trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.6129032258064515,
  "eval_steps": 500,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2796.0,
      "completions/mean_length": 328.47918701171875,
      "completions/min_length": 46.0,
      "epoch": 0.016129032258064516,
      "grad_norm": 0.2647878213954199,
      "kl": 0.0,
      "learning_rate": 8.333333333333333e-08,
      "loss": 8.76995454035523e-09,
      "memory(GiB)": 66.25,
      "reward": 0.9901389479637146,
      "reward_std": 0.2693294882774353,
      "rewards/CODEORM/mean": 0.010625001043081284,
      "rewards/CODEORM/std": 0.055468425154685974,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.01317615807056427,
      "rewards/VQAORM/mean": 0.4798611104488373,
      "rewards/VQAORM/std": 0.49976783990859985,
      "step": 1,
      "train_speed(iter/s)": 0.002513
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 4599.0,
      "completions/mean_length": 323.51251220703125,
      "completions/min_length": 46.0,
      "epoch": 0.03225806451612903,
      "grad_norm": 0.25445235632470914,
      "kl": 0.0,
      "learning_rate": 1.6666666666666665e-07,
      "loss": 2.5870072750677764e-09,
      "memory(GiB)": 67.27,
      "reward": 0.9681249856948853,
      "reward_std": 0.2566339373588562,
      "rewards/CODEORM/mean": 0.0066666672937572,
      "rewards/CODEORM/std": 0.04423702508211136,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.01317615807056427,
      "rewards/VQAORM/mean": 0.4618055522441864,
      "rewards/VQAORM/std": 0.4987122416496277,
      "step": 2,
      "train_speed(iter/s)": 0.002804
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2598.0,
      "completions/mean_length": 308.897216796875,
      "completions/min_length": 49.0,
      "epoch": 0.04838709677419355,
      "grad_norm": 0.26595615090662006,
      "kl": 0.0001987457275390625,
      "learning_rate": 2.5e-07,
      "loss": 2.0041927655256586e-06,
      "memory(GiB)": 67.27,
      "reward": 0.993263840675354,
      "reward_std": 0.24853122234344482,
      "rewards/CODEORM/mean": 0.008541665971279144,
      "rewards/CODEORM/std": 0.049912624061107635,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.5,
      "rewards/FMTORM/std": 0.0,
      "rewards/VQAORM/mean": 0.48472222685813904,
      "rewards/VQAORM/std": 0.4999401569366455,
      "step": 3,
      "train_speed(iter/s)": 0.003038
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3566.0,
      "completions/mean_length": 319.540283203125,
      "completions/min_length": 49.0,
      "epoch": 0.06451612903225806,
      "grad_norm": 0.28511858475822077,
      "kl": 0.00017852783203125,
      "learning_rate": 3.333333333333333e-07,
      "loss": 1.797711320250528e-06,
      "memory(GiB)": 67.27,
      "reward": 0.998055636882782,
      "reward_std": 0.268062561750412,
      "rewards/CODEORM/mean": 0.01020833384245634,
      "rewards/CODEORM/std": 0.05440906062722206,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4888888895511627,
      "rewards/VQAORM/std": 0.5000501871109009,
      "step": 4,
      "train_speed(iter/s)": 0.003145
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2966.0,
      "completions/mean_length": 364.01043701171875,
      "completions/min_length": 43.0,
      "epoch": 0.08064516129032258,
      "grad_norm": 0.2684251266862969,
      "kl": 0.000177764892578125,
      "learning_rate": 4.1666666666666667e-07,
      "loss": 1.7898388477988192e-06,
      "memory(GiB)": 71.39,
      "reward": 0.9475694894790649,
      "reward_std": 0.27028676867485046,
      "rewards/CODEORM/mean": 0.00937500037252903,
      "rewards/CODEORM/std": 0.05221592262387276,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.43888890743255615,
      "rewards/VQAORM/std": 0.49642378091812134,
      "step": 5,
      "train_speed(iter/s)": 0.003206
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2712.0,
      "completions/mean_length": 319.3951416015625,
      "completions/min_length": 41.0,
      "epoch": 0.0967741935483871,
      "grad_norm": 0.2404392113235966,
      "kl": 0.0001796722412109375,
      "learning_rate": 5e-07,
      "loss": 1.804373255254177e-06,
      "memory(GiB)": 71.39,
      "reward": 0.9318056106567383,
      "reward_std": 0.2407531440258026,
      "rewards/CODEORM/mean": 0.006458333693444729,
      "rewards/CODEORM/std": 0.04355579614639282,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.42569446563720703,
      "rewards/VQAORM/std": 0.4946196377277374,
      "step": 6,
      "train_speed(iter/s)": 0.003221
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2772.0,
      "completions/mean_length": 302.9992980957031,
      "completions/min_length": 54.0,
      "epoch": 0.11290322580645161,
      "grad_norm": 0.24614522756314014,
      "kl": 0.0001972198486328125,
      "learning_rate": 4.99965731410188e-07,
      "loss": 1.978319232875947e-06,
      "memory(GiB)": 71.39,
      "reward": 0.9795833230018616,
      "reward_std": 0.2439241260290146,
      "rewards/CODEORM/mean": 0.004583333153277636,
      "rewards/CODEORM/std": 0.03680942952632904,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.5,
      "rewards/FMTORM/std": 0.0,
      "rewards/VQAORM/mean": 0.4750000238418579,
      "rewards/VQAORM/std": 0.4995481073856354,
      "step": 7,
      "train_speed(iter/s)": 0.003258
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2551.0,
      "completions/mean_length": 327.8243103027344,
      "completions/min_length": 55.0,
      "epoch": 0.12903225806451613,
      "grad_norm": 0.25632964332335334,
      "kl": 0.000189208984375,
      "learning_rate": 4.998629360792966e-07,
      "loss": 1.9058469433730352e-06,
      "memory(GiB)": 71.39,
      "reward": 1.0435417890548706,
      "reward_std": 0.26714909076690674,
      "rewards/CODEORM/mean": 0.01020833384245634,
      "rewards/CODEORM/std": 0.05440906062722206,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.018627425655722618,
      "rewards/VQAORM/mean": 0.5340278148651123,
      "rewards/VQAORM/std": 0.49901407957077026,
      "step": 8,
      "train_speed(iter/s)": 0.003284
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 11576.0,
      "completions/mean_length": 324.6034851074219,
      "completions/min_length": 40.0,
      "epoch": 0.14516129032258066,
      "grad_norm": 0.2555360483438473,
      "kl": 0.00021514892578125,
      "learning_rate": 4.996916453197791e-07,
      "loss": 2.1652638224622933e-06,
      "memory(GiB)": 71.39,
      "reward": 1.0275694131851196,
      "reward_std": 0.2437307983636856,
      "rewards/CODEORM/mean": 0.009166667237877846,
      "rewards/CODEORM/std": 0.0516509935259819,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4982638955116272,
      "rewards/FMTORM/std": 0.02942180633544922,
      "rewards/VQAORM/mean": 0.5201389193534851,
      "rewards/VQAORM/std": 0.49976781010627747,
      "step": 9,
      "train_speed(iter/s)": 0.003131
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2962.0,
      "completions/mean_length": 347.196533203125,
      "completions/min_length": 58.0,
      "epoch": 0.16129032258064516,
      "grad_norm": 0.22806471451561894,
      "kl": 0.0002552032470703125,
      "learning_rate": 4.994519113084604e-07,
      "loss": 2.564860324127949e-06,
      "memory(GiB)": 71.39,
      "reward": 0.9813888669013977,
      "reward_std": 0.22997106611728668,
      "rewards/CODEORM/mean": 0.009166667237877846,
      "rewards/CODEORM/std": 0.0516509935259819,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49791666865348816,
      "rewards/FMTORM/std": 0.03221873939037323,
      "rewards/VQAORM/mean": 0.47430557012557983,
      "rewards/VQAORM/std": 0.49951285123825073,
      "step": 10,
      "train_speed(iter/s)": 0.003171
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 5360.0,
      "completions/mean_length": 360.31390380859375,
      "completions/min_length": 48.0,
      "epoch": 0.1774193548387097,
      "grad_norm": 0.2832579586762888,
      "kl": 0.0003253936767578125,
      "learning_rate": 4.991438070706428e-07,
      "loss": 3.250584086345043e-06,
      "memory(GiB)": 71.39,
      "reward": 1.0013889074325562,
      "reward_std": 0.25826162099838257,
      "rewards/CODEORM/mean": 0.01979166828095913,
      "rewards/CODEORM/std": 0.07449593394994736,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4826388955116272,
      "rewards/VQAORM/std": 0.499872088432312,
      "step": 11,
      "train_speed(iter/s)": 0.003168
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2102.0,
      "completions/mean_length": 335.2215270996094,
      "completions/min_length": 52.0,
      "epoch": 0.1935483870967742,
      "grad_norm": 0.27801393384997114,
      "kl": 0.0003223419189453125,
      "learning_rate": 4.987674264578614e-07,
      "loss": 3.231521304769558e-06,
      "memory(GiB)": 71.39,
      "reward": 0.9737499952316284,
      "reward_std": 0.27600589394569397,
      "rewards/CODEORM/mean": 0.009166666306555271,
      "rewards/CODEORM/std": 0.0516509935259819,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.4652777910232544,
      "rewards/VQAORM/std": 0.49896618723869324,
      "step": 12,
      "train_speed(iter/s)": 0.003207
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2673.0,
      "completions/mean_length": 342.41876220703125,
      "completions/min_length": 54.0,
      "epoch": 0.20967741935483872,
      "grad_norm": 0.2586540657680122,
      "kl": 0.000345611572265625,
      "learning_rate": 4.983228841192975e-07,
      "loss": 3.455609885349986e-06,
      "memory(GiB)": 71.39,
      "reward": 0.9851389527320862,
      "reward_std": 0.26006534695625305,
      "rewards/CODEORM/mean": 0.008750000037252903,
      "rewards/CODEORM/std": 0.05049958825111389,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.018627425655722618,
      "rewards/VQAORM/mean": 0.47708335518836975,
      "rewards/VQAORM/std": 0.4996480643749237,
      "step": 13,
      "train_speed(iter/s)": 0.003229
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2441.0,
      "completions/mean_length": 346.803466796875,
      "completions/min_length": 54.0,
      "epoch": 0.22580645161290322,
      "grad_norm": 0.26610442827789593,
      "kl": 0.000457763671875,
      "learning_rate": 4.978103154668533e-07,
      "loss": 4.586853265209356e-06,
      "memory(GiB)": 72.28,
      "reward": 0.9384721517562866,
      "reward_std": 0.26384806632995605,
      "rewards/CODEORM/mean": 0.008958334103226662,
      "rewards/CODEORM/std": 0.05107896029949188,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4305555522441864,
      "rewards/VQAORM/std": 0.4953259825706482,
      "step": 14,
      "train_speed(iter/s)": 0.003242
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2603.0,
      "completions/mean_length": 344.1347351074219,
      "completions/min_length": 52.0,
      "epoch": 0.24193548387096775,
      "grad_norm": 0.24930681367809324,
      "kl": 0.000681304931640625,
      "learning_rate": 4.972298766339059e-07,
      "loss": 6.819126156187849e-06,
      "memory(GiB)": 72.28,
      "reward": 0.9356943964958191,
      "reward_std": 0.2636728882789612,
      "rewards/CODEORM/mean": 0.009999999776482582,
      "rewards/CODEORM/std": 0.05387035384774208,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.4263888895511627,
      "rewards/VQAORM/std": 0.49472352862358093,
      "step": 15,
      "train_speed(iter/s)": 0.003258
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2302.0,
      "completions/mean_length": 315.2257080078125,
      "completions/min_length": 47.0,
      "epoch": 0.25806451612903225,
      "grad_norm": 0.2280556698693846,
      "kl": 0.000701904296875,
      "learning_rate": 4.965817444277468e-07,
      "loss": 7.027399533399148e-06,
      "memory(GiB)": 72.28,
      "reward": 0.9556249976158142,
      "reward_std": 0.24996338784694672,
      "rewards/CODEORM/mean": 0.008750000037252903,
      "rewards/CODEORM/std": 0.05049958825111389,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.4472222328186035,
      "rewards/VQAORM/std": 0.49737945199012756,
      "step": 16,
      "train_speed(iter/s)": 0.003286
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2913.0,
      "completions/mean_length": 346.1423645019531,
      "completions/min_length": 1.0,
      "epoch": 0.27419354838709675,
      "grad_norm": 0.9969603368489056,
      "kl": 0.001026153564453125,
      "learning_rate": 4.958661162757244e-07,
      "loss": 1.0287308214174118e-05,
      "memory(GiB)": 72.28,
      "reward": 0.9165971875190735,
      "reward_std": 0.2503140866756439,
      "rewards/CODEORM/mean": 0.008958333171904087,
      "rewards/CODEORM/std": 0.05107896029949188,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.018627425655722618,
      "rewards/VQAORM/mean": 0.40833333134651184,
      "rewards/VQAORM/std": 0.49169617891311646,
      "step": 17,
      "train_speed(iter/s)": 0.002895
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2168.0,
      "completions/mean_length": 336.17291259765625,
      "completions/min_length": 50.0,
      "epoch": 0.2903225806451613,
      "grad_norm": 0.2580204320478004,
      "kl": 0.000994873046875,
      "learning_rate": 4.950832101651062e-07,
      "loss": 9.966525794879999e-06,
      "memory(GiB)": 72.28,
      "reward": 0.9159722328186035,
      "reward_std": 0.22609728574752808,
      "rewards/CODEORM/mean": 0.0062500000931322575,
      "rewards/CODEORM/std": 0.042862728238105774,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.5,
      "rewards/FMTORM/std": 0.0,
      "rewards/VQAORM/mean": 0.409722238779068,
      "rewards/VQAORM/std": 0.4919532239437103,
      "step": 18,
      "train_speed(iter/s)": 0.002929
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2828.0,
      "completions/mean_length": 344.4798583984375,
      "completions/min_length": 56.0,
      "epoch": 0.3064516129032258,
      "grad_norm": 0.2572831613912269,
      "kl": 0.001146697998046875,
      "learning_rate": 4.94233264576678e-07,
      "loss": 1.1491146324260626e-05,
      "memory(GiB)": 72.28,
      "reward": 0.9686111807823181,
      "reward_std": 0.2488832324743271,
      "rewards/CODEORM/mean": 0.008541667833924294,
      "rewards/CODEORM/std": 0.049912624061107635,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4611111283302307,
      "rewards/VQAORM/std": 0.49865853786468506,
      "step": 19,
      "train_speed(iter/s)": 0.002957
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2005.0,
      "completions/mean_length": 339.0666809082031,
      "completions/min_length": 50.0,
      "epoch": 0.3225806451612903,
      "grad_norm": 0.25242303032801683,
      "kl": 0.0013702392578125,
      "learning_rate": 4.933165384120992e-07,
      "loss": 1.3699734154215548e-05,
      "memory(GiB)": 72.28,
      "reward": 0.9613890051841736,
      "reward_std": 0.2589081823825836,
      "rewards/CODEORM/mean": 0.006874999962747097,
      "rewards/CODEORM/std": 0.04490695893764496,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.4548611342906952,
      "rewards/VQAORM/std": 0.49813133478164673,
      "step": 20,
      "train_speed(iter/s)": 0.002987
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2956.0,
      "completions/mean_length": 347.7388916015625,
      "completions/min_length": 56.0,
      "epoch": 0.3387096774193548,
      "grad_norm": 0.2878476776805163,
      "kl": 0.001605224609375,
      "learning_rate": 4.923333109150403e-07,
      "loss": 1.6045431038946845e-05,
      "memory(GiB)": 72.28,
      "reward": 0.9531945586204529,
      "reward_std": 0.2703089118003845,
      "rewards/CODEORM/mean": 0.013958333991467953,
      "rewards/CODEORM/std": 0.06320948898792267,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4402777850627899,
      "rewards/VQAORM/std": 0.4965929090976715,
      "step": 21,
      "train_speed(iter/s)": 0.003002
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2177.0,
      "completions/mean_length": 300.4312438964844,
      "completions/min_length": 47.0,
      "epoch": 0.3548387096774194,
      "grad_norm": 0.23824468095972504,
      "kl": 0.0014373779296875,
      "learning_rate": 4.912838815861217e-07,
      "loss": 1.4388399904419202e-05,
      "memory(GiB)": 72.28,
      "reward": 0.9931944608688354,
      "reward_std": 0.24322348833084106,
      "rewards/CODEORM/mean": 0.009166667237877846,
      "rewards/CODEORM/std": 0.051650989800691605,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.48472222685813904,
      "rewards/VQAORM/std": 0.4999401569366455,
      "step": 22,
      "train_speed(iter/s)": 0.003021
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1788.0,
      "completions/mean_length": 325.1958312988281,
      "completions/min_length": 53.0,
      "epoch": 0.3709677419354839,
      "grad_norm": 0.3498689416243422,
      "kl": 0.00506134033203125,
      "learning_rate": 4.90168570091683e-07,
      "loss": 5.055519432062283e-05,
      "memory(GiB)": 72.28,
      "reward": 1.0063194036483765,
      "reward_std": 0.26852238178253174,
      "rewards/CODEORM/mean": 0.013958334922790527,
      "rewards/CODEORM/std": 0.06320948898792267,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.4930555820465088,
      "rewards/VQAORM/std": 0.5001254677772522,
      "step": 23,
      "train_speed(iter/s)": 0.003047
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2844.0,
      "completions/mean_length": 355.35833740234375,
      "completions/min_length": 59.0,
      "epoch": 0.3870967741935484,
      "grad_norm": 0.2465080561687103,
      "kl": 0.0025970458984375,
      "learning_rate": 4.889877161664096e-07,
      "loss": 2.5979932615882717e-05,
      "memory(GiB)": 72.28,
      "reward": 0.9511805176734924,
      "reward_std": 0.23766973614692688,
      "rewards/CODEORM/mean": 0.0133333345875144,
      "rewards/CODEORM/std": 0.061845600605010986,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.4381944537162781,
      "rewards/VQAORM/std": 0.49633774161338806,
      "step": 24,
      "train_speed(iter/s)": 0.00306
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2828.0,
      "completions/mean_length": 364.1534729003906,
      "completions/min_length": 55.0,
      "epoch": 0.4032258064516129,
      "grad_norm": 0.2306227358811404,
      "kl": 0.0028472900390625,
      "learning_rate": 4.877416795098463e-07,
      "loss": 2.847511314030271e-05,
      "memory(GiB)": 72.28,
      "reward": 0.9688888788223267,
      "reward_std": 0.24576471745967865,
      "rewards/CODEORM/mean": 0.01020833384245634,
      "rewards/CODEORM/std": 0.05440906062722206,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.01317615807056427,
      "rewards/VQAORM/mean": 0.45902779698371887,
      "rewards/VQAORM/std": 0.4984915852546692,
      "step": 25,
      "train_speed(iter/s)": 0.003078
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3089.0,
      "completions/mean_length": 317.5958557128906,
      "completions/min_length": 46.0,
      "epoch": 0.41935483870967744,
      "grad_norm": 0.29253210222640175,
      "kl": 0.0026397705078125,
      "learning_rate": 4.864308396768294e-07,
      "loss": 2.6405881726532243e-05,
      "memory(GiB)": 72.28,
      "reward": 0.9640277624130249,
      "reward_std": 0.24826957285404205,
      "rewards/CODEORM/mean": 0.011250000447034836,
      "rewards/CODEORM/std": 0.057014863938093185,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.5,
      "rewards/FMTORM/std": 0.0,
      "rewards/VQAORM/mean": 0.45277780294418335,
      "rewards/VQAORM/std": 0.49793797731399536,
      "step": 26,
      "train_speed(iter/s)": 0.003097
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3438.0,
      "completions/mean_length": 338.2472229003906,
      "completions/min_length": 49.0,
      "epoch": 0.43548387096774194,
      "grad_norm": 0.2644761394181819,
      "kl": 0.0032989501953125,
      "learning_rate": 4.850555959618704e-07,
      "loss": 3.298750743851997e-05,
      "memory(GiB)": 72.28,
      "reward": 1.0049306154251099,
      "reward_std": 0.2615948021411896,
      "rewards/CODEORM/mean": 0.012916668318212032,
      "rewards/CODEORM/std": 0.06091581657528877,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.4923611283302307,
      "rewards/VQAORM/std": 0.5001153349876404,
      "step": 27,
      "train_speed(iter/s)": 0.003109
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2612.0,
      "completions/mean_length": 356.2597351074219,
      "completions/min_length": 49.0,
      "epoch": 0.45161290322580644,
      "grad_norm": 0.2665651878200343,
      "kl": 0.0039642333984375,
      "learning_rate": 4.836163672775272e-07,
      "loss": 3.9662394556216896e-05,
      "memory(GiB)": 72.28,
      "reward": 0.9762500524520874,
      "reward_std": 0.25711533427238464,
      "rewards/CODEORM/mean": 0.013749999925494194,
      "rewards/CODEORM/std": 0.06275884807109833,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.46319445967674255,
      "rewards/VQAORM/std": 0.49881675839424133,
      "step": 28,
      "train_speed(iter/s)": 0.00312
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1871.0,
      "completions/mean_length": 343.495849609375,
      "completions/min_length": 39.0,
      "epoch": 0.46774193548387094,
      "grad_norm": 0.23395808716833594,
      "kl": 0.004046630859375,
      "learning_rate": 4.82113592026799e-07,
      "loss": 4.044502566102892e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9686111807823181,
      "reward_std": 0.2451702207326889,
      "rewards/CODEORM/mean": 0.011666666716337204,
      "rewards/CODEORM/std": 0.058019187301397324,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.4576388895511627,
      "rewards/VQAORM/std": 0.49837538599967957,
      "step": 29,
      "train_speed(iter/s)": 0.00314
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2888.0,
      "completions/mean_length": 346.4354248046875,
      "completions/min_length": 38.0,
      "epoch": 0.4838709677419355,
      "grad_norm": 0.25535632583207873,
      "kl": 0.0040679931640625,
      "learning_rate": 4.805477279695852e-07,
      "loss": 4.070028444402851e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9380555152893066,
      "reward_std": 0.27259209752082825,
      "rewards/CODEORM/mean": 0.012708333320915699,
      "rewards/CODEORM/std": 0.060444481670856476,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4263888895511627,
      "rewards/VQAORM/std": 0.49472349882125854,
      "step": 30,
      "train_speed(iter/s)": 0.003152
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1982.0,
      "completions/mean_length": 342.07708740234375,
      "completions/min_length": 45.0,
      "epoch": 0.5,
      "grad_norm": 0.2757470110574285,
      "kl": 0.005291748046875,
      "learning_rate": 4.789192520832462e-07,
      "loss": 5.291224078973755e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9945833086967468,
      "reward_std": 0.2432967722415924,
      "rewards/CODEORM/mean": 0.017500000074505806,
      "rewards/CODEORM/std": 0.0703362300992012,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.018627425655722618,
      "rewards/VQAORM/mean": 0.47777777910232544,
      "rewards/VQAORM/std": 0.49967944622039795,
      "step": 31,
      "train_speed(iter/s)": 0.00317
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1852.0,
      "completions/mean_length": 358.5646057128906,
      "completions/min_length": 55.0,
      "epoch": 0.5161290322580645,
      "grad_norm": 0.24973674461489612,
      "kl": 0.00535888671875,
      "learning_rate": 4.772286604173125e-07,
      "loss": 5.361745570553467e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9641666412353516,
      "reward_std": 0.24900199472904205,
      "rewards/CODEORM/mean": 0.02041666768491268,
      "rewards/CODEORM/std": 0.07557861506938934,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.018627425655722618,
      "rewards/VQAORM/mean": 0.4444444477558136,
      "rewards/VQAORM/std": 0.4970766007900238,
      "step": 32,
      "train_speed(iter/s)": 0.003173
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2454.0,
      "completions/mean_length": 316.1222229003906,
      "completions/min_length": 36.0,
      "epoch": 0.532258064516129,
      "grad_norm": 0.27765379348097485,
      "kl": 0.005023193359375,
      "learning_rate": 4.754764679423827e-07,
      "loss": 5.026329745305702e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9695139527320862,
      "reward_std": 0.23932775855064392,
      "rewards/CODEORM/mean": 0.013958333991467953,
      "rewards/CODEORM/std": 0.06320948898792267,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.5,
      "rewards/FMTORM/std": 0.0,
      "rewards/VQAORM/mean": 0.4555555582046509,
      "rewards/VQAORM/std": 0.49819380044937134,
      "step": 33,
      "train_speed(iter/s)": 0.003188
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2153.0,
      "completions/mean_length": 378.8312683105469,
      "completions/min_length": 49.0,
      "epoch": 0.5483870967741935,
      "grad_norm": 0.21472738186965287,
      "kl": 0.006378173828125,
      "learning_rate": 4.7366320839325856e-07,
      "loss": 6.384468724718317e-05,
      "memory(GiB)": 72.55,
      "reward": 1.0134028196334839,
      "reward_std": 0.2518027424812317,
      "rewards/CODEORM/mean": 0.015833334997296333,
      "rewards/CODEORM/std": 0.0671001672744751,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.49861112236976624,
      "rewards/VQAORM/std": 0.5001717805862427,
      "step": 34,
      "train_speed(iter/s)": 0.003199
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2375.0,
      "completions/mean_length": 337.39654541015625,
      "completions/min_length": 56.0,
      "epoch": 0.5645161290322581,
      "grad_norm": 0.24373439505720154,
      "kl": 0.00570068359375,
      "learning_rate": 4.71789434106364e-07,
      "loss": 5.708396929549053e-05,
      "memory(GiB)": 72.55,
      "reward": 1.0311111211776733,
      "reward_std": 0.2390429675579071,
      "rewards/CODEORM/mean": 0.013750000856816769,
      "rewards/CODEORM/std": 0.06275884807109833,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.5,
      "rewards/FMTORM/std": 0.0,
      "rewards/VQAORM/mean": 0.5173611044883728,
      "rewards/VQAORM/std": 0.499872088432312,
      "step": 35,
      "train_speed(iter/s)": 0.003203
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2472.0,
      "completions/mean_length": 378.3333435058594,
      "completions/min_length": 46.0,
      "epoch": 0.5806451612903226,
      "grad_norm": 0.2500603222917725,
      "kl": 0.008148193359375,
      "learning_rate": 4.698557158514987e-07,
      "loss": 8.145920583046973e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9934723377227783,
      "reward_std": 0.26507648825645447,
      "rewards/CODEORM/mean": 0.020208334550261497,
      "rewards/CODEORM/std": 0.07522002607584,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.01317615620791912,
      "rewards/VQAORM/mean": 0.47361111640930176,
      "rewards/VQAORM/std": 0.4994766116142273,
      "step": 36,
      "train_speed(iter/s)": 0.003211
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2245.0,
      "completions/mean_length": 382.0798645019531,
      "completions/min_length": 63.0,
      "epoch": 0.5967741935483871,
      "grad_norm": 0.21874458117115267,
      "kl": 0.009130859375,
      "learning_rate": 4.6786264265797526e-07,
      "loss": 9.136443986790255e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9631944894790649,
      "reward_std": 0.22931374609470367,
      "rewards/CODEORM/mean": 0.02083333395421505,
      "rewards/CODEORM/std": 0.07628901302814484,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49791666865348816,
      "rewards/FMTORM/std": 0.03221873939037323,
      "rewards/VQAORM/mean": 0.4444444477558136,
      "rewards/VQAORM/std": 0.4970766305923462,
      "step": 37,
      "train_speed(iter/s)": 0.003219
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2041.0,
      "completions/mean_length": 408.1896057128906,
      "completions/min_length": 49.0,
      "epoch": 0.6129032258064516,
      "grad_norm": 0.23783779914096984,
      "kl": 0.0090576171875,
      "learning_rate": 4.658108216351958e-07,
      "loss": 9.058650903170928e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9589584469795227,
      "reward_std": 0.25281450152397156,
      "rewards/CODEORM/mean": 0.01833333447575569,
      "rewards/CODEORM/std": 0.07188516855239868,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4416666626930237,
      "rewards/VQAORM/std": 0.4967580735683441,
      "step": 38,
      "train_speed(iter/s)": 0.003221
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2560.0,
      "completions/mean_length": 425.4576416015625,
      "completions/min_length": 36.0,
      "epoch": 0.6290322580645161,
      "grad_norm": 0.25328193961667955,
      "kl": 0.009716796875,
      "learning_rate": 4.6370087778772037e-07,
      "loss": 9.695239714346826e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9884028434753418,
      "reward_std": 0.2516257166862488,
      "rewards/CODEORM/mean": 0.016875000670552254,
      "rewards/CODEORM/std": 0.06914517283439636,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.026324830949306488,
      "rewards/VQAORM/mean": 0.47291669249534607,
      "rewards/VQAORM/std": 0.49943938851356506,
      "step": 39,
      "train_speed(iter/s)": 0.003227
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2677.0,
      "completions/mean_length": 380.4618225097656,
      "completions/min_length": 57.0,
      "epoch": 0.6451612903225806,
      "grad_norm": 0.24632981619112712,
      "kl": 0.009765625,
      "learning_rate": 4.6153345382488437e-07,
      "loss": 9.762972331373021e-05,
      "memory(GiB)": 72.55,
      "reward": 1.0220834016799927,
      "reward_std": 0.2817615866661072,
      "rewards/CODEORM/mean": 0.017916668206453323,
      "rewards/CODEORM/std": 0.071116141974926,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.5048611164093018,
      "rewards/VQAORM/std": 0.5001500844955444,
      "step": 40,
      "train_speed(iter/s)": 0.003233
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2209.0,
      "completions/mean_length": 347.9048767089844,
      "completions/min_length": 43.0,
      "epoch": 0.6612903225806451,
      "grad_norm": 0.2448212519488402,
      "kl": 0.007769775390625,
      "learning_rate": 4.5930920996502315e-07,
      "loss": 7.772096432745457e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9471527338027954,
      "reward_std": 0.2524837851524353,
      "rewards/CODEORM/mean": 0.012083334848284721,
      "rewards/CODEORM/std": 0.05900347977876663,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280591055750847,
      "rewards/VQAORM/mean": 0.43611112236976624,
      "rewards/VQAORM/std": 0.4960736930370331,
      "step": 41,
      "train_speed(iter/s)": 0.003239
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2924.0,
      "completions/mean_length": 363.67987060546875,
      "completions/min_length": 47.0,
      "epoch": 0.6774193548387096,
      "grad_norm": 0.22163591343338251,
      "kl": 0.0093994140625,
      "learning_rate": 4.5702882373436314e-07,
      "loss": 9.394727385370061e-05,
      "memory(GiB)": 72.55,
      "reward": 0.9790278077125549,
      "reward_std": 0.2052461951971054,
      "rewards/CODEORM/mean": 0.016875000670552254,
      "rewards/CODEORM/std": 0.06914517283439636,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.01317615807056427,
      "rewards/VQAORM/mean": 0.4625000059604645,
      "rewards/VQAORM/std": 0.4987649917602539,
      "step": 42,
      "train_speed(iter/s)": 0.003243
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2287.0,
      "completions/mean_length": 374.9638977050781,
      "completions/min_length": 49.0,
      "epoch": 0.6935483870967742,
      "grad_norm": 0.2311530259898522,
      "kl": 0.01051025390625,
      "learning_rate": 4.546929897606409e-07,
      "loss": 0.00010506305989110842,
      "memory(GiB)": 72.55,
      "reward": 0.9784722328186035,
      "reward_std": 0.2506452798843384,
      "rewards/CODEORM/mean": 0.01666666753590107,
      "rewards/CODEORM/std": 0.0687422975897789,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.018627425655722618,
      "rewards/VQAORM/mean": 0.4625000059604645,
      "rewards/VQAORM/std": 0.4987649917602539,
      "step": 43,
      "train_speed(iter/s)": 0.00325
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3019.0,
      "completions/mean_length": 345.5111083984375,
      "completions/min_length": 41.0,
      "epoch": 0.7096774193548387,
      "grad_norm": 0.25687083497515134,
      "kl": 0.0111328125,
      "learning_rate": 4.5230241956151236e-07,
      "loss": 0.00011140214337501675,
      "memory(GiB)": 72.55,
      "reward": 1.019861102104187,
      "reward_std": 0.24551907181739807,
      "rewards/CODEORM/mean": 0.017083335667848587,
      "rewards/CODEORM/std": 0.06954508274793625,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.5,
      "rewards/FMTORM/std": 0.0,
      "rewards/VQAORM/mean": 0.5027778148651123,
      "rewards/VQAORM/std": 0.5001659989356995,
      "step": 44,
      "train_speed(iter/s)": 0.00325
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2333.0,
      "completions/mean_length": 367.42779541015625,
      "completions/min_length": 50.0,
      "epoch": 0.7258064516129032,
      "grad_norm": 0.237050640265248,
      "kl": 0.010888671875,
      "learning_rate": 4.4985784132781846e-07,
      "loss": 0.00010892213322222233,
      "memory(GiB)": 72.55,
      "reward": 0.9680556058883667,
      "reward_std": 0.24429023265838623,
      "rewards/CODEORM/mean": 0.01666666753590107,
      "rewards/CODEORM/std": 0.0687422975897789,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.018627425655722618,
      "rewards/VQAORM/mean": 0.4520833492279053,
      "rewards/VQAORM/std": 0.49787160754203796,
      "step": 45,
      "train_speed(iter/s)": 0.003256
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2553.0,
      "completions/mean_length": 440.5652770996094,
      "completions/min_length": 45.0,
      "epoch": 0.7419354838709677,
      "grad_norm": 0.24164999330022277,
      "kl": 0.01668701171875,
      "learning_rate": 4.4735999970177e-07,
      "loss": 0.00016686106391716748,
      "memory(GiB)": 72.55,
      "reward": 1.0237499475479126,
      "reward_std": 0.27759116888046265,
      "rewards/CODEORM/mean": 0.0133333345875144,
      "rewards/CODEORM/std": 0.061845600605010986,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.018627425655722618,
      "rewards/VQAORM/mean": 0.5111111402511597,
      "rewards/VQAORM/std": 0.5000501871109009,
      "step": 46,
      "train_speed(iter/s)": 0.003256
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2202.0,
      "completions/mean_length": 405.51458740234375,
      "completions/min_length": 43.0,
      "epoch": 0.7580645161290323,
      "grad_norm": 0.25800115560395864,
      "kl": 0.01693115234375,
      "learning_rate": 4.4480965555012375e-07,
      "loss": 0.00016923561634030193,
      "memory(GiB)": 72.55,
      "reward": 1.0121527910232544,
      "reward_std": 0.2747058570384979,
      "rewards/CODEORM/mean": 0.01770833320915699,
      "rewards/CODEORM/std": 0.07072756439447403,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.018627425655722618,
      "rewards/VQAORM/mean": 0.49513891339302063,
      "rewards/VQAORM/std": 0.5001500844955444,
      "step": 47,
      "train_speed(iter/s)": 0.003258
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 10708.0,
      "completions/mean_length": 374.5027770996094,
      "completions/min_length": 45.0,
      "epoch": 0.7741935483870968,
      "grad_norm": 0.2341781405037742,
      "kl": 0.01322021484375,
      "learning_rate": 4.422075857324137e-07,
      "loss": 0.00013221960398368537,
      "memory(GiB)": 72.55,
      "reward": 0.9981943964958191,
      "reward_std": 0.23682290315628052,
      "rewards/CODEORM/mean": 0.015208334662020206,
      "rewards/CODEORM/std": 0.06583476066589355,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.48402780294418335,
      "rewards/VQAORM/std": 0.49991846084594727,
      "step": 48,
      "train_speed(iter/s)": 0.003224
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2973.0,
      "completions/mean_length": 388.96319580078125,
      "completions/min_length": 41.0,
      "epoch": 0.7903225806451613,
      "grad_norm": 0.23494838565416404,
      "kl": 0.01656494140625,
      "learning_rate": 4.3955458286431336e-07,
      "loss": 0.00016603765834588557,
      "memory(GiB)": 72.55,
      "reward": 0.9825693964958191,
      "reward_std": 0.25734853744506836,
      "rewards/CODEORM/mean": 0.015208334662020206,
      "rewards/CODEORM/std": 0.06583476811647415,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.026324832811951637,
      "rewards/VQAORM/mean": 0.46875,
      "rewards/VQAORM/std": 0.49919581413269043,
      "step": 49,
      "train_speed(iter/s)": 0.003224
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3018.0,
      "completions/mean_length": 397.47430419921875,
      "completions/min_length": 53.0,
      "epoch": 0.8064516129032258,
      "grad_norm": 0.22328454872730055,
      "kl": 0.01668701171875,
      "learning_rate": 4.3685145507619647e-07,
      "loss": 0.0001667440083110705,
      "memory(GiB)": 72.55,
      "reward": 0.9420834183692932,
      "reward_std": 0.23709836602210999,
      "rewards/CODEORM/mean": 0.010833334177732468,
      "rewards/CODEORM/std": 0.055989425629377365,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.018627425655722618,
      "rewards/VQAORM/mean": 0.43194445967674255,
      "rewards/VQAORM/std": 0.49551889300346375,
      "step": 50,
      "train_speed(iter/s)": 0.003226
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2100.0,
      "completions/mean_length": 375.0187683105469,
      "completions/min_length": 49.0,
      "epoch": 0.8225806451612904,
      "grad_norm": 0.2476536462434363,
      "kl": 0.01484375,
      "learning_rate": 4.3409902576697315e-07,
      "loss": 0.0001483853702666238,
      "memory(GiB)": 72.55,
      "reward": 0.9879860877990723,
      "reward_std": 0.23200082778930664,
      "rewards/CODEORM/mean": 0.01229166705161333,
      "rewards/CODEORM/std": 0.05948842316865921,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.026324830949306488,
      "rewards/VQAORM/mean": 0.47708335518836975,
      "rewards/VQAORM/std": 0.4996480941772461,
      "step": 51,
      "train_speed(iter/s)": 0.003218
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2223.0,
      "completions/mean_length": 437.4423828125,
      "completions/min_length": 49.0,
      "epoch": 0.8387096774193549,
      "grad_norm": 0.2231399513016873,
      "kl": 0.021875,
      "learning_rate": 4.3129813335327437e-07,
      "loss": 0.000219002366065979,
      "memory(GiB)": 72.55,
      "reward": 1.004652738571167,
      "reward_std": 0.22962474822998047,
      "rewards/CODEORM/mean": 0.01229166705161333,
      "rewards/CODEORM/std": 0.05948842689394951,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49791666865348816,
      "rewards/FMTORM/std": 0.03221873939037323,
      "rewards/VQAORM/mean": 0.49444445967674255,
      "rewards/VQAORM/std": 0.5001428127288818,
      "step": 52,
      "train_speed(iter/s)": 0.003217
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2176.0,
      "completions/mean_length": 407.48541259765625,
      "completions/min_length": 54.0,
      "epoch": 0.8548387096774194,
      "grad_norm": 0.22842893432608774,
      "kl": 0.020361328125,
      "learning_rate": 4.2844963101406216e-07,
      "loss": 0.00020356160530354828,
      "memory(GiB)": 72.55,
      "reward": 0.9823611378669739,
      "reward_std": 0.2580673396587372,
      "rewards/CODEORM/mean": 0.009791667573153973,
      "rewards/CODEORM/std": 0.05332539603114128,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.47291669249534607,
      "rewards/VQAORM/std": 0.49943938851356506,
      "step": 53,
      "train_speed(iter/s)": 0.003195
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 9521.0,
      "completions/mean_length": 426.1222229003906,
      "completions/min_length": 61.0,
      "epoch": 0.8709677419354839,
      "grad_norm": 0.2174558142577718,
      "kl": 0.0199951171875,
      "learning_rate": 4.255543864307431e-07,
      "loss": 0.00020004776888526976,
      "memory(GiB)": 72.55,
      "reward": 0.9699999690055847,
      "reward_std": 0.24552029371261597,
      "rewards/CODEORM/mean": 0.008541667833924294,
      "rewards/CODEORM/std": 0.049912624061107635,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4625000059604645,
      "rewards/VQAORM/std": 0.4987649917602539,
      "step": 54,
      "train_speed(iter/s)": 0.003179
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2597.0,
      "completions/mean_length": 480.8597412109375,
      "completions/min_length": 63.0,
      "epoch": 0.8870967741935484,
      "grad_norm": 0.23478429409929014,
      "kl": 0.0271484375,
      "learning_rate": 4.226132815228641e-07,
      "loss": 0.0002712236891966313,
      "memory(GiB)": 72.55,
      "reward": 0.9563194513320923,
      "reward_std": 0.26775386929512024,
      "rewards/CODEORM/mean": 0.007708333898335695,
      "rewards/CODEORM/std": 0.04748312756419182,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.02632482908666134,
      "rewards/VQAORM/mean": 0.45000001788139343,
      "rewards/VQAORM/std": 0.49766653776168823,
      "step": 55,
      "train_speed(iter/s)": 0.00318
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2971.0,
      "completions/mean_length": 435.2521057128906,
      "completions/min_length": 52.0,
      "epoch": 0.9032258064516129,
      "grad_norm": 0.22207675380665862,
      "kl": 0.0236572265625,
      "learning_rate": 4.1962721217947136e-07,
      "loss": 0.00023679251899011433,
      "memory(GiB)": 72.55,
      "reward": 0.9582638740539551,
      "reward_std": 0.2368679940700531,
      "rewards/CODEORM/mean": 0.006875000428408384,
      "rewards/CODEORM/std": 0.04490695893764496,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49791666865348816,
      "rewards/FMTORM/std": 0.03221874311566353,
      "rewards/VQAORM/mean": 0.45347222685813904,
      "rewards/VQAORM/std": 0.49800342321395874,
      "step": 56,
      "train_speed(iter/s)": 0.003181
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2578.0,
      "completions/mean_length": 446.1222229003906,
      "completions/min_length": 47.0,
      "epoch": 0.9193548387096774,
      "grad_norm": 0.23718890954848998,
      "kl": 0.0261474609375,
      "learning_rate": 4.165970879862134e-07,
      "loss": 0.0002614260884001851,
      "memory(GiB)": 72.55,
      "reward": 0.9674999713897705,
      "reward_std": 0.25433528423309326,
      "rewards/CODEORM/mean": 0.006041666492819786,
      "rewards/CODEORM/std": 0.04215723276138306,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4625000059604645,
      "rewards/VQAORM/std": 0.4987649917602539,
      "step": 57,
      "train_speed(iter/s)": 0.003184
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3103.0,
      "completions/mean_length": 439.11181640625,
      "completions/min_length": 52.0,
      "epoch": 0.9354838709677419,
      "grad_norm": 0.20743082364096332,
      "kl": 0.0252685546875,
      "learning_rate": 4.1352383194827306e-07,
      "loss": 0.00025291036581620574,
      "memory(GiB)": 72.55,
      "reward": 0.9715972542762756,
      "reward_std": 0.2394816130399704,
      "rewards/CODEORM/mean": 0.007708333898335695,
      "rewards/CODEORM/std": 0.04748312756419182,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49722224473953247,
      "rewards/FMTORM/std": 0.03717704117298126,
      "rewards/VQAORM/mean": 0.46666666865348816,
      "rewards/VQAORM/std": 0.4990609884262085,
      "step": 58,
      "train_speed(iter/s)": 0.003184
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3248.0,
      "completions/mean_length": 433.08404541015625,
      "completions/min_length": 55.0,
      "epoch": 0.9516129032258065,
      "grad_norm": 0.2068345779016102,
      "kl": 0.021630859375,
      "learning_rate": 4.1040838020921087e-07,
      "loss": 0.00021632216521538794,
      "memory(GiB)": 72.55,
      "reward": 0.9704861640930176,
      "reward_std": 0.21852527558803558,
      "rewards/CODEORM/mean": 0.0062500000931322575,
      "rewards/CODEORM/std": 0.042862728238105774,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4652777910232544,
      "rewards/VQAORM/std": 0.49896618723869324,
      "step": 59,
      "train_speed(iter/s)": 0.003185
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2551.0,
      "completions/mean_length": 426.58404541015625,
      "completions/min_length": 45.0,
      "epoch": 0.967741935483871,
      "grad_norm": 0.2064093927604063,
      "kl": 0.022265625,
      "learning_rate": 4.072516817658064e-07,
      "loss": 0.00022273208014667034,
      "memory(GiB)": 72.55,
      "reward": 0.9838889241218567,
      "reward_std": 0.22381478548049927,
      "rewards/CODEORM/mean": 0.006458332762122154,
      "rewards/CODEORM/std": 0.04355579614639282,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4975694417953491,
      "rewards/FMTORM/std": 0.03478807955980301,
      "rewards/VQAORM/mean": 0.4798611104488373,
      "rewards/VQAORM/std": 0.49976783990859985,
      "step": 60,
      "train_speed(iter/s)": 0.003161
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3516.0,
      "completions/mean_length": 447.4507141113281,
      "completions/min_length": 59.0,
      "epoch": 0.9838709677419355,
      "grad_norm": 0.2458563564310446,
      "kl": 0.02421875,
      "learning_rate": 4.0405469817898533e-07,
      "loss": 0.00024220610794145614,
      "memory(GiB)": 72.55,
      "reward": 0.9372223019599915,
      "reward_std": 0.25408098101615906,
      "rewards/CODEORM/mean": 0.004583333153277636,
      "rewards/CODEORM/std": 0.03680942580103874,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.4333333373069763,
      "rewards/VQAORM/std": 0.49570778012275696,
      "step": 61,
      "train_speed(iter/s)": 0.003159
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2159.0,
      "completions/mean_length": 462.3416748046875,
      "completions/min_length": 40.0,
      "epoch": 1.0,
      "grad_norm": 0.22387971484005992,
      "kl": 0.0287353515625,
      "learning_rate": 4.00818403280918e-07,
      "loss": 0.0002869199379347265,
      "memory(GiB)": 72.55,
      "reward": 1.0059027671813965,
      "reward_std": 0.26113608479499817,
      "rewards/CODEORM/mean": 0.0052083334885537624,
      "rewards/CODEORM/std": 0.03919745236635208,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.02632482908666134,
      "rewards/VQAORM/mean": 0.5020833611488342,
      "rewards/VQAORM/std": 0.5001693964004517,
      "step": 62,
      "train_speed(iter/s)": 0.003159
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2093.0,
      "completions/mean_length": 459.2437744140625,
      "completions/min_length": 51.0,
      "epoch": 1.0161290322580645,
      "grad_norm": 0.2176432920608644,
      "kl": 0.027392578125,
      "learning_rate": 3.9754378287838106e-07,
      "loss": 0.0002737568283919245,
      "memory(GiB)": 72.55,
      "reward": 0.9327083230018616,
      "reward_std": 0.22113071382045746,
      "rewards/CODEORM/mean": 0.003541666781529784,
      "rewards/CODEORM/std": 0.03241429105401039,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.4298611283302307,
      "rewards/VQAORM/std": 0.4952280521392822,
      "step": 63,
      "train_speed(iter/s)": 0.003162
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3442.0,
      "completions/mean_length": 413.9354248046875,
      "completions/min_length": 53.0,
      "epoch": 1.032258064516129,
      "grad_norm": 0.21234850263649102,
      "kl": 0.0212158203125,
      "learning_rate": 3.942318344524711e-07,
      "loss": 0.00021199470211286098,
      "memory(GiB)": 72.55,
      "reward": 0.9656250476837158,
      "reward_std": 0.21792437136173248,
      "rewards/CODEORM/mean": 0.00520833395421505,
      "rewards/CODEORM/std": 0.03919744864106178,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.026324830949306488,
      "rewards/VQAORM/mean": 0.4618055522441864,
      "rewards/VQAORM/std": 0.4987122416496277,
      "step": 64,
      "train_speed(iter/s)": 0.003162
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3102.0,
      "completions/mean_length": 438.6458435058594,
      "completions/min_length": 54.0,
      "epoch": 1.0483870967741935,
      "grad_norm": 0.2232724953512252,
      "kl": 0.0247314453125,
      "learning_rate": 3.908835668547622e-07,
      "loss": 0.0002475287183187902,
      "memory(GiB)": 72.55,
      "reward": 0.9796527624130249,
      "reward_std": 0.23396123945713043,
      "rewards/CODEORM/mean": 0.007083333563059568,
      "rewards/CODEORM/std": 0.04556608572602272,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.47361111640930176,
      "rewards/VQAORM/std": 0.4994766116142273,
      "step": 65,
      "train_speed(iter/s)": 0.003163
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3270.0,
      "completions/mean_length": 499.7298583984375,
      "completions/min_length": 56.0,
      "epoch": 1.064516129032258,
      "grad_norm": 0.23317861529104608,
      "kl": 0.032275390625,
      "learning_rate": 3.875e-07,
      "loss": 0.0003234185860492289,
      "memory(GiB)": 72.55,
      "reward": 0.9625694751739502,
      "reward_std": 0.27475544810295105,
      "rewards/CODEORM/mean": 0.003541666781529784,
      "rewards/CODEORM/std": 0.03241428732872009,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.026324830949306488,
      "rewards/VQAORM/mean": 0.46041667461395264,
      "rewards/VQAORM/std": 0.49860385060310364,
      "step": 66,
      "train_speed(iter/s)": 0.003163
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3353.0,
      "completions/mean_length": 418.64166259765625,
      "completions/min_length": 49.0,
      "epoch": 1.0806451612903225,
      "grad_norm": 0.23413187500533847,
      "kl": 0.0256103515625,
      "learning_rate": 3.8408216455542586e-07,
      "loss": 0.00025615975027903914,
      "memory(GiB)": 72.55,
      "reward": 1.022291660308838,
      "reward_std": 0.27045774459838867,
      "rewards/CODEORM/mean": 0.004583333618938923,
      "rewards/CODEORM/std": 0.03680942952632904,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.5180555582046509,
      "rewards/VQAORM/std": 0.4998474717140198,
      "step": 67,
      "train_speed(iter/s)": 0.003162
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2800.0,
      "completions/mean_length": 415.0632019042969,
      "completions/min_length": 54.0,
      "epoch": 1.096774193548387,
      "grad_norm": 0.2280696197690621,
      "kl": 0.0226318359375,
      "learning_rate": 3.8063110162682536e-07,
      "loss": 0.00022613290639128536,
      "memory(GiB)": 72.55,
      "reward": 0.9749305248260498,
      "reward_std": 0.23834767937660217,
      "rewards/CODEORM/mean": 0.0037500003818422556,
      "rewards/CODEORM/std": 0.033342309296131134,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.01317615620791912,
      "rewards/VQAORM/mean": 0.4715277850627899,
      "rewards/VQAORM/std": 0.4993620812892914,
      "step": 68,
      "train_speed(iter/s)": 0.003163
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3211.0,
      "completions/mean_length": 476.21875,
      "completions/min_length": 40.0,
      "epoch": 1.1129032258064515,
      "grad_norm": 0.25556590033515914,
      "kl": 0.0304443359375,
      "learning_rate": 3.77147862441398e-07,
      "loss": 0.0003040857263840735,
      "memory(GiB)": 72.55,
      "reward": 1.0324305295944214,
      "reward_std": 0.2670601010322571,
      "rewards/CODEORM/mean": 0.003958333283662796,
      "rewards/CODEORM/std": 0.03424391895532608,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.5291666984558105,
      "rewards/VQAORM/std": 0.49932199716567993,
      "step": 69,
      "train_speed(iter/s)": 0.00316
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2698.0,
      "completions/mean_length": 457.7632141113281,
      "completions/min_length": 46.0,
      "epoch": 1.129032258064516,
      "grad_norm": 0.22078914084547285,
      "kl": 0.026123046875,
      "learning_rate": 3.736335080275424e-07,
      "loss": 0.0002610915689729154,
      "memory(GiB)": 72.55,
      "reward": 0.988194465637207,
      "reward_std": 0.22356846928596497,
      "rewards/CODEORM/mean": 0.0052083334885537624,
      "rewards/CODEORM/std": 0.03919745236635208,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.48402780294418335,
      "rewards/VQAORM/std": 0.4999184310436249,
      "step": 70,
      "train_speed(iter/s)": 0.003161
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3641.0,
      "completions/mean_length": 471.3034973144531,
      "completions/min_length": 54.0,
      "epoch": 1.1451612903225807,
      "grad_norm": 0.22352257730187777,
      "kl": 0.0291015625,
      "learning_rate": 3.7008910889165734e-07,
      "loss": 0.00029054959304630756,
      "memory(GiB)": 72.55,
      "reward": 0.9690278172492981,
      "reward_std": 0.23872590065002441,
      "rewards/CODEORM/mean": 0.002708333544433117,
      "rewards/CODEORM/std": 0.028385289013385773,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.46666666865348816,
      "rewards/VQAORM/std": 0.4990609884262085,
      "step": 71,
      "train_speed(iter/s)": 0.003162
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2802.0,
      "completions/mean_length": 415.3722229003906,
      "completions/min_length": 60.0,
      "epoch": 1.1612903225806452,
      "grad_norm": 0.24022424561360972,
      "kl": 0.0269287109375,
      "learning_rate": 3.6651574469205503e-07,
      "loss": 0.0002691347326617688,
      "memory(GiB)": 72.55,
      "reward": 1.0179860591888428,
      "reward_std": 0.2555491328239441,
      "rewards/CODEORM/mean": 0.004791666753590107,
      "rewards/CODEORM/std": 0.037623435258865356,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.5138888955116272,
      "rewards/VQAORM/std": 0.4999806880950928,
      "step": 72,
      "train_speed(iter/s)": 0.003161
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2134.0,
      "completions/mean_length": 447.7007141113281,
      "completions/min_length": 52.0,
      "epoch": 1.1774193548387097,
      "grad_norm": 0.20648518009940414,
      "kl": 0.03193359375,
      "learning_rate": 3.6291450391008654e-07,
      "loss": 0.0003187285328749567,
      "memory(GiB)": 72.55,
      "reward": 1.003819465637207,
      "reward_std": 0.19375617802143097,
      "rewards/CODEORM/mean": 0.0020833334419876337,
      "rewards/CODEORM/std": 0.02492169663310051,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.5020833611488342,
      "rewards/VQAORM/std": 0.5001693367958069,
      "step": 73,
      "train_speed(iter/s)": 0.003162
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2162.0,
      "completions/mean_length": 438.4805603027344,
      "completions/min_length": 49.0,
      "epoch": 1.1935483870967742,
      "grad_norm": 0.2063264046724451,
      "kl": 0.0267333984375,
      "learning_rate": 3.592864835185802e-07,
      "loss": 0.00026784028159454465,
      "memory(GiB)": 72.55,
      "reward": 0.9955556392669678,
      "reward_std": 0.22626902163028717,
      "rewards/CODEORM/mean": 0.002500000176951289,
      "rewards/CODEORM/std": 0.027281256392598152,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.5,
      "rewards/FMTORM/std": 0.0,
      "rewards/VQAORM/mean": 0.4930555820465088,
      "rewards/VQAORM/std": 0.5001254677772522,
      "step": 74,
      "train_speed(iter/s)": 0.003162
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2497.0,
      "completions/mean_length": 390.0340270996094,
      "completions/min_length": 44.0,
      "epoch": 1.2096774193548387,
      "grad_norm": 0.2377329072915874,
      "kl": 0.0220458984375,
      "learning_rate": 3.556327886476925e-07,
      "loss": 0.00022035164874978364,
      "memory(GiB)": 72.55,
      "reward": 0.9577084183692932,
      "reward_std": 0.22338639199733734,
      "rewards/CODEORM/mean": 0.002500000409781933,
      "rewards/CODEORM/std": 0.027281254529953003,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.01317615807056427,
      "rewards/VQAORM/mean": 0.4555555582046509,
      "rewards/VQAORM/std": 0.49819377064704895,
      "step": 75,
      "train_speed(iter/s)": 0.003161
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2510.0,
      "completions/mean_length": 432.8798828125,
      "completions/min_length": 59.0,
      "epoch": 1.2258064516129032,
      "grad_norm": 0.2016163479831072,
      "kl": 0.0228759765625,
      "learning_rate": 3.5195453224827547e-07,
      "loss": 0.0002286717208335176,
      "memory(GiB)": 72.55,
      "reward": 0.9972222447395325,
      "reward_std": 0.2204340249300003,
      "rewards/CODEORM/mean": 0.0031250000465661287,
      "rewards/CODEORM/std": 0.030469313263893127,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.49513891339302063,
      "rewards/VQAORM/std": 0.5001500248908997,
      "step": 76,
      "train_speed(iter/s)": 0.00316
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2529.0,
      "completions/mean_length": 454.6923828125,
      "completions/min_length": 48.0,
      "epoch": 1.2419354838709677,
      "grad_norm": 0.22659833554073525,
      "kl": 0.0289794921875,
      "learning_rate": 3.482528347528602e-07,
      "loss": 0.00029045954579487443,
      "memory(GiB)": 72.55,
      "reward": 0.9425694942474365,
      "reward_std": 0.24315796792507172,
      "rewards/CODEORM/mean": 0.0012500000884756446,
      "rewards/CODEORM/std": 0.019331244751811028,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.44236111640930176,
      "rewards/VQAORM/std": 0.4968391954898834,
      "step": 77,
      "train_speed(iter/s)": 0.003121
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2246.0,
      "completions/mean_length": 443.509033203125,
      "completions/min_length": 39.0,
      "epoch": 1.2580645161290323,
      "grad_norm": 0.23270608066363938,
      "kl": 0.026318359375,
      "learning_rate": 3.4452882373436316e-07,
      "loss": 0.0002637436846271157,
      "memory(GiB)": 72.55,
      "reward": 0.9564583897590637,
      "reward_std": 0.2442195564508438,
      "rewards/CODEORM/mean": 0.0012500000884756446,
      "rewards/CODEORM/std": 0.019331244751811028,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.01317615620791912,
      "rewards/VQAORM/mean": 0.4555555582046509,
      "rewards/VQAORM/std": 0.49819380044937134,
      "step": 78,
      "train_speed(iter/s)": 0.003123
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2990.0,
      "completions/mean_length": 433.7701416015625,
      "completions/min_length": 53.0,
      "epoch": 1.2741935483870968,
      "grad_norm": 0.23460258143973778,
      "kl": 0.0271728515625,
      "learning_rate": 3.4078363356261576e-07,
      "loss": 0.0002716649032663554,
      "memory(GiB)": 72.55,
      "reward": 0.957777738571167,
      "reward_std": 0.22971472144126892,
      "rewards/CODEORM/mean": 0.001875000074505806,
      "rewards/CODEORM/std": 0.023651065304875374,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.45694446563720703,
      "rewards/VQAORM/std": 0.49831584095954895,
      "step": 79,
      "train_speed(iter/s)": 0.003124
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 6626.0,
      "completions/mean_length": 467.6298828125,
      "completions/min_length": 60.0,
      "epoch": 1.2903225806451613,
      "grad_norm": 0.2203448340379557,
      "kl": 0.028564453125,
      "learning_rate": 3.3701840505882476e-07,
      "loss": 0.0002859021187759936,
      "memory(GiB)": 72.55,
      "reward": 0.9722917675971985,
      "reward_std": 0.23510144650936127,
      "rewards/CODEORM/mean": 0.0014583334559574723,
      "rewards/CODEORM/std": 0.020872849971055984,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.026324830949306488,
      "rewards/VQAORM/mean": 0.472222238779068,
      "rewards/VQAORM/std": 0.4994012117385864,
      "step": 80,
      "train_speed(iter/s)": 0.003118
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2564.0,
      "completions/mean_length": 418.5590515136719,
      "completions/min_length": 52.0,
      "epoch": 1.3064516129032258,
      "grad_norm": 0.2324164062331274,
      "kl": 0.026513671875,
      "learning_rate": 3.332342851480672e-07,
      "loss": 0.00026555670774541795,
      "memory(GiB)": 72.55,
      "reward": 1.0112500190734863,
      "reward_std": 0.22868500649929047,
      "rewards/CODEORM/mean": 0.001875000074505806,
      "rewards/CODEORM/std": 0.023651065304875374,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.5104166865348816,
      "rewards/VQAORM/std": 0.5000651478767395,
      "step": 81,
      "train_speed(iter/s)": 0.003119
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3178.0,
      "completions/mean_length": 458.19097900390625,
      "completions/min_length": 40.0,
      "epoch": 1.3225806451612903,
      "grad_norm": 0.23312846161895653,
      "kl": 0.0256103515625,
      "learning_rate": 3.294324265099252e-07,
      "loss": 0.0002559431886766106,
      "memory(GiB)": 72.55,
      "reward": 0.9519445300102234,
      "reward_std": 0.23361359536647797,
      "rewards/CODEORM/mean": 0.0012500000884756446,
      "rewards/CODEORM/std": 0.019331244751811028,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.4513888955116272,
      "rewards/VQAORM/std": 0.4978042244911194,
      "step": 82,
      "train_speed(iter/s)": 0.003119
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2874.0,
      "completions/mean_length": 389.6625061035156,
      "completions/min_length": 54.0,
      "epoch": 1.3387096774193548,
      "grad_norm": 0.21650063881691448,
      "kl": 0.0203857421875,
      "learning_rate": 3.256139872273696e-07,
      "loss": 0.00020405216491781175,
      "memory(GiB)": 72.55,
      "reward": 0.9847917556762695,
      "reward_std": 0.2361350655555725,
      "rewards/CODEORM/mean": 0.002500000176951289,
      "rewards/CODEORM/std": 0.027281256392598152,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.4833333492279053,
      "rewards/VQAORM/std": 0.49989575147628784,
      "step": 83,
      "train_speed(iter/s)": 0.003118
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3981.0,
      "completions/mean_length": 438.977783203125,
      "completions/min_length": 64.0,
      "epoch": 1.3548387096774195,
      "grad_norm": 0.2017820781256394,
      "kl": 0.025146484375,
      "learning_rate": 3.217801304339958e-07,
      "loss": 0.0002516305830795318,
      "memory(GiB)": 72.55,
      "reward": 0.9712501168251038,
      "reward_std": 0.23558129370212555,
      "rewards/CODEORM/mean": 0.0014583334559574723,
      "rewards/CODEORM/std": 0.020872849971055984,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.47083333134651184,
      "rewards/VQAORM/std": 0.49932199716567993,
      "step": 84,
      "train_speed(iter/s)": 0.003115
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3085.0,
      "completions/mean_length": 471.2632141113281,
      "completions/min_length": 40.0,
      "epoch": 1.370967741935484,
      "grad_norm": 0.23160273619017643,
      "kl": 0.027685546875,
      "learning_rate": 3.1793202395972255e-07,
      "loss": 0.000276345235761255,
      "memory(GiB)": 72.55,
      "reward": 0.9824305176734924,
      "reward_std": 0.2479579597711563,
      "rewards/CODEORM/mean": 0.00083333341171965,
      "rewards/CODEORM/std": 0.015794897451996803,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4982638955116272,
      "rewards/FMTORM/std": 0.02942180261015892,
      "rewards/VQAORM/mean": 0.4833333492279053,
      "rewards/VQAORM/std": 0.49989572167396545,
      "step": 85,
      "train_speed(iter/s)": 0.003114
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3080.0,
      "completions/mean_length": 438.87640380859375,
      "completions/min_length": 61.0,
      "epoch": 1.3870967741935485,
      "grad_norm": 0.194152810560773,
      "kl": 0.0279052734375,
      "learning_rate": 3.1407083997505936e-07,
      "loss": 0.0002790595463011414,
      "memory(GiB)": 72.55,
      "reward": 0.9933332800865173,
      "reward_std": 0.1885756254196167,
      "rewards/CODEORM/mean": 0.0016666668234393,
      "rewards/CODEORM/std": 0.022306226193904877,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.02632482908666134,
      "rewards/VQAORM/mean": 0.4930555820465088,
      "rewards/VQAORM/std": 0.5001254677772522,
      "step": 86,
      "train_speed(iter/s)": 0.003113
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1936.0,
      "completions/mean_length": 384.3055725097656,
      "completions/min_length": 45.0,
      "epoch": 1.403225806451613,
      "grad_norm": 0.24459951745327296,
      "kl": 0.0214599609375,
      "learning_rate": 3.1019775463405193e-07,
      "loss": 0.0002145486796507612,
      "memory(GiB)": 72.55,
      "reward": 0.9838194251060486,
      "reward_std": 0.25452497601509094,
      "rewards/CODEORM/mean": 0.001875000074505806,
      "rewards/CODEORM/std": 0.023651065304875374,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.4826388955116272,
      "rewards/VQAORM/std": 0.4998720586299896,
      "step": 87,
      "train_speed(iter/s)": 0.003116
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2332.0,
      "completions/mean_length": 427.3319396972656,
      "completions/min_length": 46.0,
      "epoch": 1.4193548387096775,
      "grad_norm": 0.22526758870168345,
      "kl": 0.026171875,
      "learning_rate": 3.063139477160147e-07,
      "loss": 0.00026203077868558466,
      "memory(GiB)": 72.55,
      "reward": 1.0167361497879028,
      "reward_std": 0.27220967411994934,
      "rewards/CODEORM/mean": 0.002500000176951289,
      "rewards/CODEORM/std": 0.027281256392598152,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.5145833492279053,
      "rewards/VQAORM/std": 0.49996086955070496,
      "step": 88,
      "train_speed(iter/s)": 0.003118
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2392.0,
      "completions/mean_length": 386.8798828125,
      "completions/min_length": 48.0,
      "epoch": 1.435483870967742,
      "grad_norm": 0.22614274989113003,
      "kl": 0.019384765625,
      "learning_rate": 3.024206022661582e-07,
      "loss": 0.0001937642809934914,
      "memory(GiB)": 72.55,
      "reward": 0.9733334183692932,
      "reward_std": 0.23752889037132263,
      "rewards/CODEORM/mean": 0.0014583334559574723,
      "rewards/CODEORM/std": 0.020872849971055984,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.47291669249534607,
      "rewards/VQAORM/std": 0.49943938851356506,
      "step": 89,
      "train_speed(iter/s)": 0.003118
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2021.0,
      "completions/mean_length": 363.97222900390625,
      "completions/min_length": 41.0,
      "epoch": 1.4516129032258065,
      "grad_norm": 0.2398906373001323,
      "kl": 0.0197509765625,
      "learning_rate": 2.98518904235222e-07,
      "loss": 0.00019763951422646642,
      "memory(GiB)": 72.55,
      "reward": 1.0004167556762695,
      "reward_std": 0.23650936782360077,
      "rewards/CODEORM/mean": 0.0014583334559574723,
      "rewards/CODEORM/std": 0.020872849971055984,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.4993055760860443,
      "rewards/VQAORM/std": 0.5001731514930725,
      "step": 90,
      "train_speed(iter/s)": 0.00312
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2460.0,
      "completions/mean_length": 369.35418701171875,
      "completions/min_length": 44.0,
      "epoch": 1.467741935483871,
      "grad_norm": 0.23682693521156525,
      "kl": 0.020166015625,
      "learning_rate": 2.946100421182231e-07,
      "loss": 0.00020152229990344495,
      "memory(GiB)": 72.55,
      "reward": 1.0318750143051147,
      "reward_std": 0.24585361778736115,
      "rewards/CODEORM/mean": 0.0016666668234393,
      "rewards/CODEORM/std": 0.022306226193904877,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.5305555462837219,
      "rewards/VQAORM/std": 0.49923887848854065,
      "step": 91,
      "train_speed(iter/s)": 0.003123
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3620.0,
      "completions/mean_length": 414.87432861328125,
      "completions/min_length": 58.0,
      "epoch": 1.4838709677419355,
      "grad_norm": 0.23857713339439449,
      "kl": 0.0223388671875,
      "learning_rate": 2.906952065924282e-07,
      "loss": 0.00022341267322190106,
      "memory(GiB)": 72.55,
      "reward": 0.9906944632530212,
      "reward_std": 0.2329772561788559,
      "rewards/CODEORM/mean": 0.0014583334559574723,
      "rewards/CODEORM/std": 0.020872849971055984,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.4895833432674408,
      "rewards/VQAORM/std": 0.5000651478767395,
      "step": 92,
      "train_speed(iter/s)": 0.003121
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2573.0,
      "completions/mean_length": 390.6229248046875,
      "completions/min_length": 56.0,
      "epoch": 1.5,
      "grad_norm": 0.2533005916156193,
      "kl": 0.0206298828125,
      "learning_rate": 2.867755901546624e-07,
      "loss": 0.00020597083494067192,
      "memory(GiB)": 72.55,
      "reward": 1.0013195276260376,
      "reward_std": 0.2565329372882843,
      "rewards/CODEORM/mean": 0.0006250000442378223,
      "rewards/CODEORM/std": 0.013683545403182507,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.02632482908666134,
      "rewards/VQAORM/mean": 0.5020833611488342,
      "rewards/VQAORM/std": 0.5001693964004517,
      "step": 93,
      "train_speed(iter/s)": 0.003122
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3188.0,
      "completions/mean_length": 376.4840393066406,
      "completions/min_length": 55.0,
      "epoch": 1.5161290322580645,
      "grad_norm": 0.20748704361554995,
      "kl": 0.018408203125,
      "learning_rate": 2.8285238675806277e-07,
      "loss": 0.00018427410395815969,
      "memory(GiB)": 72.55,
      "reward": 0.932847261428833,
      "reward_std": 0.21737442910671234,
      "rewards/CODEORM/mean": 0.0012500000884756446,
      "rewards/CODEORM/std": 0.01933124288916588,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.43263891339302063,
      "rewards/VQAORM/std": 0.49561378359794617,
      "step": 94,
      "train_speed(iter/s)": 0.003121
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3588.0,
      "completions/mean_length": 389.97015380859375,
      "completions/min_length": 50.0,
      "epoch": 1.532258064516129,
      "grad_norm": 0.21959352013259997,
      "kl": 0.0175537109375,
      "learning_rate": 2.7892679144838883e-07,
      "loss": 0.0001756453566486016,
      "memory(GiB)": 72.55,
      "reward": 0.9204861521720886,
      "reward_std": 0.2225702553987503,
      "rewards/CODEORM/mean": 0.0010416667209938169,
      "rewards/CODEORM/std": 0.0176530834287405,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49861112236976624,
      "rewards/FMTORM/std": 0.026324830949306488,
      "rewards/VQAORM/mean": 0.4208333492279053,
      "rewards/VQAORM/std": 0.49386435747146606,
      "step": 95,
      "train_speed(iter/s)": 0.003119
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2273.0,
      "completions/mean_length": 411.7618103027344,
      "completions/min_length": 48.0,
      "epoch": 1.5483870967741935,
      "grad_norm": 0.22487601694634113,
      "kl": 0.0216552734375,
      "learning_rate": 2.75e-07,
      "loss": 0.00021632894640788436,
      "memory(GiB)": 72.55,
      "reward": 1.0133333206176758,
      "reward_std": 0.2392030954360962,
      "rewards/CODEORM/mean": 0.001875000074505806,
      "rewards/CODEORM/std": 0.023651065304875374,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4989583492279053,
      "rewards/FMTORM/std": 0.02280590869486332,
      "rewards/VQAORM/mean": 0.512499988079071,
      "rewards/VQAORM/std": 0.5000174045562744,
      "step": 96,
      "train_speed(iter/s)": 0.003118
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2213.0,
      "completions/mean_length": 377.1131896972656,
      "completions/min_length": 59.0,
      "epoch": 1.564516129032258,
      "grad_norm": 0.22674694657896516,
      "kl": 0.018505859375,
      "learning_rate": 2.710732085516112e-07,
      "loss": 0.0001852215063991025,
      "memory(GiB)": 72.55,
      "reward": 0.9863194823265076,
      "reward_std": 0.23460441827774048,
      "rewards/CODEORM/mean": 0.0012500000884756446,
      "rewards/CODEORM/std": 0.019331244751811028,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.4854166805744171,
      "rewards/VQAORM/std": 0.49996092915534973,
      "step": 97,
      "train_speed(iter/s)": 0.003119
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2510.0,
      "completions/mean_length": 392.0382080078125,
      "completions/min_length": 40.0,
      "epoch": 1.5806451612903225,
      "grad_norm": 0.21844911939638736,
      "kl": 0.02021484375,
      "learning_rate": 2.6714761324193736e-07,
      "loss": 0.00020227438653819263,
      "memory(GiB)": 72.55,
      "reward": 0.962569534778595,
      "reward_std": 0.22012458741664886,
      "rewards/CODEORM/mean": 0.0014583334559574723,
      "rewards/CODEORM/std": 0.020872848108410835,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.4993055760860443,
      "rewards/FMTORM/std": 0.01862742379307747,
      "rewards/VQAORM/mean": 0.4618055522441864,
      "rewards/VQAORM/std": 0.4987122416496277,
      "step": 98,
      "train_speed(iter/s)": 0.00312
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2929.0,
      "completions/mean_length": 382.51458740234375,
      "completions/min_length": 46.0,
      "epoch": 1.596774193548387,
      "grad_norm": 0.21209971181854506,
      "kl": 0.018798828125,
      "learning_rate": 2.632244098453376e-07,
      "loss": 0.0001878054899862036,
      "memory(GiB)": 72.55,
      "reward": 0.9566666483879089,
      "reward_std": 0.22168388962745667,
      "rewards/CODEORM/mean": 0.0014583333395421505,
      "rewards/CODEORM/std": 0.020872848108410835,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49965280294418335,
      "rewards/FMTORM/std": 0.013176157139241695,
      "rewards/VQAORM/mean": 0.4555555582046509,
      "rewards/VQAORM/std": 0.49819380044937134,
      "step": 99,
      "train_speed(iter/s)": 0.003121
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3312.0,
      "completions/mean_length": 380.2840270996094,
      "completions/min_length": 49.0,
      "epoch": 1.6129032258064515,
      "grad_norm": 0.21940988188704383,
      "kl": 0.01865234375,
      "learning_rate": 2.593047934075718e-07,
      "loss": 0.0001866342208813876,
      "memory(GiB)": 72.55,
      "reward": 0.9885416626930237,
      "reward_std": 0.22178694605827332,
      "rewards/CODEORM/mean": 0.0010416667209938169,
      "rewards/CODEORM/std": 0.0176530834287405,
      "rewards/CSTORM/mean": 0.0,
      "rewards/CSTORM/std": 0.0,
      "rewards/FMTORM/mean": 0.49791666865348816,
      "rewards/FMTORM/std": 0.03221873939037323,
      "rewards/VQAORM/mean": 0.4895833432674408,
      "rewards/VQAORM/std": 0.5000651478767395,
      "step": 100,
      "train_speed(iter/s)": 0.003122
    }
  ],
  "logging_steps": 1,
  "max_steps": 186,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}