| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 2.0, |
| "eval_steps": 500, |
| "global_step": 1492, |
| "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": 0.05, |
| "completions/max_length": 246.3, |
| "completions/max_terminated_length": 240.26, |
| "completions/mean_length": 195.0275, |
| "completions/mean_terminated_length": 191.87969146728517, |
| "completions/min_length": 147.76, |
| "completions/min_terminated_length": 147.76, |
| "entropy": 0.06807037293910981, |
| "epoch": 0.06702412868632708, |
| "frac_reward_zero_std": 0.4475, |
| "grad_norm": 0.1978774070739746, |
| "learning_rate": 1e-05, |
| "loss": -0.0022, |
| "num_tokens": 6268258.0, |
| "reward": 12.489985446929932, |
| "reward_std": 1.05244723290205, |
| "rewards/event_reward_fn/mean": 11.62375, |
| "rewards/event_reward_fn/std": 7.598931360244751, |
| "rewards/format_reward_fn/mean": 0.8662354218959808, |
| "rewards/format_reward_fn/std": 0.24084076710045338, |
| "step": 50, |
| "step_time": 24.881226640827954 |
| }, |
| { |
| "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": 0.043125, |
| "completions/max_length": 249.38, |
| "completions/max_terminated_length": 244.26, |
| "completions/mean_length": 198.4925, |
| "completions/mean_terminated_length": 195.8674203491211, |
| "completions/min_length": 155.1, |
| "completions/min_terminated_length": 155.1, |
| "entropy": 0.07096008479595184, |
| "epoch": 0.13404825737265416, |
| "frac_reward_zero_std": 0.42, |
| "grad_norm": 0.31616032123565674, |
| "learning_rate": 1e-05, |
| "loss": -0.0052, |
| "num_tokens": 12603730.0, |
| "reward": 11.722552404403686, |
| "reward_std": 1.104598103761673, |
| "rewards/event_reward_fn/mean": 10.865, |
| "rewards/event_reward_fn/std": 7.203483366966248, |
| "rewards/format_reward_fn/mean": 0.8575523483753205, |
| "rewards/format_reward_fn/std": 0.25920433282852173, |
| "step": 100, |
| "step_time": 23.881343694739044 |
| }, |
| { |
| "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": 0.069375, |
| "completions/max_length": 251.04, |
| "completions/max_terminated_length": 245.08, |
| "completions/mean_length": 201.58125, |
| "completions/mean_terminated_length": 197.60445678710937, |
| "completions/min_length": 157.96, |
| "completions/min_terminated_length": 157.96, |
| "entropy": 0.07228697955608368, |
| "epoch": 0.20107238605898123, |
| "frac_reward_zero_std": 0.41, |
| "grad_norm": 0.1767224669456482, |
| "learning_rate": 1e-05, |
| "loss": 0.002, |
| "num_tokens": 19236102.0, |
| "reward": 11.989666719436645, |
| "reward_std": 1.2850025883316993, |
| "rewards/event_reward_fn/mean": 11.1225, |
| "rewards/event_reward_fn/std": 7.3152674865722656, |
| "rewards/format_reward_fn/mean": 0.8671666479110718, |
| "rewards/format_reward_fn/std": 0.24983404949307442, |
| "step": 150, |
| "step_time": 27.783113366477192 |
| }, |
| { |
| "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": 250.62, |
| "completions/max_terminated_length": 244.54, |
| "completions/mean_length": 201.1025, |
| "completions/mean_terminated_length": 197.25198516845703, |
| "completions/min_length": 156.4, |
| "completions/min_terminated_length": 156.4, |
| "entropy": 0.06773373357951641, |
| "epoch": 0.2680965147453083, |
| "frac_reward_zero_std": 0.415, |
| "grad_norm": 0.13261352479457855, |
| "learning_rate": 1e-05, |
| "loss": -0.0029, |
| "num_tokens": 25426958.0, |
| "reward": 12.467143926620484, |
| "reward_std": 1.1554639112949372, |
| "rewards/event_reward_fn/mean": 11.59875, |
| "rewards/event_reward_fn/std": 7.149877543449402, |
| "rewards/format_reward_fn/mean": 0.8683938610553742, |
| "rewards/format_reward_fn/std": 0.24253679752349855, |
| "step": 200, |
| "step_time": 24.421198091395198 |
| }, |
| { |
| "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": 0.058125, |
| "completions/max_length": 250.22, |
| "completions/max_terminated_length": 243.8, |
| "completions/mean_length": 200.303125, |
| "completions/mean_terminated_length": 196.90248321533204, |
| "completions/min_length": 162.42, |
| "completions/min_terminated_length": 162.42, |
| "entropy": 0.06486415289342404, |
| "epoch": 0.3351206434316354, |
| "frac_reward_zero_std": 0.385, |
| "grad_norm": 0.49442073702812195, |
| "learning_rate": 1e-05, |
| "loss": -0.0036, |
| "num_tokens": 31582342.0, |
| "reward": 12.355808296203612, |
| "reward_std": 1.1142808997631073, |
| "rewards/event_reward_fn/mean": 11.48875, |
| "rewards/event_reward_fn/std": 7.448825697898865, |
| "rewards/format_reward_fn/mean": 0.8670582604408265, |
| "rewards/format_reward_fn/std": 0.24978963822126388, |
| "step": 250, |
| "step_time": 25.453000083304943 |
| }, |
| { |
| "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": 0.04875, |
| "completions/max_length": 248.68, |
| "completions/max_terminated_length": 244.22, |
| "completions/mean_length": 198.759375, |
| "completions/mean_terminated_length": 196.16592681884765, |
| "completions/min_length": 156.2, |
| "completions/min_terminated_length": 156.2, |
| "entropy": 0.0681518343836069, |
| "epoch": 0.40214477211796246, |
| "frac_reward_zero_std": 0.39, |
| "grad_norm": 0.48775437474250793, |
| "learning_rate": 1e-05, |
| "loss": -0.0057, |
| "num_tokens": 37800719.0, |
| "reward": 12.434584522247315, |
| "reward_std": 1.183589797616005, |
| "rewards/event_reward_fn/mean": 11.56375, |
| "rewards/event_reward_fn/std": 7.52141658782959, |
| "rewards/format_reward_fn/mean": 0.8708344352245331, |
| "rewards/format_reward_fn/std": 0.23306368254125118, |
| "step": 300, |
| "step_time": 25.360634116120636 |
| }, |
| { |
| "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": 0.034375, |
| "completions/max_length": 248.34, |
| "completions/max_terminated_length": 245.28, |
| "completions/mean_length": 203.264375, |
| "completions/mean_terminated_length": 201.32774475097656, |
| "completions/min_length": 157.54, |
| "completions/min_terminated_length": 157.54, |
| "entropy": 0.06739457175135613, |
| "epoch": 0.4691689008042895, |
| "frac_reward_zero_std": 0.3525, |
| "grad_norm": 0.33356958627700806, |
| "learning_rate": 1e-05, |
| "loss": -0.004, |
| "num_tokens": 44150011.0, |
| "reward": 13.173797435760498, |
| "reward_std": 1.2946509444713592, |
| "rewards/event_reward_fn/mean": 12.28875, |
| "rewards/event_reward_fn/std": 7.145490102767944, |
| "rewards/format_reward_fn/mean": 0.885047378540039, |
| "rewards/format_reward_fn/std": 0.22108205765485764, |
| "step": 350, |
| "step_time": 26.940150288008155 |
| }, |
| { |
| "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": 0.064375, |
| "completions/max_length": 252.9, |
| "completions/max_terminated_length": 247.8, |
| "completions/mean_length": 203.064375, |
| "completions/mean_terminated_length": 199.5350747680664, |
| "completions/min_length": 158.26, |
| "completions/min_terminated_length": 158.26, |
| "entropy": 0.0657703248411417, |
| "epoch": 0.5361930294906166, |
| "frac_reward_zero_std": 0.435, |
| "grad_norm": 0.26359474658966064, |
| "learning_rate": 1e-05, |
| "loss": -0.0021, |
| "num_tokens": 50384400.0, |
| "reward": 12.238037357330322, |
| "reward_std": 1.057584773004055, |
| "rewards/event_reward_fn/mean": 11.37, |
| "rewards/event_reward_fn/std": 7.154304637908935, |
| "rewards/format_reward_fn/mean": 0.8680373668670655, |
| "rewards/format_reward_fn/std": 0.26109003871679304, |
| "step": 400, |
| "step_time": 25.59800311360508 |
| }, |
| { |
| "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": 0.049375, |
| "completions/max_length": 249.06, |
| "completions/max_terminated_length": 244.9, |
| "completions/mean_length": 203.706875, |
| "completions/mean_terminated_length": 200.99220581054686, |
| "completions/min_length": 161.06, |
| "completions/min_terminated_length": 161.06, |
| "entropy": 0.06626586891710758, |
| "epoch": 0.6032171581769437, |
| "frac_reward_zero_std": 0.3775, |
| "grad_norm": 0.48660293221473694, |
| "learning_rate": 1e-05, |
| "loss": -0.004, |
| "num_tokens": 56771056.0, |
| "reward": 13.009743461608887, |
| "reward_std": 1.2429037857055665, |
| "rewards/event_reward_fn/mean": 12.130625, |
| "rewards/event_reward_fn/std": 7.234463820457458, |
| "rewards/format_reward_fn/mean": 0.8791184043884277, |
| "rewards/format_reward_fn/std": 0.23800445690751076, |
| "step": 450, |
| "step_time": 25.550446799769997 |
| }, |
| { |
| "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": 0.066875, |
| "completions/max_length": 251.92, |
| "completions/max_terminated_length": 246.12, |
| "completions/mean_length": 204.07625, |
| "completions/mean_terminated_length": 200.35590240478516, |
| "completions/min_length": 160.74, |
| "completions/min_terminated_length": 160.74, |
| "entropy": 0.06663089752197265, |
| "epoch": 0.6702412868632708, |
| "frac_reward_zero_std": 0.4025, |
| "grad_norm": 0.6319305300712585, |
| "learning_rate": 1e-05, |
| "loss": -0.0042, |
| "num_tokens": 63078757.0, |
| "reward": 12.313038005828858, |
| "reward_std": 1.1368902394175529, |
| "rewards/event_reward_fn/mean": 11.4575, |
| "rewards/event_reward_fn/std": 6.7143393945693965, |
| "rewards/format_reward_fn/mean": 0.8555380630493165, |
| "rewards/format_reward_fn/std": 0.2657873314619064, |
| "step": 500, |
| "step_time": 26.24973841637373 |
| }, |
| { |
| "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": 0.091875, |
| "completions/max_length": 252.82, |
| "completions/max_terminated_length": 246.88, |
| "completions/mean_length": 203.815, |
| "completions/mean_terminated_length": 198.69242126464843, |
| "completions/min_length": 161.16, |
| "completions/min_terminated_length": 161.16, |
| "entropy": 0.06187104433774948, |
| "epoch": 0.7372654155495979, |
| "frac_reward_zero_std": 0.425, |
| "grad_norm": 0.40395304560661316, |
| "learning_rate": 1e-05, |
| "loss": -0.0025, |
| "num_tokens": 69170452.0, |
| "reward": 12.482298536300659, |
| "reward_std": 1.0457301473617553, |
| "rewards/event_reward_fn/mean": 11.64625, |
| "rewards/event_reward_fn/std": 7.317771224975586, |
| "rewards/format_reward_fn/mean": 0.8360484623908997, |
| "rewards/format_reward_fn/std": 0.2895883430540562, |
| "step": 550, |
| "step_time": 24.193240740820766 |
| }, |
| { |
| "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": 0.110625, |
| "completions/max_length": 252.62, |
| "completions/max_terminated_length": 246.8, |
| "completions/mean_length": 208.275625, |
| "completions/mean_terminated_length": 202.49910614013672, |
| "completions/min_length": 165.54, |
| "completions/min_terminated_length": 165.54, |
| "entropy": 0.0649487990140915, |
| "epoch": 0.8042895442359249, |
| "frac_reward_zero_std": 0.38, |
| "grad_norm": 0.37119486927986145, |
| "learning_rate": 1e-05, |
| "loss": 0.0006, |
| "num_tokens": 75499314.0, |
| "reward": 12.80059557914734, |
| "reward_std": 1.1889909988641738, |
| "rewards/event_reward_fn/mean": 11.97375, |
| "rewards/event_reward_fn/std": 7.475857477188111, |
| "rewards/format_reward_fn/mean": 0.8268455564975739, |
| "rewards/format_reward_fn/std": 0.29714462146162984, |
| "step": 600, |
| "step_time": 24.3176869976148 |
| }, |
| { |
| "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": 0.05625, |
| "completions/max_length": 249.28, |
| "completions/max_terminated_length": 244.8, |
| "completions/mean_length": 202.789375, |
| "completions/mean_terminated_length": 199.91522064208985, |
| "completions/min_length": 161.74, |
| "completions/min_terminated_length": 161.74, |
| "entropy": 0.06481640346348286, |
| "epoch": 0.871313672922252, |
| "frac_reward_zero_std": 0.3975, |
| "grad_norm": 0.08866075426340103, |
| "learning_rate": 1e-05, |
| "loss": -0.0023, |
| "num_tokens": 81673001.0, |
| "reward": 12.689926280975342, |
| "reward_std": 1.2458794575929641, |
| "rewards/event_reward_fn/mean": 11.815625, |
| "rewards/event_reward_fn/std": 7.275726590156555, |
| "rewards/format_reward_fn/mean": 0.8743013119697571, |
| "rewards/format_reward_fn/std": 0.23756251022219657, |
| "step": 650, |
| "step_time": 25.04028965227306 |
| }, |
| { |
| "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": 0.100625, |
| "completions/max_length": 253.72, |
| "completions/max_terminated_length": 248.28, |
| "completions/mean_length": 205.536875, |
| "completions/mean_terminated_length": 200.1349432373047, |
| "completions/min_length": 162.16, |
| "completions/min_terminated_length": 162.16, |
| "entropy": 0.0658975774794817, |
| "epoch": 0.938337801608579, |
| "frac_reward_zero_std": 0.3975, |
| "grad_norm": 0.2268964648246765, |
| "learning_rate": 1e-05, |
| "loss": -0.0008, |
| "num_tokens": 87934795.0, |
| "reward": 12.72035478591919, |
| "reward_std": 1.1722034803032875, |
| "rewards/event_reward_fn/mean": 11.888125, |
| "rewards/event_reward_fn/std": 7.583159003257752, |
| "rewards/format_reward_fn/mean": 0.8322297859191895, |
| "rewards/format_reward_fn/std": 0.29026631206274034, |
| "step": 700, |
| "step_time": 24.744350045956672 |
| }, |
| { |
| "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": 0.110625, |
| "completions/max_length": 253.56, |
| "completions/max_terminated_length": 248.12, |
| "completions/mean_length": 205.06125, |
| "completions/mean_terminated_length": 198.67494079589844, |
| "completions/min_length": 155.32, |
| "completions/min_terminated_length": 155.32, |
| "entropy": 0.06425789251923561, |
| "epoch": 1.0053619302949062, |
| "frac_reward_zero_std": 0.4325, |
| "grad_norm": 0.6108524799346924, |
| "learning_rate": 1e-05, |
| "loss": -0.0016, |
| "num_tokens": 94007519.0, |
| "reward": 12.451499423980714, |
| "reward_std": 1.17638818860054, |
| "rewards/event_reward_fn/mean": 11.630625, |
| "rewards/event_reward_fn/std": 7.498000311851501, |
| "rewards/format_reward_fn/mean": 0.8208743929862976, |
| "rewards/format_reward_fn/std": 0.3162671920657158, |
| "step": 750, |
| "step_time": 25.037732787020506 |
| }, |
| { |
| "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": 0.11625, |
| "completions/max_length": 252.56, |
| "completions/max_terminated_length": 246.16, |
| "completions/mean_length": 205.586875, |
| "completions/mean_terminated_length": 199.2539434814453, |
| "completions/min_length": 159.14, |
| "completions/min_terminated_length": 159.14, |
| "entropy": 0.06738013252615929, |
| "epoch": 1.0723860589812333, |
| "frac_reward_zero_std": 0.4275, |
| "grad_norm": 0.1556655317544937, |
| "learning_rate": 1e-05, |
| "loss": -0.0027, |
| "num_tokens": 100291983.0, |
| "reward": 12.259195594787597, |
| "reward_std": 1.0736777836084366, |
| "rewards/event_reward_fn/mean": 11.440625, |
| "rewards/event_reward_fn/std": 7.4294485759735105, |
| "rewards/format_reward_fn/mean": 0.8185705220699311, |
| "rewards/format_reward_fn/std": 0.30390245616436007, |
| "step": 800, |
| "step_time": 25.602326317727567 |
| }, |
| { |
| "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": 0.124375, |
| "completions/max_length": 253.96, |
| "completions/max_terminated_length": 248.22, |
| "completions/mean_length": 207.960625, |
| "completions/mean_terminated_length": 201.18785888671874, |
| "completions/min_length": 163.16, |
| "completions/min_terminated_length": 163.16, |
| "entropy": 0.06615202344954013, |
| "epoch": 1.1394101876675604, |
| "frac_reward_zero_std": 0.4075, |
| "grad_norm": 0.34817561507225037, |
| "learning_rate": 1e-05, |
| "loss": -0.0008, |
| "num_tokens": 106570401.0, |
| "reward": 13.199323978424072, |
| "reward_std": 1.0555983385443688, |
| "rewards/event_reward_fn/mean": 12.388125, |
| "rewards/event_reward_fn/std": 7.955943965911866, |
| "rewards/format_reward_fn/mean": 0.8111989140510559, |
| "rewards/format_reward_fn/std": 0.318571160286665, |
| "step": 850, |
| "step_time": 25.24974968288094 |
| }, |
| { |
| "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": 0.091875, |
| "completions/max_length": 254.2, |
| "completions/max_terminated_length": 248.06, |
| "completions/mean_length": 204.960625, |
| "completions/mean_terminated_length": 199.68820373535155, |
| "completions/min_length": 160.64, |
| "completions/min_terminated_length": 160.64, |
| "entropy": 0.06304333973675966, |
| "epoch": 1.2064343163538873, |
| "frac_reward_zero_std": 0.4475, |
| "grad_norm": 0.70001620054245, |
| "learning_rate": 1e-05, |
| "loss": -0.0014, |
| "num_tokens": 112801556.0, |
| "reward": 12.732025756835938, |
| "reward_std": 1.1594133710861205, |
| "rewards/event_reward_fn/mean": 11.89, |
| "rewards/event_reward_fn/std": 7.514157109260559, |
| "rewards/format_reward_fn/mean": 0.8420258605480194, |
| "rewards/format_reward_fn/std": 0.28636473283171654, |
| "step": 900, |
| "step_time": 26.33290139209479 |
| }, |
| { |
| "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": 0.093125, |
| "completions/max_length": 253.24, |
| "completions/max_terminated_length": 248.26, |
| "completions/mean_length": 208.87875, |
| "completions/mean_terminated_length": 204.2489434814453, |
| "completions/min_length": 162.6, |
| "completions/min_terminated_length": 162.6, |
| "entropy": 0.06733124047517776, |
| "epoch": 1.2734584450402144, |
| "frac_reward_zero_std": 0.3875, |
| "grad_norm": 0.42013463377952576, |
| "learning_rate": 1e-05, |
| "loss": -0.0016, |
| "num_tokens": 119241388.0, |
| "reward": 13.07864158630371, |
| "reward_std": 1.2291508412361145, |
| "rewards/event_reward_fn/mean": 12.24375, |
| "rewards/event_reward_fn/std": 7.096493782997132, |
| "rewards/format_reward_fn/mean": 0.8348916172981262, |
| "rewards/format_reward_fn/std": 0.29221967339515686, |
| "step": 950, |
| "step_time": 26.795288713425396 |
| }, |
| { |
| "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": 0.115, |
| "completions/max_length": 254.4, |
| "completions/max_terminated_length": 248.28, |
| "completions/mean_length": 208.54375, |
| "completions/mean_terminated_length": 202.61700103759765, |
| "completions/min_length": 164.5, |
| "completions/min_terminated_length": 164.5, |
| "entropy": 0.0674970293045044, |
| "epoch": 1.3404825737265416, |
| "frac_reward_zero_std": 0.4125, |
| "grad_norm": 0.22458799183368683, |
| "learning_rate": 1e-05, |
| "loss": -0.001, |
| "num_tokens": 125595072.0, |
| "reward": 13.164627075195312, |
| "reward_std": 1.1512372946739198, |
| "rewards/event_reward_fn/mean": 12.338125, |
| "rewards/event_reward_fn/std": 7.617123994827271, |
| "rewards/format_reward_fn/mean": 0.8265021049976349, |
| "rewards/format_reward_fn/std": 0.3076726086437702, |
| "step": 1000, |
| "step_time": 26.641401502527298 |
| }, |
| { |
| "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": 0.1175, |
| "completions/max_length": 254.24, |
| "completions/max_terminated_length": 247.84, |
| "completions/mean_length": 208.166875, |
| "completions/mean_terminated_length": 202.1624597167969, |
| "completions/min_length": 165.56, |
| "completions/min_terminated_length": 165.56, |
| "entropy": 0.06616773374378682, |
| "epoch": 1.4075067024128687, |
| "frac_reward_zero_std": 0.37, |
| "grad_norm": 0.34188446402549744, |
| "learning_rate": 1e-05, |
| "loss": -0.0015, |
| "num_tokens": 131697642.0, |
| "reward": 12.04176643371582, |
| "reward_std": 1.14944515645504, |
| "rewards/event_reward_fn/mean": 11.2275, |
| "rewards/event_reward_fn/std": 6.602086658477783, |
| "rewards/format_reward_fn/mean": 0.8142664790153503, |
| "rewards/format_reward_fn/std": 0.31338023476302623, |
| "step": 1050, |
| "step_time": 23.740581118687988 |
| }, |
| { |
| "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": 0.14625, |
| "completions/max_length": 254.62, |
| "completions/max_terminated_length": 246.4, |
| "completions/mean_length": 210.695, |
| "completions/mean_terminated_length": 203.1378680419922, |
| "completions/min_length": 165.5, |
| "completions/min_terminated_length": 165.5, |
| "entropy": 0.06749342102557421, |
| "epoch": 1.4745308310991958, |
| "frac_reward_zero_std": 0.405, |
| "grad_norm": 0.052085984498262405, |
| "learning_rate": 1e-05, |
| "loss": 0.0009, |
| "num_tokens": 137994889.0, |
| "reward": 12.919311103820801, |
| "reward_std": 1.1956829646229743, |
| "rewards/event_reward_fn/mean": 12.114375, |
| "rewards/event_reward_fn/std": 7.290353746414184, |
| "rewards/format_reward_fn/mean": 0.8049361062049866, |
| "rewards/format_reward_fn/std": 0.34213557571172715, |
| "step": 1100, |
| "step_time": 25.770162526927887 |
| }, |
| { |
| "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": 0.089375, |
| "completions/max_length": 253.1, |
| "completions/max_terminated_length": 248.18, |
| "completions/mean_length": 208.190625, |
| "completions/mean_terminated_length": 203.62595764160156, |
| "completions/min_length": 165.04, |
| "completions/min_terminated_length": 165.04, |
| "entropy": 0.06642140716314315, |
| "epoch": 1.5415549597855227, |
| "frac_reward_zero_std": 0.4025, |
| "grad_norm": 0.2945052981376648, |
| "learning_rate": 1e-05, |
| "loss": -0.0024, |
| "num_tokens": 144185904.0, |
| "reward": 12.36389796257019, |
| "reward_std": 1.0678041917830705, |
| "rewards/event_reward_fn/mean": 11.530625, |
| "rewards/event_reward_fn/std": 6.954381022453308, |
| "rewards/format_reward_fn/mean": 0.8332730257511138, |
| "rewards/format_reward_fn/std": 0.2778134834766388, |
| "step": 1150, |
| "step_time": 25.22423570729792 |
| }, |
| { |
| "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": 0.08375, |
| "completions/max_length": 251.84, |
| "completions/max_terminated_length": 248.02, |
| "completions/mean_length": 207.27625, |
| "completions/mean_terminated_length": 202.8922332763672, |
| "completions/min_length": 162.06, |
| "completions/min_terminated_length": 162.06, |
| "entropy": 0.06358764633536339, |
| "epoch": 1.6085790884718498, |
| "frac_reward_zero_std": 0.4425, |
| "grad_norm": 0.18330521881580353, |
| "learning_rate": 1e-05, |
| "loss": 0.0013, |
| "num_tokens": 150418096.0, |
| "reward": 13.024293642044068, |
| "reward_std": 1.0176034840941428, |
| "rewards/event_reward_fn/mean": 12.1875, |
| "rewards/event_reward_fn/std": 7.625103950500488, |
| "rewards/format_reward_fn/mean": 0.8367935848236084, |
| "rewards/format_reward_fn/std": 0.2789352157711983, |
| "step": 1200, |
| "step_time": 25.085855303443967 |
| }, |
| { |
| "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": 0.061875, |
| "completions/max_length": 252.08, |
| "completions/max_terminated_length": 246.72, |
| "completions/mean_length": 203.59125, |
| "completions/mean_terminated_length": 200.21913879394532, |
| "completions/min_length": 163.32, |
| "completions/min_terminated_length": 163.32, |
| "entropy": 0.062210914567112925, |
| "epoch": 1.675603217158177, |
| "frac_reward_zero_std": 0.47, |
| "grad_norm": 0.40231508016586304, |
| "learning_rate": 1e-05, |
| "loss": -0.0028, |
| "num_tokens": 156497077.0, |
| "reward": 13.417193460464478, |
| "reward_std": 0.9512623021006584, |
| "rewards/event_reward_fn/mean": 12.55625, |
| "rewards/event_reward_fn/std": 7.378204255104065, |
| "rewards/format_reward_fn/mean": 0.8609435141086579, |
| "rewards/format_reward_fn/std": 0.2591093431413174, |
| "step": 1250, |
| "step_time": 24.335955093875526 |
| }, |
| { |
| "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": 0.083125, |
| "completions/max_length": 252.32, |
| "completions/max_terminated_length": 248.26, |
| "completions/mean_length": 208.861875, |
| "completions/mean_terminated_length": 204.65392211914062, |
| "completions/min_length": 167.5, |
| "completions/min_terminated_length": 167.5, |
| "entropy": 0.06511906914412975, |
| "epoch": 1.742627345844504, |
| "frac_reward_zero_std": 0.4125, |
| "grad_norm": 0.5480676889419556, |
| "learning_rate": 1e-05, |
| "loss": 0.0038, |
| "num_tokens": 162925421.0, |
| "reward": 13.32497908592224, |
| "reward_std": 1.179861811459996, |
| "rewards/event_reward_fn/mean": 12.480625, |
| "rewards/event_reward_fn/std": 7.050848822593689, |
| "rewards/format_reward_fn/mean": 0.8443540966510773, |
| "rewards/format_reward_fn/std": 0.27325084805488586, |
| "step": 1300, |
| "step_time": 25.89649803057313 |
| }, |
| { |
| "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": 0.09, |
| "completions/max_length": 253.5, |
| "completions/max_terminated_length": 248.1, |
| "completions/mean_length": 208.64, |
| "completions/mean_terminated_length": 204.13995147705077, |
| "completions/min_length": 161.3, |
| "completions/min_terminated_length": 161.3, |
| "entropy": 0.06706013955175877, |
| "epoch": 1.8096514745308312, |
| "frac_reward_zero_std": 0.375, |
| "grad_norm": 0.2488502711057663, |
| "learning_rate": 1e-05, |
| "loss": -0.0035, |
| "num_tokens": 169272835.0, |
| "reward": 12.847478866577148, |
| "reward_std": 1.1458918780088425, |
| "rewards/event_reward_fn/mean": 11.995, |
| "rewards/event_reward_fn/std": 7.542719440460205, |
| "rewards/format_reward_fn/mean": 0.852478951215744, |
| "rewards/format_reward_fn/std": 0.28421372681856155, |
| "step": 1350, |
| "step_time": 25.208012702837586 |
| }, |
| { |
| "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": 0.10375, |
| "completions/max_length": 253.5, |
| "completions/max_terminated_length": 248.62, |
| "completions/mean_length": 209.326875, |
| "completions/mean_terminated_length": 204.015185546875, |
| "completions/min_length": 164.82, |
| "completions/min_terminated_length": 164.82, |
| "entropy": 0.06814022108912468, |
| "epoch": 1.876675603217158, |
| "frac_reward_zero_std": 0.3525, |
| "grad_norm": 0.16522294282913208, |
| "learning_rate": 1e-05, |
| "loss": -0.0006, |
| "num_tokens": 175702725.0, |
| "reward": 13.004317455291748, |
| "reward_std": 1.250138430595398, |
| "rewards/event_reward_fn/mean": 12.18, |
| "rewards/event_reward_fn/std": 7.555402121543884, |
| "rewards/format_reward_fn/mean": 0.8243174004554749, |
| "rewards/format_reward_fn/std": 0.2965914273262024, |
| "step": 1400, |
| "step_time": 26.651869118101896 |
| }, |
| { |
| "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": 0.066875, |
| "completions/max_length": 252.98, |
| "completions/max_terminated_length": 249.98, |
| "completions/mean_length": 208.22625, |
| "completions/mean_terminated_length": 204.81213348388673, |
| "completions/min_length": 164.02, |
| "completions/min_terminated_length": 164.02, |
| "entropy": 0.0681792851537466, |
| "epoch": 1.9436997319034852, |
| "frac_reward_zero_std": 0.3775, |
| "grad_norm": 0.3131242096424103, |
| "learning_rate": 1e-05, |
| "loss": -0.0045, |
| "num_tokens": 181968342.0, |
| "reward": 12.955870590209962, |
| "reward_std": 1.2048768895864486, |
| "rewards/event_reward_fn/mean": 12.0975, |
| "rewards/event_reward_fn/std": 7.502232184410095, |
| "rewards/format_reward_fn/mean": 0.8583705246448516, |
| "rewards/format_reward_fn/std": 0.2481384778022766, |
| "step": 1450, |
| "step_time": 23.915284326784313 |
| } |
| ], |
| "logging_steps": 50, |
| "max_steps": 7460, |
| "num_input_tokens_seen": 187096042, |
| "num_train_epochs": 10, |
| "save_steps": 500, |
| "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 |
| } |
|
|
