{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.26666666666666666,
  "eval_steps": 500,
  "global_step": 200,
  "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,
      "completion_length": 1369.5,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 1369.5,
      "completions/max_terminated_length": 967.4,
      "completions/mean_length": 349.9875,
      "completions/mean_terminated_length": 310.6795928955078,
      "completions/min_length": 19.7,
      "completions/min_terminated_length": 19.7,
      "epoch": 0.013333333333333334,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 175.513671875,
      "kl": 0.406684020392967,
      "learning_rate": 9.000000000000001e-07,
      "loss": 0.0004,
      "num_tokens": 73366.0,
      "reward": -1.9828404784202576,
      "reward_std": 0.43738200813531875,
      "rewards/reward_func/mean": -1.9828404903411865,
      "rewards/reward_func/std": 0.5390992075204849,
      "step": 10
    },
    {
      "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,
      "completion_length": 1114.4,
      "completions/clipped_ratio": 0.025,
      "completions/max_length": 1114.4,
      "completions/max_terminated_length": 878.8,
      "completions/mean_length": 285.65,
      "completions/mean_terminated_length": 253.98928833007812,
      "completions/min_length": 21.2,
      "completions/min_terminated_length": 21.2,
      "epoch": 0.02666666666666667,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 84.8028335571289,
      "kl": 0.06926110610365868,
      "learning_rate": 1.9000000000000002e-06,
      "loss": 0.0001,
      "num_tokens": 135618.0,
      "reward": -2.0299901008605956,
      "reward_std": 0.34289126843214035,
      "rewards/reward_func/mean": -2.0299901247024534,
      "rewards/reward_func/std": 0.36671974062919616,
      "step": 20
    },
    {
      "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,
      "completion_length": 1320.9,
      "completions/clipped_ratio": 0.0375,
      "completions/max_length": 1320.9,
      "completions/max_terminated_length": 1137.8,
      "completions/mean_length": 378.5125,
      "completions/mean_terminated_length": 332.92625885009767,
      "completions/min_length": 32.1,
      "completions/min_terminated_length": 32.1,
      "epoch": 0.04,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 171.62295532226562,
      "kl": 0.6327452264726162,
      "learning_rate": 2.9e-06,
      "loss": 0.0006,
      "num_tokens": 212112.0,
      "reward": -1.8967698693275452,
      "reward_std": 0.5322311833500862,
      "rewards/reward_func/mean": -1.8967698216438293,
      "rewards/reward_func/std": 0.6287578240036964,
      "step": 30
    },
    {
      "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,
      "completion_length": 1158.5,
      "completions/clipped_ratio": 0.04375,
      "completions/max_length": 1158.5,
      "completions/max_terminated_length": 814.9,
      "completions/mean_length": 302.51875,
      "completions/mean_terminated_length": 245.66839904785155,
      "completions/min_length": 21.4,
      "completions/min_terminated_length": 21.4,
      "epoch": 0.05333333333333334,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 155.87843322753906,
      "kl": 20.71284626722336,
      "learning_rate": 3.900000000000001e-06,
      "loss": 0.0207,
      "num_tokens": 278219.0,
      "reward": -1.8949542403221131,
      "reward_std": 0.5414254635572433,
      "rewards/reward_func/mean": -1.8949542760848999,
      "rewards/reward_func/std": 0.6462407022714615,
      "step": 40
    },
    {
      "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,
      "completion_length": 1438.0,
      "completions/clipped_ratio": 0.05625,
      "completions/max_length": 1438.0,
      "completions/max_terminated_length": 957.8,
      "completions/mean_length": 368.88125,
      "completions/mean_terminated_length": 298.481379699707,
      "completions/min_length": 25.8,
      "completions/min_terminated_length": 25.8,
      "epoch": 0.06666666666666667,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 246.18365478515625,
      "kl": 81.723886013031,
      "learning_rate": 4.9000000000000005e-06,
      "loss": 0.0817,
      "num_tokens": 354520.0,
      "reward": -1.9963937759399415,
      "reward_std": 0.4492200046777725,
      "rewards/reward_func/mean": -1.9963937878608704,
      "rewards/reward_func/std": 0.5674424737691879,
      "step": 50
    },
    {
      "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,
      "completion_length": 1460.3,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 1460.3,
      "completions/max_terminated_length": 968.9,
      "completions/mean_length": 354.225,
      "completions/mean_terminated_length": 275.43447265625,
      "completions/min_length": 24.4,
      "completions/min_terminated_length": 24.4,
      "epoch": 0.08,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 101.76862335205078,
      "kl": 7.8479931354522705,
      "learning_rate": 4.9000000000000005e-06,
      "loss": 0.0078,
      "num_tokens": 428700.0,
      "reward": -2.0827497005462647,
      "reward_std": 0.33130341917276385,
      "rewards/reward_func/mean": -2.0827497005462647,
      "rewards/reward_func/std": 0.4056150123476982,
      "step": 60
    },
    {
      "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,
      "completion_length": 1161.2,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 1161.2,
      "completions/max_terminated_length": 900.7,
      "completions/mean_length": 321.36875,
      "completions/mean_terminated_length": 282.4824447631836,
      "completions/min_length": 11.5,
      "completions/min_terminated_length": 11.5,
      "epoch": 0.09333333333333334,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 140.64456176757812,
      "kl": 15.888973093032837,
      "learning_rate": 4.7888888888888894e-06,
      "loss": 0.0159,
      "num_tokens": 497455.0,
      "reward": -1.948601198196411,
      "reward_std": 0.4984253913164139,
      "rewards/reward_func/mean": -1.9486011862754822,
      "rewards/reward_func/std": 0.587457275390625,
      "step": 70
    },
    {
      "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,
      "completion_length": 1319.1,
      "completions/clipped_ratio": 0.075,
      "completions/max_length": 1319.1,
      "completions/max_terminated_length": 917.3,
      "completions/mean_length": 363.85,
      "completions/mean_terminated_length": 271.0082809448242,
      "completions/min_length": 31.9,
      "completions/min_terminated_length": 31.9,
      "epoch": 0.10666666666666667,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 151.1342010498047,
      "kl": 612.0117300510407,
      "learning_rate": 4.677777777777778e-06,
      "loss": 0.612,
      "num_tokens": 572911.0,
      "reward": -2.036101925373077,
      "reward_std": 0.3775564029812813,
      "rewards/reward_func/mean": -2.0361019134521485,
      "rewards/reward_func/std": 0.44103380143642423,
      "step": 80
    },
    {
      "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,
      "completion_length": 1144.8,
      "completions/clipped_ratio": 0.025,
      "completions/max_length": 1144.8,
      "completions/max_terminated_length": 842.0,
      "completions/mean_length": 297.375,
      "completions/mean_terminated_length": 264.98167114257814,
      "completions/min_length": 13.7,
      "completions/min_terminated_length": 13.7,
      "epoch": 0.12,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 76.42510223388672,
      "kl": 29.60899884700775,
      "learning_rate": 4.566666666666667e-06,
      "loss": 0.0296,
      "num_tokens": 637651.0,
      "reward": -2.0117875218391417,
      "reward_std": 0.48156112879514695,
      "rewards/reward_func/mean": -2.011787533760071,
      "rewards/reward_func/std": 0.6340822532773018,
      "step": 90
    },
    {
      "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,
      "completion_length": 1019.2,
      "completions/clipped_ratio": 0.0375,
      "completions/max_length": 1019.2,
      "completions/max_terminated_length": 805.0,
      "completions/mean_length": 292.1625,
      "completions/mean_terminated_length": 243.4895050048828,
      "completions/min_length": 21.2,
      "completions/min_terminated_length": 21.2,
      "epoch": 0.13333333333333333,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 52.52836990356445,
      "kl": 5.190238785743714,
      "learning_rate": 4.455555555555555e-06,
      "loss": 0.0052,
      "num_tokens": 701457.0,
      "reward": -2.0103227376937864,
      "reward_std": 0.3981830716133118,
      "rewards/reward_func/mean": -2.010322690010071,
      "rewards/reward_func/std": 0.46229155659675597,
      "step": 100
    },
    {
      "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,
      "completion_length": 1195.6,
      "completions/clipped_ratio": 0.04375,
      "completions/max_length": 1195.6,
      "completions/max_terminated_length": 707.5,
      "completions/mean_length": 273.6875,
      "completions/mean_terminated_length": 217.00977020263673,
      "completions/min_length": 18.2,
      "completions/min_terminated_length": 18.2,
      "epoch": 0.14666666666666667,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 111.76286315917969,
      "kl": 11.202206754684449,
      "learning_rate": 4.344444444444445e-06,
      "loss": 0.0112,
      "num_tokens": 762523.0,
      "reward": -1.9468143582344055,
      "reward_std": 0.541467210650444,
      "rewards/reward_func/mean": -1.9468143582344055,
      "rewards/reward_func/std": 0.7460459738969802,
      "step": 110
    },
    {
      "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,
      "completion_length": 964.2,
      "completions/clipped_ratio": 0.01875,
      "completions/max_length": 964.2,
      "completions/max_terminated_length": 663.5,
      "completions/mean_length": 232.1,
      "completions/mean_terminated_length": 206.75458526611328,
      "completions/min_length": 26.5,
      "completions/min_terminated_length": 26.5,
      "epoch": 0.16,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 85.74720764160156,
      "kl": 9.81872730255127,
      "learning_rate": 4.233333333333334e-06,
      "loss": 0.0098,
      "num_tokens": 816111.0,
      "reward": -1.8584201455116272,
      "reward_std": 0.6271616354584694,
      "rewards/reward_func/mean": -1.8584201216697693,
      "rewards/reward_func/std": 0.8233610987663269,
      "step": 120
    },
    {
      "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,
      "completion_length": 1176.6,
      "completions/clipped_ratio": 0.05,
      "completions/max_length": 1176.6,
      "completions/max_terminated_length": 865.2,
      "completions/mean_length": 337.5375,
      "completions/mean_terminated_length": 275.18006286621096,
      "completions/min_length": 22.3,
      "completions/min_terminated_length": 22.3,
      "epoch": 0.17333333333333334,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 26.790021896362305,
      "kl": 8.61414155960083,
      "learning_rate": 4.122222222222222e-06,
      "loss": 0.0086,
      "num_tokens": 887825.0,
      "reward": -2.0355328917503357,
      "reward_std": 0.3970488652586937,
      "rewards/reward_func/mean": -2.0355329513549805,
      "rewards/reward_func/std": 0.5011579066514968,
      "step": 130
    },
    {
      "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,
      "completion_length": 1129.0,
      "completions/clipped_ratio": 0.01875,
      "completions/max_length": 1129.0,
      "completions/max_terminated_length": 1012.8,
      "completions/mean_length": 338.84375,
      "completions/mean_terminated_length": 316.7058380126953,
      "completions/min_length": 29.9,
      "completions/min_terminated_length": 29.9,
      "epoch": 0.18666666666666668,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 42.169864654541016,
      "kl": 3.7959092617034913,
      "learning_rate": 4.011111111111111e-06,
      "loss": 0.0038,
      "num_tokens": 960112.0,
      "reward": -1.9997008681297301,
      "reward_std": 0.42410101294517516,
      "rewards/reward_func/mean": -1.9997008919715882,
      "rewards/reward_func/std": 0.5003743663430213,
      "step": 140
    },
    {
      "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,
      "completion_length": 1149.0,
      "completions/clipped_ratio": 0.0375,
      "completions/max_length": 1149.0,
      "completions/max_terminated_length": 891.2,
      "completions/mean_length": 321.8125,
      "completions/mean_terminated_length": 275.2519104003906,
      "completions/min_length": 28.0,
      "completions/min_terminated_length": 28.0,
      "epoch": 0.2,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 68.53196716308594,
      "kl": 5.498324632644653,
      "learning_rate": 3.900000000000001e-06,
      "loss": 0.0055,
      "num_tokens": 1028238.0,
      "reward": -1.9284112930297852,
      "reward_std": 0.5184856146574021,
      "rewards/reward_func/mean": -1.928411316871643,
      "rewards/reward_func/std": 0.6621677219867707,
      "step": 150
    },
    {
      "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,
      "completion_length": 1062.3,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 1062.3,
      "completions/max_terminated_length": 784.8,
      "completions/mean_length": 272.725,
      "completions/mean_terminated_length": 233.19851989746093,
      "completions/min_length": 23.0,
      "completions/min_terminated_length": 23.0,
      "epoch": 0.21333333333333335,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 60.397335052490234,
      "kl": 6.759932923316955,
      "learning_rate": 3.7888888888888893e-06,
      "loss": 0.0068,
      "num_tokens": 1088610.0,
      "reward": -1.950536346435547,
      "reward_std": 0.536798107624054,
      "rewards/reward_func/mean": -1.9505364656448365,
      "rewards/reward_func/std": 0.6081681102514267,
      "step": 160
    },
    {
      "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,
      "completion_length": 1134.6,
      "completions/clipped_ratio": 0.025,
      "completions/max_length": 1134.6,
      "completions/max_terminated_length": 823.1,
      "completions/mean_length": 282.64375,
      "completions/mean_terminated_length": 250.63875427246094,
      "completions/min_length": 20.3,
      "completions/min_terminated_length": 20.3,
      "epoch": 0.22666666666666666,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 123.81739044189453,
      "kl": 27.248933839797974,
      "learning_rate": 3.6777777777777778e-06,
      "loss": 0.0272,
      "num_tokens": 1150729.0,
      "reward": -1.9780633449554443,
      "reward_std": 0.4464616924524307,
      "rewards/reward_func/mean": -1.9780633687973022,
      "rewards/reward_func/std": 0.5329610571265221,
      "step": 170
    },
    {
      "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,
      "completion_length": 1139.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 1139.0,
      "completions/max_terminated_length": 907.5,
      "completions/mean_length": 327.10625,
      "completions/mean_terminated_length": 287.2654800415039,
      "completions/min_length": 21.8,
      "completions/min_terminated_length": 21.8,
      "epoch": 0.24,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 93.7626953125,
      "kl": 31.076939582824707,
      "learning_rate": 3.566666666666667e-06,
      "loss": 0.0311,
      "num_tokens": 1220558.0,
      "reward": -1.978137457370758,
      "reward_std": 0.43074723780155183,
      "rewards/reward_func/mean": -1.978137445449829,
      "rewards/reward_func/std": 0.5373590663075447,
      "step": 180
    },
    {
      "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,
      "completion_length": 1082.6,
      "completions/clipped_ratio": 0.0375,
      "completions/max_length": 1082.6,
      "completions/max_terminated_length": 809.4,
      "completions/mean_length": 307.125,
      "completions/mean_terminated_length": 260.3149475097656,
      "completions/min_length": 23.9,
      "completions/min_terminated_length": 23.9,
      "epoch": 0.25333333333333335,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 121.53104400634766,
      "kl": 30.77431230545044,
      "learning_rate": 3.455555555555556e-06,
      "loss": 0.0308,
      "num_tokens": 1286870.0,
      "reward": -1.9268421530723572,
      "reward_std": 0.4559009030461311,
      "rewards/reward_func/mean": -1.926842176914215,
      "rewards/reward_func/std": 0.5596317678689957,
      "step": 190
    },
    {
      "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,
      "completion_length": 1152.0,
      "completions/clipped_ratio": 0.0125,
      "completions/max_length": 1152.0,
      "completions/max_terminated_length": 1065.2,
      "completions/mean_length": 340.56875,
      "completions/mean_terminated_length": 326.0941680908203,
      "completions/min_length": 21.5,
      "completions/min_terminated_length": 21.5,
      "epoch": 0.26666666666666666,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 68.84746551513672,
      "kl": 13.801477253437042,
      "learning_rate": 3.3444444444444445e-06,
      "loss": 0.0138,
      "num_tokens": 1358797.0,
      "reward": -1.9825191140174865,
      "reward_std": 0.45144801288843156,
      "rewards/reward_func/mean": -1.9825191378593445,
      "rewards/reward_func/std": 0.5391437739133835,
      "step": 200
    }
  ],
  "logging_steps": 10,
  "max_steps": 500,
  "num_input_tokens_seen": 1358797,
  "num_train_epochs": 1,
  "save_steps": 100,
  "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": 16,
  "trial_name": null,
  "trial_params": null
}