Files
supergames-grpo/checkpoint-80/trainer_state.json
ModelHub XC a53def3b08 初始化项目,由ModelHub XC社区提供模型
Model: pvs333/supergames-grpo
Source: Original Platform
2026-06-16 08:06:17 +08:00

2355 lines
86 KiB
JSON

{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.4,
"eval_steps": 500,
"global_step": 80,
"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": 478.0,
"completions/max_terminated_length": 478.0,
"completions/mean_length": 191.0,
"completions/mean_terminated_length": 191.0,
"completions/min_length": 66.0,
"completions/min_terminated_length": 66.0,
"entropy": 0.5566893219947815,
"epoch": 0.005,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.370361328125,
"learning_rate": 1e-06,
"loss": -0.24658073484897614,
"num_tokens": 10832.0,
"reward": -3.3000030517578125,
"reward_std": 85.62333679199219,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -3.4000015258789062,
"rewards/supergames_reward/std": 85.62333679199219,
"step": 1,
"step_time": 22.779711072013015
},
{
"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": 216.0,
"completions/max_terminated_length": 216.0,
"completions/mean_length": 131.5,
"completions/mean_terminated_length": 131.5,
"completions/min_length": 60.0,
"completions/min_terminated_length": 60.0,
"entropy": 0.6395841240882874,
"epoch": 0.01,
"frac_reward_zero_std": 0.0,
"grad_norm": 8.09774398803711,
"learning_rate": 9.95e-07,
"loss": 0.20567649602890015,
"num_tokens": 16404.0,
"reward": -12.422499656677246,
"reward_std": 7.134707450866699,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -12.522500038146973,
"rewards/supergames_reward/std": 7.134707927703857,
"step": 2,
"step_time": 9.960156448010821
},
{
"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": 287.0,
"completions/max_terminated_length": 287.0,
"completions/mean_length": 156.625,
"completions/mean_terminated_length": 156.625,
"completions/min_length": 77.0,
"completions/min_terminated_length": 77.0,
"entropy": 0.562222421169281,
"epoch": 0.015,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.9053616523742676,
"learning_rate": 9.9e-07,
"loss": 0.09602774679660797,
"num_tokens": 26953.0,
"reward": -74.94999694824219,
"reward_std": 70.73094177246094,
"rewards/json_format_reward/mean": 0.05000000074505806,
"rewards/json_format_reward/std": 0.09258200973272324,
"rewards/supergames_reward/mean": -75.0,
"rewards/supergames_reward/std": 70.71067810058594,
"step": 3,
"step_time": 13.686320498993155
},
{
"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": 268.0,
"completions/max_terminated_length": 268.0,
"completions/mean_length": 154.125,
"completions/mean_terminated_length": 154.125,
"completions/min_length": 53.0,
"completions/min_terminated_length": 53.0,
"entropy": 0.7323317527770996,
"epoch": 0.02,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.933310031890869,
"learning_rate": 9.849999999999999e-07,
"loss": 0.28249427676200867,
"num_tokens": 36514.0,
"reward": -40.7599983215332,
"reward_std": 81.73140716552734,
"rewards/json_format_reward/mean": 0.05000000074505806,
"rewards/json_format_reward/std": 0.09258200973272324,
"rewards/supergames_reward/mean": -40.80999755859375,
"rewards/supergames_reward/std": 81.68995666503906,
"step": 4,
"step_time": 12.805880262021674
},
{
"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": 401.0,
"completions/max_terminated_length": 401.0,
"completions/mean_length": 205.25,
"completions/mean_terminated_length": 205.25,
"completions/min_length": 115.0,
"completions/min_terminated_length": 115.0,
"entropy": 0.6167430281639099,
"epoch": 0.025,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.344135046005249,
"learning_rate": 9.8e-07,
"loss": 0.1433994024991989,
"num_tokens": 47476.0,
"reward": -97.6612548828125,
"reward_std": 6.3321428298950195,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -97.76124572753906,
"rewards/supergames_reward/std": 6.3321428298950195,
"step": 5,
"step_time": 18.789364666008623
},
{
"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": 276.0,
"completions/max_terminated_length": 276.0,
"completions/mean_length": 167.25,
"completions/mean_terminated_length": 167.25,
"completions/min_length": 73.0,
"completions/min_terminated_length": 73.0,
"entropy": 0.7583790421485901,
"epoch": 0.03,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.172371864318848,
"learning_rate": 9.75e-07,
"loss": -0.12127404659986496,
"num_tokens": 55246.0,
"reward": -58.687503814697266,
"reward_std": 58.64059829711914,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -58.76250457763672,
"rewards/supergames_reward/std": 58.620460510253906,
"step": 6,
"step_time": 12.948570273991209
},
{
"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": 261.0,
"completions/max_terminated_length": 261.0,
"completions/mean_length": 133.875,
"completions/mean_terminated_length": 133.875,
"completions/min_length": 57.0,
"completions/min_terminated_length": 57.0,
"entropy": 0.7447654604911804,
"epoch": 0.035,
"frac_reward_zero_std": 0.0,
"grad_norm": 6.428595066070557,
"learning_rate": 9.7e-07,
"loss": 0.10135584324598312,
"num_tokens": 60837.0,
"reward": -32.45000076293945,
"reward_std": 41.754486083984375,
"rewards/json_format_reward/mean": 0.05000000074505806,
"rewards/json_format_reward/std": 0.09258200973272324,
"rewards/supergames_reward/mean": -32.5,
"rewards/supergames_reward/std": 41.66190719604492,
"step": 7,
"step_time": 12.058315072004916
},
{
"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": 414.0,
"completions/max_terminated_length": 414.0,
"completions/mean_length": 209.125,
"completions/mean_terminated_length": 209.125,
"completions/min_length": 104.0,
"completions/min_terminated_length": 104.0,
"entropy": 0.3860666751861572,
"epoch": 0.04,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.806255578994751,
"learning_rate": 9.649999999999999e-07,
"loss": -0.33238139748573303,
"num_tokens": 70838.0,
"reward": -80.50375366210938,
"reward_std": 38.82380294799805,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -80.60375213623047,
"rewards/supergames_reward/std": 38.82379913330078,
"step": 8,
"step_time": 19.323370319994865
},
{
"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": 0.25,
"completions/max_length": 512.0,
"completions/max_terminated_length": 199.0,
"completions/mean_length": 247.125,
"completions/mean_terminated_length": 158.83334350585938,
"completions/min_length": 106.0,
"completions/min_terminated_length": 106.0,
"entropy": 0.5816237330436707,
"epoch": 0.045,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.3209943771362305,
"learning_rate": 9.6e-07,
"loss": 0.36752766370773315,
"num_tokens": 82135.0,
"reward": -27.688751220703125,
"reward_std": 82.8358154296875,
"rewards/json_format_reward/mean": 0.05000000074505806,
"rewards/json_format_reward/std": 0.09258200973272324,
"rewards/supergames_reward/mean": -27.738750457763672,
"rewards/supergames_reward/std": 82.78590393066406,
"step": 9,
"step_time": 21.948575104994234
},
{
"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": 240.0,
"completions/max_terminated_length": 240.0,
"completions/mean_length": 153.0,
"completions/mean_terminated_length": 153.0,
"completions/min_length": 51.0,
"completions/min_terminated_length": 51.0,
"entropy": 0.6301521062850952,
"epoch": 0.05,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.125359058380127,
"learning_rate": 9.55e-07,
"loss": 0.06129350885748863,
"num_tokens": 89807.0,
"reward": -78.39624786376953,
"reward_std": 40.17042922973633,
"rewards/json_format_reward/mean": 0.05000000074505806,
"rewards/json_format_reward/std": 0.09258200973272324,
"rewards/supergames_reward/mean": -78.44625091552734,
"rewards/supergames_reward/std": 40.1396484375,
"step": 10,
"step_time": 9.098202993016457
},
{
"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": 133.0,
"completions/max_terminated_length": 133.0,
"completions/mean_length": 112.0,
"completions/mean_terminated_length": 112.0,
"completions/min_length": 68.0,
"completions/min_terminated_length": 68.0,
"entropy": 0.5521990656852722,
"epoch": 0.055,
"frac_reward_zero_std": 0.0,
"grad_norm": 6.616060256958008,
"learning_rate": 9.499999999999999e-07,
"loss": -0.006313305348157883,
"num_tokens": 95231.0,
"reward": -12.161249160766602,
"reward_std": 6.39578104019165,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -12.261249542236328,
"rewards/supergames_reward/std": 6.39578104019165,
"step": 11,
"step_time": 5.136311663984088
},
{
"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": 396.0,
"completions/max_terminated_length": 396.0,
"completions/mean_length": 214.875,
"completions/mean_terminated_length": 214.875,
"completions/min_length": 55.0,
"completions/min_terminated_length": 55.0,
"entropy": 0.5505905151367188,
"epoch": 0.06,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.289583206176758,
"learning_rate": 9.45e-07,
"loss": -0.29802003502845764,
"num_tokens": 106238.0,
"reward": -91.66999816894531,
"reward_std": 23.2779541015625,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -91.77000427246094,
"rewards/supergames_reward/std": 23.277956008911133,
"step": 12,
"step_time": 14.945365622988902
},
{
"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": 269.0,
"completions/max_terminated_length": 269.0,
"completions/mean_length": 167.375,
"completions/mean_terminated_length": 167.375,
"completions/min_length": 113.0,
"completions/min_terminated_length": 113.0,
"entropy": 0.540415346622467,
"epoch": 0.065,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.858291149139404,
"learning_rate": 9.399999999999999e-07,
"loss": 0.24004128575325012,
"num_tokens": 116937.0,
"reward": -20.38249969482422,
"reward_std": 87.55204010009766,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -20.457500457763672,
"rewards/supergames_reward/std": 87.52605438232422,
"step": 13,
"step_time": 10.482285185018554
},
{
"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": 254.0,
"completions/max_terminated_length": 254.0,
"completions/mean_length": 156.875,
"completions/mean_terminated_length": 156.875,
"completions/min_length": 114.0,
"completions/min_terminated_length": 114.0,
"entropy": 0.7237679958343506,
"epoch": 0.07,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.083980560302734,
"learning_rate": 9.35e-07,
"loss": -0.19025824964046478,
"num_tokens": 124640.0,
"reward": -51.397499084472656,
"reward_std": 54.486454010009766,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -51.49749755859375,
"rewards/supergames_reward/std": 54.486454010009766,
"step": 14,
"step_time": 9.597231683001155
},
{
"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": 236.0,
"completions/max_terminated_length": 236.0,
"completions/mean_length": 141.375,
"completions/mean_terminated_length": 141.375,
"completions/min_length": 91.0,
"completions/min_terminated_length": 91.0,
"entropy": 0.754334568977356,
"epoch": 0.075,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.720428943634033,
"learning_rate": 9.3e-07,
"loss": -0.16830675303936005,
"num_tokens": 132219.0,
"reward": -56.226253509521484,
"reward_std": 60.115657806396484,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -56.32625198364258,
"rewards/supergames_reward/std": 60.11566162109375,
"step": 15,
"step_time": 8.998362872982398
},
{
"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": 209.0,
"completions/max_terminated_length": 209.0,
"completions/mean_length": 136.375,
"completions/mean_terminated_length": 136.375,
"completions/min_length": 97.0,
"completions/min_terminated_length": 97.0,
"entropy": 0.6280568838119507,
"epoch": 0.08,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.9681806564331055,
"learning_rate": 9.25e-07,
"loss": -0.01910916529595852,
"num_tokens": 139726.0,
"reward": -68.24000549316406,
"reward_std": 52.238914489746094,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -68.33999633789062,
"rewards/supergames_reward/std": 52.238914489746094,
"step": 16,
"step_time": 8.034480294008972
},
{
"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": 365.0,
"completions/max_terminated_length": 365.0,
"completions/mean_length": 211.875,
"completions/mean_terminated_length": 211.875,
"completions/min_length": 100.0,
"completions/min_terminated_length": 100.0,
"entropy": 0.899176836013794,
"epoch": 0.085,
"frac_reward_zero_std": 0.0,
"grad_norm": 6.691746711730957,
"learning_rate": 9.2e-07,
"loss": 0.25513291358947754,
"num_tokens": 147869.0,
"reward": -99.92500305175781,
"reward_std": 0.07070883363485336,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -100.0,
"rewards/supergames_reward/std": 0.0,
"step": 17,
"step_time": 13.655792869016295
},
{
"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": 387.0,
"completions/max_terminated_length": 387.0,
"completions/mean_length": 238.0,
"completions/mean_terminated_length": 238.0,
"completions/min_length": 87.0,
"completions/min_terminated_length": 87.0,
"entropy": 0.5596873760223389,
"epoch": 0.09,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.6711347103118896,
"learning_rate": 9.15e-07,
"loss": -0.2628335654735565,
"num_tokens": 159101.0,
"reward": -24.900001525878906,
"reward_std": 103.50983428955078,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -25.0,
"rewards/supergames_reward/std": 103.50983428955078,
"step": 18,
"step_time": 14.784329420013819
},
{
"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": 0.125,
"completions/max_length": 512.0,
"completions/max_terminated_length": 277.0,
"completions/mean_length": 219.75,
"completions/mean_terminated_length": 178.00001525878906,
"completions/min_length": 90.0,
"completions/min_terminated_length": 90.0,
"entropy": 0.9245517253875732,
"epoch": 0.095,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.55502986907959,
"learning_rate": 9.1e-07,
"loss": 0.13060179352760315,
"num_tokens": 167299.0,
"reward": -86.89125061035156,
"reward_std": 36.95603561401367,
"rewards/json_format_reward/mean": 0.05000000074505806,
"rewards/json_format_reward/std": 0.09258200973272324,
"rewards/supergames_reward/mean": -86.94125366210938,
"rewards/supergames_reward/std": 36.93572235107422,
"step": 19,
"step_time": 18.832684700988466
},
{
"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": 267.0,
"completions/max_terminated_length": 267.0,
"completions/mean_length": 177.25,
"completions/mean_terminated_length": 177.25,
"completions/min_length": 93.0,
"completions/min_terminated_length": 93.0,
"entropy": 0.7331254482269287,
"epoch": 0.1,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.538497447967529,
"learning_rate": 9.05e-07,
"loss": -0.16418980062007904,
"num_tokens": 175141.0,
"reward": -68.52874755859375,
"reward_std": 58.48719787597656,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -68.62875366210938,
"rewards/supergames_reward/std": 58.48720169067383,
"step": 20,
"step_time": 10.090975169994636
},
{
"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": 300.0,
"completions/max_terminated_length": 300.0,
"completions/mean_length": 190.625,
"completions/mean_terminated_length": 190.625,
"completions/min_length": 103.0,
"completions/min_terminated_length": 103.0,
"entropy": 0.6659425497055054,
"epoch": 0.105,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.485373497009277,
"learning_rate": 9e-07,
"loss": -0.10106432437896729,
"num_tokens": 183098.0,
"reward": -71.48875427246094,
"reward_std": 55.33565139770508,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -71.5887451171875,
"rewards/supergames_reward/std": 55.33565139770508,
"step": 21,
"step_time": 11.255001295008697
},
{
"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": 279.0,
"completions/max_terminated_length": 279.0,
"completions/mean_length": 179.0,
"completions/mean_terminated_length": 179.0,
"completions/min_length": 103.0,
"completions/min_terminated_length": 103.0,
"entropy": 0.49365949630737305,
"epoch": 0.11,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.829419136047363,
"learning_rate": 8.95e-07,
"loss": 0.10587231814861298,
"num_tokens": 192826.0,
"reward": -75.62000274658203,
"reward_std": 45.01959991455078,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -75.69499969482422,
"rewards/supergames_reward/std": 45.00411605834961,
"step": 22,
"step_time": 10.611246599000879
},
{
"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": 315.0,
"completions/max_terminated_length": 315.0,
"completions/mean_length": 212.625,
"completions/mean_terminated_length": 212.625,
"completions/min_length": 100.0,
"completions/min_terminated_length": 100.0,
"entropy": 0.43159469962120056,
"epoch": 0.115,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.541125774383545,
"learning_rate": 8.9e-07,
"loss": -0.1681259125471115,
"num_tokens": 202815.0,
"reward": 23.67624855041504,
"reward_std": 67.74340057373047,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 23.576250076293945,
"rewards/supergames_reward/std": 67.743408203125,
"step": 23,
"step_time": 11.731771531980485
},
{
"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": 168.0,
"completions/max_terminated_length": 168.0,
"completions/mean_length": 101.625,
"completions/mean_terminated_length": 101.625,
"completions/min_length": 65.0,
"completions/min_terminated_length": 65.0,
"entropy": 0.3972298502922058,
"epoch": 0.12,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.8285417556762695,
"learning_rate": 8.85e-07,
"loss": -0.011585958302021027,
"num_tokens": 208188.0,
"reward": 58.849998474121094,
"reward_std": 56.9304084777832,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 58.75,
"rewards/supergames_reward/std": 56.9304084777832,
"step": 24,
"step_time": 6.187504950998118
},
{
"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": 303.0,
"completions/max_terminated_length": 303.0,
"completions/mean_length": 165.375,
"completions/mean_terminated_length": 165.375,
"completions/min_length": 69.0,
"completions/min_terminated_length": 69.0,
"entropy": 0.6339899301528931,
"epoch": 0.125,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.823799133300781,
"learning_rate": 8.799999999999999e-07,
"loss": 0.05159185826778412,
"num_tokens": 215935.0,
"reward": -23.521251678466797,
"reward_std": 74.23067474365234,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -23.621248245239258,
"rewards/supergames_reward/std": 74.23067474365234,
"step": 25,
"step_time": 11.059416300005978
},
{
"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": 349.0,
"completions/max_terminated_length": 349.0,
"completions/mean_length": 204.0,
"completions/mean_terminated_length": 204.0,
"completions/min_length": 143.0,
"completions/min_terminated_length": 143.0,
"entropy": 0.6462895274162292,
"epoch": 0.13,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.969292640686035,
"learning_rate": 8.75e-07,
"loss": 0.0115677984431386,
"num_tokens": 223999.0,
"reward": -74.8175048828125,
"reward_std": 40.94221115112305,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -74.89250183105469,
"rewards/supergames_reward/std": 40.924625396728516,
"step": 26,
"step_time": 12.666237785975682
},
{
"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": 299.0,
"completions/max_terminated_length": 299.0,
"completions/mean_length": 166.5,
"completions/mean_terminated_length": 166.5,
"completions/min_length": 87.0,
"completions/min_terminated_length": 87.0,
"entropy": 0.4756850600242615,
"epoch": 0.135,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.915069103240967,
"learning_rate": 8.699999999999999e-07,
"loss": -0.13321346044540405,
"num_tokens": 233643.0,
"reward": -52.125,
"reward_std": 19.405282974243164,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -52.20000076293945,
"rewards/supergames_reward/std": 19.33464813232422,
"step": 27,
"step_time": 11.172539752995363
},
{
"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": 184.0,
"completions/max_terminated_length": 184.0,
"completions/mean_length": 137.125,
"completions/mean_terminated_length": 137.125,
"completions/min_length": 100.0,
"completions/min_terminated_length": 100.0,
"entropy": 0.5834778547286987,
"epoch": 0.14,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.146666526794434,
"learning_rate": 8.65e-07,
"loss": 0.060454584658145905,
"num_tokens": 241204.0,
"reward": -68.99250793457031,
"reward_std": 51.010581970214844,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -69.09249877929688,
"rewards/supergames_reward/std": 51.010581970214844,
"step": 28,
"step_time": 7.025046669004951
},
{
"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": 331.0,
"completions/max_terminated_length": 331.0,
"completions/mean_length": 182.375,
"completions/mean_terminated_length": 182.375,
"completions/min_length": 106.0,
"completions/min_terminated_length": 106.0,
"entropy": 0.6368551254272461,
"epoch": 0.145,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.15535306930542,
"learning_rate": 8.599999999999999e-07,
"loss": -0.3323673903942108,
"num_tokens": 249087.0,
"reward": -44.17000198364258,
"reward_std": 80.5373306274414,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -44.27000045776367,
"rewards/supergames_reward/std": 80.5373306274414,
"step": 29,
"step_time": 12.079259724996518
},
{
"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": 190.0,
"completions/max_terminated_length": 190.0,
"completions/mean_length": 127.75,
"completions/mean_terminated_length": 127.75,
"completions/min_length": 80.0,
"completions/min_terminated_length": 80.0,
"entropy": 0.4491196870803833,
"epoch": 0.15,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.896186828613281,
"learning_rate": 8.55e-07,
"loss": -0.12887656688690186,
"num_tokens": 259445.0,
"reward": 57.446250915527344,
"reward_std": 38.78837585449219,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 57.34625244140625,
"rewards/supergames_reward/std": 38.78837966918945,
"step": 30,
"step_time": 7.483972909016302
},
{
"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": 166.0,
"completions/max_terminated_length": 166.0,
"completions/mean_length": 112.875,
"completions/mean_terminated_length": 112.875,
"completions/min_length": 85.0,
"completions/min_terminated_length": 85.0,
"entropy": 0.4858454465866089,
"epoch": 0.155,
"frac_reward_zero_std": 0.0,
"grad_norm": 6.742906093597412,
"learning_rate": 8.499999999999999e-07,
"loss": 0.06930096447467804,
"num_tokens": 264924.0,
"reward": -21.174999237060547,
"reward_std": 31.89051628112793,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -21.25,
"rewards/supergames_reward/std": 31.819807052612305,
"step": 31,
"step_time": 6.1866529019898735
},
{
"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": 148.0,
"completions/max_terminated_length": 148.0,
"completions/mean_length": 118.125,
"completions/mean_terminated_length": 118.125,
"completions/min_length": 98.0,
"completions/min_terminated_length": 98.0,
"entropy": 0.45428749918937683,
"epoch": 0.16,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.663414478302002,
"learning_rate": 8.45e-07,
"loss": 0.05735419690608978,
"num_tokens": 270381.0,
"reward": 0.4687504768371582,
"reward_std": 38.66371154785156,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 0.36875057220458984,
"rewards/supergames_reward/std": 38.6637077331543,
"step": 32,
"step_time": 5.577042757999152
},
{
"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": 345.0,
"completions/max_terminated_length": 345.0,
"completions/mean_length": 166.875,
"completions/mean_terminated_length": 166.875,
"completions/min_length": 86.0,
"completions/min_terminated_length": 86.0,
"entropy": 0.6107293367385864,
"epoch": 0.165,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.75006628036499,
"learning_rate": 8.399999999999999e-07,
"loss": -0.14760127663612366,
"num_tokens": 278156.0,
"reward": -87.84750366210938,
"reward_std": 26.84708023071289,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -87.94749450683594,
"rewards/supergames_reward/std": 26.847076416015625,
"step": 33,
"step_time": 12.552876825997373
},
{
"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": 0.125,
"completions/max_length": 512.0,
"completions/max_terminated_length": 360.0,
"completions/mean_length": 231.0,
"completions/mean_terminated_length": 190.85714721679688,
"completions/min_length": 74.0,
"completions/min_terminated_length": 74.0,
"entropy": 0.6089653372764587,
"epoch": 0.17,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.781267166137695,
"learning_rate": 8.349999999999999e-07,
"loss": 0.4294504225254059,
"num_tokens": 289364.0,
"reward": -99.92500305175781,
"reward_std": 0.07070883363485336,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -100.0,
"rewards/supergames_reward/std": 0.0,
"step": 34,
"step_time": 18.961819477990502
},
{
"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": 218.0,
"completions/max_terminated_length": 218.0,
"completions/mean_length": 134.125,
"completions/mean_terminated_length": 134.125,
"completions/min_length": 104.0,
"completions/min_terminated_length": 104.0,
"entropy": 0.46158918738365173,
"epoch": 0.175,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.016353607177734,
"learning_rate": 8.299999999999999e-07,
"loss": 0.09251818805932999,
"num_tokens": 298733.0,
"reward": -35.558753967285156,
"reward_std": 15.881204605102539,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -35.65875244140625,
"rewards/supergames_reward/std": 15.881203651428223,
"step": 35,
"step_time": 8.38491243700264
},
{
"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": 207.0,
"completions/max_terminated_length": 207.0,
"completions/mean_length": 156.875,
"completions/mean_terminated_length": 156.875,
"completions/min_length": 111.0,
"completions/min_terminated_length": 111.0,
"entropy": 0.5104788541793823,
"epoch": 0.18,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.592893123626709,
"learning_rate": 8.249999999999999e-07,
"loss": -0.0410832017660141,
"num_tokens": 306420.0,
"reward": 5.4212493896484375,
"reward_std": 93.50404357910156,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": 5.346250534057617,
"rewards/supergames_reward/std": 93.4718246459961,
"step": 36,
"step_time": 7.811868985998444
},
{
"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": 241.0,
"completions/max_terminated_length": 241.0,
"completions/mean_length": 178.0,
"completions/mean_terminated_length": 178.0,
"completions/min_length": 106.0,
"completions/min_terminated_length": 106.0,
"entropy": 0.7527878880500793,
"epoch": 0.185,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.8812546730041504,
"learning_rate": 8.199999999999999e-07,
"loss": 0.12539231777191162,
"num_tokens": 314252.0,
"reward": -89.58500671386719,
"reward_std": 29.256120681762695,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -89.66000366210938,
"rewards/supergames_reward/std": 29.24593734741211,
"step": 37,
"step_time": 8.959742914012168
},
{
"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": 199.0,
"completions/max_terminated_length": 199.0,
"completions/mean_length": 133.375,
"completions/mean_terminated_length": 133.375,
"completions/min_length": 87.0,
"completions/min_terminated_length": 87.0,
"entropy": 0.3347901403903961,
"epoch": 0.19,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.205414295196533,
"learning_rate": 8.149999999999999e-07,
"loss": -0.06991486996412277,
"num_tokens": 319839.0,
"reward": -14.498749732971191,
"reward_std": 13.007229804992676,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -14.598750114440918,
"rewards/supergames_reward/std": 13.007229804992676,
"step": 38,
"step_time": 7.377183554985095
},
{
"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": 181.0,
"completions/max_terminated_length": 181.0,
"completions/mean_length": 114.375,
"completions/mean_terminated_length": 114.375,
"completions/min_length": 58.0,
"completions/min_terminated_length": 58.0,
"entropy": 0.4445965886116028,
"epoch": 0.195,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.561298847198486,
"learning_rate": 8.1e-07,
"loss": -0.06235164776444435,
"num_tokens": 325290.0,
"reward": -7.425000190734863,
"reward_std": 53.701602935791016,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -7.5,
"rewards/supergames_reward/std": 53.652320861816406,
"step": 39,
"step_time": 6.753862089011818
},
{
"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": 370.0,
"completions/max_terminated_length": 370.0,
"completions/mean_length": 187.625,
"completions/mean_terminated_length": 187.625,
"completions/min_length": 60.0,
"completions/min_terminated_length": 60.0,
"entropy": 0.3950234353542328,
"epoch": 0.2,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.5988712310791016,
"learning_rate": 8.05e-07,
"loss": -0.1381440907716751,
"num_tokens": 335119.0,
"reward": -33.68375015258789,
"reward_std": 42.069435119628906,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -33.783748626708984,
"rewards/supergames_reward/std": 42.069435119628906,
"step": 40,
"step_time": 13.718958162004128
},
{
"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": 0.125,
"completions/max_length": 512.0,
"completions/max_terminated_length": 371.0,
"completions/mean_length": 251.125,
"completions/mean_terminated_length": 213.85714721679688,
"completions/min_length": 121.0,
"completions/min_terminated_length": 121.0,
"entropy": 0.507067084312439,
"epoch": 0.205,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.3898658752441406,
"learning_rate": 8e-07,
"loss": 0.0078964838758111,
"num_tokens": 346464.0,
"reward": -86.48625183105469,
"reward_std": 38.020694732666016,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -86.56124877929688,
"rewards/supergames_reward/std": 38.010528564453125,
"step": 41,
"step_time": 18.90960379401804
},
{
"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": 152.0,
"completions/max_terminated_length": 152.0,
"completions/mean_length": 116.375,
"completions/mean_terminated_length": 116.375,
"completions/min_length": 93.0,
"completions/min_terminated_length": 93.0,
"entropy": 0.3801707923412323,
"epoch": 0.21,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 7.95e-07,
"loss": 0.0,
"num_tokens": 351955.0,
"reward": -9.899999618530273,
"reward_std": 0.0,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -10.0,
"rewards/supergames_reward/std": 0.0,
"step": 42,
"step_time": 5.711630532023264
},
{
"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": 308.0,
"completions/max_terminated_length": 308.0,
"completions/mean_length": 247.375,
"completions/mean_terminated_length": 247.375,
"completions/min_length": 116.0,
"completions/min_terminated_length": 116.0,
"entropy": 0.3185043931007385,
"epoch": 0.215,
"frac_reward_zero_std": 0.0,
"grad_norm": 2.168762445449829,
"learning_rate": 7.9e-07,
"loss": 0.1970764398574829,
"num_tokens": 363278.0,
"reward": -53.89249801635742,
"reward_std": 63.02619934082031,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -53.99250030517578,
"rewards/supergames_reward/std": 63.02619934082031,
"step": 43,
"step_time": 11.661934906995157
},
{
"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": 330.0,
"completions/max_terminated_length": 330.0,
"completions/mean_length": 212.5,
"completions/mean_terminated_length": 212.5,
"completions/min_length": 106.0,
"completions/min_terminated_length": 106.0,
"entropy": 0.4446268379688263,
"epoch": 0.22,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.305844306945801,
"learning_rate": 7.85e-07,
"loss": -0.06735197454690933,
"num_tokens": 374306.0,
"reward": 18.059999465942383,
"reward_std": 98.9417953491211,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 17.959999084472656,
"rewards/supergames_reward/std": 98.9417953491211,
"step": 44,
"step_time": 12.492453911982011
},
{
"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": 316.0,
"completions/max_terminated_length": 316.0,
"completions/mean_length": 183.0,
"completions/mean_terminated_length": 183.0,
"completions/min_length": 103.0,
"completions/min_terminated_length": 103.0,
"entropy": 0.5269614458084106,
"epoch": 0.225,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.8006370067596436,
"learning_rate": 7.799999999999999e-07,
"loss": -0.08870815485715866,
"num_tokens": 382210.0,
"reward": -54.928749084472656,
"reward_std": 72.5127944946289,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -55.02874755859375,
"rewards/supergames_reward/std": 72.5127944946289,
"step": 45,
"step_time": 11.7636467939883
},
{
"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": 244.0,
"completions/max_terminated_length": 244.0,
"completions/mean_length": 136.625,
"completions/mean_terminated_length": 136.625,
"completions/min_length": 85.0,
"completions/min_terminated_length": 85.0,
"entropy": 0.34820204973220825,
"epoch": 0.23,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.911736011505127,
"learning_rate": 7.75e-07,
"loss": 0.04397330805659294,
"num_tokens": 391583.0,
"reward": -19.10375213623047,
"reward_std": 23.967702865600586,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -19.203750610351562,
"rewards/supergames_reward/std": 23.967702865600586,
"step": 46,
"step_time": 10.119426151999505
},
{
"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": 218.0,
"completions/max_terminated_length": 218.0,
"completions/mean_length": 160.375,
"completions/mean_terminated_length": 160.375,
"completions/min_length": 93.0,
"completions/min_terminated_length": 93.0,
"entropy": 0.611914336681366,
"epoch": 0.235,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.571913242340088,
"learning_rate": 7.699999999999999e-07,
"loss": -0.044293053448200226,
"num_tokens": 399290.0,
"reward": -25.186250686645508,
"reward_std": 71.46305084228516,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -25.286251068115234,
"rewards/supergames_reward/std": 71.46305084228516,
"step": 47,
"step_time": 8.563868903991533
},
{
"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": 167.0,
"completions/max_terminated_length": 167.0,
"completions/mean_length": 111.125,
"completions/mean_terminated_length": 111.125,
"completions/min_length": 56.0,
"completions/min_terminated_length": 56.0,
"entropy": 0.46585777401924133,
"epoch": 0.24,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.654109954833984,
"learning_rate": 7.65e-07,
"loss": -0.035151124000549316,
"num_tokens": 404707.0,
"reward": -0.408750057220459,
"reward_std": 31.939748764038086,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -0.5087499618530273,
"rewards/supergames_reward/std": 31.93975257873535,
"step": 48,
"step_time": 6.438482580007985
},
{
"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": 306.0,
"completions/max_terminated_length": 306.0,
"completions/mean_length": 216.625,
"completions/mean_terminated_length": 216.625,
"completions/min_length": 132.0,
"completions/min_terminated_length": 132.0,
"entropy": 0.33054471015930176,
"epoch": 0.245,
"frac_reward_zero_std": 0.0,
"grad_norm": 2.700303792953491,
"learning_rate": 7.599999999999999e-07,
"loss": -0.04393656551837921,
"num_tokens": 414736.0,
"reward": -97.26000213623047,
"reward_std": 1.6294406652450562,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -97.36000061035156,
"rewards/supergames_reward/std": 1.629441738128662,
"step": 49,
"step_time": 11.636040106008295
},
{
"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": 378.0,
"completions/max_terminated_length": 378.0,
"completions/mean_length": 218.375,
"completions/mean_terminated_length": 218.375,
"completions/min_length": 104.0,
"completions/min_terminated_length": 104.0,
"entropy": 0.48711177706718445,
"epoch": 0.25,
"frac_reward_zero_std": 0.0,
"grad_norm": 2.9650564193725586,
"learning_rate": 7.55e-07,
"loss": -0.3529520332813263,
"num_tokens": 424795.0,
"reward": 47.408748626708984,
"reward_std": 72.72083282470703,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 47.30875015258789,
"rewards/supergames_reward/std": 72.72083282470703,
"step": 50,
"step_time": 14.258096842997475
},
{
"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": 0.125,
"completions/max_length": 512.0,
"completions/max_terminated_length": 492.0,
"completions/mean_length": 267.625,
"completions/mean_terminated_length": 232.71429443359375,
"completions/min_length": 90.0,
"completions/min_terminated_length": 90.0,
"entropy": 0.4482860863208771,
"epoch": 0.255,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.2998197078704834,
"learning_rate": 7.5e-07,
"loss": 0.2493654191493988,
"num_tokens": 436288.0,
"reward": 50.61125183105469,
"reward_std": 75.3349838256836,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": 50.53624725341797,
"rewards/supergames_reward/std": 75.27783966064453,
"step": 51,
"step_time": 19.20827590499539
},
{
"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": 176.0,
"completions/max_terminated_length": 176.0,
"completions/mean_length": 144.5,
"completions/mean_terminated_length": 144.5,
"completions/min_length": 99.0,
"completions/min_terminated_length": 99.0,
"entropy": 0.5378735661506653,
"epoch": 0.26,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.114461898803711,
"learning_rate": 7.45e-07,
"loss": 0.0036597400903701782,
"num_tokens": 441980.0,
"reward": -15.20625114440918,
"reward_std": 45.35526657104492,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -15.28125,
"rewards/supergames_reward/std": 45.301815032958984,
"step": 52,
"step_time": 6.660627231001854
},
{
"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": 136.0,
"completions/max_terminated_length": 136.0,
"completions/mean_length": 119.0,
"completions/mean_terminated_length": 119.0,
"completions/min_length": 105.0,
"completions/min_terminated_length": 105.0,
"entropy": 0.41796669363975525,
"epoch": 0.265,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.059505462646484,
"learning_rate": 7.4e-07,
"loss": -0.014973883517086506,
"num_tokens": 447468.0,
"reward": 15.170000076293945,
"reward_std": 46.658329010009766,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 15.070000648498535,
"rewards/supergames_reward/std": 46.658329010009766,
"step": 53,
"step_time": 5.271571868011961
},
{
"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": 259.0,
"completions/max_terminated_length": 259.0,
"completions/mean_length": 191.125,
"completions/mean_terminated_length": 191.125,
"completions/min_length": 91.0,
"completions/min_terminated_length": 91.0,
"entropy": 0.45421987771987915,
"epoch": 0.27,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.4454867839813232,
"learning_rate": 7.35e-07,
"loss": -0.1128973588347435,
"num_tokens": 457301.0,
"reward": 38.89875030517578,
"reward_std": 62.1772575378418,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 38.79874801635742,
"rewards/supergames_reward/std": 62.1772575378418,
"step": 54,
"step_time": 10.524528659996577
},
{
"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": 319.0,
"completions/max_terminated_length": 319.0,
"completions/mean_length": 233.375,
"completions/mean_terminated_length": 233.375,
"completions/min_length": 136.0,
"completions/min_terminated_length": 136.0,
"entropy": 0.4449213445186615,
"epoch": 0.275,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.2065398693084717,
"learning_rate": 7.3e-07,
"loss": 0.0016277075046673417,
"num_tokens": 467520.0,
"reward": 27.372499465942383,
"reward_std": 81.23126983642578,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 27.27250099182129,
"rewards/supergames_reward/std": 81.23126983642578,
"step": 55,
"step_time": 12.592280682991259
},
{
"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": 0.125,
"completions/max_length": 512.0,
"completions/max_terminated_length": 357.0,
"completions/mean_length": 261.0,
"completions/mean_terminated_length": 225.1428680419922,
"completions/min_length": 95.0,
"completions/min_terminated_length": 95.0,
"entropy": 0.506920337677002,
"epoch": 0.28,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.058388710021973,
"learning_rate": 7.249999999999999e-07,
"loss": 0.3289242386817932,
"num_tokens": 476056.0,
"reward": 47.05249786376953,
"reward_std": 80.97753143310547,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": 46.977500915527344,
"rewards/supergames_reward/std": 80.9256362915039,
"step": 56,
"step_time": 18.845177220006008
},
{
"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": 250.0,
"completions/max_terminated_length": 250.0,
"completions/mean_length": 192.75,
"completions/mean_terminated_length": 192.75,
"completions/min_length": 122.0,
"completions/min_terminated_length": 122.0,
"entropy": 0.5197336077690125,
"epoch": 0.285,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.230815410614014,
"learning_rate": 7.2e-07,
"loss": -0.15186546742916107,
"num_tokens": 484014.0,
"reward": -64.5050048828125,
"reward_std": 50.35145568847656,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -64.60499572753906,
"rewards/supergames_reward/std": 50.3514518737793,
"step": 57,
"step_time": 9.576232638006331
},
{
"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": 324.0,
"completions/max_terminated_length": 324.0,
"completions/mean_length": 249.0,
"completions/mean_terminated_length": 249.0,
"completions/min_length": 129.0,
"completions/min_terminated_length": 129.0,
"entropy": 0.42913514375686646,
"epoch": 0.29,
"frac_reward_zero_std": 0.0,
"grad_norm": 2.8544528484344482,
"learning_rate": 7.149999999999999e-07,
"loss": -0.056405920535326004,
"num_tokens": 494390.0,
"reward": 40.123748779296875,
"reward_std": 84.164794921875,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 40.02375030517578,
"rewards/supergames_reward/std": 84.164794921875,
"step": 58,
"step_time": 12.385353341000155
},
{
"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": 138.0,
"completions/max_terminated_length": 138.0,
"completions/mean_length": 107.25,
"completions/mean_terminated_length": 107.25,
"completions/min_length": 81.0,
"completions/min_terminated_length": 81.0,
"entropy": 0.34386488795280457,
"epoch": 0.295,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.6915154457092285,
"learning_rate": 7.1e-07,
"loss": 0.0865321159362793,
"num_tokens": 499816.0,
"reward": -8.244999885559082,
"reward_std": 4.681046962738037,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -8.345000267028809,
"rewards/supergames_reward/std": 4.681046962738037,
"step": 59,
"step_time": 5.42325339600211
},
{
"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": 187.0,
"completions/max_terminated_length": 187.0,
"completions/mean_length": 127.25,
"completions/mean_terminated_length": 127.25,
"completions/min_length": 97.0,
"completions/min_terminated_length": 97.0,
"entropy": 0.5430045127868652,
"epoch": 0.3,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.635311126708984,
"learning_rate": 7.049999999999999e-07,
"loss": -0.05412375554442406,
"num_tokens": 505410.0,
"reward": 20.005001068115234,
"reward_std": 55.156005859375,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 19.904998779296875,
"rewards/supergames_reward/std": 55.156005859375,
"step": 60,
"step_time": 7.068010902003152
},
{
"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": 286.0,
"completions/max_terminated_length": 286.0,
"completions/mean_length": 192.125,
"completions/mean_terminated_length": 192.125,
"completions/min_length": 95.0,
"completions/min_terminated_length": 95.0,
"entropy": 0.4213869869709015,
"epoch": 0.305,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.4664618968963623,
"learning_rate": 7e-07,
"loss": -0.0495159812271595,
"num_tokens": 516283.0,
"reward": -47.212501525878906,
"reward_std": 91.2253646850586,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -47.3125,
"rewards/supergames_reward/std": 91.2253646850586,
"step": 61,
"step_time": 11.176304235996213
},
{
"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": 324.0,
"completions/max_terminated_length": 324.0,
"completions/mean_length": 171.0,
"completions/mean_terminated_length": 171.0,
"completions/min_length": 92.0,
"completions/min_terminated_length": 92.0,
"entropy": 0.4626811146736145,
"epoch": 0.31,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.949252128601074,
"learning_rate": 6.949999999999999e-07,
"loss": 0.192047581076622,
"num_tokens": 522251.0,
"reward": -15.092500686645508,
"reward_std": 49.82962417602539,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -15.167499542236328,
"rewards/supergames_reward/std": 49.78091049194336,
"step": 62,
"step_time": 12.075224861997413
},
{
"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": 0.125,
"completions/max_length": 512.0,
"completions/max_terminated_length": 429.0,
"completions/mean_length": 283.5,
"completions/mean_terminated_length": 250.85714721679688,
"completions/min_length": 85.0,
"completions/min_terminated_length": 85.0,
"entropy": 0.4814216196537018,
"epoch": 0.315,
"frac_reward_zero_std": 0.0,
"grad_norm": 2.7918753623962402,
"learning_rate": 6.9e-07,
"loss": 0.15706156194210052,
"num_tokens": 533815.0,
"reward": -20.86625099182129,
"reward_std": 85.99250030517578,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -20.94124984741211,
"rewards/supergames_reward/std": 85.96620178222656,
"step": 63,
"step_time": 19.448832260008203
},
{
"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": 189.0,
"completions/max_terminated_length": 189.0,
"completions/mean_length": 131.25,
"completions/mean_terminated_length": 131.25,
"completions/min_length": 106.0,
"completions/min_terminated_length": 106.0,
"entropy": 0.4929414987564087,
"epoch": 0.32,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.774101734161377,
"learning_rate": 6.85e-07,
"loss": -0.060691747814416885,
"num_tokens": 539465.0,
"reward": -21.0674991607666,
"reward_std": 6.892730236053467,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -21.167499542236328,
"rewards/supergames_reward/std": 6.892730236053467,
"step": 64,
"step_time": 7.141569407976931
},
{
"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": 454.0,
"completions/max_terminated_length": 454.0,
"completions/mean_length": 281.375,
"completions/mean_terminated_length": 281.375,
"completions/min_length": 107.0,
"completions/min_terminated_length": 107.0,
"entropy": 0.45866572856903076,
"epoch": 0.325,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.092717409133911,
"learning_rate": 6.800000000000001e-07,
"loss": 0.11694959551095963,
"num_tokens": 551012.0,
"reward": 49.938751220703125,
"reward_std": 79.1865005493164,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 49.8387451171875,
"rewards/supergames_reward/std": 79.1865005493164,
"step": 65,
"step_time": 17.06170882002334
},
{
"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": 203.0,
"completions/max_terminated_length": 203.0,
"completions/mean_length": 128.625,
"completions/mean_terminated_length": 128.625,
"completions/min_length": 88.0,
"completions/min_terminated_length": 88.0,
"entropy": 0.5005782246589661,
"epoch": 0.33,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.77616024017334,
"learning_rate": 6.75e-07,
"loss": 0.11242837458848953,
"num_tokens": 556577.0,
"reward": 48.875,
"reward_std": 49.13539123535156,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 48.775001525878906,
"rewards/supergames_reward/std": 49.13539123535156,
"step": 66,
"step_time": 7.695410129002994
},
{
"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": 436.0,
"completions/max_terminated_length": 436.0,
"completions/mean_length": 281.875,
"completions/mean_terminated_length": 281.875,
"completions/min_length": 133.0,
"completions/min_terminated_length": 133.0,
"entropy": 0.507097065448761,
"epoch": 0.335,
"frac_reward_zero_std": 0.0,
"grad_norm": 2.911184787750244,
"learning_rate": 6.7e-07,
"loss": -0.08151137083768845,
"num_tokens": 567120.0,
"reward": -12.886249542236328,
"reward_std": 84.04269409179688,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -12.986251831054688,
"rewards/supergames_reward/std": 84.04269409179688,
"step": 67,
"step_time": 16.409127942984924
},
{
"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": 0.125,
"completions/max_length": 512.0,
"completions/max_terminated_length": 501.0,
"completions/mean_length": 318.375,
"completions/mean_terminated_length": 290.71429443359375,
"completions/min_length": 65.0,
"completions/min_terminated_length": 65.0,
"entropy": 0.486402690410614,
"epoch": 0.34,
"frac_reward_zero_std": 0.0,
"grad_norm": 2.9461607933044434,
"learning_rate": 6.65e-07,
"loss": 0.2442178875207901,
"num_tokens": 578995.0,
"reward": 2.8387489318847656,
"reward_std": 98.47579956054688,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": 2.7637500762939453,
"rewards/supergames_reward/std": 98.44595336914062,
"step": 68,
"step_time": 19.32364999302081
},
{
"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": 291.0,
"completions/max_terminated_length": 291.0,
"completions/mean_length": 219.625,
"completions/mean_terminated_length": 219.625,
"completions/min_length": 117.0,
"completions/min_terminated_length": 117.0,
"entropy": 0.37856727838516235,
"epoch": 0.345,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.0204732418060303,
"learning_rate": 6.6e-07,
"loss": -0.08764916658401489,
"num_tokens": 589072.0,
"reward": -54.18375015258789,
"reward_std": 71.45012664794922,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -54.283748626708984,
"rewards/supergames_reward/std": 71.45013427734375,
"step": 69,
"step_time": 11.126611230982235
},
{
"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": 0.125,
"completions/max_length": 512.0,
"completions/max_terminated_length": 305.0,
"completions/mean_length": 254.75,
"completions/mean_terminated_length": 218.00001525878906,
"completions/min_length": 104.0,
"completions/min_terminated_length": 104.0,
"entropy": 0.46423614025115967,
"epoch": 0.35,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.0140154361724854,
"learning_rate": 6.55e-07,
"loss": 0.26797160506248474,
"num_tokens": 600438.0,
"reward": -32.04875183105469,
"reward_std": 69.86957550048828,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -32.12375259399414,
"rewards/supergames_reward/std": 69.84178161621094,
"step": 70,
"step_time": 19.12617294798838
},
{
"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": 372.0,
"completions/max_terminated_length": 372.0,
"completions/mean_length": 194.125,
"completions/mean_terminated_length": 194.125,
"completions/min_length": 94.0,
"completions/min_terminated_length": 94.0,
"entropy": 0.3427293300628662,
"epoch": 0.355,
"frac_reward_zero_std": 0.0,
"grad_norm": 2.781383991241455,
"learning_rate": 6.5e-07,
"loss": 0.354524701833725,
"num_tokens": 610327.0,
"reward": -6.913749694824219,
"reward_std": 87.56071472167969,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": -6.988750457763672,
"rewards/supergames_reward/std": 87.53033447265625,
"step": 71,
"step_time": 14.041668199002743
},
{
"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": 197.0,
"completions/max_terminated_length": 197.0,
"completions/mean_length": 124.75,
"completions/mean_terminated_length": 124.75,
"completions/min_length": 93.0,
"completions/min_terminated_length": 93.0,
"entropy": 0.4455118477344513,
"epoch": 0.36,
"frac_reward_zero_std": 0.0,
"grad_norm": 5.726032733917236,
"learning_rate": 6.45e-07,
"loss": 0.18206289410591125,
"num_tokens": 615853.0,
"reward": 41.95124816894531,
"reward_std": 60.640811920166016,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 41.85124969482422,
"rewards/supergames_reward/std": 60.640811920166016,
"step": 72,
"step_time": 7.416493425989756
},
{
"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": 323.0,
"completions/max_terminated_length": 323.0,
"completions/mean_length": 147.375,
"completions/mean_terminated_length": 147.375,
"completions/min_length": 94.0,
"completions/min_terminated_length": 94.0,
"entropy": 0.480733186006546,
"epoch": 0.365,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.853621482849121,
"learning_rate": 6.4e-07,
"loss": 0.04558124020695686,
"num_tokens": 621584.0,
"reward": 14.884998321533203,
"reward_std": 46.25857162475586,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 14.78499984741211,
"rewards/supergames_reward/std": 46.25857162475586,
"step": 73,
"step_time": 11.95379540900467
},
{
"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": 283.0,
"completions/max_terminated_length": 283.0,
"completions/mean_length": 188.125,
"completions/mean_terminated_length": 188.125,
"completions/min_length": 116.0,
"completions/min_terminated_length": 116.0,
"entropy": 0.5526013374328613,
"epoch": 0.37,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.729883909225464,
"learning_rate": 6.35e-07,
"loss": 0.06268303096294403,
"num_tokens": 629545.0,
"reward": -68.05750274658203,
"reward_std": 59.44808578491211,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -68.15750122070312,
"rewards/supergames_reward/std": 59.44808578491211,
"step": 74,
"step_time": 10.64124580900534
},
{
"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": 141.0,
"completions/max_terminated_length": 141.0,
"completions/mean_length": 117.375,
"completions/mean_terminated_length": 117.375,
"completions/min_length": 100.0,
"completions/min_terminated_length": 100.0,
"entropy": 0.3744319677352905,
"epoch": 0.375,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.775154113769531,
"learning_rate": 6.3e-07,
"loss": -0.023284845054149628,
"num_tokens": 635052.0,
"reward": 3.9374990463256836,
"reward_std": 61.85771179199219,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": 3.862499237060547,
"rewards/supergames_reward/std": 61.80967712402344,
"step": 75,
"step_time": 5.432648951013107
},
{
"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": 268.0,
"completions/max_terminated_length": 268.0,
"completions/mean_length": 157.0,
"completions/mean_terminated_length": 157.0,
"completions/min_length": 104.0,
"completions/min_terminated_length": 104.0,
"entropy": 0.4822140336036682,
"epoch": 0.38,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.715704917907715,
"learning_rate": 6.249999999999999e-07,
"loss": -0.04918142035603523,
"num_tokens": 644644.0,
"reward": 48.64125061035156,
"reward_std": 78.0205078125,
"rewards/json_format_reward/mean": 0.07500000298023224,
"rewards/json_format_reward/std": 0.0707106813788414,
"rewards/supergames_reward/mean": 48.56624984741211,
"rewards/supergames_reward/std": 77.96605682373047,
"step": 76,
"step_time": 10.348935816989979
},
{
"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": 380.0,
"completions/max_terminated_length": 380.0,
"completions/mean_length": 210.875,
"completions/mean_terminated_length": 210.875,
"completions/min_length": 126.0,
"completions/min_terminated_length": 126.0,
"entropy": 0.45722317695617676,
"epoch": 0.385,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.6057567596435547,
"learning_rate": 6.2e-07,
"loss": 0.058352649211883545,
"num_tokens": 652771.0,
"reward": -8.283750534057617,
"reward_std": 77.68436431884766,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -8.383749008178711,
"rewards/supergames_reward/std": 77.68437194824219,
"step": 77,
"step_time": 14.142948755004909
},
{
"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": 324.0,
"completions/max_terminated_length": 324.0,
"completions/mean_length": 188.25,
"completions/mean_terminated_length": 188.25,
"completions/min_length": 68.0,
"completions/min_terminated_length": 68.0,
"entropy": 0.33380764722824097,
"epoch": 0.39,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.2239863872528076,
"learning_rate": 6.149999999999999e-07,
"loss": -0.04592633992433548,
"num_tokens": 662653.0,
"reward": 11.153749465942383,
"reward_std": 89.88329315185547,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 11.053749084472656,
"rewards/supergames_reward/std": 89.88329315185547,
"step": 78,
"step_time": 12.277474621019792
},
{
"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": 143.0,
"completions/max_terminated_length": 143.0,
"completions/mean_length": 108.875,
"completions/mean_terminated_length": 108.875,
"completions/min_length": 85.0,
"completions/min_terminated_length": 85.0,
"entropy": 0.3518812656402588,
"epoch": 0.395,
"frac_reward_zero_std": 0.0,
"grad_norm": 3.967776298522949,
"learning_rate": 6.1e-07,
"loss": -0.10800496488809586,
"num_tokens": 668076.0,
"reward": 70.33250427246094,
"reward_std": 49.9145393371582,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": 70.23249816894531,
"rewards/supergames_reward/std": 49.91453552246094,
"step": 79,
"step_time": 5.52714040101273
},
{
"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": 184.0,
"completions/max_terminated_length": 184.0,
"completions/mean_length": 152.125,
"completions/mean_terminated_length": 152.125,
"completions/min_length": 104.0,
"completions/min_terminated_length": 104.0,
"entropy": 0.4699896574020386,
"epoch": 0.4,
"frac_reward_zero_std": 0.0,
"grad_norm": 4.561161994934082,
"learning_rate": 6.049999999999999e-07,
"loss": 0.09505834430456161,
"num_tokens": 675701.0,
"reward": -20.25625228881836,
"reward_std": 53.78542709350586,
"rewards/json_format_reward/mean": 0.10000000149011612,
"rewards/json_format_reward/std": 0.0,
"rewards/supergames_reward/mean": -20.35624885559082,
"rewards/supergames_reward/std": 53.78542709350586,
"step": 80,
"step_time": 7.155081019998761
}
],
"logging_steps": 1,
"max_steps": 200,
"num_input_tokens_seen": 675701,
"num_train_epochs": 1,
"save_steps": 20,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}