{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.9861932938856017,
  "eval_steps": 500,
  "global_step": 252,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 129.85416984558105,
      "epoch": 0.007889546351084813,
      "grad_norm": 3.8763763904571533,
      "kl": 0.0,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "reward": 0.9994155764579773,
      "reward_std": 0.7669541835784912,
      "rewards/format_reward": 0.5937500149011612,
      "rewards/segmentation_reward": 0.4056655764579773,
      "step": 1
    },
    {
      "completion_length": 136.31250381469727,
      "epoch": 0.015779092702169626,
      "grad_norm": 15.093283653259277,
      "kl": 0.0006837844848632812,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "reward": 1.158398061990738,
      "reward_std": 0.7105055451393127,
      "rewards/format_reward": 0.6979166865348816,
      "rewards/segmentation_reward": 0.4604813829064369,
      "step": 2
    },
    {
      "completion_length": 141.71875381469727,
      "epoch": 0.023668639053254437,
      "grad_norm": 6.756906032562256,
      "kl": 0.000885009765625,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "reward": 1.1017816215753555,
      "reward_std": 0.5777320563793182,
      "rewards/format_reward": 0.6770833507180214,
      "rewards/segmentation_reward": 0.4246982932090759,
      "step": 3
    },
    {
      "completion_length": 133.3958396911621,
      "epoch": 0.03155818540433925,
      "grad_norm": 10.368873596191406,
      "kl": 0.0015869140625,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "reward": 1.040566012263298,
      "reward_std": 0.7502522468566895,
      "rewards/format_reward": 0.6041666716337204,
      "rewards/segmentation_reward": 0.43639934808015823,
      "step": 4
    },
    {
      "completion_length": 143.71875762939453,
      "epoch": 0.03944773175542406,
      "grad_norm": 4.0498151779174805,
      "kl": 0.00193023681640625,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "reward": 1.1338110864162445,
      "reward_std": 0.6782345250248909,
      "rewards/format_reward": 0.708333358168602,
      "rewards/segmentation_reward": 0.4254777356982231,
      "step": 5
    },
    {
      "completion_length": 124.4062557220459,
      "epoch": 0.047337278106508875,
      "grad_norm": 20.274686813354492,
      "kl": 0.0040283203125,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "reward": 1.3649465143680573,
      "reward_std": 0.4273431599140167,
      "rewards/format_reward": 0.7916666865348816,
      "rewards/segmentation_reward": 0.5732797980308533,
      "step": 6
    },
    {
      "completion_length": 131.6666717529297,
      "epoch": 0.055226824457593686,
      "grad_norm": 13.247719764709473,
      "kl": 0.00662994384765625,
      "learning_rate": 1e-06,
      "loss": 0.0003,
      "reward": 1.2582235038280487,
      "reward_std": 0.5580896884202957,
      "rewards/format_reward": 0.7500000149011612,
      "rewards/segmentation_reward": 0.5082234740257263,
      "step": 7
    },
    {
      "completion_length": 131.52083778381348,
      "epoch": 0.0631163708086785,
      "grad_norm": 11.869627952575684,
      "kl": 0.0071258544921875,
      "learning_rate": 1e-06,
      "loss": 0.0003,
      "reward": 1.4533603489398956,
      "reward_std": 0.5124156475067139,
      "rewards/format_reward": 0.8541666716337204,
      "rewards/segmentation_reward": 0.5991936326026917,
      "step": 8
    },
    {
      "completion_length": 132.90625381469727,
      "epoch": 0.07100591715976332,
      "grad_norm": 3.0018997192382812,
      "kl": 0.00543212890625,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "reward": 1.2137048244476318,
      "reward_std": 0.5586739107966423,
      "rewards/format_reward": 0.708333358168602,
      "rewards/segmentation_reward": 0.5053714141249657,
      "step": 9
    },
    {
      "completion_length": 129.46875381469727,
      "epoch": 0.07889546351084813,
      "grad_norm": 15.782857894897461,
      "kl": 0.0078582763671875,
      "learning_rate": 1e-06,
      "loss": 0.0003,
      "reward": 1.3436061143875122,
      "reward_std": 0.46356433629989624,
      "rewards/format_reward": 0.8229166865348816,
      "rewards/segmentation_reward": 0.5206894353032112,
      "step": 10
    },
    {
      "completion_length": 125.33333587646484,
      "epoch": 0.08678500986193294,
      "grad_norm": 5.636582851409912,
      "kl": 0.0071868896484375,
      "learning_rate": 1e-06,
      "loss": 0.0003,
      "reward": 1.5922125577926636,
      "reward_std": 0.32092406041920185,
      "rewards/format_reward": 0.9166666716337204,
      "rewards/segmentation_reward": 0.6755458265542984,
      "step": 11
    },
    {
      "completion_length": 120.16666793823242,
      "epoch": 0.09467455621301775,
      "grad_norm": 17.472490310668945,
      "kl": 0.00865936279296875,
      "learning_rate": 1e-06,
      "loss": 0.0003,
      "reward": 1.4907157123088837,
      "reward_std": 0.3035526014864445,
      "rewards/format_reward": 0.9062500149011612,
      "rewards/segmentation_reward": 0.5844657346606255,
      "step": 12
    },
    {
      "completion_length": 125.35416984558105,
      "epoch": 0.10256410256410256,
      "grad_norm": 3.8183038234710693,
      "kl": 0.0095672607421875,
      "learning_rate": 1e-06,
      "loss": 0.0004,
      "reward": 1.4928309321403503,
      "reward_std": 0.46883198618888855,
      "rewards/format_reward": 0.8750000149011612,
      "rewards/segmentation_reward": 0.6178309172391891,
      "step": 13
    },
    {
      "completion_length": 117.92708587646484,
      "epoch": 0.11045364891518737,
      "grad_norm": 6.6655073165893555,
      "kl": 0.0110626220703125,
      "learning_rate": 1e-06,
      "loss": 0.0004,
      "reward": 1.5008599758148193,
      "reward_std": 0.32115011289715767,
      "rewards/format_reward": 0.9479167014360428,
      "rewards/segmentation_reward": 0.5529432520270348,
      "step": 14
    },
    {
      "completion_length": 123.93750190734863,
      "epoch": 0.11834319526627218,
      "grad_norm": 5.591380596160889,
      "kl": 0.012359619140625,
      "learning_rate": 1e-06,
      "loss": 0.0005,
      "reward": 1.5433619320392609,
      "reward_std": 0.24045727029442787,
      "rewards/format_reward": 0.9479167014360428,
      "rewards/segmentation_reward": 0.5954452455043793,
      "step": 15
    },
    {
      "completion_length": 119.31250190734863,
      "epoch": 0.126232741617357,
      "grad_norm": 4.104308605194092,
      "kl": 0.0131378173828125,
      "learning_rate": 1e-06,
      "loss": 0.0005,
      "reward": 1.5899344384670258,
      "reward_std": 0.2260684370994568,
      "rewards/format_reward": 0.9479167014360428,
      "rewards/segmentation_reward": 0.6420177668333054,
      "step": 16
    },
    {
      "completion_length": 121.08333778381348,
      "epoch": 0.1341222879684418,
      "grad_norm": 5.432396411895752,
      "kl": 0.0205535888671875,
      "learning_rate": 1e-06,
      "loss": 0.0008,
      "reward": 1.5686425566673279,
      "reward_std": 0.3436120003461838,
      "rewards/format_reward": 0.9062500149011612,
      "rewards/segmentation_reward": 0.6623925417661667,
      "step": 17
    },
    {
      "completion_length": 113.20833587646484,
      "epoch": 0.14201183431952663,
      "grad_norm": 5.052847862243652,
      "kl": 0.0161285400390625,
      "learning_rate": 1e-06,
      "loss": 0.0006,
      "reward": 1.618556797504425,
      "reward_std": 0.22547486051917076,
      "rewards/format_reward": 0.9583333432674408,
      "rewards/segmentation_reward": 0.6602234989404678,
      "step": 18
    },
    {
      "completion_length": 119.66666793823242,
      "epoch": 0.14990138067061143,
      "grad_norm": 4.373110771179199,
      "kl": 0.0154266357421875,
      "learning_rate": 1e-06,
      "loss": 0.0006,
      "reward": 1.5658972263336182,
      "reward_std": 0.2384246401488781,
      "rewards/format_reward": 0.9479166865348816,
      "rewards/segmentation_reward": 0.617980569601059,
      "step": 19
    },
    {
      "completion_length": 114.64583778381348,
      "epoch": 0.15779092702169625,
      "grad_norm": 5.60256290435791,
      "kl": 0.022735595703125,
      "learning_rate": 1e-06,
      "loss": 0.0009,
      "reward": 1.6209804117679596,
      "reward_std": 0.16078345850110054,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.641813725233078,
      "step": 20
    },
    {
      "completion_length": 112.28125190734863,
      "epoch": 0.16568047337278108,
      "grad_norm": 7.399819850921631,
      "kl": 0.0194091796875,
      "learning_rate": 1e-06,
      "loss": 0.0008,
      "reward": 1.6231780648231506,
      "reward_std": 0.16180265322327614,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.644011452794075,
      "step": 21
    },
    {
      "completion_length": 118.37500190734863,
      "epoch": 0.17357001972386588,
      "grad_norm": 5.146966934204102,
      "kl": 0.020172119140625,
      "learning_rate": 1e-06,
      "loss": 0.0008,
      "reward": 1.5712138712406158,
      "reward_std": 0.18034345097839832,
      "rewards/format_reward": 0.9687500298023224,
      "rewards/segmentation_reward": 0.602463886141777,
      "step": 22
    },
    {
      "completion_length": 109.30208587646484,
      "epoch": 0.1814595660749507,
      "grad_norm": 10.886876106262207,
      "kl": 0.0229644775390625,
      "learning_rate": 1e-06,
      "loss": 0.0009,
      "reward": 1.6121678054332733,
      "reward_std": 0.24894177541136742,
      "rewards/format_reward": 0.927083358168602,
      "rewards/segmentation_reward": 0.6850844025611877,
      "step": 23
    },
    {
      "completion_length": 115.40625190734863,
      "epoch": 0.1893491124260355,
      "grad_norm": 7.059934139251709,
      "kl": 0.0250244140625,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.603627324104309,
      "reward_std": 0.2590697519481182,
      "rewards/format_reward": 0.9479166865348816,
      "rewards/segmentation_reward": 0.6557105779647827,
      "step": 24
    },
    {
      "completion_length": 105.81250190734863,
      "epoch": 0.19723865877712032,
      "grad_norm": 6.3141889572143555,
      "kl": 0.02276611328125,
      "learning_rate": 1e-06,
      "loss": 0.0009,
      "reward": 1.609460562467575,
      "reward_std": 0.21640464290976524,
      "rewards/format_reward": 0.958333358168602,
      "rewards/segmentation_reward": 0.651127278804779,
      "step": 25
    },
    {
      "completion_length": 106.47916793823242,
      "epoch": 0.20512820512820512,
      "grad_norm": 5.775827884674072,
      "kl": 0.0296630859375,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.5251965522766113,
      "reward_std": 0.21586111560463905,
      "rewards/format_reward": 0.958333358168602,
      "rewards/segmentation_reward": 0.566863164305687,
      "step": 26
    },
    {
      "completion_length": 107.38541984558105,
      "epoch": 0.21301775147928995,
      "grad_norm": 6.3134379386901855,
      "kl": 0.0269775390625,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.7112098634243011,
      "reward_std": 0.08758194698020816,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.711209774017334,
      "step": 27
    },
    {
      "completion_length": 106.05208778381348,
      "epoch": 0.22090729783037474,
      "grad_norm": 8.249260902404785,
      "kl": 0.02569580078125,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.737591177225113,
      "reward_std": 0.03949332144111395,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7375911623239517,
      "step": 28
    },
    {
      "completion_length": 107.68750190734863,
      "epoch": 0.22879684418145957,
      "grad_norm": 21.619091033935547,
      "kl": 0.025360107421875,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.7143003344535828,
      "reward_std": 0.14945492893457413,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7143002450466156,
      "step": 29
    },
    {
      "completion_length": 113.18750381469727,
      "epoch": 0.23668639053254437,
      "grad_norm": 11.503852844238281,
      "kl": 0.027099609375,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.6730247139930725,
      "reward_std": 0.17038158606737852,
      "rewards/format_reward": 0.9687500298023224,
      "rewards/segmentation_reward": 0.7042748034000397,
      "step": 30
    },
    {
      "completion_length": 107.75000190734863,
      "epoch": 0.2445759368836292,
      "grad_norm": 6.8054704666137695,
      "kl": 0.0262451171875,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.7297659516334534,
      "reward_std": 0.0851635541766882,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7401826083660126,
      "step": 31
    },
    {
      "completion_length": 109.43750190734863,
      "epoch": 0.252465483234714,
      "grad_norm": 7.1466593742370605,
      "kl": 0.03009033203125,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.6955690681934357,
      "reward_std": 0.13199757551774383,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7164023518562317,
      "step": 32
    },
    {
      "completion_length": 112.22916984558105,
      "epoch": 0.2603550295857988,
      "grad_norm": 4.616764068603516,
      "kl": 0.031646728515625,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.6858111023902893,
      "reward_std": 0.15736100357025862,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7066444158554077,
      "step": 33
    },
    {
      "completion_length": 105.96875190734863,
      "epoch": 0.2682445759368836,
      "grad_norm": 4.323781490325928,
      "kl": 0.0306396484375,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.692588210105896,
      "reward_std": 0.18604120332747698,
      "rewards/format_reward": 0.9479166865348816,
      "rewards/segmentation_reward": 0.7446715235710144,
      "step": 34
    },
    {
      "completion_length": 107.27083587646484,
      "epoch": 0.27613412228796846,
      "grad_norm": 6.824800968170166,
      "kl": 0.038543701171875,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.6549388766288757,
      "reward_std": 0.17549242451786995,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.686188817024231,
      "step": 35
    },
    {
      "completion_length": 109.16666793823242,
      "epoch": 0.28402366863905326,
      "grad_norm": 7.346378803253174,
      "kl": 0.025604248046875,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.6843693554401398,
      "reward_std": 0.13274440541863441,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.6947860270738602,
      "step": 36
    },
    {
      "completion_length": 113.21875190734863,
      "epoch": 0.29191321499013806,
      "grad_norm": 10.401052474975586,
      "kl": 0.03076171875,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.6937299370765686,
      "reward_std": 0.08907313086092472,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7041465640068054,
      "step": 37
    },
    {
      "completion_length": 115.80208587646484,
      "epoch": 0.29980276134122286,
      "grad_norm": 8.589919090270996,
      "kl": 0.0274658203125,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.7341200113296509,
      "reward_std": 0.12931476812809706,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7549533247947693,
      "step": 38
    },
    {
      "completion_length": 118.71875190734863,
      "epoch": 0.3076923076923077,
      "grad_norm": 12.668815612792969,
      "kl": 0.02984619140625,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.6736123263835907,
      "reward_std": 0.09443093463778496,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.6840289980173111,
      "step": 39
    },
    {
      "completion_length": 121.02083587646484,
      "epoch": 0.3155818540433925,
      "grad_norm": 6.178909778594971,
      "kl": 0.022613525390625,
      "learning_rate": 1e-06,
      "loss": 0.0009,
      "reward": 1.7086512446403503,
      "reward_std": 0.13165022525936365,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7190678864717484,
      "step": 40
    },
    {
      "completion_length": 121.04167175292969,
      "epoch": 0.3234714003944773,
      "grad_norm": 5.155839443206787,
      "kl": 0.022979736328125,
      "learning_rate": 1e-06,
      "loss": 0.0009,
      "reward": 1.6464684903621674,
      "reward_std": 0.16541135124862194,
      "rewards/format_reward": 0.958333358168602,
      "rewards/segmentation_reward": 0.6881351172924042,
      "step": 41
    },
    {
      "completion_length": 123.53125381469727,
      "epoch": 0.33136094674556216,
      "grad_norm": 6.814601898193359,
      "kl": 0.01806640625,
      "learning_rate": 1e-06,
      "loss": 0.0007,
      "reward": 1.7974173426628113,
      "reward_std": 0.06069173011928797,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7974172979593277,
      "step": 42
    },
    {
      "completion_length": 114.95833587646484,
      "epoch": 0.33925049309664695,
      "grad_norm": 5.867588520050049,
      "kl": 0.030914306640625,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.6036024987697601,
      "reward_std": 0.14355931896716356,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.6244357973337173,
      "step": 43
    },
    {
      "completion_length": 119.41666793823242,
      "epoch": 0.34714003944773175,
      "grad_norm": 23.10361099243164,
      "kl": 0.02508544921875,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.7223087847232819,
      "reward_std": 0.1400171834975481,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.7431419938802719,
      "step": 44
    },
    {
      "completion_length": 120.78125190734863,
      "epoch": 0.35502958579881655,
      "grad_norm": 8.673208236694336,
      "kl": 0.030670166015625,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.5970956683158875,
      "reward_std": 0.20673380978405476,
      "rewards/format_reward": 0.958333358168602,
      "rewards/segmentation_reward": 0.6387622952461243,
      "step": 45
    },
    {
      "completion_length": 122.50000190734863,
      "epoch": 0.3629191321499014,
      "grad_norm": 7.40756893157959,
      "kl": 0.032196044921875,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.6095560789108276,
      "reward_std": 0.24808678403496742,
      "rewards/format_reward": 0.9479167014360428,
      "rewards/segmentation_reward": 0.6616393625736237,
      "step": 46
    },
    {
      "completion_length": 119.16666984558105,
      "epoch": 0.3708086785009862,
      "grad_norm": 7.48468017578125,
      "kl": 0.02667236328125,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.6872661709785461,
      "reward_std": 0.1560370922088623,
      "rewards/format_reward": 0.958333358168602,
      "rewards/segmentation_reward": 0.7289328575134277,
      "step": 47
    },
    {
      "completion_length": 121.70833587646484,
      "epoch": 0.378698224852071,
      "grad_norm": 6.0176005363464355,
      "kl": 0.02667236328125,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.705275982618332,
      "reward_std": 0.12693702802062035,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7156926393508911,
      "step": 48
    },
    {
      "completion_length": 119.76041793823242,
      "epoch": 0.3865877712031558,
      "grad_norm": 19.220964431762695,
      "kl": 0.03289794921875,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.67743381857872,
      "reward_std": 0.11258519534021616,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.6982671469449997,
      "step": 49
    },
    {
      "completion_length": 124.2812557220459,
      "epoch": 0.39447731755424065,
      "grad_norm": 5.315231800079346,
      "kl": 0.033966064453125,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.6300671994686127,
      "reward_std": 0.2049922477453947,
      "rewards/format_reward": 0.9583333432674408,
      "rewards/segmentation_reward": 0.6717338263988495,
      "step": 50
    },
    {
      "completion_length": 120.82291984558105,
      "epoch": 0.40236686390532544,
      "grad_norm": 6.098005771636963,
      "kl": 0.028411865234375,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.7469660639762878,
      "reward_std": 0.137971930205822,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7573827505111694,
      "step": 51
    },
    {
      "completion_length": 125.18750381469727,
      "epoch": 0.41025641025641024,
      "grad_norm": 8.745720863342285,
      "kl": 0.02703857421875,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.6644043028354645,
      "reward_std": 0.13505547679960728,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.6852375864982605,
      "step": 52
    },
    {
      "completion_length": 119.29166984558105,
      "epoch": 0.4181459566074951,
      "grad_norm": 7.582885265350342,
      "kl": 0.02655029296875,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.7572149634361267,
      "reward_std": 0.05950829852372408,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7572149932384491,
      "step": 53
    },
    {
      "completion_length": 122.39583587646484,
      "epoch": 0.4260355029585799,
      "grad_norm": 5.690119743347168,
      "kl": 0.02685546875,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.7116829752922058,
      "reward_std": 0.1341047789901495,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7325162440538406,
      "step": 54
    },
    {
      "completion_length": 123.38541793823242,
      "epoch": 0.4339250493096647,
      "grad_norm": 4.931894779205322,
      "kl": 0.026763916015625,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.7832928001880646,
      "reward_std": 0.14402389503084123,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.8041260987520218,
      "step": 55
    },
    {
      "completion_length": 123.33333587646484,
      "epoch": 0.4418145956607495,
      "grad_norm": 6.575313091278076,
      "kl": 0.026123046875,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.7617658376693726,
      "reward_std": 0.08360383100807667,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7617658227682114,
      "step": 56
    },
    {
      "completion_length": 117.22916793823242,
      "epoch": 0.44970414201183434,
      "grad_norm": 5.90615701675415,
      "kl": 0.02032470703125,
      "learning_rate": 1e-06,
      "loss": 0.0008,
      "reward": 1.733703076839447,
      "reward_std": 0.08622908592224121,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7441196739673615,
      "step": 57
    },
    {
      "completion_length": 120.58333587646484,
      "epoch": 0.45759368836291914,
      "grad_norm": 5.734189510345459,
      "kl": 0.0242919921875,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.765779048204422,
      "reward_std": 0.05810722103342414,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7657789587974548,
      "step": 58
    },
    {
      "completion_length": 122.66666984558105,
      "epoch": 0.46548323471400394,
      "grad_norm": 4.0810956954956055,
      "kl": 0.02264404296875,
      "learning_rate": 1e-06,
      "loss": 0.0009,
      "reward": 1.7729882895946503,
      "reward_std": 0.06873794272542,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7729882448911667,
      "step": 59
    },
    {
      "completion_length": 122.95833587646484,
      "epoch": 0.47337278106508873,
      "grad_norm": 8.775988578796387,
      "kl": 0.02838134765625,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.6857317984104156,
      "reward_std": 0.20080409944057465,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.7169817835092545,
      "step": 60
    },
    {
      "completion_length": 122.13541984558105,
      "epoch": 0.4812623274161736,
      "grad_norm": 12.763365745544434,
      "kl": 0.024200439453125,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.7577170729637146,
      "reward_std": 0.1079479455947876,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7681336998939514,
      "step": 61
    },
    {
      "completion_length": 119.69791793823242,
      "epoch": 0.4891518737672584,
      "grad_norm": 9.827467918395996,
      "kl": 0.026092529296875,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.7310417592525482,
      "reward_std": 0.16975797526538372,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.7518750727176666,
      "step": 62
    },
    {
      "completion_length": 123.47916984558105,
      "epoch": 0.4970414201183432,
      "grad_norm": 11.630967140197754,
      "kl": 0.03076171875,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.6405887305736542,
      "reward_std": 0.20058110356330872,
      "rewards/format_reward": 0.9583333432674408,
      "rewards/segmentation_reward": 0.6822553277015686,
      "step": 63
    },
    {
      "completion_length": 116.79166793823242,
      "epoch": 0.504930966469428,
      "grad_norm": 4.396234035491943,
      "kl": 0.02423095703125,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.7577708065509796,
      "reward_std": 0.056545618921518326,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.757770836353302,
      "step": 64
    },
    {
      "completion_length": 116.90625381469727,
      "epoch": 0.5128205128205128,
      "grad_norm": 8.891778945922852,
      "kl": 0.024566650390625,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.7180190980434418,
      "reward_std": 0.1297496184706688,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.738852322101593,
      "step": 65
    },
    {
      "completion_length": 115.18750190734863,
      "epoch": 0.5207100591715976,
      "grad_norm": 5.857807159423828,
      "kl": 0.026397705078125,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.6948060393333435,
      "reward_std": 0.09921691659837961,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7052226364612579,
      "step": 66
    },
    {
      "completion_length": 121.09375381469727,
      "epoch": 0.5285996055226825,
      "grad_norm": 6.213242053985596,
      "kl": 0.022308349609375,
      "learning_rate": 1e-06,
      "loss": 0.0009,
      "reward": 1.7551944553852081,
      "reward_std": 0.15169572597369552,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7760278284549713,
      "step": 67
    },
    {
      "completion_length": 115.51041793823242,
      "epoch": 0.5364891518737672,
      "grad_norm": 8.733428001403809,
      "kl": 0.027435302734375,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.6302316784858704,
      "reward_std": 0.229142090305686,
      "rewards/format_reward": 0.9375,
      "rewards/segmentation_reward": 0.6927317082881927,
      "step": 68
    },
    {
      "completion_length": 127.31250381469727,
      "epoch": 0.5443786982248521,
      "grad_norm": 8.592183113098145,
      "kl": 0.026947021484375,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.5972277224063873,
      "reward_std": 0.1787685491144657,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.6284777224063873,
      "step": 69
    },
    {
      "completion_length": 121.78125190734863,
      "epoch": 0.5522682445759369,
      "grad_norm": 4.904063701629639,
      "kl": 0.032379150390625,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.6003088057041168,
      "reward_std": 0.14930523186922073,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.6315587908029556,
      "step": 70
    },
    {
      "completion_length": 116.23958587646484,
      "epoch": 0.5601577909270217,
      "grad_norm": 7.675012588500977,
      "kl": 0.048187255859375,
      "learning_rate": 1e-06,
      "loss": 0.0019,
      "reward": 1.6342654526233673,
      "reward_std": 0.24084181897342205,
      "rewards/format_reward": 0.9479166865348816,
      "rewards/segmentation_reward": 0.6863488405942917,
      "step": 71
    },
    {
      "completion_length": 110.20833587646484,
      "epoch": 0.5680473372781065,
      "grad_norm": 6.364027500152588,
      "kl": 0.0234375,
      "learning_rate": 1e-06,
      "loss": 0.0009,
      "reward": 1.6845062971115112,
      "reward_std": 0.07788908947259188,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.705339640378952,
      "step": 72
    },
    {
      "completion_length": 114.71875190734863,
      "epoch": 0.5759368836291914,
      "grad_norm": 12.653519630432129,
      "kl": 0.03668212890625,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.655396193265915,
      "reward_std": 0.16850140318274498,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.6866461485624313,
      "step": 73
    },
    {
      "completion_length": 114.77083587646484,
      "epoch": 0.5838264299802761,
      "grad_norm": 7.866998195648193,
      "kl": 0.025146484375,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.6978517174720764,
      "reward_std": 0.14674336509779096,
      "rewards/format_reward": 0.96875,
      "rewards/segmentation_reward": 0.7291017025709152,
      "step": 74
    },
    {
      "completion_length": 113.13541984558105,
      "epoch": 0.591715976331361,
      "grad_norm": 9.050918579101562,
      "kl": 0.0279541015625,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.6342374980449677,
      "reward_std": 0.09696032106876373,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.6342374533414841,
      "step": 75
    },
    {
      "completion_length": 110.10417175292969,
      "epoch": 0.5996055226824457,
      "grad_norm": 11.760872840881348,
      "kl": 0.02703857421875,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.7879691123962402,
      "reward_std": 0.08564800582826138,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7983856648206711,
      "step": 76
    },
    {
      "completion_length": 116.18750381469727,
      "epoch": 0.6074950690335306,
      "grad_norm": 8.666062355041504,
      "kl": 0.0250244140625,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.6772170960903168,
      "reward_std": 0.1316035082563758,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.6980504393577576,
      "step": 77
    },
    {
      "completion_length": 113.64583396911621,
      "epoch": 0.6153846153846154,
      "grad_norm": 10.047959327697754,
      "kl": 0.0338134765625,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.7389695942401886,
      "reward_std": 0.06903288420289755,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7389696538448334,
      "step": 78
    },
    {
      "completion_length": 115.52083587646484,
      "epoch": 0.6232741617357002,
      "grad_norm": 10.192525863647461,
      "kl": 0.0306396484375,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.7115220725536346,
      "reward_std": 0.13409498427063227,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7323553562164307,
      "step": 79
    },
    {
      "completion_length": 110.38541984558105,
      "epoch": 0.631163708086785,
      "grad_norm": 6.273956298828125,
      "kl": 0.0413818359375,
      "learning_rate": 1e-06,
      "loss": 0.0017,
      "reward": 1.5747724771499634,
      "reward_std": 0.12117741256952286,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.5747724398970604,
      "step": 80
    },
    {
      "completion_length": 109.43750381469727,
      "epoch": 0.6390532544378699,
      "grad_norm": 6.993807315826416,
      "kl": 0.030975341796875,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.7259157598018646,
      "reward_std": 0.09234390081837773,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7363324463367462,
      "step": 81
    },
    {
      "completion_length": 110.15625190734863,
      "epoch": 0.6469428007889546,
      "grad_norm": 5.614496231079102,
      "kl": 0.030731201171875,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.7176712453365326,
      "reward_std": 0.15571256913244724,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7385045886039734,
      "step": 82
    },
    {
      "completion_length": 109.75000190734863,
      "epoch": 0.6548323471400395,
      "grad_norm": 25.766254425048828,
      "kl": 0.0313720703125,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.641405701637268,
      "reward_std": 0.14323315117508173,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.6726557016372681,
      "step": 83
    },
    {
      "completion_length": 115.44791793823242,
      "epoch": 0.6627218934911243,
      "grad_norm": 5.393517971038818,
      "kl": 0.028839111328125,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.7041931748390198,
      "reward_std": 0.07750691333785653,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.714609831571579,
      "step": 84
    },
    {
      "completion_length": 111.65625190734863,
      "epoch": 0.6706114398422091,
      "grad_norm": 4.121002674102783,
      "kl": 0.02972412109375,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.7423414289951324,
      "reward_std": 0.10370620153844357,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7527580559253693,
      "step": 85
    },
    {
      "completion_length": 116.79166984558105,
      "epoch": 0.6785009861932939,
      "grad_norm": 18.222322463989258,
      "kl": 0.02703857421875,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.7806086838245392,
      "reward_std": 0.07675650157034397,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.791025385260582,
      "step": 86
    },
    {
      "completion_length": 115.37500381469727,
      "epoch": 0.6863905325443787,
      "grad_norm": 7.765015125274658,
      "kl": 0.02838134765625,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.693938434123993,
      "reward_std": 0.15568579966202378,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.7251883894205093,
      "step": 87
    },
    {
      "completion_length": 115.29166984558105,
      "epoch": 0.6942800788954635,
      "grad_norm": 18.34151840209961,
      "kl": 0.026214599609375,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "reward": 1.691928893327713,
      "reward_std": 0.140016196295619,
      "rewards/format_reward": 0.9687500298023224,
      "rewards/segmentation_reward": 0.7231789231300354,
      "step": 88
    },
    {
      "completion_length": 121.05208396911621,
      "epoch": 0.7021696252465484,
      "grad_norm": 10.368853569030762,
      "kl": 0.021881103515625,
      "learning_rate": 1e-06,
      "loss": 0.0009,
      "reward": 1.8064870834350586,
      "reward_std": 0.09992434550076723,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.816903680562973,
      "step": 89
    },
    {
      "completion_length": 114.10416984558105,
      "epoch": 0.7100591715976331,
      "grad_norm": 4.952389240264893,
      "kl": 0.03314208984375,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.7042470872402191,
      "reward_std": 0.06299979891628027,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7042470574378967,
      "step": 90
    },
    {
      "completion_length": 117.68750190734863,
      "epoch": 0.717948717948718,
      "grad_norm": 9.378170013427734,
      "kl": 0.03192138671875,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.7310494482517242,
      "reward_std": 0.09936824138276279,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7414660751819611,
      "step": 91
    },
    {
      "completion_length": 119.51041793823242,
      "epoch": 0.7258382642998028,
      "grad_norm": 14.939569473266602,
      "kl": 0.029388427734375,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.8184643685817719,
      "reward_std": 0.07522716512903571,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.818464383482933,
      "step": 92
    },
    {
      "completion_length": 128.58333778381348,
      "epoch": 0.7337278106508875,
      "grad_norm": 8.10946273803711,
      "kl": 0.032073974609375,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.6870926320552826,
      "reward_std": 0.14518141373991966,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.707925945520401,
      "step": 93
    },
    {
      "completion_length": 116.75000190734863,
      "epoch": 0.7416173570019724,
      "grad_norm": 4.120285511016846,
      "kl": 0.02752685546875,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.6995415091514587,
      "reward_std": 0.1466266419738531,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.720374807715416,
      "step": 94
    },
    {
      "completion_length": 123.09375190734863,
      "epoch": 0.7495069033530573,
      "grad_norm": 16.04999351501465,
      "kl": 0.032073974609375,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.7072420418262482,
      "reward_std": 0.12071913667023182,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7280752956867218,
      "step": 95
    },
    {
      "completion_length": 125.27083778381348,
      "epoch": 0.757396449704142,
      "grad_norm": 4.702588081359863,
      "kl": 0.029693603515625,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.629258632659912,
      "reward_std": 0.13004819490015507,
      "rewards/format_reward": 0.9687500298023224,
      "rewards/segmentation_reward": 0.6605086177587509,
      "step": 96
    },
    {
      "completion_length": 120.84375190734863,
      "epoch": 0.7652859960552268,
      "grad_norm": 4.365908622741699,
      "kl": 0.03076171875,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.6611265540122986,
      "reward_std": 0.12717730086296797,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.6819598078727722,
      "step": 97
    },
    {
      "completion_length": 121.59375381469727,
      "epoch": 0.7731755424063116,
      "grad_norm": 9.223386764526367,
      "kl": 0.02752685546875,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "reward": 1.7275770008563995,
      "reward_std": 0.08318935800343752,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7275769412517548,
      "step": 98
    },
    {
      "completion_length": 126.01041793823242,
      "epoch": 0.7810650887573964,
      "grad_norm": 12.06326961517334,
      "kl": 0.032257080078125,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.7239615619182587,
      "reward_std": 0.11263703089207411,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7343782037496567,
      "step": 99
    },
    {
      "completion_length": 128.11458587646484,
      "epoch": 0.7889546351084813,
      "grad_norm": 16.42841148376465,
      "kl": 0.028778076171875,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.8079151809215546,
      "reward_std": 0.06341414572671056,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.818331703543663,
      "step": 100
    },
    {
      "completion_length": 118.48958587646484,
      "epoch": 0.796844181459566,
      "grad_norm": 6.883551597595215,
      "kl": 0.0294189453125,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.7349311411380768,
      "reward_std": 0.12644702824763954,
      "rewards/format_reward": 0.9583333432674408,
      "rewards/segmentation_reward": 0.7765978276729584,
      "step": 101
    },
    {
      "completion_length": 127.86458778381348,
      "epoch": 0.8047337278106509,
      "grad_norm": 9.619441032409668,
      "kl": 0.03240966796875,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.7535338401794434,
      "reward_std": 0.1224900484085083,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7639505118131638,
      "step": 102
    },
    {
      "completion_length": 124.44792175292969,
      "epoch": 0.8126232741617357,
      "grad_norm": 4.861848831176758,
      "kl": 0.03369140625,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.7802566289901733,
      "reward_std": 0.06228824611753225,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7906733006238937,
      "step": 103
    },
    {
      "completion_length": 118.57291793823242,
      "epoch": 0.8205128205128205,
      "grad_norm": 12.132479667663574,
      "kl": 0.03399658203125,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.6861309111118317,
      "reward_std": 0.06774610374122858,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.6861308366060257,
      "step": 104
    },
    {
      "completion_length": 122.41666793823242,
      "epoch": 0.8284023668639053,
      "grad_norm": 4.027884483337402,
      "kl": 0.03985595703125,
      "learning_rate": 1e-06,
      "loss": 0.0016,
      "reward": 1.7455581724643707,
      "reward_std": 0.11986150033771992,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7663915157318115,
      "step": 105
    },
    {
      "completion_length": 121.88542175292969,
      "epoch": 0.8362919132149902,
      "grad_norm": 8.1727933883667,
      "kl": 0.03948974609375,
      "learning_rate": 1e-06,
      "loss": 0.0016,
      "reward": 1.7751370966434479,
      "reward_std": 0.05637869983911514,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7751370966434479,
      "step": 106
    },
    {
      "completion_length": 120.05208587646484,
      "epoch": 0.8441814595660749,
      "grad_norm": 30.948610305786133,
      "kl": 0.0394287109375,
      "learning_rate": 1e-06,
      "loss": 0.0016,
      "reward": 1.6998119950294495,
      "reward_std": 0.13076962530612946,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7206452935934067,
      "step": 107
    },
    {
      "completion_length": 125.39583396911621,
      "epoch": 0.8520710059171598,
      "grad_norm": 10.170028686523438,
      "kl": 0.03857421875,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.7232375144958496,
      "reward_std": 0.051960008684545755,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7336541265249252,
      "step": 108
    },
    {
      "completion_length": 120.46875190734863,
      "epoch": 0.8599605522682445,
      "grad_norm": 35.49694061279297,
      "kl": 0.04034423828125,
      "learning_rate": 1e-06,
      "loss": 0.0016,
      "reward": 1.7316677272319794,
      "reward_std": 0.07910315738990903,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7420843094587326,
      "step": 109
    },
    {
      "completion_length": 120.42708587646484,
      "epoch": 0.8678500986193294,
      "grad_norm": 12.206290245056152,
      "kl": 0.036865234375,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.7922857403755188,
      "reward_std": 0.05224635393824428,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7922856956720352,
      "step": 110
    },
    {
      "completion_length": 123.88541793823242,
      "epoch": 0.8757396449704142,
      "grad_norm": 7.881491184234619,
      "kl": 0.038055419921875,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.6793254017829895,
      "reward_std": 0.2003505825996399,
      "rewards/format_reward": 0.958333358168602,
      "rewards/segmentation_reward": 0.7209920138120651,
      "step": 111
    },
    {
      "completion_length": 117.27083778381348,
      "epoch": 0.883629191321499,
      "grad_norm": 7.963736057281494,
      "kl": 0.048980712890625,
      "learning_rate": 1e-06,
      "loss": 0.002,
      "reward": 1.6975261867046356,
      "reward_std": 0.16515915468335152,
      "rewards/format_reward": 0.9687500298023224,
      "rewards/segmentation_reward": 0.728776216506958,
      "step": 112
    },
    {
      "completion_length": 115.54166793823242,
      "epoch": 0.8915187376725838,
      "grad_norm": 4.596273899078369,
      "kl": 0.0467529296875,
      "learning_rate": 1e-06,
      "loss": 0.0019,
      "reward": 1.7152385115623474,
      "reward_std": 0.09552549291402102,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7256551682949066,
      "step": 113
    },
    {
      "completion_length": 120.69791793823242,
      "epoch": 0.8994082840236687,
      "grad_norm": 26.24955177307129,
      "kl": 0.03668212890625,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.673191875219345,
      "reward_std": 0.07764662057161331,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.6731918603181839,
      "step": 114
    },
    {
      "completion_length": 117.08333587646484,
      "epoch": 0.9072978303747534,
      "grad_norm": 4.7493696212768555,
      "kl": 0.03692626953125,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.7461699545383453,
      "reward_std": 0.08548186905682087,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.756586566567421,
      "step": 115
    },
    {
      "completion_length": 115.64583587646484,
      "epoch": 0.9151873767258383,
      "grad_norm": 3.8763372898101807,
      "kl": 0.03564453125,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.7519335448741913,
      "reward_std": 0.07535870576975867,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7623501420021057,
      "step": 116
    },
    {
      "completion_length": 116.66666984558105,
      "epoch": 0.9230769230769231,
      "grad_norm": 7.987335681915283,
      "kl": 0.034820556640625,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.752055674791336,
      "reward_std": 0.04146988596767187,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7520556747913361,
      "step": 117
    },
    {
      "completion_length": 112.97916984558105,
      "epoch": 0.9309664694280079,
      "grad_norm": 7.9439802169799805,
      "kl": 0.033843994140625,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.7821557521820068,
      "reward_std": 0.04594178858678788,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.782155767083168,
      "step": 118
    },
    {
      "completion_length": 112.35416793823242,
      "epoch": 0.9388560157790927,
      "grad_norm": 14.11174488067627,
      "kl": 0.036865234375,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.7454370260238647,
      "reward_std": 0.07132519292645156,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7558536529541016,
      "step": 119
    },
    {
      "completion_length": 116.09375190734863,
      "epoch": 0.9467455621301775,
      "grad_norm": 6.1710309982299805,
      "kl": 0.04046630859375,
      "learning_rate": 1e-06,
      "loss": 0.0016,
      "reward": 1.6930521726608276,
      "reward_std": 0.08492282312363386,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7034688144922256,
      "step": 120
    },
    {
      "completion_length": 110.62500190734863,
      "epoch": 0.9546351084812623,
      "grad_norm": 5.6831254959106445,
      "kl": 0.03900146484375,
      "learning_rate": 1e-06,
      "loss": 0.0016,
      "reward": 1.7358277440071106,
      "reward_std": 0.03763490542769432,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7358278036117554,
      "step": 121
    },
    {
      "completion_length": 110.30208396911621,
      "epoch": 0.9625246548323472,
      "grad_norm": 6.731239318847656,
      "kl": 0.0379638671875,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.7100496292114258,
      "reward_std": 0.11989523191004992,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7308829575777054,
      "step": 122
    },
    {
      "completion_length": 121.90625190734863,
      "epoch": 0.9704142011834319,
      "grad_norm": 6.267615795135498,
      "kl": 0.03106689453125,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.6641573905944824,
      "reward_std": 0.06016020243987441,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.674574002623558,
      "step": 123
    },
    {
      "completion_length": 115.28125190734863,
      "epoch": 0.9783037475345168,
      "grad_norm": 11.331254959106445,
      "kl": 0.03729248046875,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.7366459667682648,
      "reward_std": 0.1369774853810668,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.76789590716362,
      "step": 124
    },
    {
      "completion_length": 115.06250190734863,
      "epoch": 0.9861932938856016,
      "grad_norm": 7.267125606536865,
      "kl": 0.034027099609375,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.765406996011734,
      "reward_std": 0.06693344842642546,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7758236229419708,
      "step": 125
    },
    {
      "completion_length": 120.85416984558105,
      "epoch": 0.9940828402366864,
      "grad_norm": 24.98333168029785,
      "kl": 0.03643798828125,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.6338366866111755,
      "reward_std": 0.11646552104502916,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.654670000076294,
      "step": 126
    },
    {
      "completion_length": 117.5138931274414,
      "epoch": 1.0,
      "grad_norm": 24.98333168029785,
      "kl": 0.038167317708333336,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "reward": 1.8023079633712769,
      "reward_std": 0.0641996127863725,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.8023078640302023,
      "step": 127
    },
    {
      "completion_length": 118.71875190734863,
      "epoch": 1.0078895463510849,
      "grad_norm": 3.1590914726257324,
      "kl": 0.036834716796875,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.74771448969841,
      "reward_std": 0.12587013002485037,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7685477286577225,
      "step": 128
    },
    {
      "completion_length": 117.43750190734863,
      "epoch": 1.0157790927021697,
      "grad_norm": 3.237163543701172,
      "kl": 0.0372314453125,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.7158082127571106,
      "reward_std": 0.08808104507625103,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7262248396873474,
      "step": 129
    },
    {
      "completion_length": 116.84375190734863,
      "epoch": 1.0236686390532543,
      "grad_norm": 2.3869004249572754,
      "kl": 0.045928955078125,
      "learning_rate": 1e-06,
      "loss": 0.0018,
      "reward": 1.732060581445694,
      "reward_std": 0.05624359124340117,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7320606112480164,
      "step": 130
    },
    {
      "completion_length": 115.43750190734863,
      "epoch": 1.0315581854043392,
      "grad_norm": 2.4740047454833984,
      "kl": 0.032684326171875,
      "learning_rate": 1e-06,
      "loss": 0.0013,
      "reward": 1.6838627457618713,
      "reward_std": 0.1397713664919138,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.7151127308607101,
      "step": 131
    },
    {
      "completion_length": 121.77083778381348,
      "epoch": 1.039447731755424,
      "grad_norm": 4.081189155578613,
      "kl": 0.0347900390625,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.6779897511005402,
      "reward_std": 0.12438491266220808,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.6988230645656586,
      "step": 132
    },
    {
      "completion_length": 124.55208778381348,
      "epoch": 1.047337278106509,
      "grad_norm": 2.9231972694396973,
      "kl": 0.036865234375,
      "learning_rate": 1e-06,
      "loss": 0.0015,
      "reward": 1.7227727770805359,
      "reward_std": 0.09910841938108206,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7436061203479767,
      "step": 133
    },
    {
      "completion_length": 116.89583778381348,
      "epoch": 1.0552268244575937,
      "grad_norm": 1.8719556331634521,
      "kl": 0.03448486328125,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.7760542333126068,
      "reward_std": 0.05843149218708277,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.776054173707962,
      "step": 134
    },
    {
      "completion_length": 125.28125190734863,
      "epoch": 1.0631163708086786,
      "grad_norm": 4.612652778625488,
      "kl": 0.03936767578125,
      "learning_rate": 1e-06,
      "loss": 0.0016,
      "reward": 1.7310554683208466,
      "reward_std": 0.0841823904775083,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.741472139954567,
      "step": 135
    },
    {
      "completion_length": 125.41667175292969,
      "epoch": 1.0710059171597632,
      "grad_norm": 2.9326233863830566,
      "kl": 0.03546142578125,
      "learning_rate": 1e-06,
      "loss": 0.0014,
      "reward": 1.7228147089481354,
      "reward_std": 0.09376791398972273,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7228147238492966,
      "step": 136
    },
    {
      "completion_length": 124.57291793823242,
      "epoch": 1.078895463510848,
      "grad_norm": 2.6695199012756348,
      "kl": 0.04388427734375,
      "learning_rate": 1e-06,
      "loss": 0.0018,
      "reward": 1.7047127783298492,
      "reward_std": 0.08051302284002304,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7151293754577637,
      "step": 137
    },
    {
      "completion_length": 118.53125381469727,
      "epoch": 1.086785009861933,
      "grad_norm": 5.622714996337891,
      "kl": 0.042724609375,
      "learning_rate": 1e-06,
      "loss": 0.0017,
      "reward": 1.682199478149414,
      "reward_std": 0.05199452117085457,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.6821994781494141,
      "step": 138
    },
    {
      "completion_length": 127.53125762939453,
      "epoch": 1.0946745562130178,
      "grad_norm": 2.1606082916259766,
      "kl": 0.0406494140625,
      "learning_rate": 1e-06,
      "loss": 0.0016,
      "reward": 1.7326411008834839,
      "reward_std": 0.18974016141146421,
      "rewards/format_reward": 0.9583333432674408,
      "rewards/segmentation_reward": 0.7743077427148819,
      "step": 139
    },
    {
      "completion_length": 121.75000381469727,
      "epoch": 1.1025641025641026,
      "grad_norm": 2.8073818683624268,
      "kl": 0.041961669921875,
      "learning_rate": 1e-06,
      "loss": 0.0017,
      "reward": 1.7088265120983124,
      "reward_std": 0.0832697688601911,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7192431837320328,
      "step": 140
    },
    {
      "completion_length": 117.55208587646484,
      "epoch": 1.1104536489151873,
      "grad_norm": 1.8191330432891846,
      "kl": 0.041748046875,
      "learning_rate": 1e-06,
      "loss": 0.0017,
      "reward": 1.721179723739624,
      "reward_std": 0.05665353685617447,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7211797088384628,
      "step": 141
    },
    {
      "completion_length": 110.60416984558105,
      "epoch": 1.1183431952662721,
      "grad_norm": 1.681236743927002,
      "kl": 0.04168701171875,
      "learning_rate": 1e-06,
      "loss": 0.0017,
      "reward": 1.8110205233097076,
      "reward_std": 0.0791134424507618,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.8214371502399445,
      "step": 142
    },
    {
      "completion_length": 112.11458587646484,
      "epoch": 1.126232741617357,
      "grad_norm": 1.9118095636367798,
      "kl": 0.0439453125,
      "learning_rate": 1e-06,
      "loss": 0.0018,
      "reward": 1.7839480638504028,
      "reward_std": 0.08729123277589679,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7943646758794785,
      "step": 143
    },
    {
      "completion_length": 114.46875381469727,
      "epoch": 1.1341222879684418,
      "grad_norm": 2.732917308807373,
      "kl": 0.044921875,
      "learning_rate": 1e-06,
      "loss": 0.0018,
      "reward": 1.7890871465206146,
      "reward_std": 0.08393839001655579,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7995037585496902,
      "step": 144
    },
    {
      "completion_length": 113.70833587646484,
      "epoch": 1.1420118343195267,
      "grad_norm": 4.204154014587402,
      "kl": 0.05718994140625,
      "learning_rate": 1e-06,
      "loss": 0.0023,
      "reward": 1.718978464603424,
      "reward_std": 0.11887417268007994,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.7398117780685425,
      "step": 145
    },
    {
      "completion_length": 112.35416984558105,
      "epoch": 1.1499013806706113,
      "grad_norm": 7.688730239868164,
      "kl": 0.048583984375,
      "learning_rate": 1e-06,
      "loss": 0.0019,
      "reward": 1.7795838117599487,
      "reward_std": 0.06728760804980993,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7795837968587875,
      "step": 146
    },
    {
      "completion_length": 116.03125190734863,
      "epoch": 1.1577909270216962,
      "grad_norm": 1.9673783779144287,
      "kl": 0.05224609375,
      "learning_rate": 1e-06,
      "loss": 0.0021,
      "reward": 1.740858942270279,
      "reward_std": 0.08058743178844452,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7408589273691177,
      "step": 147
    },
    {
      "completion_length": 114.85416984558105,
      "epoch": 1.165680473372781,
      "grad_norm": 3.3340156078338623,
      "kl": 0.05194091796875,
      "learning_rate": 1e-06,
      "loss": 0.0021,
      "reward": 1.72911536693573,
      "reward_std": 0.09798233769834042,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7395320534706116,
      "step": 148
    },
    {
      "completion_length": 119.91667175292969,
      "epoch": 1.1735700197238659,
      "grad_norm": 2.7571516036987305,
      "kl": 0.0528564453125,
      "learning_rate": 1e-06,
      "loss": 0.0021,
      "reward": 1.6843095421791077,
      "reward_std": 0.172153084538877,
      "rewards/format_reward": 0.958333358168602,
      "rewards/segmentation_reward": 0.7259761840105057,
      "step": 149
    },
    {
      "completion_length": 111.11458587646484,
      "epoch": 1.1814595660749507,
      "grad_norm": 5.544634819030762,
      "kl": 0.0479736328125,
      "learning_rate": 1e-06,
      "loss": 0.0019,
      "reward": 1.7362055480480194,
      "reward_std": 0.12884071236476302,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7570388317108154,
      "step": 150
    },
    {
      "completion_length": 114.02083587646484,
      "epoch": 1.1893491124260356,
      "grad_norm": 1.6093388795852661,
      "kl": 0.05059814453125,
      "learning_rate": 1e-06,
      "loss": 0.002,
      "reward": 1.7716272175312042,
      "reward_std": 0.09435711428523064,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.782043844461441,
      "step": 151
    },
    {
      "completion_length": 106.68750190734863,
      "epoch": 1.1972386587771204,
      "grad_norm": 2.3275067806243896,
      "kl": 0.06378173828125,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.7265048921108246,
      "reward_std": 0.08734399639070034,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7369215190410614,
      "step": 152
    },
    {
      "completion_length": 114.94791984558105,
      "epoch": 1.205128205128205,
      "grad_norm": 3.726362705230713,
      "kl": 0.06097412109375,
      "learning_rate": 1e-06,
      "loss": 0.0024,
      "reward": 1.7594355046749115,
      "reward_std": 0.12504899874329567,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7802688479423523,
      "step": 153
    },
    {
      "completion_length": 117.42708778381348,
      "epoch": 1.21301775147929,
      "grad_norm": 4.784036636352539,
      "kl": 0.052490234375,
      "learning_rate": 1e-06,
      "loss": 0.0021,
      "reward": 1.795677661895752,
      "reward_std": 0.0523160370066762,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.795677661895752,
      "step": 154
    },
    {
      "completion_length": 116.28125190734863,
      "epoch": 1.2209072978303748,
      "grad_norm": 4.563464641571045,
      "kl": 0.05303955078125,
      "learning_rate": 1e-06,
      "loss": 0.0021,
      "reward": 1.7723517715930939,
      "reward_std": 0.0657765488140285,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7723517417907715,
      "step": 155
    },
    {
      "completion_length": 115.67708587646484,
      "epoch": 1.2287968441814596,
      "grad_norm": 5.471823215484619,
      "kl": 0.0528564453125,
      "learning_rate": 1e-06,
      "loss": 0.0021,
      "reward": 1.7356767654418945,
      "reward_std": 0.12338399747386575,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.7565100789070129,
      "step": 156
    },
    {
      "completion_length": 115.55208778381348,
      "epoch": 1.2366863905325443,
      "grad_norm": 1.571742057800293,
      "kl": 0.0523681640625,
      "learning_rate": 1e-06,
      "loss": 0.0021,
      "reward": 1.7827491760253906,
      "reward_std": 0.06495401542633772,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7931658029556274,
      "step": 157
    },
    {
      "completion_length": 114.43750190734863,
      "epoch": 1.244575936883629,
      "grad_norm": 2.303530216217041,
      "kl": 0.0565185546875,
      "learning_rate": 1e-06,
      "loss": 0.0023,
      "reward": 1.7282366454601288,
      "reward_std": 0.04335784679278731,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7282366305589676,
      "step": 158
    },
    {
      "completion_length": 108.57291793823242,
      "epoch": 1.252465483234714,
      "grad_norm": 2.374271869659424,
      "kl": 0.06158447265625,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "reward": 1.7437051832675934,
      "reward_std": 0.07279739435762167,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.743705153465271,
      "step": 159
    },
    {
      "completion_length": 107.39583396911621,
      "epoch": 1.2603550295857988,
      "grad_norm": 6.32477331161499,
      "kl": 0.0670166015625,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.726712703704834,
      "reward_std": 0.1198442354798317,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.747546061873436,
      "step": 160
    },
    {
      "completion_length": 106.02083587646484,
      "epoch": 1.2682445759368837,
      "grad_norm": 22.797327041625977,
      "kl": 0.0625,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "reward": 1.7934371829032898,
      "reward_std": 0.06572614051401615,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.8038537800312042,
      "step": 161
    },
    {
      "completion_length": 108.35416984558105,
      "epoch": 1.2761341222879685,
      "grad_norm": 2.8181421756744385,
      "kl": 0.1036376953125,
      "learning_rate": 1e-06,
      "loss": 0.0042,
      "reward": 1.793130248785019,
      "reward_std": 0.06474862340837717,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.8035468757152557,
      "step": 162
    },
    {
      "completion_length": 105.84375190734863,
      "epoch": 1.2840236686390534,
      "grad_norm": 2.300621509552002,
      "kl": 0.0782470703125,
      "learning_rate": 1e-06,
      "loss": 0.0031,
      "reward": 1.7725840210914612,
      "reward_std": 0.09849729388952255,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7830007076263428,
      "step": 163
    },
    {
      "completion_length": 103.26041984558105,
      "epoch": 1.291913214990138,
      "grad_norm": 19.949954986572266,
      "kl": 0.072998046875,
      "learning_rate": 1e-06,
      "loss": 0.0029,
      "reward": 1.7890552282333374,
      "reward_std": 0.06625080667436123,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7994718849658966,
      "step": 164
    },
    {
      "completion_length": 103.05208587646484,
      "epoch": 1.2998027613412229,
      "grad_norm": 2.4026191234588623,
      "kl": 0.0733642578125,
      "learning_rate": 1e-06,
      "loss": 0.0029,
      "reward": 1.7272422015666962,
      "reward_std": 0.07955199759453535,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7376587837934494,
      "step": 165
    },
    {
      "completion_length": 96.37500381469727,
      "epoch": 1.3076923076923077,
      "grad_norm": 3.9920060634613037,
      "kl": 0.0792236328125,
      "learning_rate": 1e-06,
      "loss": 0.0032,
      "reward": 1.7014544606208801,
      "reward_std": 0.10399023815989494,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7222877889871597,
      "step": 166
    },
    {
      "completion_length": 96.16666793823242,
      "epoch": 1.3155818540433926,
      "grad_norm": 3.3116116523742676,
      "kl": 0.08984375,
      "learning_rate": 1e-06,
      "loss": 0.0036,
      "reward": 1.795901119709015,
      "reward_std": 0.06413333816453815,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7959010004997253,
      "step": 167
    },
    {
      "completion_length": 95.82291984558105,
      "epoch": 1.3234714003944772,
      "grad_norm": 1.6592262983322144,
      "kl": 0.090576171875,
      "learning_rate": 1e-06,
      "loss": 0.0036,
      "reward": 1.6837320029735565,
      "reward_std": 0.14411024749279022,
      "rewards/format_reward": 0.9687500298023224,
      "rewards/segmentation_reward": 0.7149820476770401,
      "step": 168
    },
    {
      "completion_length": 94.01041984558105,
      "epoch": 1.331360946745562,
      "grad_norm": 2.873749256134033,
      "kl": 0.080078125,
      "learning_rate": 1e-06,
      "loss": 0.0032,
      "reward": 1.7172091603279114,
      "reward_std": 0.10507537145167589,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7276259064674377,
      "step": 169
    },
    {
      "completion_length": 92.86458587646484,
      "epoch": 1.339250493096647,
      "grad_norm": 8.259775161743164,
      "kl": 0.0858154296875,
      "learning_rate": 1e-06,
      "loss": 0.0034,
      "reward": 1.7761318981647491,
      "reward_std": 0.041504258988425136,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.776131808757782,
      "step": 170
    },
    {
      "completion_length": 90.58333587646484,
      "epoch": 1.3471400394477318,
      "grad_norm": 2.346843719482422,
      "kl": 0.119384765625,
      "learning_rate": 1e-06,
      "loss": 0.0048,
      "reward": 1.759489893913269,
      "reward_std": 0.0735306958667934,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7699065655469894,
      "step": 171
    },
    {
      "completion_length": 93.85416984558105,
      "epoch": 1.3550295857988166,
      "grad_norm": 1.9567958116531372,
      "kl": 0.11865234375,
      "learning_rate": 1e-06,
      "loss": 0.0047,
      "reward": 1.7193129658699036,
      "reward_std": 0.07020795345306396,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7193129658699036,
      "step": 172
    },
    {
      "completion_length": 92.85416984558105,
      "epoch": 1.3629191321499015,
      "grad_norm": 2.2362473011016846,
      "kl": 0.1275634765625,
      "learning_rate": 1e-06,
      "loss": 0.0051,
      "reward": 1.7795703411102295,
      "reward_std": 0.07625795714557171,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7899868935346603,
      "step": 173
    },
    {
      "completion_length": 101.35416984558105,
      "epoch": 1.3708086785009863,
      "grad_norm": 3.968324899673462,
      "kl": 0.125732421875,
      "learning_rate": 1e-06,
      "loss": 0.005,
      "reward": 1.7042430341243744,
      "reward_std": 0.10214248253032565,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7250763475894928,
      "step": 174
    },
    {
      "completion_length": 95.37500190734863,
      "epoch": 1.378698224852071,
      "grad_norm": 3.955090045928955,
      "kl": 0.1297607421875,
      "learning_rate": 1e-06,
      "loss": 0.0052,
      "reward": 1.7905175387859344,
      "reward_std": 0.038271178025752306,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7905174940824509,
      "step": 175
    },
    {
      "completion_length": 95.66666984558105,
      "epoch": 1.3865877712031558,
      "grad_norm": 2.1160168647766113,
      "kl": 0.1488037109375,
      "learning_rate": 1e-06,
      "loss": 0.006,
      "reward": 1.7364437878131866,
      "reward_std": 0.06818734481930733,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7364437729120255,
      "step": 176
    },
    {
      "completion_length": 96.19791984558105,
      "epoch": 1.3944773175542406,
      "grad_norm": 2.239210367202759,
      "kl": 0.121826171875,
      "learning_rate": 1e-06,
      "loss": 0.0049,
      "reward": 1.710487574338913,
      "reward_std": 0.12724610278382897,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.7417375445365906,
      "step": 177
    },
    {
      "completion_length": 97.14583587646484,
      "epoch": 1.4023668639053255,
      "grad_norm": 1.8484389781951904,
      "kl": 0.1024169921875,
      "learning_rate": 1e-06,
      "loss": 0.0041,
      "reward": 1.706442952156067,
      "reward_std": 0.13373982720077038,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.7376929074525833,
      "step": 178
    },
    {
      "completion_length": 97.36458587646484,
      "epoch": 1.4102564102564101,
      "grad_norm": 4.035000801086426,
      "kl": 0.0792236328125,
      "learning_rate": 1e-06,
      "loss": 0.0032,
      "reward": 1.675961673259735,
      "reward_std": 0.07547118235379457,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.6863782703876495,
      "step": 179
    },
    {
      "completion_length": 95.48958396911621,
      "epoch": 1.418145956607495,
      "grad_norm": 2.940195322036743,
      "kl": 0.078369140625,
      "learning_rate": 1e-06,
      "loss": 0.0031,
      "reward": 1.6918656527996063,
      "reward_std": 0.15424280427396297,
      "rewards/format_reward": 0.9583333432674408,
      "rewards/segmentation_reward": 0.7335322201251984,
      "step": 180
    },
    {
      "completion_length": 98.84375381469727,
      "epoch": 1.4260355029585798,
      "grad_norm": 2.4502110481262207,
      "kl": 0.06298828125,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "reward": 1.7267792224884033,
      "reward_std": 0.08856830094009638,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7371958494186401,
      "step": 181
    },
    {
      "completion_length": 99.81250190734863,
      "epoch": 1.4339250493096647,
      "grad_norm": 3.8718204498291016,
      "kl": 0.067626953125,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.7854253649711609,
      "reward_std": 0.08994585368782282,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7958420068025589,
      "step": 182
    },
    {
      "completion_length": 97.03125190734863,
      "epoch": 1.4418145956607495,
      "grad_norm": 3.1483190059661865,
      "kl": 0.05712890625,
      "learning_rate": 1e-06,
      "loss": 0.0023,
      "reward": 1.7442137897014618,
      "reward_std": 0.07469187397509813,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.754630520939827,
      "step": 183
    },
    {
      "completion_length": 97.07291984558105,
      "epoch": 1.4497041420118344,
      "grad_norm": 3.180529832839966,
      "kl": 0.0655517578125,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.7026579678058624,
      "reward_std": 0.20012306701391935,
      "rewards/format_reward": 0.9583333432674408,
      "rewards/segmentation_reward": 0.7443246394395828,
      "step": 184
    },
    {
      "completion_length": 100.80208587646484,
      "epoch": 1.4575936883629192,
      "grad_norm": 3.37025785446167,
      "kl": 0.06414794921875,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.7813129425048828,
      "reward_std": 0.06885035336017609,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7917295545339584,
      "step": 185
    },
    {
      "completion_length": 97.64583778381348,
      "epoch": 1.4654832347140039,
      "grad_norm": 3.9030656814575195,
      "kl": 0.06085205078125,
      "learning_rate": 1e-06,
      "loss": 0.0024,
      "reward": 1.715784877538681,
      "reward_std": 0.08244315255433321,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7262014895677567,
      "step": 186
    },
    {
      "completion_length": 101.83333396911621,
      "epoch": 1.4733727810650887,
      "grad_norm": 4.519697666168213,
      "kl": 0.06439208984375,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.6724973022937775,
      "reward_std": 0.14969274401664734,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.6933306306600571,
      "step": 187
    },
    {
      "completion_length": 99.75,
      "epoch": 1.4812623274161736,
      "grad_norm": 2.1684651374816895,
      "kl": 0.0518798828125,
      "learning_rate": 1e-06,
      "loss": 0.0021,
      "reward": 1.7607315480709076,
      "reward_std": 0.028905052226036787,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7607315182685852,
      "step": 188
    },
    {
      "completion_length": 97.66666984558105,
      "epoch": 1.4891518737672584,
      "grad_norm": 2.300941228866577,
      "kl": 0.06951904296875,
      "learning_rate": 1e-06,
      "loss": 0.0028,
      "reward": 1.655644565820694,
      "reward_std": 0.12620826810598373,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.6764779090881348,
      "step": 189
    },
    {
      "completion_length": 99.77083587646484,
      "epoch": 1.497041420118343,
      "grad_norm": 2.674388885498047,
      "kl": 0.06103515625,
      "learning_rate": 1e-06,
      "loss": 0.0024,
      "reward": 1.7588639855384827,
      "reward_std": 0.11651718989014626,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7796972990036011,
      "step": 190
    },
    {
      "completion_length": 98.09375190734863,
      "epoch": 1.504930966469428,
      "grad_norm": 11.82987117767334,
      "kl": 0.0665283203125,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.7967063784599304,
      "reward_std": 0.030512763187289238,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7967063188552856,
      "step": 191
    },
    {
      "completion_length": 97.13541793823242,
      "epoch": 1.5128205128205128,
      "grad_norm": 1.5385335683822632,
      "kl": 0.0543212890625,
      "learning_rate": 1e-06,
      "loss": 0.0022,
      "reward": 1.7657526135444641,
      "reward_std": 0.08876698836684227,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7761692702770233,
      "step": 192
    },
    {
      "completion_length": 96.39583587646484,
      "epoch": 1.5207100591715976,
      "grad_norm": 1.428580641746521,
      "kl": 0.06072998046875,
      "learning_rate": 1e-06,
      "loss": 0.0024,
      "reward": 1.7331555485725403,
      "reward_std": 0.08605837309733033,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7435721755027771,
      "step": 193
    },
    {
      "completion_length": 96.07291984558105,
      "epoch": 1.5285996055226825,
      "grad_norm": 2.108811616897583,
      "kl": 0.071044921875,
      "learning_rate": 1e-06,
      "loss": 0.0028,
      "reward": 1.757724791765213,
      "reward_std": 0.07616918394342065,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7681414484977722,
      "step": 194
    },
    {
      "completion_length": 94.59375190734863,
      "epoch": 1.5364891518737673,
      "grad_norm": 4.254746913909912,
      "kl": 0.111572265625,
      "learning_rate": 1e-06,
      "loss": 0.0045,
      "reward": 1.7280253171920776,
      "reward_std": 0.0810533631592989,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7384419143199921,
      "step": 195
    },
    {
      "completion_length": 99.73958587646484,
      "epoch": 1.5443786982248522,
      "grad_norm": 2.411263942718506,
      "kl": 0.072509765625,
      "learning_rate": 1e-06,
      "loss": 0.0029,
      "reward": 1.7581177949905396,
      "reward_std": 0.04941168939694762,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7581177949905396,
      "step": 196
    },
    {
      "completion_length": 97.29166793823242,
      "epoch": 1.552268244575937,
      "grad_norm": 1.9083694219589233,
      "kl": 0.0650634765625,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.7459802627563477,
      "reward_std": 0.09188527404330671,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.766813650727272,
      "step": 197
    },
    {
      "completion_length": 93.73958587646484,
      "epoch": 1.5601577909270217,
      "grad_norm": 1.550249457359314,
      "kl": 0.0621337890625,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "reward": 1.7677803039550781,
      "reward_std": 0.04235605616122484,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7677803039550781,
      "step": 198
    },
    {
      "completion_length": 93.08333587646484,
      "epoch": 1.5680473372781065,
      "grad_norm": 9.880169868469238,
      "kl": 0.0809326171875,
      "learning_rate": 1e-06,
      "loss": 0.0032,
      "reward": 1.7984677851200104,
      "reward_std": 0.024913502275012434,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7984676957130432,
      "step": 199
    },
    {
      "completion_length": 94.59375381469727,
      "epoch": 1.5759368836291914,
      "grad_norm": 1.7292006015777588,
      "kl": 0.0706787109375,
      "learning_rate": 1e-06,
      "loss": 0.0028,
      "reward": 1.6630052626132965,
      "reward_std": 0.10535286273807287,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.6734219044446945,
      "step": 200
    },
    {
      "completion_length": 89.83333587646484,
      "epoch": 1.583826429980276,
      "grad_norm": 3.4253365993499756,
      "kl": 0.0667724609375,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.7671616673469543,
      "reward_std": 0.04377884138375521,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7671616673469543,
      "step": 201
    },
    {
      "completion_length": 89.06250190734863,
      "epoch": 1.5917159763313609,
      "grad_norm": 3.8319993019104004,
      "kl": 0.0673828125,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.7373066842556,
      "reward_std": 0.05676834611222148,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.747723326086998,
      "step": 202
    },
    {
      "completion_length": 90.77083587646484,
      "epoch": 1.5996055226824457,
      "grad_norm": 2.4027304649353027,
      "kl": 0.0836181640625,
      "learning_rate": 1e-06,
      "loss": 0.0033,
      "reward": 1.7197685837745667,
      "reward_std": 0.048274436965584755,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7197685688734055,
      "step": 203
    },
    {
      "completion_length": 91.70833587646484,
      "epoch": 1.6074950690335306,
      "grad_norm": 2.952907085418701,
      "kl": 0.079345703125,
      "learning_rate": 1e-06,
      "loss": 0.0032,
      "reward": 1.7424035966396332,
      "reward_std": 0.08384070452302694,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.75282022356987,
      "step": 204
    },
    {
      "completion_length": 84.82292175292969,
      "epoch": 1.6153846153846154,
      "grad_norm": 3.295057773590088,
      "kl": 0.0828857421875,
      "learning_rate": 1e-06,
      "loss": 0.0033,
      "reward": 1.7684763371944427,
      "reward_std": 0.042786918580532074,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7684763222932816,
      "step": 205
    },
    {
      "completion_length": 86.11458587646484,
      "epoch": 1.6232741617357003,
      "grad_norm": 5.877788543701172,
      "kl": 0.08740234375,
      "learning_rate": 1e-06,
      "loss": 0.0035,
      "reward": 1.7584342658519745,
      "reward_std": 0.04845387488603592,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7584342360496521,
      "step": 206
    },
    {
      "completion_length": 87.69791984558105,
      "epoch": 1.6311637080867851,
      "grad_norm": 11.630145072937012,
      "kl": 0.074462890625,
      "learning_rate": 1e-06,
      "loss": 0.003,
      "reward": 1.7452757060527802,
      "reward_std": 0.048048963537439704,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7452756613492966,
      "step": 207
    },
    {
      "completion_length": 87.88541984558105,
      "epoch": 1.63905325443787,
      "grad_norm": 2.092057228088379,
      "kl": 0.102294921875,
      "learning_rate": 1e-06,
      "loss": 0.0041,
      "reward": 1.7092933058738708,
      "reward_std": 0.12242361903190613,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7301265597343445,
      "step": 208
    },
    {
      "completion_length": 88.29166984558105,
      "epoch": 1.6469428007889546,
      "grad_norm": 1.6916812658309937,
      "kl": 0.0902099609375,
      "learning_rate": 1e-06,
      "loss": 0.0036,
      "reward": 1.7132295370101929,
      "reward_std": 0.17004388198256493,
      "rewards/format_reward": 0.9479167014360428,
      "rewards/segmentation_reward": 0.7653128653764725,
      "step": 209
    },
    {
      "completion_length": 86.42708396911621,
      "epoch": 1.6548323471400395,
      "grad_norm": 2.2493884563446045,
      "kl": 0.08447265625,
      "learning_rate": 1e-06,
      "loss": 0.0034,
      "reward": 1.7234868705272675,
      "reward_std": 0.06308982148766518,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7339034825563431,
      "step": 210
    },
    {
      "completion_length": 92.68750190734863,
      "epoch": 1.6627218934911243,
      "grad_norm": 2.1522436141967773,
      "kl": 0.07208251953125,
      "learning_rate": 1e-06,
      "loss": 0.0029,
      "reward": 1.753316581249237,
      "reward_std": 0.127250699326396,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.774149939417839,
      "step": 211
    },
    {
      "completion_length": 90.13541984558105,
      "epoch": 1.670611439842209,
      "grad_norm": 1.7701236009597778,
      "kl": 0.06756591796875,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.7777323722839355,
      "reward_std": 0.04612369416281581,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7777323573827744,
      "step": 212
    },
    {
      "completion_length": 89.05208587646484,
      "epoch": 1.6785009861932938,
      "grad_norm": 9.215665817260742,
      "kl": 0.06768798828125,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.8172322511672974,
      "reward_std": 0.058106679469347,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.827648863196373,
      "step": 213
    },
    {
      "completion_length": 92.09375190734863,
      "epoch": 1.6863905325443787,
      "grad_norm": 1.429556131362915,
      "kl": 0.0614013671875,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "reward": 1.7814797163009644,
      "reward_std": 0.05074140336364508,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7814796715974808,
      "step": 214
    },
    {
      "completion_length": 95.375,
      "epoch": 1.6942800788954635,
      "grad_norm": 1.654097080230713,
      "kl": 0.06536865234375,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.798690527677536,
      "reward_std": 0.054276286624372005,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.79869045317173,
      "step": 215
    },
    {
      "completion_length": 91.86458587646484,
      "epoch": 1.7021696252465484,
      "grad_norm": 2.162991523742676,
      "kl": 0.063720703125,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.816997766494751,
      "reward_std": 0.045502190943807364,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.816997766494751,
      "step": 216
    },
    {
      "completion_length": 102.71875381469727,
      "epoch": 1.7100591715976332,
      "grad_norm": 2.4174466133117676,
      "kl": 0.0667724609375,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.7929243445396423,
      "reward_std": 0.057674878276884556,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7929242998361588,
      "step": 217
    },
    {
      "completion_length": 100.30208778381348,
      "epoch": 1.717948717948718,
      "grad_norm": 2.792175531387329,
      "kl": 0.08599853515625,
      "learning_rate": 1e-06,
      "loss": 0.0034,
      "reward": 1.7294350564479828,
      "reward_std": 0.12824528431519866,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.7502683401107788,
      "step": 218
    },
    {
      "completion_length": 101.83333778381348,
      "epoch": 1.725838264299803,
      "grad_norm": 2.035576820373535,
      "kl": 0.0584716796875,
      "learning_rate": 1e-06,
      "loss": 0.0023,
      "reward": 1.765143632888794,
      "reward_std": 0.12963381642475724,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7859769463539124,
      "step": 219
    },
    {
      "completion_length": 101.50000190734863,
      "epoch": 1.7337278106508875,
      "grad_norm": 2.4070231914520264,
      "kl": 0.07916259765625,
      "learning_rate": 1e-06,
      "loss": 0.0032,
      "reward": 1.7128539979457855,
      "reward_std": 0.11889935936778784,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7336873859167099,
      "step": 220
    },
    {
      "completion_length": 101.09375190734863,
      "epoch": 1.7416173570019724,
      "grad_norm": 2.5801444053649902,
      "kl": 0.06781005859375,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.7691087424755096,
      "reward_std": 0.07278524897992611,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7691086679697037,
      "step": 221
    },
    {
      "completion_length": 104.84375381469727,
      "epoch": 1.7495069033530573,
      "grad_norm": 3.1142055988311768,
      "kl": 0.06622314453125,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.726320058107376,
      "reward_std": 0.09467954037245363,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7367367446422577,
      "step": 222
    },
    {
      "completion_length": 101.51041984558105,
      "epoch": 1.7573964497041419,
      "grad_norm": 2.2835278511047363,
      "kl": 0.0704345703125,
      "learning_rate": 1e-06,
      "loss": 0.0028,
      "reward": 1.697280466556549,
      "reward_std": 0.08086889609694481,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7076971083879471,
      "step": 223
    },
    {
      "completion_length": 105.22916984558105,
      "epoch": 1.7652859960552267,
      "grad_norm": 1.3429104089736938,
      "kl": 0.07025146484375,
      "learning_rate": 1e-06,
      "loss": 0.0028,
      "reward": 1.6932333409786224,
      "reward_std": 0.058003371581435204,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7036499530076981,
      "step": 224
    },
    {
      "completion_length": 104.63541793823242,
      "epoch": 1.7731755424063116,
      "grad_norm": 2.0300707817077637,
      "kl": 0.06463623046875,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.758153110742569,
      "reward_std": 0.049405214842408895,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.758153110742569,
      "step": 225
    },
    {
      "completion_length": 109.54166984558105,
      "epoch": 1.7810650887573964,
      "grad_norm": 2.2276878356933594,
      "kl": 0.05450439453125,
      "learning_rate": 1e-06,
      "loss": 0.0022,
      "reward": 1.7898038029670715,
      "reward_std": 0.07037154771387577,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.8002204447984695,
      "step": 226
    },
    {
      "completion_length": 110.67708587646484,
      "epoch": 1.7889546351084813,
      "grad_norm": 2.3131182193756104,
      "kl": 0.05841064453125,
      "learning_rate": 1e-06,
      "loss": 0.0023,
      "reward": 1.810005784034729,
      "reward_std": 0.02927766414359212,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.8100057393312454,
      "step": 227
    },
    {
      "completion_length": 109.03125190734863,
      "epoch": 1.7968441814595661,
      "grad_norm": 5.591848373413086,
      "kl": 0.063232421875,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "reward": 1.7375752925872803,
      "reward_std": 0.06387870013713837,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7375752925872803,
      "step": 228
    },
    {
      "completion_length": 115.06250381469727,
      "epoch": 1.804733727810651,
      "grad_norm": 2.8504865169525146,
      "kl": 0.0628662109375,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "reward": 1.6701695919036865,
      "reward_std": 0.09486013744026423,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.6805862784385681,
      "step": 229
    },
    {
      "completion_length": 114.32291793823242,
      "epoch": 1.8126232741617359,
      "grad_norm": 2.7374796867370605,
      "kl": 0.06536865234375,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.7931168377399445,
      "reward_std": 0.06577130127698183,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7931168079376221,
      "step": 230
    },
    {
      "completion_length": 116.29167175292969,
      "epoch": 1.8205128205128205,
      "grad_norm": 3.255202054977417,
      "kl": 0.06494140625,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.728683203458786,
      "reward_std": 0.09193603880703449,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7390998601913452,
      "step": 231
    },
    {
      "completion_length": 121.97916793823242,
      "epoch": 1.8284023668639053,
      "grad_norm": 3.000345468521118,
      "kl": 0.0560302734375,
      "learning_rate": 1e-06,
      "loss": 0.0022,
      "reward": 1.741953283548355,
      "reward_std": 0.10554883629083633,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.7627865821123123,
      "step": 232
    },
    {
      "completion_length": 118.30208587646484,
      "epoch": 1.8362919132149902,
      "grad_norm": 2.914006471633911,
      "kl": 0.07244873046875,
      "learning_rate": 1e-06,
      "loss": 0.0029,
      "reward": 1.7272584438323975,
      "reward_std": 0.1506546987220645,
      "rewards/format_reward": 0.9687500149011612,
      "rewards/segmentation_reward": 0.7585083991289139,
      "step": 233
    },
    {
      "completion_length": 121.15625381469727,
      "epoch": 1.8441814595660748,
      "grad_norm": 2.076815128326416,
      "kl": 0.069091796875,
      "learning_rate": 1e-06,
      "loss": 0.0028,
      "reward": 1.6870795786380768,
      "reward_std": 0.10884091723710299,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7079129219055176,
      "step": 234
    },
    {
      "completion_length": 115.11458587646484,
      "epoch": 1.8520710059171597,
      "grad_norm": 2.715985059738159,
      "kl": 0.0792236328125,
      "learning_rate": 1e-06,
      "loss": 0.0032,
      "reward": 1.7684183418750763,
      "reward_std": 0.09033020678907633,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7788349986076355,
      "step": 235
    },
    {
      "completion_length": 119.62500381469727,
      "epoch": 1.8599605522682445,
      "grad_norm": 3.4143686294555664,
      "kl": 0.0587158203125,
      "learning_rate": 1e-06,
      "loss": 0.0024,
      "reward": 1.8009620606899261,
      "reward_std": 0.046604559291154146,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.8009620159864426,
      "step": 236
    },
    {
      "completion_length": 126.77083778381348,
      "epoch": 1.8678500986193294,
      "grad_norm": 2.546739339828491,
      "kl": 0.062255859375,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "reward": 1.761908084154129,
      "reward_std": 0.07723418436944485,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7723246961832047,
      "step": 237
    },
    {
      "completion_length": 122.57291793823242,
      "epoch": 1.8757396449704142,
      "grad_norm": 11.713898658752441,
      "kl": 0.06842041015625,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.7244771420955658,
      "reward_std": 0.04781572869978845,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7244771271944046,
      "step": 238
    },
    {
      "completion_length": 120.54166793823242,
      "epoch": 1.883629191321499,
      "grad_norm": 4.24526834487915,
      "kl": 0.0626220703125,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "reward": 1.7059939503669739,
      "reward_std": 0.13842293061316013,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7268272489309311,
      "step": 239
    },
    {
      "completion_length": 117.65625381469727,
      "epoch": 1.891518737672584,
      "grad_norm": 4.1007561683654785,
      "kl": 0.06787109375,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.7405757904052734,
      "reward_std": 0.04020647844299674,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.740575760602951,
      "step": 240
    },
    {
      "completion_length": 117.26041984558105,
      "epoch": 1.8994082840236688,
      "grad_norm": 2.303384780883789,
      "kl": 0.07501220703125,
      "learning_rate": 1e-06,
      "loss": 0.003,
      "reward": 1.6991797089576721,
      "reward_std": 0.19727707654237747,
      "rewards/format_reward": 0.9479167014360428,
      "rewards/segmentation_reward": 0.7512629479169846,
      "step": 241
    },
    {
      "completion_length": 121.5000057220459,
      "epoch": 1.9072978303747534,
      "grad_norm": 1.9164984226226807,
      "kl": 0.0572509765625,
      "learning_rate": 1e-06,
      "loss": 0.0023,
      "reward": 1.7220526039600372,
      "reward_std": 0.06677288189530373,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.722052663564682,
      "step": 242
    },
    {
      "completion_length": 118.84375190734863,
      "epoch": 1.9151873767258383,
      "grad_norm": 4.73820161819458,
      "kl": 0.05950927734375,
      "learning_rate": 1e-06,
      "loss": 0.0024,
      "reward": 1.7459237277507782,
      "reward_std": 0.041601293021813035,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7459237277507782,
      "step": 243
    },
    {
      "completion_length": 113.8125,
      "epoch": 1.9230769230769231,
      "grad_norm": 3.414625406265259,
      "kl": 0.0687255859375,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.750750720500946,
      "reward_std": 0.10410196101292968,
      "rewards/format_reward": 0.9791666865348816,
      "rewards/segmentation_reward": 0.7715839594602585,
      "step": 244
    },
    {
      "completion_length": 114.04166793823242,
      "epoch": 1.9309664694280078,
      "grad_norm": 3.6483190059661865,
      "kl": 0.05694580078125,
      "learning_rate": 1e-06,
      "loss": 0.0023,
      "reward": 1.850210815668106,
      "reward_std": 0.030932937748730183,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.8502108156681061,
      "step": 245
    },
    {
      "completion_length": 111.08333587646484,
      "epoch": 1.9388560157790926,
      "grad_norm": 2.375276803970337,
      "kl": 0.06561279296875,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.741035521030426,
      "reward_std": 0.10635390551760793,
      "rewards/format_reward": 0.9791666716337204,
      "rewards/segmentation_reward": 0.761868804693222,
      "step": 246
    },
    {
      "completion_length": 112.56250381469727,
      "epoch": 1.9467455621301775,
      "grad_norm": 3.1703550815582275,
      "kl": 0.06463623046875,
      "learning_rate": 1e-06,
      "loss": 0.0026,
      "reward": 1.743516057729721,
      "reward_std": 0.054085819981992245,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7435160130262375,
      "step": 247
    },
    {
      "completion_length": 112.88541984558105,
      "epoch": 1.9546351084812623,
      "grad_norm": 19.166675567626953,
      "kl": 0.069091796875,
      "learning_rate": 1e-06,
      "loss": 0.0028,
      "reward": 1.6917358338832855,
      "reward_std": 0.08896347414702177,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7021525055170059,
      "step": 248
    },
    {
      "completion_length": 113.57291984558105,
      "epoch": 1.9625246548323472,
      "grad_norm": 1.4388498067855835,
      "kl": 0.0714111328125,
      "learning_rate": 1e-06,
      "loss": 0.0029,
      "reward": 1.7879726886749268,
      "reward_std": 0.04161839000880718,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.787972629070282,
      "step": 249
    },
    {
      "completion_length": 113.20833587646484,
      "epoch": 1.970414201183432,
      "grad_norm": 4.911606311798096,
      "kl": 0.068603515625,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.675138384103775,
      "reward_std": 0.07887607486918569,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.6855549961328506,
      "step": 250
    },
    {
      "completion_length": 107.64583396911621,
      "epoch": 1.9783037475345169,
      "grad_norm": 2.6165802478790283,
      "kl": 0.06707763671875,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "reward": 1.7642135918140411,
      "reward_std": 0.05668263230472803,
      "rewards/format_reward": 1.0,
      "rewards/segmentation_reward": 0.7642135620117188,
      "step": 251
    },
    {
      "completion_length": 112.51041984558105,
      "epoch": 1.9861932938856017,
      "grad_norm": 2.5721254348754883,
      "kl": 0.0810546875,
      "learning_rate": 1e-06,
      "loss": 0.0032,
      "reward": 1.7676794826984406,
      "reward_std": 0.05829272768460214,
      "rewards/format_reward": 0.9895833432674408,
      "rewards/segmentation_reward": 0.7780960947275162,
      "step": 252
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 252,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 200,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}