      "learning_rate": 4.268789675247029e-05,
      "loss": 0.03236865624785423,
      "memory(GiB)": 79.48,
      "step": 118,
      "train_speed(iter/s)": 0.002695
    },
    {
      "clip_ratio/high_max": 0.0012886597542092204,
      "clip_ratio/high_mean": 0.0012886597542092204,
      "clip_ratio/low_mean": 0.0051683366764336824,
      "clip_ratio/low_min": 0.0051683366764336824,
      "clip_ratio/region_mean": 0.006456996430642903,
      "completions/clipped_ratio": 0.0703125,
      "completions/max_length": 22.0,
      "completions/mean_length": 5.6171875,
      "completions/min_length": 3.0,
      "epoch": 1.4417910447761195,
      "grad_norm": 3.2126971808505513,
      "kl": 1.1720350683217475,
      "learning_rate": 4.105970708717244e-05,
      "loss": 0.018675971776247025,
      "memory(GiB)": 79.48,
      "reward": 0.5625,
      "reward_std": 0.22125522792339325,
      "rewards/Response_no_think_reward_1/mean": 0.5625,
      "rewards/Response_no_think_reward_1/std": 0.6490453481674194,
      "step": 119,
      "train_speed(iter/s)": 0.002674
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.020605165394954383,
      "clip_ratio/low_min": 0.020605165394954383,
      "clip_ratio/region_mean": 0.020605165394954383,
      "epoch": 1.4537313432835821,
      "grad_norm": 1.102830732170957,
      "kl": 1.583265769013451,
      "learning_rate": 3.945511667394719e-05,
      "loss": 0.003417772939428687,
      "memory(GiB)": 79.48,
      "step": 120,
      "train_speed(iter/s)": 0.002685
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.007155258092097938,
      "clip_ratio/low_min": 0.007155258092097938,
      "clip_ratio/region_mean": 0.007155258092097938,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 23.0,
      "completions/mean_length": 5.6640625,
      "completions/min_length": 3.0,
      "epoch": 1.4656716417910447,
      "grad_norm": 2.9093250272424034,
      "kl": 1.8708921101060696,
      "learning_rate": 3.787476797878459e-05,
      "loss": 0.01941162720322609,
      "memory(GiB)": 79.48,
      "reward": 0.59375,
      "reward_std": 0.3198433816432953,
      "rewards/Response_no_think_reward_1/mean": 0.59375,
      "rewards/Response_no_think_reward_1/std": 0.692337155342102,
      "step": 121,
      "train_speed(iter/s)": 0.002659
    },
    {
      "clip_ratio/high_max": 0.02566964237485081,
      "clip_ratio/high_mean": 0.02566964237485081,
      "clip_ratio/low_mean": 0.01106150820851326,
      "clip_ratio/low_min": 0.01106150820851326,
      "clip_ratio/region_mean": 0.03673115174751729,
      "epoch": 1.4776119402985075,
      "grad_norm": 1.6206572311785468,
      "kl": 2.092067622463219,
      "learning_rate": 3.631929376147207e-05,
      "loss": 0.00795831624418497,
      "memory(GiB)": 79.48,
      "step": 122,
      "train_speed(iter/s)": 0.00267
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 16.0,
      "completions/mean_length": 5.140625,
      "completions/min_length": 3.0,
      "epoch": 1.48955223880597,
      "grad_norm": 1.1898236332850007,
      "kl": 0.5545606576924911,
      "learning_rate": 3.47893168222425e-05,
      "loss": 0.005211046896874905,
      "memory(GiB)": 79.48,
      "reward": 0.5078125,
      "reward_std": 0.2012200653553009,
      "rewards/Response_no_think_reward_1/mean": 0.5078125,
      "rewards/Response_no_think_reward_1/std": 0.763292670249939,
      "step": 123,
      "train_speed(iter/s)": 0.002638
    },
    {
      "clip_ratio/high_max": 0.015350877307355404,
      "clip_ratio/high_mean": 0.015350877307355404,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.015350877307355404,
      "epoch": 1.5014925373134327,
      "grad_norm": 0.3409132689994619,
      "kl": 0.5577427046882804,
      "learning_rate": 3.328544975240932e-05,
      "loss": -0.0002708420215640217,
      "memory(GiB)": 79.48,
      "step": 124,
      "train_speed(iter/s)": 0.002649
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.00601851858664304,
      "clip_ratio/low_min": 0.00601851858664304,
      "clip_ratio/region_mean": 0.00601851858664304,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 30.0,
      "completions/mean_length": 5.8671875,
      "completions/min_length": 3.0,
      "epoch": 1.5134328358208955,
      "grad_norm": 5.906541393453947,
      "kl": 6.567164011168643,
      "learning_rate": 3.180829468908986e-05,
      "loss": 0.07308901846408844,
      "memory(GiB)": 79.48,
      "reward": 0.3671875,
      "reward_std": 0.1412346363067627,
      "rewards/Response_no_think_reward_1/mean": 0.3671875,
      "rewards/Response_no_think_reward_1/std": 0.8406137228012085,
      "step": 125,
      "train_speed(iter/s)": 0.002642
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.010185185354202986,
      "clip_ratio/low_min": 0.010185185354202986,
      "clip_ratio/region_mean": 0.010185185354202986,
      "epoch": 1.5253731343283583,
      "grad_norm": 1.637348051569756,
      "kl": 3.2303040275146486,
      "learning_rate": 3.035844307411384e-05,
      "loss": 0.04529271274805069,
      "memory(GiB)": 79.48,
      "step": 126,
      "train_speed(iter/s)": 0.002652
    },
    {
      "clip_ratio/high_max": 0.00046296295477077365,
      "clip_ratio/high_mean": 0.00046296295477077365,
      "clip_ratio/low_mean": 0.004759002127684653,
      "clip_ratio/low_min": 0.004759002127684653,
      "clip_ratio/region_mean": 0.0052219650824554265,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 29.0,
      "completions/mean_length": 6.4453125,
      "completions/min_length": 3.0,
      "epoch": 1.537313432835821,
      "grad_norm": 1.5814533572332585,
      "kl": 1.9979073457579943,
      "learning_rate": 2.8936475417214794e-05,
      "loss": 0.022465700283646584,
      "memory(GiB)": 79.48,
      "reward": 0.3359375,
      "reward_std": 0.17123225331306458,
      "rewards/Response_no_think_reward_1/mean": 0.3359375,
      "rewards/Response_no_think_reward_1/std": 0.7453514337539673,
      "step": 127,
      "train_speed(iter/s)": 0.00264
    },
    {
      "clip_ratio/high_max": 0.0014710274408571422,
      "clip_ratio/high_mean": 0.0014710274408571422,
      "clip_ratio/low_mean": 0.022759197046980262,
      "clip_ratio/low_min": 0.022759197046980262,
      "clip_ratio/region_mean": 0.024230224487837404,
      "epoch": 1.5492537313432835,
      "grad_norm": 0.9057768060895434,
      "kl": 1.9242111706989817,
      "learning_rate": 2.7542961063598104e-05,
      "loss": 0.013444737531244755,
      "memory(GiB)": 79.48,
      "step": 128,
      "train_speed(iter/s)": 0.002651
    },
    {
      "clip_ratio/high_max": 0.0026041667442768812,
      "clip_ratio/high_mean": 0.0026041667442768812,
      "clip_ratio/low_mean": 0.010856111068278551,
      "clip_ratio/low_min": 0.010856111068278551,
      "clip_ratio/region_mean": 0.013460277812555432,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 28.0,
      "completions/mean_length": 5.6875,
      "completions/min_length": 3.0,
      "epoch": 1.5611940298507463,
      "grad_norm": 2.279027191303148,
      "kl": 1.264391661035006,
      "learning_rate": 2.617845796597954e-05,
      "loss": 0.005756520200520754,
      "memory(GiB)": 79.48,
      "reward": 0.3671875,
      "reward_std": 0.15756267309188843,
      "rewards/Response_no_think_reward_1/mean": 0.3671875,
      "rewards/Response_no_think_reward_1/std": 0.7410472631454468,
      "step": 129,
      "train_speed(iter/s)": 0.002635
    },
    {
      "clip_ratio/high_max": 0.0026041667442768812,
      "clip_ratio/high_mean": 0.0026041667442768812,
      "clip_ratio/low_mean": 0.06580501515418291,
      "clip_ratio/low_min": 0.06580501515418291,
      "clip_ratio/region_mean": 0.06840918306261301,
      "epoch": 1.573134328358209,
      "grad_norm": 1.198012219937442,
      "kl": 1.689065290265944,
      "learning_rate": 2.484351246118507e-05,
      "loss": -0.000346488319337368,
      "memory(GiB)": 79.48,
      "step": 130,
      "train_speed(iter/s)": 0.002646
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0059523810632526875,
      "clip_ratio/low_min": 0.0059523810632526875,
      "clip_ratio/region_mean": 0.0059523810632526875,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 25.0,
      "completions/mean_length": 5.15625,
      "completions/min_length": 3.0,
      "epoch": 1.5850746268656717,
      "grad_norm": 2.595184247375665,
      "kl": 0.8083602132865053,
      "learning_rate": 2.353865905140187e-05,
      "loss": 0.009201270528137684,
      "memory(GiB)": 79.48,
      "reward": 0.4296875,
      "reward_std": 0.13553348183631897,
      "rewards/Response_no_think_reward_1/mean": 0.4296875,
      "rewards/Response_no_think_reward_1/std": 0.7496308088302612,
      "step": 131,
      "train_speed(iter/s)": 0.002635
    },
    {
      "clip_ratio/high_max": 0.0019841270986944437,
      "clip_ratio/high_mean": 0.0019841270986944437,
      "clip_ratio/low_mean": 0.01378506887704134,
      "clip_ratio/low_min": 0.01378506887704134,
      "clip_ratio/region_mean": 0.01576919574290514,
      "epoch": 1.5970149253731343,
      "grad_norm": 1.0694122716373746,
      "kl": 0.8849439112927939,
      "learning_rate": 2.226442019016739e-05,
      "loss": 0.0032178200781345367,
      "memory(GiB)": 79.48,
      "step": 132,
      "train_speed(iter/s)": 0.002646
    },
    {
      "clip_ratio/high_max": 0.0014880952658131719,
      "clip_ratio/high_mean": 0.0014880952658131719,
      "clip_ratio/low_mean": 0.004777232185006142,
      "clip_ratio/low_min": 0.004777232185006142,
      "clip_ratio/region_mean": 0.0062653274508193135,
      "completions/clipped_ratio": 0.0703125,
      "completions/max_length": 51.0,
      "completions/mean_length": 7.703125,
      "completions/min_length": 3.0,
      "epoch": 1.608955223880597,
      "grad_norm": 8.857193213468985,
      "kl": 6.557242290626164,
      "learning_rate": 2.1021306073183167e-05,
      "loss": 0.09046173840761185,
      "memory(GiB)": 79.48,
      "reward": 0.5625,
      "reward_std": 0.3117782771587372,
      "rewards/Response_no_think_reward_1/mean": 0.5625,
      "rewards/Response_no_think_reward_1/std": 0.6728714108467102,
      "step": 133,
      "train_speed(iter/s)": 0.002636
    },
    {
      "clip_ratio/high_max": 0.0043966451194137335,
      "clip_ratio/high_mean": 0.0043966451194137335,
      "clip_ratio/low_mean": 0.0072958533419296145,
      "clip_ratio/low_min": 0.0072958533419296145,
      "clip_ratio/region_mean": 0.01169249857775867,
      "epoch": 1.6208955223880597,
      "grad_norm": 5.293380513533871,
      "kl": 4.1661525671806885,
      "learning_rate": 1.9809814434036e-05,
      "loss": 0.05591786280274391,
      "memory(GiB)": 79.48,
      "step": 134,
      "train_speed(iter/s)": 0.002646
    },
    {
      "clip_ratio/high_max": 0.0006793478387407959,
      "clip_ratio/high_mean": 0.0006793478387407959,
      "clip_ratio/low_mean": 0.008976120705483481,
      "clip_ratio/low_min": 0.008976120705483481,
      "clip_ratio/region_mean": 0.009655468544224277,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 29.0,
      "completions/mean_length": 6.125,
      "completions/min_length": 3.0,
      "epoch": 1.6328358208955223,
      "grad_norm": 2.599750460948482,
      "kl": 2.9140283690503566,
      "learning_rate": 1.863043034490938e-05,
      "loss": 0.026728810742497444,
      "memory(GiB)": 79.48,
      "reward": 0.3828125,
      "reward_std": 0.2688143849372864,
      "rewards/Response_no_think_reward_1/mean": 0.3828125,
      "rewards/Response_no_think_reward_1/std": 0.7542122602462769,
      "step": 135,
      "train_speed(iter/s)": 0.002641
    },
    {
      "clip_ratio/high_max": 0.0006793478387407959,
      "clip_ratio/high_mean": 0.0006793478387407959,
      "clip_ratio/low_mean": 0.01892346324166283,
      "clip_ratio/low_min": 0.01892346324166283,
      "clip_ratio/region_mean": 0.019602811022195965,
      "epoch": 1.644776119402985,
      "grad_norm": 1.7776881222618184,
      "kl": 2.936662324536883,
      "learning_rate": 1.748362602236403e-05,
      "loss": 0.021627021953463554,
      "memory(GiB)": 79.48,
      "step": 136,
      "train_speed(iter/s)": 0.002651
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0003906250058207661,
      "clip_ratio/low_min": 0.0003906250058207661,
      "clip_ratio/region_mean": 0.0003906250058207661,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 24.0,
      "completions/mean_length": 6.6953125,
      "completions/min_length": 3.0,
      "epoch": 1.6567164179104479,
      "grad_norm": 2.8008543153134777,
      "kl": 1.5049916390344151,
      "learning_rate": 1.63698606382661e-05,
      "loss": 0.022346211597323418,
      "memory(GiB)": 79.48,
      "reward": 0.3828125,
      "reward_std": 0.15467959642410278,
      "rewards/Response_no_think_reward_1/mean": 0.3828125,
      "rewards/Response_no_think_reward_1/std": 0.7436988949775696,
      "step": 137,
      "train_speed(iter/s)": 0.002647
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.004379986581625417,
      "clip_ratio/low_min": 0.004379986581625417,
      "clip_ratio/region_mean": 0.004379986581625417,
      "epoch": 1.6686567164179105,
      "grad_norm": 1.5764765403480743,
      "kl": 1.4290150789947802,
      "learning_rate": 1.528958013593801e-05,
      "loss": 0.0181417278945446,
      "memory(GiB)": 79.48,
      "step": 138,
      "train_speed(iter/s)": 0.002657
    },
    {
      "clip_ratio/high_max": 0.0009469697251915932,
      "clip_ratio/high_mean": 0.0009469697251915932,
      "clip_ratio/low_mean": 0.009301686193794012,
      "clip_ratio/low_min": 0.009301686193794012,
      "clip_ratio/region_mean": 0.010248655918985605,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 25.0,
      "completions/mean_length": 5.15625,
      "completions/min_length": 3.0,
      "epoch": 1.680597014925373,
      "grad_norm": 2.987112951446729,
      "kl": 2.2646760795032606,
      "learning_rate": 1.4243217051606283e-05,
      "loss": 0.022022899240255356,
      "memory(GiB)": 79.48,
      "reward": 0.5625,
      "reward_std": 0.3713865876197815,
      "rewards/Response_no_think_reward_1/mean": 0.5625,
      "rewards/Response_no_think_reward_1/std": 0.6958821415901184,
      "step": 139,
      "train_speed(iter/s)": 0.002652
    },
    {
      "clip_ratio/high_max": 0.0010245901066809893,
      "clip_ratio/high_mean": 0.0010245901066809893,
      "clip_ratio/low_mean": 0.008475731243379414,
      "clip_ratio/low_min": 0.008475731243379414,
      "clip_ratio/region_mean": 0.009500321350060403,
      "epoch": 1.6925373134328359,
      "grad_norm": 2.122472644745234,
      "kl": 2.2440224051242694,
      "learning_rate": 1.3231190341217081e-05,
      "loss": 0.017425674945116043,
      "memory(GiB)": 79.48,
      "step": 140,
      "train_speed(iter/s)": 0.002662
    }
  ],
  "logging_steps": 1,
  "max_steps": 166,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 5,
  "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": 8,
  "trial_name": null,
  "trial_params": null
}