{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 1.2070006035003018, "eval_steps": 250, "global_step": 2000, "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": 0.12125, "completions/max_length": 255.34, "completions/max_terminated_length": 252.14, "completions/mean_length": 221.534375, "completions/mean_terminated_length": 216.93697082519532, "completions/min_length": 173.54, "completions/min_terminated_length": 173.54, "entropy": 0.10048629969358444, "epoch": 0.030175015087507542, "frac_reward_zero_std": 0.3225, "grad_norm": 0.46380576491355896, "learning_rate": 5e-05, "loss": 0.004, "num_tokens": 8142396.0, "reward": 7.30375, "reward_std": 1.5006456315517425, "rewards/event_reward_fn/mean": 7.30375, "rewards/event_reward_fn/std": 6.278198585510254, "step": 50, "step_time": 40.824848868116966 }, { "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": 0.068125, "completions/max_length": 251.74, "completions/max_terminated_length": 248.06, "completions/mean_length": 215.08625, "completions/mean_terminated_length": 212.25316284179686, "completions/min_length": 171.76, "completions/min_terminated_length": 171.76, "entropy": 0.10318506792187691, "epoch": 0.060350030175015085, "frac_reward_zero_std": 0.325, "grad_norm": 0.21978232264518738, "learning_rate": 5e-05, "loss": -0.0025, "num_tokens": 16421719.0, "reward": 7.36875, "reward_std": 1.3263894939422607, "rewards/event_reward_fn/mean": 7.36875, "rewards/event_reward_fn/std": 6.119045643806458, "step": 100, "step_time": 38.99798643006128 }, { "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": 0.4825, "completions/max_length": 256.0, "completions/max_terminated_length": 251.32, "completions/mean_length": 238.104375, "completions/mean_terminated_length": 221.8957485961914, "completions/min_length": 191.34, "completions/min_terminated_length": 191.34, "entropy": 0.10444845259189606, "epoch": 0.09052504526252263, "frac_reward_zero_std": 0.2925, "grad_norm": 0.5579063892364502, "learning_rate": 5e-05, "loss": -0.0006, "num_tokens": 24885844.0, "reward": 7.74625, "reward_std": 1.5345598912239076, "rewards/event_reward_fn/mean": 7.74625, "rewards/event_reward_fn/std": 6.464660973548889, "step": 150, "step_time": 41.26081488572061 }, { "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": 0.7925, "completions/max_length": 256.0, "completions/max_terminated_length": 202.92, "completions/mean_length": 245.916875, "completions/mean_terminated_length": 184.6587713623047, "completions/min_length": 199.94, "completions/min_terminated_length": 169.22, "entropy": 0.10581055819988251, "epoch": 0.12070006035003017, "frac_reward_zero_std": 0.33, "grad_norm": 0.31808722019195557, "learning_rate": 5e-05, "loss": 0.0003, "num_tokens": 33226966.0, "reward": 7.19125, "reward_std": 1.4298825466632843, "rewards/event_reward_fn/mean": 7.19125, "rewards/event_reward_fn/std": 5.8599746036529545, "step": 200, "step_time": 41.91275953448203 }, { "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": 0.825, "completions/max_length": 256.0, "completions/max_terminated_length": 181.12, "completions/mean_length": 245.851875, "completions/mean_terminated_length": 163.72261688232422, "completions/min_length": 198.46, "completions/min_terminated_length": 152.38, "entropy": 0.10499135926365852, "epoch": 0.15087507543753773, "frac_reward_zero_std": 0.2875, "grad_norm": 0.2646925449371338, "learning_rate": 5e-05, "loss": 0.0005, "num_tokens": 41523308.0, "reward": 7.9475, "reward_std": 1.5300491595268249, "rewards/event_reward_fn/mean": 7.9475, "rewards/event_reward_fn/std": 6.3965685844421385, "step": 250, "step_time": 41.663273623897695 }, { "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": 0.898125, "completions/max_length": 256.0, "completions/max_terminated_length": 149.62, "completions/mean_length": 250.625625, "completions/mean_terminated_length": 144.78653198242188, "completions/min_length": 215.68, "completions/min_terminated_length": 138.88, "entropy": 0.10884671121835708, "epoch": 0.18105009052504525, "frac_reward_zero_std": 0.3325, "grad_norm": 0.5418329834938049, "learning_rate": 5e-05, "loss": -0.0002, "num_tokens": 49889481.0, "reward": 7.489375, "reward_std": 1.5504147619009019, "rewards/event_reward_fn/mean": 7.489375, "rewards/event_reward_fn/std": 6.099679977893829, "step": 300, "step_time": 40.817094522019616 }, { "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": 0.9275, "completions/max_length": 256.0, "completions/max_terminated_length": 131.48, "completions/mean_length": 253.1625, "completions/mean_terminated_length": 125.53590209960937, "completions/min_length": 228.04, "completions/min_terminated_length": 120.52, "entropy": 0.10796756476163864, "epoch": 0.2112251056125528, "frac_reward_zero_std": 0.3175, "grad_norm": 0.4433981776237488, "learning_rate": 5e-05, "loss": 0.0019, "num_tokens": 58206892.0, "reward": 7.89625, "reward_std": 1.573977051973343, "rewards/event_reward_fn/mean": 7.89625, "rewards/event_reward_fn/std": 6.586006484031677, "step": 350, "step_time": 42.12015992245928 }, { "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": 0.945625, "completions/max_length": 256.0, "completions/max_terminated_length": 151.42, "completions/mean_length": 254.76625, "completions/mean_terminated_length": 146.34000091552736, "completions/min_length": 238.42, "completions/min_terminated_length": 141.14, "entropy": 0.11530103281140328, "epoch": 0.24140012070006034, "frac_reward_zero_std": 0.29, "grad_norm": 0.3932775855064392, "learning_rate": 5e-05, "loss": 0.0001, "num_tokens": 66513664.0, "reward": 7.304375, "reward_std": 1.552179645895958, "rewards/event_reward_fn/mean": 7.304375, "rewards/event_reward_fn/std": 5.687906408309937, "step": 400, "step_time": 40.78123372233997 }, { "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": 0.92375, "completions/max_length": 256.0, "completions/max_terminated_length": 185.1, "completions/mean_length": 254.35875, "completions/mean_terminated_length": 178.61883544921875, "completions/min_length": 232.56, "completions/min_terminated_length": 171.12, "entropy": 0.13443249970674515, "epoch": 0.27157513578756787, "frac_reward_zero_std": 0.315, "grad_norm": 0.2284364551305771, "learning_rate": 5e-05, "loss": -0.0013, "num_tokens": 74493599.0, "reward": 7.766875, "reward_std": 1.5890911322832109, "rewards/event_reward_fn/mean": 7.766875, "rewards/event_reward_fn/std": 6.074563751220703, "step": 450, "step_time": 40.8964025861409 }, { "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": 0.983125, "completions/max_length": 256.0, "completions/max_terminated_length": 73.22, "completions/mean_length": 255.728125, "completions/mean_terminated_length": 72.48666687011719, "completions/min_length": 250.14, "completions/min_terminated_length": 70.94, "entropy": 0.1348781806230545, "epoch": 0.30175015087507545, "frac_reward_zero_std": 0.32, "grad_norm": 0.44683775305747986, "learning_rate": 5e-05, "loss": 0.0006, "num_tokens": 82766712.0, "reward": 7.835625, "reward_std": 1.6530324041843414, "rewards/event_reward_fn/mean": 7.835625, "rewards/event_reward_fn/std": 6.139980282783508, "step": 500, "step_time": 41.13054014526191 }, { "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": 0.996875, "completions/max_length": 256.0, "completions/max_terminated_length": 24.36, "completions/mean_length": 255.96125, "completions/mean_terminated_length": 24.36, "completions/min_length": 254.76, "completions/min_terminated_length": 24.36, "entropy": 0.13759294494986535, "epoch": 0.331925165962583, "frac_reward_zero_std": 0.265, "grad_norm": 0.40625813603401184, "learning_rate": 5e-05, "loss": -0.0, "num_tokens": 91249822.0, "reward": 7.55375, "reward_std": 1.706419097185135, "rewards/event_reward_fn/mean": 7.55375, "rewards/event_reward_fn/std": 5.799948143959045, "step": 550, "step_time": 42.41818935459829 }, { "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": 0.9975, "completions/max_length": 256.0, "completions/max_terminated_length": 15.02, "completions/mean_length": 255.981875, "completions/mean_terminated_length": 14.98, "completions/min_length": 255.58, "completions/min_terminated_length": 14.94, "entropy": 0.135696639418602, "epoch": 0.3621001810500905, "frac_reward_zero_std": 0.27, "grad_norm": 0.34435781836509705, "learning_rate": 5e-05, "loss": 0.0, "num_tokens": 99526582.0, "reward": 8.249375, "reward_std": 1.7093309688568115, "rewards/event_reward_fn/mean": 8.249375, "rewards/event_reward_fn/std": 6.437053818702697, "step": 600, "step_time": 42.50718544923991 }, { "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": 1.0, "completions/max_length": 256.0, "completions/max_terminated_length": 0.0, "completions/mean_length": 256.0, "completions/mean_terminated_length": 0.0, "completions/min_length": 256.0, "completions/min_terminated_length": 0.0, "entropy": 0.14394851058721542, "epoch": 0.3922751961375981, "frac_reward_zero_std": 0.3125, "grad_norm": 0.2780283987522125, "learning_rate": 5e-05, "loss": 0.0, "num_tokens": 107729490.0, "reward": 8.115625, "reward_std": 1.5425574934482575, "rewards/event_reward_fn/mean": 8.115625, "rewards/event_reward_fn/std": 6.014267163276672, "step": 650, "step_time": 60.873589005278774 }, { "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": 1.0, "completions/max_length": 256.0, "completions/max_terminated_length": 0.0, "completions/mean_length": 256.0, "completions/mean_terminated_length": 0.0, "completions/min_length": 256.0, "completions/min_terminated_length": 0.0, "entropy": 0.1508351384103298, "epoch": 0.4224502112251056, "frac_reward_zero_std": 0.275, "grad_norm": 0.21461114287376404, "learning_rate": 5e-05, "loss": 0.0, "num_tokens": 116090467.0, "reward": 8.166875, "reward_std": 1.743121521472931, "rewards/event_reward_fn/mean": 8.166875, "rewards/event_reward_fn/std": 6.155384964942932, "step": 700, "step_time": 41.52520179868036 }, { "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": 0.99875, "completions/max_length": 256.0, "completions/max_terminated_length": 9.66, "completions/mean_length": 255.981875, "completions/mean_terminated_length": 9.66, "completions/min_length": 255.42, "completions/min_terminated_length": 9.66, "entropy": 0.16852732509374618, "epoch": 0.45262522631261315, "frac_reward_zero_std": 0.2975, "grad_norm": 0.30243417620658875, "learning_rate": 5e-05, "loss": 0.0, "num_tokens": 124432282.0, "reward": 7.599375, "reward_std": 1.4733531725406648, "rewards/event_reward_fn/mean": 7.599375, "rewards/event_reward_fn/std": 5.72325975894928, "step": 750, "step_time": 45.15769186520076 }, { "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": 1.0, "completions/max_length": 256.0, "completions/max_terminated_length": 0.0, "completions/mean_length": 256.0, "completions/mean_terminated_length": 0.0, "completions/min_length": 256.0, "completions/min_terminated_length": 0.0, "entropy": 0.18111263811588288, "epoch": 0.4828002414001207, "frac_reward_zero_std": 0.28, "grad_norm": 0.3852519989013672, "learning_rate": 5e-05, "loss": 0.0, "num_tokens": 132857345.0, "reward": 8.109375, "reward_std": 1.542456374168396, "rewards/event_reward_fn/mean": 8.109375, "rewards/event_reward_fn/std": 5.9935719728469845, "step": 800, "step_time": 42.812528482141204 }, { "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": 0.99875, "completions/max_length": 256.0, "completions/max_terminated_length": 10.0, "completions/mean_length": 255.9925, "completions/mean_terminated_length": 10.0, "completions/min_length": 255.76, "completions/min_terminated_length": 10.0, "entropy": 0.21682359665632248, "epoch": 0.5129752564876282, "frac_reward_zero_std": 0.2425, "grad_norm": 0.34055572748184204, "learning_rate": 5e-05, "loss": -0.0001, "num_tokens": 141348381.0, "reward": 7.83125, "reward_std": 1.7194657081365585, "rewards/event_reward_fn/mean": 7.83125, "rewards/event_reward_fn/std": 6.176298160552978, "step": 850, "step_time": 42.62233325715526 }, { "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": 0.960625, "completions/max_length": 256.0, "completions/max_terminated_length": 101.5, "completions/mean_length": 254.8675, "completions/mean_terminated_length": 95.82615142822266, "completions/min_length": 237.06, "completions/min_terminated_length": 88.58, "entropy": 1.2534486263990403, "epoch": 0.5431502715751357, "frac_reward_zero_std": 0.22, "grad_norm": 1.9650917053222656, "learning_rate": 5e-05, "loss": 0.0044, "num_tokens": 149758372.0, "reward": 7.47875, "reward_std": 1.8692259776592255, "rewards/event_reward_fn/mean": 7.47875, "rewards/event_reward_fn/std": 5.956067395210266, "step": 900, "step_time": 41.407225477900354 }, { "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": 0.944375, "completions/max_length": 256.0, "completions/max_terminated_length": 140.22, "completions/mean_length": 253.330625, "completions/mean_terminated_length": 132.0480009460449, "completions/min_length": 209.16, "completions/min_terminated_length": 122.12, "entropy": 2.884652135372162, "epoch": 0.5733252866626434, "frac_reward_zero_std": 0.3375, "grad_norm": 2.167388916015625, "learning_rate": 5e-05, "loss": -0.0175, "num_tokens": 158149273.0, "reward": 6.6125, "reward_std": 1.6964978063106537, "rewards/event_reward_fn/mean": 6.6125, "rewards/event_reward_fn/std": 6.0062398338317875, "step": 950, "step_time": 41.2563528472418 }, { "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": 0.994375, "completions/max_length": 256.0, "completions/max_terminated_length": 31.2, "completions/mean_length": 255.541875, "completions/mean_terminated_length": 31.16, "completions/min_length": 246.16, "completions/min_terminated_length": 31.12, "entropy": 1.2761903527379035, "epoch": 0.6035003017501509, "frac_reward_zero_std": 0.3825, "grad_norm": 0.4628017842769623, "learning_rate": 5e-05, "loss": -0.0033, "num_tokens": 166538998.0, "reward": 6.94875, "reward_std": 1.378657329082489, "rewards/event_reward_fn/mean": 6.94875, "rewards/event_reward_fn/std": 6.216048922538757, "step": 1000, "step_time": 42.322889749883906 }, { "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": 0.984375, "completions/max_length": 256.0, "completions/max_terminated_length": 61.4, "completions/mean_length": 255.64625, "completions/mean_terminated_length": 60.46533386230469, "completions/min_length": 248.46, "completions/min_terminated_length": 59.02, "entropy": 0.4890740931034088, "epoch": 0.6336753168376584, "frac_reward_zero_std": 0.395, "grad_norm": 0.33829060196876526, "learning_rate": 5e-05, "loss": -0.0003, "num_tokens": 174874216.0, "reward": 6.793125, "reward_std": 1.1867496293783188, "rewards/event_reward_fn/mean": 6.793125, "rewards/event_reward_fn/std": 5.404484539031983, "step": 1050, "step_time": 42.91726479450008 }, { "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": 0.993125, "completions/max_length": 256.0, "completions/max_terminated_length": 28.78, "completions/mean_length": 255.443125, "completions/mean_terminated_length": 28.65, "completions/min_length": 243.56, "completions/min_terminated_length": 28.52, "entropy": 0.7336188541352748, "epoch": 0.663850331925166, "frac_reward_zero_std": 0.3925, "grad_norm": 0.5986895561218262, "learning_rate": 5e-05, "loss": -0.0053, "num_tokens": 183293167.0, "reward": 6.371875, "reward_std": 1.27341972053051, "rewards/event_reward_fn/mean": 6.371875, "rewards/event_reward_fn/std": 5.160589256286621, "step": 1100, "step_time": 42.857495513077595 }, { "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": 0.99875, "completions/max_length": 256.0, "completions/max_terminated_length": 8.74, "completions/mean_length": 255.953125, "completions/mean_terminated_length": 8.74, "completions/min_length": 254.5, "completions/min_terminated_length": 8.74, "entropy": 0.41359326869249347, "epoch": 0.6940253470126735, "frac_reward_zero_std": 0.365, "grad_norm": 0.714463472366333, "learning_rate": 5e-05, "loss": 0.0001, "num_tokens": 191713508.0, "reward": 6.96625, "reward_std": 1.294020129442215, "rewards/event_reward_fn/mean": 6.96625, "rewards/event_reward_fn/std": 5.736661648750305, "step": 1150, "step_time": 41.747385050542654 }, { "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": 0.98875, "completions/max_length": 256.0, "completions/max_terminated_length": 64.84, "completions/mean_length": 255.530625, "completions/mean_terminated_length": 64.83, "completions/min_length": 244.02, "completions/min_terminated_length": 64.82, "entropy": 0.7842025232315063, "epoch": 0.724200362100181, "frac_reward_zero_std": 0.3825, "grad_norm": 2.366915225982666, "learning_rate": 5e-05, "loss": 0.0001, "num_tokens": 199636639.0, "reward": 7.296875, "reward_std": 1.2271459007263184, "rewards/event_reward_fn/mean": 7.296875, "rewards/event_reward_fn/std": 5.948999562263489, "step": 1200, "step_time": 41.03189678700059 }, { "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": 0.99875, "completions/max_length": 256.0, "completions/max_terminated_length": 8.64, "completions/mean_length": 255.95, "completions/mean_terminated_length": 8.64, "completions/min_length": 254.4, "completions/min_terminated_length": 8.64, "entropy": 1.2313472920656203, "epoch": 0.7543753771876885, "frac_reward_zero_std": 0.35, "grad_norm": 2.256929636001587, "learning_rate": 5e-05, "loss": 0.0001, "num_tokens": 207997095.0, "reward": 6.69375, "reward_std": 1.492494255900383, "rewards/event_reward_fn/mean": 6.69375, "rewards/event_reward_fn/std": 5.9429325056076046, "step": 1250, "step_time": 42.135132130276176 }, { "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": 0.98875, "completions/max_length": 256.0, "completions/max_terminated_length": 45.68, "completions/mean_length": 255.489375, "completions/mean_terminated_length": 45.05666687011719, "completions/min_length": 244.34, "completions/min_terminated_length": 44.66, "entropy": 1.172162665054202, "epoch": 0.7845503922751962, "frac_reward_zero_std": 0.36, "grad_norm": 0.8358303904533386, "learning_rate": 5e-05, "loss": -0.0024, "num_tokens": 216307264.0, "reward": 6.914375, "reward_std": 1.4417870903015138, "rewards/event_reward_fn/mean": 6.914375, "rewards/event_reward_fn/std": 5.760623688697815, "step": 1300, "step_time": 41.30298829050036 }, { "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": 0.996875, "completions/max_length": 256.0, "completions/max_terminated_length": 18.28, "completions/mean_length": 255.91, "completions/mean_terminated_length": 18.09, "completions/min_length": 253.42, "completions/min_terminated_length": 17.9, "entropy": 1.4694241133332253, "epoch": 0.8147254073627037, "frac_reward_zero_std": 0.34, "grad_norm": 6.394040107727051, "learning_rate": 5e-05, "loss": -0.0003, "num_tokens": 224510279.0, "reward": 6.273125, "reward_std": 1.626619552373886, "rewards/event_reward_fn/mean": 6.273125, "rewards/event_reward_fn/std": 5.248045358657837, "step": 1350, "step_time": 40.860976390804865 }, { "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": 1.0, "completions/max_length": 256.0, "completions/max_terminated_length": 0.0, "completions/mean_length": 256.0, "completions/mean_terminated_length": 0.0, "completions/min_length": 256.0, "completions/min_terminated_length": 0.0, "entropy": 2.0628131467103956, "epoch": 0.8449004224502112, "frac_reward_zero_std": 0.3375, "grad_norm": 1.5476884841918945, "learning_rate": 5e-05, "loss": 0.0, "num_tokens": 232969228.0, "reward": 6.45875, "reward_std": 1.7393629193305968, "rewards/event_reward_fn/mean": 6.45875, "rewards/event_reward_fn/std": 5.68415337562561, "step": 1400, "step_time": 43.151147363660854 }, { "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": 0.999375, "completions/max_length": 256.0, "completions/max_terminated_length": 0.04, "completions/mean_length": 255.84125, "completions/mean_terminated_length": 0.04, "completions/min_length": 250.92, "completions/min_terminated_length": 0.04, "entropy": 1.3854397583007811, "epoch": 0.8750754375377188, "frac_reward_zero_std": 0.3375, "grad_norm": 3.6241378784179688, "learning_rate": 5e-05, "loss": -0.001, "num_tokens": 241277888.0, "reward": 6.719375, "reward_std": 1.752496111392975, "rewards/event_reward_fn/mean": 6.719375, "rewards/event_reward_fn/std": 6.006656441688538, "step": 1450, "step_time": 41.94062339906115 }, { "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": 0.994375, "completions/max_length": 256.0, "completions/max_terminated_length": 6.86, "completions/mean_length": 254.820625, "completions/mean_terminated_length": 6.8333333349227905, "completions/min_length": 237.22, "completions/min_terminated_length": 6.82, "entropy": 1.1917432191967965, "epoch": 0.9052504526252263, "frac_reward_zero_std": 0.36, "grad_norm": 7.937113285064697, "learning_rate": 5e-05, "loss": -0.0277, "num_tokens": 249570792.0, "reward": 6.343125, "reward_std": 1.5732547068595886, "rewards/event_reward_fn/mean": 6.343125, "rewards/event_reward_fn/std": 5.813223929405212, "step": 1500, "step_time": 43.07453210723819 }, { "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": 0.926875, "completions/max_length": 256.0, "completions/max_terminated_length": 131.5, "completions/mean_length": 253.175, "completions/mean_terminated_length": 125.9899688720703, "completions/min_length": 228.9, "completions/min_terminated_length": 121.38, "entropy": 0.5228479199111462, "epoch": 0.9354254677127338, "frac_reward_zero_std": 0.3875, "grad_norm": 1.1253968477249146, "learning_rate": 5e-05, "loss": 0.0002, "num_tokens": 257734094.0, "reward": 7.513125, "reward_std": 1.2554410457611085, "rewards/event_reward_fn/mean": 7.513125, "rewards/event_reward_fn/std": 5.942786226272583, "step": 1550, "step_time": 42.567975415034454 }, { "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": 0.926875, "completions/max_length": 256.0, "completions/max_terminated_length": 131.88, "completions/mean_length": 253.125625, "completions/mean_terminated_length": 128.5973336791992, "completions/min_length": 223.42, "completions/min_terminated_length": 126.14, "entropy": 0.8324106151610613, "epoch": 0.9656004828002414, "frac_reward_zero_std": 0.35, "grad_norm": 0.8266918659210205, "learning_rate": 5e-05, "loss": -0.0003, "num_tokens": 266238921.0, "reward": 6.545625, "reward_std": 1.3680988204479219, "rewards/event_reward_fn/mean": 6.545625, "rewards/event_reward_fn/std": 5.431567845344543, "step": 1600, "step_time": 42.35816644520266 }, { "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": 0.979375, "completions/max_length": 256.0, "completions/max_terminated_length": 52.06, "completions/mean_length": 255.3625, "completions/mean_terminated_length": 50.906666870117185, "completions/min_length": 244.64, "completions/min_terminated_length": 50.08, "entropy": 1.4035920506715776, "epoch": 0.995775497887749, "frac_reward_zero_std": 0.3475, "grad_norm": 11.164112091064453, "learning_rate": 5e-05, "loss": -0.0041, "num_tokens": 274713769.0, "reward": 6.30375, "reward_std": 1.4735591614246368, "rewards/event_reward_fn/mean": 6.30375, "rewards/event_reward_fn/std": 5.667113132476807, "step": 1650, "step_time": 41.3417172247381 }, { "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": 0.99875, "completions/max_length": 256.0, "completions/max_terminated_length": 4.16, "completions/mean_length": 255.81, "completions/mean_terminated_length": 4.16, "completions/min_length": 249.92, "completions/min_terminated_length": 4.16, "entropy": 1.0942185708135366, "epoch": 1.0259505129752564, "frac_reward_zero_std": 0.4125, "grad_norm": 2.0238513946533203, "learning_rate": 5e-05, "loss": -0.0042, "num_tokens": 282976766.0, "reward": 7.19375, "reward_std": 1.4945010322332382, "rewards/event_reward_fn/mean": 7.19375, "rewards/event_reward_fn/std": 5.62495879650116, "step": 1700, "step_time": 41.58126259226352 }, { "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": 0.995, "completions/max_length": 256.0, "completions/max_terminated_length": 27.9, "completions/mean_length": 255.874375, "completions/mean_terminated_length": 27.873333435058594, "completions/min_length": 253.14, "completions/min_terminated_length": 27.86, "entropy": 0.6756551740318537, "epoch": 1.056125528062764, "frac_reward_zero_std": 0.4125, "grad_norm": 0.9028272032737732, "learning_rate": 5e-05, "loss": 0.0002, "num_tokens": 291424307.0, "reward": 6.863125, "reward_std": 1.4053943872451782, "rewards/event_reward_fn/mean": 6.863125, "rewards/event_reward_fn/std": 5.494638476371765, "step": 1750, "step_time": 41.3957015178015 }, { "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": 0.985, "completions/max_length": 256.0, "completions/max_terminated_length": 71.64, "completions/mean_length": 255.53375, "completions/mean_terminated_length": 71.1, "completions/min_length": 244.52, "completions/min_terminated_length": 70.44, "entropy": 0.8187148047238588, "epoch": 1.0863005431502715, "frac_reward_zero_std": 0.3525, "grad_norm": 4.235013484954834, "learning_rate": 5e-05, "loss": -0.0009, "num_tokens": 300026957.0, "reward": 7.29375, "reward_std": 1.5134036219120026, "rewards/event_reward_fn/mean": 7.29375, "rewards/event_reward_fn/std": 6.304830470085144, "step": 1800, "step_time": 42.75003189910087 }, { "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": 0.9275, "completions/max_length": 256.0, "completions/max_terminated_length": 154.7, "completions/mean_length": 253.389375, "completions/mean_terminated_length": 145.82189025878907, "completions/min_length": 221.36, "completions/min_terminated_length": 134.32, "entropy": 0.8770341634750366, "epoch": 1.1164755582377792, "frac_reward_zero_std": 0.42, "grad_norm": 1.056412696838379, "learning_rate": 5e-05, "loss": 0.0016, "num_tokens": 308294562.0, "reward": 7.195, "reward_std": 1.413882914185524, "rewards/event_reward_fn/mean": 7.195, "rewards/event_reward_fn/std": 5.875013728141784, "step": 1850, "step_time": 42.22391830024426 }, { "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": 0.9925, "completions/max_length": 256.0, "completions/max_terminated_length": 50.76, "completions/mean_length": 255.740625, "completions/mean_terminated_length": 49.8, "completions/min_length": 248.52, "completions/min_terminated_length": 48.84, "entropy": 0.7402717351168394, "epoch": 1.1466505733252867, "frac_reward_zero_std": 0.3725, "grad_norm": 1.3885732889175415, "learning_rate": 5e-05, "loss": 0.0007, "num_tokens": 316652055.0, "reward": 7.290625, "reward_std": 1.5660697519779205, "rewards/event_reward_fn/mean": 7.290625, "rewards/event_reward_fn/std": 6.019521760940552, "step": 1900, "step_time": 42.37986288452754 }, { "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": 0.99375, "completions/max_length": 256.0, "completions/max_terminated_length": 35.46, "completions/mean_length": 255.766875, "completions/mean_terminated_length": 35.0, "completions/min_length": 249.58, "completions/min_terminated_length": 34.54, "entropy": 1.1067684018611907, "epoch": 1.1768255884127943, "frac_reward_zero_std": 0.3925, "grad_norm": 1.0810959339141846, "learning_rate": 5e-05, "loss": 0.0014, "num_tokens": 325016448.0, "reward": 6.665, "reward_std": 1.538350248336792, "rewards/event_reward_fn/mean": 6.665, "rewards/event_reward_fn/std": 6.030116739273072, "step": 1950, "step_time": 40.92805422256584 }, { "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": 0.990625, "completions/max_length": 256.0, "completions/max_terminated_length": 63.46, "completions/mean_length": 255.716875, "completions/mean_terminated_length": 63.21, "completions/min_length": 247.28, "completions/min_terminated_length": 62.96, "entropy": 0.7298687703162432, "epoch": 1.2070006035003018, "frac_reward_zero_std": 0.3375, "grad_norm": 0.858845591545105, "learning_rate": 5e-05, "loss": 0.0002, "num_tokens": 333296057.0, "reward": 7.73125, "reward_std": 1.5106933176517487, "rewards/event_reward_fn/mean": 7.73125, "rewards/event_reward_fn/std": 6.448639197349548, "step": 2000, "step_time": 41.15218346009729 } ], "logging_steps": 50, "max_steps": 16570, "num_input_tokens_seen": 333296057, "num_train_epochs": 10, "save_steps": 250, "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": 4, "trial_name": null, "trial_params": null }