| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984, | |
| "eval_steps": 100, | |
| "global_step": 468, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.328125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 807.5, | |
| "completions/mean_length": 603.8828125, | |
| "completions/mean_terminated_length": 396.44896697998047, | |
| "completions/min_length": 85.25, | |
| "completions/min_terminated_length": 85.25, | |
| "epoch": 0.0021333333333333334, | |
| "grad_norm": 0.29046246524602504, | |
| "kl": 0.0, | |
| "learning_rate": 0.0, | |
| "loss": -0.0123, | |
| "num_tokens": 96033.0, | |
| "reward": 0.109375, | |
| "reward_std": 0.18543372116982937, | |
| "rewards/accuracy_reward/mean": 0.109375, | |
| "rewards/accuracy_reward/std": 0.23456304892897606, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.310546875, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 809.125, | |
| "completions/mean_length": 579.490234375, | |
| "completions/mean_terminated_length": 381.3663749694824, | |
| "completions/min_length": 69.09375, | |
| "completions/min_terminated_length": 69.09375, | |
| "epoch": 0.010666666666666666, | |
| "grad_norm": 0.32341920946567715, | |
| "kl": 0.000206679105758667, | |
| "learning_rate": 2.553191489361702e-07, | |
| "loss": 0.0139, | |
| "num_tokens": 463812.0, | |
| "reward": 0.107421875, | |
| "reward_std": 0.18129237182438374, | |
| "rewards/accuracy_reward/mean": 0.107421875, | |
| "rewards/accuracy_reward/std": 0.23479509353637695, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 5 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2828125, | |
| "completions/max_length": 1022.75, | |
| "completions/max_terminated_length": 800.325, | |
| "completions/mean_length": 570.1828125, | |
| "completions/mean_terminated_length": 391.86217346191404, | |
| "completions/min_length": 65.05, | |
| "completions/min_terminated_length": 65.05, | |
| "epoch": 0.021333333333333333, | |
| "grad_norm": 0.31691394403397916, | |
| "kl": 0.00028862953186035154, | |
| "learning_rate": 5.74468085106383e-07, | |
| "loss": 0.0054, | |
| "num_tokens": 917313.0, | |
| "reward": 0.1046875, | |
| "reward_std": 0.19559106044471264, | |
| "rewards/accuracy_reward/mean": 0.1046875, | |
| "rewards/accuracy_reward/std": 0.2545704640448093, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 10 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.278125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 825.475, | |
| "completions/mean_length": 561.0359375, | |
| "completions/mean_terminated_length": 383.8768829345703, | |
| "completions/min_length": 75.625, | |
| "completions/min_terminated_length": 75.625, | |
| "epoch": 0.032, | |
| "grad_norm": 0.2712721137539722, | |
| "kl": 0.0003191232681274414, | |
| "learning_rate": 8.936170212765958e-07, | |
| "loss": 0.0012, | |
| "num_tokens": 1370600.0, | |
| "reward": 0.090625, | |
| "reward_std": 0.16034209839999675, | |
| "rewards/accuracy_reward/mean": 0.090625, | |
| "rewards/accuracy_reward/std": 0.20271828323602675, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 15 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2703125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 810.4, | |
| "completions/mean_length": 557.6, | |
| "completions/mean_terminated_length": 387.25951232910154, | |
| "completions/min_length": 68.8, | |
| "completions/min_terminated_length": 68.8, | |
| "epoch": 0.042666666666666665, | |
| "grad_norm": 0.22807478508268822, | |
| "kl": 0.0006281852722167969, | |
| "learning_rate": 1.2127659574468085e-06, | |
| "loss": 0.0119, | |
| "num_tokens": 1827680.0, | |
| "reward": 0.0953125, | |
| "reward_std": 0.1787047166377306, | |
| "rewards/accuracy_reward/mean": 0.0953125, | |
| "rewards/accuracy_reward/std": 0.23710305467247964, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 20 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.221875, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 842.425, | |
| "completions/mean_length": 573.5546875, | |
| "completions/mean_terminated_length": 449.37009353637694, | |
| "completions/min_length": 123.775, | |
| "completions/min_terminated_length": 123.775, | |
| "epoch": 0.05333333333333334, | |
| "grad_norm": 0.44388579378313797, | |
| "kl": 0.0013070106506347656, | |
| "learning_rate": 1.5319148936170212e-06, | |
| "loss": -0.0003, | |
| "num_tokens": 2288499.0, | |
| "reward": 0.115625, | |
| "reward_std": 0.19537889324128627, | |
| "rewards/accuracy_reward/mean": 0.115625, | |
| "rewards/accuracy_reward/std": 0.24712119549512862, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 25 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2125, | |
| "completions/max_length": 1013.025, | |
| "completions/max_terminated_length": 822.925, | |
| "completions/mean_length": 571.5015625, | |
| "completions/mean_terminated_length": 456.50114212036135, | |
| "completions/min_length": 126.625, | |
| "completions/min_terminated_length": 126.625, | |
| "epoch": 0.064, | |
| "grad_norm": 0.29459049413117055, | |
| "kl": 0.0028873443603515624, | |
| "learning_rate": 1.851063829787234e-06, | |
| "loss": 0.0168, | |
| "num_tokens": 2747108.0, | |
| "reward": 0.1703125, | |
| "reward_std": 0.23525591157376766, | |
| "rewards/accuracy_reward/mean": 0.1703125, | |
| "rewards/accuracy_reward/std": 0.30535352900624274, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 30 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1921875, | |
| "completions/max_length": 983.125, | |
| "completions/max_terminated_length": 856.15, | |
| "completions/mean_length": 569.5109375, | |
| "completions/mean_terminated_length": 468.74621734619143, | |
| "completions/min_length": 140.675, | |
| "completions/min_terminated_length": 140.675, | |
| "epoch": 0.07466666666666667, | |
| "grad_norm": 0.312902146600806, | |
| "kl": 0.004683685302734375, | |
| "learning_rate": 2.170212765957447e-06, | |
| "loss": -0.0096, | |
| "num_tokens": 3203995.0, | |
| "reward": 0.2921875, | |
| "reward_std": 0.2818045925348997, | |
| "rewards/accuracy_reward/mean": 0.2921875, | |
| "rewards/accuracy_reward/std": 0.35747046694159507, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 35 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.20625, | |
| "completions/max_length": 1015.0, | |
| "completions/max_terminated_length": 860.75, | |
| "completions/mean_length": 610.028125, | |
| "completions/mean_terminated_length": 512.4429527282715, | |
| "completions/min_length": 243.85, | |
| "completions/min_terminated_length": 243.85, | |
| "epoch": 0.08533333333333333, | |
| "grad_norm": 0.3492387121634339, | |
| "kl": 0.006721115112304688, | |
| "learning_rate": 2.4893617021276598e-06, | |
| "loss": 0.0155, | |
| "num_tokens": 3683013.0, | |
| "reward": 0.4015625, | |
| "reward_std": 0.30027358755469324, | |
| "rewards/accuracy_reward/mean": 0.4015625, | |
| "rewards/accuracy_reward/std": 0.405949179828167, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 40 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.18125, | |
| "completions/max_length": 1003.275, | |
| "completions/max_terminated_length": 807.4, | |
| "completions/mean_length": 574.1125, | |
| "completions/mean_terminated_length": 482.73938674926757, | |
| "completions/min_length": 233.7, | |
| "completions/min_terminated_length": 233.7, | |
| "epoch": 0.096, | |
| "grad_norm": 0.34838562088741554, | |
| "kl": 0.009783935546875, | |
| "learning_rate": 2.8085106382978724e-06, | |
| "loss": 0.0274, | |
| "num_tokens": 4141981.0, | |
| "reward": 0.453125, | |
| "reward_std": 0.26892418451607225, | |
| "rewards/accuracy_reward/mean": 0.453125, | |
| "rewards/accuracy_reward/std": 0.428996454924345, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 45 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.24375, | |
| "completions/max_length": 1007.625, | |
| "completions/max_terminated_length": 854.4, | |
| "completions/mean_length": 660.3234375, | |
| "completions/mean_terminated_length": 554.0355278015137, | |
| "completions/min_length": 317.9, | |
| "completions/min_terminated_length": 317.9, | |
| "epoch": 0.10666666666666667, | |
| "grad_norm": 0.3583613636029149, | |
| "kl": 0.009557342529296875, | |
| "learning_rate": 2.9998329491279003e-06, | |
| "loss": 0.0343, | |
| "num_tokens": 4648684.0, | |
| "reward": 0.4359375, | |
| "reward_std": 0.26149683743715285, | |
| "rewards/accuracy_reward/mean": 0.4359375, | |
| "rewards/accuracy_reward/std": 0.40705749318003653, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 50 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2875, | |
| "completions/max_length": 1015.175, | |
| "completions/max_terminated_length": 881.525, | |
| "completions/mean_length": 678.659375, | |
| "completions/mean_terminated_length": 557.4395378112793, | |
| "completions/min_length": 286.05, | |
| "completions/min_terminated_length": 286.05, | |
| "epoch": 0.11733333333333333, | |
| "grad_norm": 0.2916953090873945, | |
| "kl": 0.01027984619140625, | |
| "learning_rate": 2.9979540541011378e-06, | |
| "loss": 0.0193, | |
| "num_tokens": 5163018.0, | |
| "reward": 0.4265625, | |
| "reward_std": 0.2812244530767202, | |
| "rewards/accuracy_reward/mean": 0.4265625, | |
| "rewards/accuracy_reward/std": 0.39905172660946847, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 55 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.31875, | |
| "completions/max_length": 1011.0, | |
| "completions/max_terminated_length": 856.575, | |
| "completions/mean_length": 680.340625, | |
| "completions/mean_terminated_length": 527.7020431518555, | |
| "completions/min_length": 285.125, | |
| "completions/min_terminated_length": 285.125, | |
| "epoch": 0.128, | |
| "grad_norm": 0.3097602989893739, | |
| "kl": 0.01049346923828125, | |
| "learning_rate": 2.993990074515455e-06, | |
| "loss": 0.0374, | |
| "num_tokens": 5687748.0, | |
| "reward": 0.4953125, | |
| "reward_std": 0.27433369904756544, | |
| "rewards/accuracy_reward/mean": 0.4953125, | |
| "rewards/accuracy_reward/std": 0.42715574726462363, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 60 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.378125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 860.525, | |
| "completions/mean_length": 737.828125, | |
| "completions/mean_terminated_length": 572.9206932067871, | |
| "completions/min_length": 346.95, | |
| "completions/min_terminated_length": 346.95, | |
| "epoch": 0.13866666666666666, | |
| "grad_norm": 0.3757665149243777, | |
| "kl": 0.0093902587890625, | |
| "learning_rate": 2.987946528051855e-06, | |
| "loss": 0.0336, | |
| "num_tokens": 6255822.0, | |
| "reward": 0.3546875, | |
| "reward_std": 0.26096867546439173, | |
| "rewards/accuracy_reward/mean": 0.3546875, | |
| "rewards/accuracy_reward/std": 0.3680890344083309, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 65 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.3265625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 794.05, | |
| "completions/mean_length": 666.8078125, | |
| "completions/mean_terminated_length": 498.4326515197754, | |
| "completions/min_length": 271.175, | |
| "completions/min_terminated_length": 271.175, | |
| "epoch": 0.14933333333333335, | |
| "grad_norm": 0.3067868868446902, | |
| "kl": 0.0102752685546875, | |
| "learning_rate": 2.9798318270547456e-06, | |
| "loss": 0.0687, | |
| "num_tokens": 6760139.0, | |
| "reward": 0.5265625, | |
| "reward_std": 0.3408231448382139, | |
| "rewards/accuracy_reward/mean": 0.5265625, | |
| "rewards/accuracy_reward/std": 0.4369465745985508, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 70 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2609375, | |
| "completions/max_length": 1020.325, | |
| "completions/max_terminated_length": 863.3, | |
| "completions/mean_length": 670.8140625, | |
| "completions/mean_terminated_length": 550.6407440185546, | |
| "completions/min_length": 294.25, | |
| "completions/min_terminated_length": 294.25, | |
| "epoch": 0.16, | |
| "grad_norm": 0.29577449728572786, | |
| "kl": 0.00943450927734375, | |
| "learning_rate": 2.96965726682234e-06, | |
| "loss": 0.0437, | |
| "num_tokens": 7277532.0, | |
| "reward": 0.359375, | |
| "reward_std": 0.26728419214487076, | |
| "rewards/accuracy_reward/mean": 0.359375, | |
| "rewards/accuracy_reward/std": 0.3770089760422707, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 75 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2140625, | |
| "completions/max_length": 992.175, | |
| "completions/max_terminated_length": 824.875, | |
| "completions/mean_length": 624.7859375, | |
| "completions/mean_terminated_length": 528.7103263854981, | |
| "completions/min_length": 286.725, | |
| "completions/min_terminated_length": 286.725, | |
| "epoch": 0.17066666666666666, | |
| "grad_norm": 0.25358782513135053, | |
| "kl": 0.01270294189453125, | |
| "learning_rate": 2.9574370098841073e-06, | |
| "loss": 0.0206, | |
| "num_tokens": 7770987.0, | |
| "reward": 0.446875, | |
| "reward_std": 0.25387620851397513, | |
| "rewards/accuracy_reward/mean": 0.446875, | |
| "rewards/accuracy_reward/std": 0.41064497008919715, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 80 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2296875, | |
| "completions/max_length": 1019.075, | |
| "completions/max_terminated_length": 863.475, | |
| "completions/mean_length": 652.8296875, | |
| "completions/mean_terminated_length": 547.2898578643799, | |
| "completions/min_length": 284.5, | |
| "completions/min_terminated_length": 284.5, | |
| "epoch": 0.18133333333333335, | |
| "grad_norm": 0.24694213995347897, | |
| "kl": 0.012436676025390624, | |
| "learning_rate": 2.9431880662871697e-06, | |
| "loss": 0.0418, | |
| "num_tokens": 8277326.0, | |
| "reward": 0.43125, | |
| "reward_std": 0.27223431766033174, | |
| "rewards/accuracy_reward/mean": 0.43125, | |
| "rewards/accuracy_reward/std": 0.40274464786052705, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 85 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2734375, | |
| "completions/max_length": 1016.775, | |
| "completions/max_terminated_length": 867.3, | |
| "completions/mean_length": 672.5609375, | |
| "completions/mean_terminated_length": 555.7766807556152, | |
| "completions/min_length": 290.45, | |
| "completions/min_terminated_length": 290.45, | |
| "epoch": 0.192, | |
| "grad_norm": 0.43538797427920595, | |
| "kl": 0.01246337890625, | |
| "learning_rate": 2.926930269919085e-06, | |
| "loss": 0.0397, | |
| "num_tokens": 8791085.0, | |
| "reward": 0.484375, | |
| "reward_std": 0.32977913729846475, | |
| "rewards/accuracy_reward/mean": 0.484375, | |
| "rewards/accuracy_reward/std": 0.4205139525234699, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 90 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.21875, | |
| "completions/max_length": 1015.15, | |
| "completions/max_terminated_length": 810.625, | |
| "completions/mean_length": 604.1609375, | |
| "completions/mean_terminated_length": 492.36522674560547, | |
| "completions/min_length": 240.85, | |
| "completions/min_terminated_length": 240.85, | |
| "epoch": 0.20266666666666666, | |
| "grad_norm": 0.363946563847488, | |
| "kl": 0.014809417724609374, | |
| "learning_rate": 2.908686250899966e-06, | |
| "loss": 0.0326, | |
| "num_tokens": 9263084.0, | |
| "reward": 0.4546875, | |
| "reward_std": 0.2643970146775246, | |
| "rewards/accuracy_reward/mean": 0.4546875, | |
| "rewards/accuracy_reward/std": 0.37832572385668756, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.21333333333333335, | |
| "grad_norm": 0.40119715842853293, | |
| "learning_rate": 2.8884814040823755e-06, | |
| "loss": 0.0642, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.21333333333333335, | |
| "eval_clip_ratio": 0.0, | |
| "eval_completions/clipped_ratio": 0.254325, | |
| "eval_completions/max_length": 1017.8164, | |
| "eval_completions/max_terminated_length": 844.72, | |
| "eval_completions/mean_length": 641.31905, | |
| "eval_completions/mean_terminated_length": 518.6613545043946, | |
| "eval_completions/min_length": 271.8056, | |
| "eval_completions/min_terminated_length": 271.8056, | |
| "eval_kl": 0.01416409912109375, | |
| "eval_loss": 0.043244268745183945, | |
| "eval_num_tokens": 9739068.0, | |
| "eval_reward": 0.402925, | |
| "eval_reward_std": 0.2727512115895748, | |
| "eval_rewards/accuracy_reward/mean": 0.402925, | |
| "eval_rewards/accuracy_reward/std": 0.4037328534960747, | |
| "eval_rewards/format_reward/mean": 0.0, | |
| "eval_rewards/format_reward/std": 0.0, | |
| "eval_runtime": 17198.7184, | |
| "eval_samples_per_second": 0.291, | |
| "eval_steps_per_second": 0.018, | |
| "step": 100 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2265625, | |
| "completions/max_length": 1017.7875, | |
| "completions/max_terminated_length": 829.6, | |
| "completions/mean_length": 614.565625, | |
| "completions/mean_terminated_length": 500.5110694885254, | |
| "completions/min_length": 247.775, | |
| "completions/min_terminated_length": 247.775, | |
| "epoch": 0.224, | |
| "grad_norm": 0.40852771135479854, | |
| "kl": 0.0152984619140625, | |
| "learning_rate": 2.866343853702835e-06, | |
| "loss": 0.0384, | |
| "num_tokens": 10216536.0, | |
| "reward": 0.48515625, | |
| "reward_std": 0.2992698045447469, | |
| "rewards/accuracy_reward/mean": 0.48515625, | |
| "rewards/accuracy_reward/std": 0.43572295978665354, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 105 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2515625, | |
| "completions/max_length": 1018.075, | |
| "completions/max_terminated_length": 817.625, | |
| "completions/mean_length": 635.7125, | |
| "completions/mean_terminated_length": 516.0587692260742, | |
| "completions/min_length": 288.85, | |
| "completions/min_terminated_length": 288.85, | |
| "epoch": 0.23466666666666666, | |
| "grad_norm": 0.3358851233027883, | |
| "kl": 0.015325927734375, | |
| "learning_rate": 2.842304414234153e-06, | |
| "loss": 0.0292, | |
| "num_tokens": 10712328.0, | |
| "reward": 0.46875, | |
| "reward_std": 0.254452820494771, | |
| "rewards/accuracy_reward/mean": 0.46875, | |
| "rewards/accuracy_reward/std": 0.3950736179947853, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 110 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2625, | |
| "completions/max_length": 1012.575, | |
| "completions/max_terminated_length": 838.725, | |
| "completions/mean_length": 643.7859375, | |
| "completions/mean_terminated_length": 518.7130432128906, | |
| "completions/min_length": 279.9, | |
| "completions/min_terminated_length": 279.9, | |
| "epoch": 0.24533333333333332, | |
| "grad_norm": 0.3712248783220928, | |
| "kl": 0.01651611328125, | |
| "learning_rate": 2.8163965474930673e-06, | |
| "loss": 0.0588, | |
| "num_tokens": 11223503.0, | |
| "reward": 0.4390625, | |
| "reward_std": 0.30547982417047026, | |
| "rewards/accuracy_reward/mean": 0.4390625, | |
| "rewards/accuracy_reward/std": 0.4317807413637638, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 115 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.3046875, | |
| "completions/max_length": 1014.375, | |
| "completions/max_terminated_length": 813.85, | |
| "completions/mean_length": 662.615625, | |
| "completions/mean_terminated_length": 519.530436706543, | |
| "completions/min_length": 310.8, | |
| "completions/min_terminated_length": 310.8, | |
| "epoch": 0.256, | |
| "grad_norm": 0.24811201734525618, | |
| "kl": 0.019573974609375, | |
| "learning_rate": 2.7886563160629058e-06, | |
| "loss": 0.0372, | |
| "num_tokens": 11742665.0, | |
| "reward": 0.4453125, | |
| "reward_std": 0.23867536522448063, | |
| "rewards/accuracy_reward/mean": 0.4453125, | |
| "rewards/accuracy_reward/std": 0.417168840020895, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 120 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2359375, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 827.8, | |
| "completions/mean_length": 629.021875, | |
| "completions/mean_terminated_length": 511.2683464050293, | |
| "completions/min_length": 278.975, | |
| "completions/min_terminated_length": 278.975, | |
| "epoch": 0.26666666666666666, | |
| "grad_norm": 0.344688397601857, | |
| "kl": 0.02204742431640625, | |
| "learning_rate": 2.759122333096093e-06, | |
| "loss": 0.0357, | |
| "num_tokens": 12236375.0, | |
| "reward": 0.4375, | |
| "reward_std": 0.2976874437183142, | |
| "rewards/accuracy_reward/mean": 0.4375, | |
| "rewards/accuracy_reward/std": 0.40365294441580774, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 125 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2265625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 866.45, | |
| "completions/mean_length": 639.3734375, | |
| "completions/mean_terminated_length": 530.9625923156739, | |
| "completions/min_length": 275.025, | |
| "completions/min_terminated_length": 275.025, | |
| "epoch": 0.2773333333333333, | |
| "grad_norm": 0.3373017645799004, | |
| "kl": 0.024224853515625, | |
| "learning_rate": 2.727835708566381e-06, | |
| "loss": 0.0511, | |
| "num_tokens": 12729606.0, | |
| "reward": 0.3921875, | |
| "reward_std": 0.2933184415102005, | |
| "rewards/accuracy_reward/mean": 0.3921875, | |
| "rewards/accuracy_reward/std": 0.3842721916735172, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 130 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2015625, | |
| "completions/max_length": 1013.65, | |
| "completions/max_terminated_length": 827.25, | |
| "completions/mean_length": 609.1671875, | |
| "completions/mean_terminated_length": 517.1776489257812, | |
| "completions/min_length": 257.25, | |
| "completions/min_terminated_length": 257.25, | |
| "epoch": 0.288, | |
| "grad_norm": 0.47904116965848426, | |
| "kl": 0.02641754150390625, | |
| "learning_rate": 2.6948399920456223e-06, | |
| "loss": 0.0707, | |
| "num_tokens": 13210177.0, | |
| "reward": 0.4015625, | |
| "reward_std": 0.29006172008812425, | |
| "rewards/accuracy_reward/mean": 0.4015625, | |
| "rewards/accuracy_reward/std": 0.3616370469331741, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 135 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2703125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 813.775, | |
| "completions/mean_length": 630.115625, | |
| "completions/mean_terminated_length": 490.09764099121094, | |
| "completions/min_length": 231.425, | |
| "completions/min_terminated_length": 231.425, | |
| "epoch": 0.2986666666666667, | |
| "grad_norm": 0.4924178489635141, | |
| "kl": 0.0541534423828125, | |
| "learning_rate": 2.6601811120847203e-06, | |
| "loss": 0.1306, | |
| "num_tokens": 13698515.0, | |
| "reward": 0.3859375, | |
| "reward_std": 0.3275134950876236, | |
| "rewards/accuracy_reward/mean": 0.3859375, | |
| "rewards/accuracy_reward/std": 0.4226386234164238, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 140 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.3296875, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 825.4, | |
| "completions/mean_length": 655.7046875, | |
| "completions/mean_terminated_length": 479.4415023803711, | |
| "completions/min_length": 163.15, | |
| "completions/min_terminated_length": 163.15, | |
| "epoch": 0.30933333333333335, | |
| "grad_norm": 2.123406819046491, | |
| "kl": 0.1634521484375, | |
| "learning_rate": 2.6239073122831634e-06, | |
| "loss": 0.1303, | |
| "num_tokens": 14208934.0, | |
| "reward": 0.2640625, | |
| "reward_std": 0.31472352109849455, | |
| "rewards/accuracy_reward/mean": 0.2640625, | |
| "rewards/accuracy_reward/std": 0.38663889542222024, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 145 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.109375, | |
| "completions/max_length": 876.975, | |
| "completions/max_terminated_length": 738.675, | |
| "completions/mean_length": 424.2921875, | |
| "completions/mean_terminated_length": 355.7382415771484, | |
| "completions/min_length": 43.55, | |
| "completions/min_terminated_length": 43.55, | |
| "epoch": 0.32, | |
| "grad_norm": 1.0873205027125876, | |
| "kl": 0.9138916015625, | |
| "learning_rate": 2.586069084136102e-06, | |
| "loss": 0.013, | |
| "num_tokens": 14563945.0, | |
| "reward": 0.203125, | |
| "reward_std": 0.27037961296737195, | |
| "rewards/accuracy_reward/mean": 0.203125, | |
| "rewards/accuracy_reward/std": 0.35853438526391984, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 150 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.015625, | |
| "completions/max_length": 791.55, | |
| "completions/max_terminated_length": 750.375, | |
| "completions/mean_length": 370.10625, | |
| "completions/mean_terminated_length": 359.8678153991699, | |
| "completions/min_length": 44.675, | |
| "completions/min_terminated_length": 44.675, | |
| "epoch": 0.33066666666666666, | |
| "grad_norm": 3.705700983890623, | |
| "kl": 0.6580078125, | |
| "learning_rate": 2.5467190967524655e-06, | |
| "loss": -0.034, | |
| "num_tokens": 14891197.0, | |
| "reward": 0.315625, | |
| "reward_std": 0.3082596454769373, | |
| "rewards/accuracy_reward/mean": 0.315625, | |
| "rewards/accuracy_reward/std": 0.37598869428038595, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 155 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.0140625, | |
| "completions/max_length": 730.85, | |
| "completions/max_terminated_length": 714.8, | |
| "completions/mean_length": 380.8421875, | |
| "completions/mean_terminated_length": 372.76271896362306, | |
| "completions/min_length": 87.425, | |
| "completions/min_terminated_length": 87.425, | |
| "epoch": 0.3413333333333333, | |
| "grad_norm": 1.6894860441577937, | |
| "kl": 2.70595703125, | |
| "learning_rate": 2.5059121235419343e-06, | |
| "loss": 0.0725, | |
| "num_tokens": 15224776.0, | |
| "reward": 0.2046875, | |
| "reward_std": 0.27738223411142826, | |
| "rewards/accuracy_reward/mean": 0.2046875, | |
| "rewards/accuracy_reward/std": 0.3607805661857128, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 160 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1, | |
| "completions/max_length": 938.025, | |
| "completions/max_terminated_length": 841.95, | |
| "completions/mean_length": 528.2921875, | |
| "completions/mean_terminated_length": 475.2810600280762, | |
| "completions/min_length": 129.55, | |
| "completions/min_terminated_length": 129.55, | |
| "epoch": 0.352, | |
| "grad_norm": 1.528288841175706, | |
| "kl": 1.4619140625, | |
| "learning_rate": 2.4637049659728207e-06, | |
| "loss": 0.065, | |
| "num_tokens": 15654979.0, | |
| "reward": 0.1921875, | |
| "reward_std": 0.28522051945328714, | |
| "rewards/accuracy_reward/mean": 0.1921875, | |
| "rewards/accuracy_reward/std": 0.3629676692187786, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 165 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2359375, | |
| "completions/max_length": 1019.875, | |
| "completions/max_terminated_length": 885.0, | |
| "completions/mean_length": 598.4390625, | |
| "completions/mean_terminated_length": 470.77597427368164, | |
| "completions/min_length": 82.1, | |
| "completions/min_terminated_length": 82.1, | |
| "epoch": 0.3626666666666667, | |
| "grad_norm": 2.803962527847425, | |
| "kl": 3.966015625, | |
| "learning_rate": 2.4201563745069844e-06, | |
| "loss": 0.2138, | |
| "num_tokens": 16128140.0, | |
| "reward": 0.159375, | |
| "reward_std": 0.2523420058190823, | |
| "rewards/accuracy_reward/mean": 0.159375, | |
| "rewards/accuracy_reward/std": 0.29368442595005034, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 170 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.259375, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 878.375, | |
| "completions/mean_length": 601.2546875, | |
| "completions/mean_terminated_length": 457.18211555480957, | |
| "completions/min_length": 76.55, | |
| "completions/min_terminated_length": 76.55, | |
| "epoch": 0.37333333333333335, | |
| "grad_norm": 5.829506705682438, | |
| "kl": 2.53173828125, | |
| "learning_rate": 2.3753269668218372e-06, | |
| "loss": 0.1753, | |
| "num_tokens": 16595127.0, | |
| "reward": 0.2046875, | |
| "reward_std": 0.3113111235201359, | |
| "rewards/accuracy_reward/mean": 0.2046875, | |
| "rewards/accuracy_reward/std": 0.36455870419740677, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 175 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.259375, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 896.85, | |
| "completions/mean_length": 621.9390625, | |
| "completions/mean_terminated_length": 485.3581092834473, | |
| "completions/min_length": 118.325, | |
| "completions/min_terminated_length": 118.325, | |
| "epoch": 0.384, | |
| "grad_norm": 1.5424634966107815, | |
| "kl": 3.1536865234375, | |
| "learning_rate": 2.3292791434332643e-06, | |
| "loss": 0.2008, | |
| "num_tokens": 17087648.0, | |
| "reward": 0.228125, | |
| "reward_std": 0.2769108280539513, | |
| "rewards/accuracy_reward/mean": 0.228125, | |
| "rewards/accuracy_reward/std": 0.36085558757185937, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 180 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2984375, | |
| "completions/max_length": 1019.65, | |
| "completions/max_terminated_length": 893.25, | |
| "completions/mean_length": 664.053125, | |
| "completions/mean_terminated_length": 521.8890960693359, | |
| "completions/min_length": 193.275, | |
| "completions/min_terminated_length": 193.275, | |
| "epoch": 0.39466666666666667, | |
| "grad_norm": 3.157457940002755, | |
| "kl": 2.6540283203125, | |
| "learning_rate": 2.2820770008369208e-06, | |
| "loss": 0.2153, | |
| "num_tokens": 17601938.0, | |
| "reward": 0.2890625, | |
| "reward_std": 0.2875839114189148, | |
| "rewards/accuracy_reward/mean": 0.2890625, | |
| "rewards/accuracy_reward/std": 0.37700623869895933, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 185 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.303125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 877.9, | |
| "completions/mean_length": 662.246875, | |
| "completions/mean_terminated_length": 510.23935470581057, | |
| "completions/min_length": 191.375, | |
| "completions/min_terminated_length": 191.375, | |
| "epoch": 0.4053333333333333, | |
| "grad_norm": 5.343900565343268, | |
| "kl": 3.46298828125, | |
| "learning_rate": 2.233786242288795e-06, | |
| "loss": 0.2422, | |
| "num_tokens": 18114840.0, | |
| "reward": 0.259375, | |
| "reward_std": 0.3033699918538332, | |
| "rewards/accuracy_reward/mean": 0.259375, | |
| "rewards/accuracy_reward/std": 0.3962884694337845, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 190 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2765625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 803.6, | |
| "completions/mean_length": 610.2328125, | |
| "completions/mean_terminated_length": 449.8131446838379, | |
| "completions/min_length": 150.925, | |
| "completions/min_terminated_length": 150.925, | |
| "epoch": 0.416, | |
| "grad_norm": 1.4123262482389565, | |
| "kl": 4.473828125, | |
| "learning_rate": 2.184474086349237e-06, | |
| "loss": 0.3007, | |
| "num_tokens": 18590405.0, | |
| "reward": 0.2953125, | |
| "reward_std": 0.329832362011075, | |
| "rewards/accuracy_reward/mean": 0.2953125, | |
| "rewards/accuracy_reward/std": 0.3952501378953457, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.4266666666666667, | |
| "grad_norm": 1.335408537990007, | |
| "learning_rate": 2.1342091733177504e-06, | |
| "loss": 0.1653, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4266666666666667, | |
| "eval_clip_ratio": 0.0, | |
| "eval_completions/clipped_ratio": 0.249725, | |
| "eval_completions/max_length": 1020.764, | |
| "eval_completions/max_terminated_length": 864.5468, | |
| "eval_completions/mean_length": 603.358575, | |
| "eval_completions/mean_terminated_length": 465.70587670288086, | |
| "eval_completions/min_length": 101.0624, | |
| "eval_completions/min_terminated_length": 101.0624, | |
| "eval_kl": 4.10953203125, | |
| "eval_loss": 0.22873078286647797, | |
| "eval_num_tokens": 19080613.0, | |
| "eval_reward": 0.181575, | |
| "eval_reward_std": 0.25937469556927684, | |
| "eval_rewards/accuracy_reward/mean": 0.181575, | |
| "eval_rewards/accuracy_reward/std": 0.3261967261195183, | |
| "eval_rewards/format_reward/mean": 0.0, | |
| "eval_rewards/format_reward/std": 0.0, | |
| "eval_runtime": 17076.072, | |
| "eval_samples_per_second": 0.293, | |
| "eval_steps_per_second": 0.018, | |
| "step": 200 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.26640625, | |
| "completions/max_length": 1023.3875, | |
| "completions/max_terminated_length": 867.1625, | |
| "completions/mean_length": 619.5703125, | |
| "completions/mean_terminated_length": 475.73999614715575, | |
| "completions/min_length": 108.0875, | |
| "completions/min_terminated_length": 108.0875, | |
| "epoch": 0.43733333333333335, | |
| "grad_norm": 1.9587392151855365, | |
| "kl": 2.8674636840820313, | |
| "learning_rate": 2.0830614696887893e-06, | |
| "loss": 0.1883, | |
| "num_tokens": 19562599.0, | |
| "reward": 0.19609375, | |
| "reward_std": 0.26347289625555276, | |
| "rewards/accuracy_reward/mean": 0.19609375, | |
| "rewards/accuracy_reward/std": 0.34130338206887245, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 205 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.225, | |
| "completions/max_length": 1015.2, | |
| "completions/max_terminated_length": 865.525, | |
| "completions/mean_length": 601.5953125, | |
| "completions/mean_terminated_length": 481.98564224243165, | |
| "completions/min_length": 99.225, | |
| "completions/min_terminated_length": 99.225, | |
| "epoch": 0.448, | |
| "grad_norm": 2.694092158079111, | |
| "kl": 2.51298828125, | |
| "learning_rate": 2.0311021707615474e-06, | |
| "loss": 0.1151, | |
| "num_tokens": 20033692.0, | |
| "reward": 0.2125, | |
| "reward_std": 0.2790681302547455, | |
| "rewards/accuracy_reward/mean": 0.2125, | |
| "rewards/accuracy_reward/std": 0.3386077769100666, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 210 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.2375, | |
| "completions/max_length": 1010.2, | |
| "completions/max_terminated_length": 868.775, | |
| "completions/mean_length": 599.2375, | |
| "completions/mean_terminated_length": 471.63191413879395, | |
| "completions/min_length": 124.35, | |
| "completions/min_terminated_length": 124.35, | |
| "epoch": 0.45866666666666667, | |
| "grad_norm": 1.5861291482690305, | |
| "kl": 3.3865478515625, | |
| "learning_rate": 1.9784036015393108e-06, | |
| "loss": 0.221, | |
| "num_tokens": 20505108.0, | |
| "reward": 0.203125, | |
| "reward_std": 0.26243593245744706, | |
| "rewards/accuracy_reward/mean": 0.203125, | |
| "rewards/accuracy_reward/std": 0.3315103515982628, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 215 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1875, | |
| "completions/max_length": 1001.85, | |
| "completions/max_terminated_length": 826.225, | |
| "completions/mean_length": 562.4015625, | |
| "completions/mean_terminated_length": 460.1457885742187, | |
| "completions/min_length": 126.6, | |
| "completions/min_terminated_length": 126.6, | |
| "epoch": 0.4693333333333333, | |
| "grad_norm": 1.7695739456560913, | |
| "kl": 2.744598388671875, | |
| "learning_rate": 1.9250391160563114e-06, | |
| "loss": 0.1631, | |
| "num_tokens": 20947373.0, | |
| "reward": 0.315625, | |
| "reward_std": 0.3029405642300844, | |
| "rewards/accuracy_reward/mean": 0.315625, | |
| "rewards/accuracy_reward/std": 0.3904874354600906, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 220 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1828125, | |
| "completions/max_length": 1019.125, | |
| "completions/max_terminated_length": 890.5, | |
| "completions/mean_length": 564.0375, | |
| "completions/mean_terminated_length": 460.37510833740237, | |
| "completions/min_length": 125.95, | |
| "completions/min_terminated_length": 125.95, | |
| "epoch": 0.48, | |
| "grad_norm": 3.4078593954773204, | |
| "kl": 3.49443359375, | |
| "learning_rate": 1.8710829952722163e-06, | |
| "loss": 0.2053, | |
| "num_tokens": 21400653.0, | |
| "reward": 0.3140625, | |
| "reward_std": 0.32277651987969874, | |
| "rewards/accuracy_reward/mean": 0.3140625, | |
| "rewards/accuracy_reward/std": 0.4073892831802368, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 225 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1421875, | |
| "completions/max_length": 984.2, | |
| "completions/max_terminated_length": 844.9, | |
| "completions/mean_length": 538.209375, | |
| "completions/mean_terminated_length": 461.8832672119141, | |
| "completions/min_length": 74.375, | |
| "completions/min_terminated_length": 74.375, | |
| "epoch": 0.49066666666666664, | |
| "grad_norm": 4.5323959558225395, | |
| "kl": 2.24453125, | |
| "learning_rate": 1.8166103436763777e-06, | |
| "loss": 0.1171, | |
| "num_tokens": 21837691.0, | |
| "reward": 0.3359375, | |
| "reward_std": 0.3033065766096115, | |
| "rewards/accuracy_reward/mean": 0.3359375, | |
| "rewards/accuracy_reward/std": 0.42303009480237963, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 230 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1921875, | |
| "completions/max_length": 1003.35, | |
| "completions/max_terminated_length": 889.725, | |
| "completions/mean_length": 567.85625, | |
| "completions/mean_terminated_length": 464.1244514465332, | |
| "completions/min_length": 105.775, | |
| "completions/min_terminated_length": 105.775, | |
| "epoch": 0.5013333333333333, | |
| "grad_norm": 5.127794348866437, | |
| "kl": 2.8025146484375, | |
| "learning_rate": 1.7616969847457666e-06, | |
| "loss": 0.1524, | |
| "num_tokens": 22288055.0, | |
| "reward": 0.2484375, | |
| "reward_std": 0.265233264490962, | |
| "rewards/accuracy_reward/mean": 0.2484375, | |
| "rewards/accuracy_reward/std": 0.3456104569137096, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 235 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1625, | |
| "completions/max_length": 1004.375, | |
| "completions/max_terminated_length": 885.5, | |
| "completions/mean_length": 578.2203125, | |
| "completions/mean_terminated_length": 495.75377731323243, | |
| "completions/min_length": 116.125, | |
| "completions/min_terminated_length": 116.125, | |
| "epoch": 0.512, | |
| "grad_norm": 2.2253978317000347, | |
| "kl": 2.90908203125, | |
| "learning_rate": 1.7064193554021108e-06, | |
| "loss": 0.138, | |
| "num_tokens": 22747892.0, | |
| "reward": 0.25, | |
| "reward_std": 0.3034702904522419, | |
| "rewards/accuracy_reward/mean": 0.25, | |
| "rewards/accuracy_reward/std": 0.39005310013890265, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 240 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1578125, | |
| "completions/max_length": 1012.0, | |
| "completions/max_terminated_length": 860.35, | |
| "completions/mean_length": 565.81875, | |
| "completions/mean_terminated_length": 479.0156246185303, | |
| "completions/min_length": 136.125, | |
| "completions/min_terminated_length": 136.125, | |
| "epoch": 0.5226666666666666, | |
| "grad_norm": 1.4256050986060915, | |
| "kl": 3.39208984375, | |
| "learning_rate": 1.650854399615142e-06, | |
| "loss": 0.2056, | |
| "num_tokens": 23206816.0, | |
| "reward": 0.2375, | |
| "reward_std": 0.2754390574991703, | |
| "rewards/accuracy_reward/mean": 0.2375, | |
| "rewards/accuracy_reward/std": 0.3435414247214794, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 245 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1953125, | |
| "completions/max_length": 1013.175, | |
| "completions/max_terminated_length": 904.125, | |
| "completions/mean_length": 600.421875, | |
| "completions/mean_terminated_length": 501.47122344970705, | |
| "completions/min_length": 146.425, | |
| "completions/min_terminated_length": 146.425, | |
| "epoch": 0.5333333333333333, | |
| "grad_norm": 8.99884757175663, | |
| "kl": 3.725, | |
| "learning_rate": 1.5950794613000567e-06, | |
| "loss": 0.2221, | |
| "num_tokens": 23687462.0, | |
| "reward": 0.1796875, | |
| "reward_std": 0.24051060006022454, | |
| "rewards/accuracy_reward/mean": 0.1796875, | |
| "rewards/accuracy_reward/std": 0.2946265310049057, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 250 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1890625, | |
| "completions/max_length": 1019.725, | |
| "completions/max_terminated_length": 886.175, | |
| "completions/mean_length": 576.121875, | |
| "completions/mean_terminated_length": 469.11180267333987, | |
| "completions/min_length": 121.925, | |
| "completions/min_terminated_length": 121.925, | |
| "epoch": 0.544, | |
| "grad_norm": 2.729106221857157, | |
| "kl": 3.590283203125, | |
| "learning_rate": 1.539172176658271e-06, | |
| "loss": 0.1995, | |
| "num_tokens": 24149060.0, | |
| "reward": 0.1828125, | |
| "reward_std": 0.25070299617946146, | |
| "rewards/accuracy_reward/mean": 0.1828125, | |
| "rewards/accuracy_reward/std": 0.3266736909747124, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 255 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1359375, | |
| "completions/max_length": 993.8, | |
| "completions/max_terminated_length": 849.3, | |
| "completions/mean_length": 526.68125, | |
| "completions/mean_terminated_length": 446.4154727935791, | |
| "completions/min_length": 120.825, | |
| "completions/min_terminated_length": 120.825, | |
| "epoch": 0.5546666666666666, | |
| "grad_norm": 1.4990062784258935, | |
| "kl": 2.2947998046875, | |
| "learning_rate": 1.4832103661113236e-06, | |
| "loss": 0.1452, | |
| "num_tokens": 24577128.0, | |
| "reward": 0.2078125, | |
| "reward_std": 0.27170360907912255, | |
| "rewards/accuracy_reward/mean": 0.2078125, | |
| "rewards/accuracy_reward/std": 0.35073904544115064, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 260 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1390625, | |
| "completions/max_length": 1008.375, | |
| "completions/max_terminated_length": 875.0, | |
| "completions/mean_length": 541.9234375, | |
| "completions/mean_terminated_length": 467.3186508178711, | |
| "completions/min_length": 130.6, | |
| "completions/min_terminated_length": 130.6, | |
| "epoch": 0.5653333333333334, | |
| "grad_norm": 9.148886998248635, | |
| "kl": 4.06611328125, | |
| "learning_rate": 1.4272719259783586e-06, | |
| "loss": 0.2278, | |
| "num_tokens": 25023999.0, | |
| "reward": 0.221875, | |
| "reward_std": 0.2942699518054724, | |
| "rewards/accuracy_reward/mean": 0.221875, | |
| "rewards/accuracy_reward/std": 0.3605892524123192, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 265 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.128125, | |
| "completions/max_length": 978.775, | |
| "completions/max_terminated_length": 867.6, | |
| "completions/mean_length": 528.0546875, | |
| "completions/mean_terminated_length": 461.0882026672363, | |
| "completions/min_length": 116.375, | |
| "completions/min_terminated_length": 116.375, | |
| "epoch": 0.576, | |
| "grad_norm": 1.904370064809824, | |
| "kl": 4.56875, | |
| "learning_rate": 1.3714347200479572e-06, | |
| "loss": 0.216, | |
| "num_tokens": 25454978.0, | |
| "reward": 0.23125, | |
| "reward_std": 0.3002675145864487, | |
| "rewards/accuracy_reward/mean": 0.23125, | |
| "rewards/accuracy_reward/std": 0.37225412726402285, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 270 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1484375, | |
| "completions/max_length": 1009.9, | |
| "completions/max_terminated_length": 880.575, | |
| "completions/mean_length": 583.6515625, | |
| "completions/mean_terminated_length": 508.4187866210938, | |
| "completions/min_length": 192.575, | |
| "completions/min_terminated_length": 192.575, | |
| "epoch": 0.5866666666666667, | |
| "grad_norm": 2.863201246372455, | |
| "kl": 1.67100830078125, | |
| "learning_rate": 1.3157764711952488e-06, | |
| "loss": 0.1031, | |
| "num_tokens": 25916763.0, | |
| "reward": 0.296875, | |
| "reward_std": 0.2727500643581152, | |
| "rewards/accuracy_reward/mean": 0.296875, | |
| "rewards/accuracy_reward/std": 0.36488829776644705, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 275 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.159375, | |
| "completions/max_length": 1002.75, | |
| "completions/max_terminated_length": 857.4, | |
| "completions/mean_length": 561.0359375, | |
| "completions/mean_terminated_length": 478.21215591430666, | |
| "completions/min_length": 133.1, | |
| "completions/min_terminated_length": 133.1, | |
| "epoch": 0.5973333333333334, | |
| "grad_norm": 6.145508296791858, | |
| "kl": 2.973114013671875, | |
| "learning_rate": 1.2603746531951753e-06, | |
| "loss": 0.1455, | |
| "num_tokens": 26363146.0, | |
| "reward": 0.240625, | |
| "reward_std": 0.27465322241187096, | |
| "rewards/accuracy_reward/mean": 0.240625, | |
| "rewards/accuracy_reward/std": 0.3454093523323536, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 280 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.175, | |
| "completions/max_length": 1017.1, | |
| "completions/max_terminated_length": 848.325, | |
| "completions/mean_length": 533.7046875, | |
| "completions/mean_terminated_length": 434.07742347717283, | |
| "completions/min_length": 123.6, | |
| "completions/min_terminated_length": 123.6, | |
| "epoch": 0.608, | |
| "grad_norm": 7.437800094717786, | |
| "kl": 5.8435546875, | |
| "learning_rate": 1.2053063828824825e-06, | |
| "loss": 0.2914, | |
| "num_tokens": 26800613.0, | |
| "reward": 0.1484375, | |
| "reward_std": 0.22240859605371952, | |
| "rewards/accuracy_reward/mean": 0.1484375, | |
| "rewards/accuracy_reward/std": 0.3025800295174122, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 285 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1546875, | |
| "completions/max_length": 1008.075, | |
| "completions/max_terminated_length": 873.875, | |
| "completions/mean_length": 522.96875, | |
| "completions/mean_terminated_length": 434.07748947143557, | |
| "completions/min_length": 101.3, | |
| "completions/min_terminated_length": 101.3, | |
| "epoch": 0.6186666666666667, | |
| "grad_norm": 4.1540581465241475, | |
| "kl": 2.5693359375, | |
| "learning_rate": 1.1506483128085632e-06, | |
| "loss": 0.1233, | |
| "num_tokens": 27224441.0, | |
| "reward": 0.1328125, | |
| "reward_std": 0.20410939157009125, | |
| "rewards/accuracy_reward/mean": 0.1328125, | |
| "rewards/accuracy_reward/std": 0.27071708589792254, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 290 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.134375, | |
| "completions/max_length": 987.3, | |
| "completions/max_terminated_length": 832.7, | |
| "completions/mean_length": 508.940625, | |
| "completions/mean_terminated_length": 430.16086196899414, | |
| "completions/min_length": 114.6, | |
| "completions/min_terminated_length": 114.6, | |
| "epoch": 0.6293333333333333, | |
| "grad_norm": 3.767225289300394, | |
| "kl": 1.0184326171875, | |
| "learning_rate": 1.0964765245445572e-06, | |
| "loss": 0.0948, | |
| "num_tokens": 27641059.0, | |
| "reward": 0.1984375, | |
| "reward_std": 0.27595735117793085, | |
| "rewards/accuracy_reward/mean": 0.1984375, | |
| "rewards/accuracy_reward/std": 0.3417211093008518, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 1.6353961367471301, | |
| "learning_rate": 1.042866422779233e-06, | |
| "loss": 0.1214, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_clip_ratio": 0.0, | |
| "eval_completions/clipped_ratio": 0.1834, | |
| "eval_completions/max_length": 1016.308, | |
| "eval_completions/max_terminated_length": 858.028, | |
| "eval_completions/mean_length": 556.6972, | |
| "eval_completions/mean_terminated_length": 453.3012262512207, | |
| "eval_completions/min_length": 128.7608, | |
| "eval_completions/min_terminated_length": 128.7608, | |
| "eval_kl": 3.4017, | |
| "eval_loss": 0.18172034621238708, | |
| "eval_num_tokens": 28075018.0, | |
| "eval_reward": 0.144625, | |
| "eval_reward_std": 0.2177023424565792, | |
| "eval_rewards/accuracy_reward/mean": 0.144625, | |
| "eval_rewards/accuracy_reward/std": 0.28106210635900497, | |
| "eval_rewards/format_reward/mean": 0.0, | |
| "eval_rewards/format_reward/std": 0.0, | |
| "eval_runtime": 16581.1407, | |
| "eval_samples_per_second": 0.302, | |
| "eval_steps_per_second": 0.019, | |
| "step": 300 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.18046875, | |
| "completions/max_length": 1021.1625, | |
| "completions/max_terminated_length": 858.15, | |
| "completions/mean_length": 553.5671875, | |
| "completions/mean_terminated_length": 452.57582130432127, | |
| "completions/min_length": 137.275, | |
| "completions/min_terminated_length": 137.275, | |
| "epoch": 0.6506666666666666, | |
| "grad_norm": 1.5919395722688163, | |
| "kl": 2.91702880859375, | |
| "learning_rate": 9.898926303590562e-07, | |
| "loss": 0.2225, | |
| "num_tokens": 28530073.0, | |
| "reward": 0.17890625, | |
| "reward_std": 0.25828637965023515, | |
| "rewards/accuracy_reward/mean": 0.17890625, | |
| "rewards/accuracy_reward/std": 0.32390750013291836, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 305 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.15625, | |
| "completions/max_length": 1006.85, | |
| "completions/max_terminated_length": 822.9, | |
| "completions/mean_length": 533.6984375, | |
| "completions/mean_terminated_length": 447.3911735534668, | |
| "completions/min_length": 134.1, | |
| "completions/min_terminated_length": 134.1, | |
| "epoch": 0.6613333333333333, | |
| "grad_norm": 2.975928575760629, | |
| "kl": 2.889697265625, | |
| "learning_rate": 9.376288844165526e-07, | |
| "loss": 0.1673, | |
| "num_tokens": 28955912.0, | |
| "reward": 0.2828125, | |
| "reward_std": 0.3022126518189907, | |
| "rewards/accuracy_reward/mean": 0.2828125, | |
| "rewards/accuracy_reward/std": 0.3880471274256706, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 310 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1875, | |
| "completions/max_length": 1015.55, | |
| "completions/max_terminated_length": 834.25, | |
| "completions/mean_length": 571.515625, | |
| "completions/mean_terminated_length": 467.75330581665037, | |
| "completions/min_length": 144.75, | |
| "completions/min_terminated_length": 144.75, | |
| "epoch": 0.672, | |
| "grad_norm": 6.74008756789244, | |
| "kl": 3.23857421875, | |
| "learning_rate": 8.861479337315364e-07, | |
| "loss": 0.1582, | |
| "num_tokens": 29408978.0, | |
| "reward": 0.2046875, | |
| "reward_std": 0.24277526065707206, | |
| "rewards/accuracy_reward/mean": 0.2046875, | |
| "rewards/accuracy_reward/std": 0.32984848618507384, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 315 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1640625, | |
| "completions/max_length": 998.125, | |
| "completions/max_terminated_length": 869.1, | |
| "completions/mean_length": 573.5921875, | |
| "completions/mean_terminated_length": 491.0239372253418, | |
| "completions/min_length": 142.975, | |
| "completions/min_terminated_length": 142.975, | |
| "epoch": 0.6826666666666666, | |
| "grad_norm": 1.9324158149075346, | |
| "kl": 2.8765625, | |
| "learning_rate": 8.355214374680916e-07, | |
| "loss": 0.148, | |
| "num_tokens": 29870653.0, | |
| "reward": 0.1828125, | |
| "reward_std": 0.2596026983112097, | |
| "rewards/accuracy_reward/mean": 0.1828125, | |
| "rewards/accuracy_reward/std": 0.3082869052886963, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 320 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.175, | |
| "completions/max_length": 1012.875, | |
| "completions/max_terminated_length": 871.675, | |
| "completions/mean_length": 568.49375, | |
| "completions/mean_terminated_length": 474.15330467224123, | |
| "completions/min_length": 137.95, | |
| "completions/min_terminated_length": 137.95, | |
| "epoch": 0.6933333333333334, | |
| "grad_norm": 2.9206893390038653, | |
| "kl": 2.656689453125, | |
| "learning_rate": 7.858198654282411e-07, | |
| "loss": 0.1795, | |
| "num_tokens": 30318865.0, | |
| "reward": 0.2359375, | |
| "reward_std": 0.28795090727508066, | |
| "rewards/accuracy_reward/mean": 0.2359375, | |
| "rewards/accuracy_reward/std": 0.37537991255521774, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 325 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1640625, | |
| "completions/max_length": 1020.825, | |
| "completions/max_terminated_length": 861.275, | |
| "completions/mean_length": 556.4671875, | |
| "completions/mean_terminated_length": 466.5027053833008, | |
| "completions/min_length": 126.9, | |
| "completions/min_terminated_length": 126.9, | |
| "epoch": 0.704, | |
| "grad_norm": 1.2056546338068763, | |
| "kl": 3.62041015625, | |
| "learning_rate": 7.371123999611557e-07, | |
| "loss": 0.2013, | |
| "num_tokens": 30762364.0, | |
| "reward": 0.23125, | |
| "reward_std": 0.30185016691684724, | |
| "rewards/accuracy_reward/mean": 0.23125, | |
| "rewards/accuracy_reward/std": 0.3574558347463608, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 330 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.184375, | |
| "completions/max_length": 1005.925, | |
| "completions/max_terminated_length": 822.675, | |
| "completions/mean_length": 565.3328125, | |
| "completions/mean_terminated_length": 462.0029693603516, | |
| "completions/min_length": 116.5, | |
| "completions/min_terminated_length": 116.5, | |
| "epoch": 0.7146666666666667, | |
| "grad_norm": 1.3742588150692108, | |
| "kl": 3.824609375, | |
| "learning_rate": 6.894668396644439e-07, | |
| "loss": 0.1776, | |
| "num_tokens": 31211257.0, | |
| "reward": 0.215625, | |
| "reward_std": 0.2738583661615849, | |
| "rewards/accuracy_reward/mean": 0.215625, | |
| "rewards/accuracy_reward/std": 0.3528005562722683, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 335 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1828125, | |
| "completions/max_length": 1019.475, | |
| "completions/max_terminated_length": 851.825, | |
| "completions/mean_length": 577.940625, | |
| "completions/mean_terminated_length": 479.3160934448242, | |
| "completions/min_length": 123.675, | |
| "completions/min_terminated_length": 123.675, | |
| "epoch": 0.7253333333333334, | |
| "grad_norm": 1.4758651208465319, | |
| "kl": 2.486328125, | |
| "learning_rate": 6.429495050115576e-07, | |
| "loss": 0.1256, | |
| "num_tokens": 31673435.0, | |
| "reward": 0.2078125, | |
| "reward_std": 0.2131578464061022, | |
| "rewards/accuracy_reward/mean": 0.2078125, | |
| "rewards/accuracy_reward/std": 0.32094074562191965, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 340 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1515625, | |
| "completions/max_length": 1000.475, | |
| "completions/max_terminated_length": 834.775, | |
| "completions/mean_length": 517.709375, | |
| "completions/mean_terminated_length": 429.16680297851565, | |
| "completions/min_length": 92.775, | |
| "completions/min_terminated_length": 92.775, | |
| "epoch": 0.736, | |
| "grad_norm": 2.818679663631946, | |
| "kl": 1.72802734375, | |
| "learning_rate": 5.976251460366778e-07, | |
| "loss": 0.1104, | |
| "num_tokens": 32104025.0, | |
| "reward": 0.1953125, | |
| "reward_std": 0.24928856678307057, | |
| "rewards/accuracy_reward/mean": 0.1953125, | |
| "rewards/accuracy_reward/std": 0.30930979251861573, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 345 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.140625, | |
| "completions/max_length": 1007.825, | |
| "completions/max_terminated_length": 848.925, | |
| "completions/mean_length": 525.865625, | |
| "completions/mean_terminated_length": 444.4057159423828, | |
| "completions/min_length": 111.3, | |
| "completions/min_terminated_length": 111.3, | |
| "epoch": 0.7466666666666667, | |
| "grad_norm": 2.579459444619993, | |
| "kl": 2.3644287109375, | |
| "learning_rate": 5.535568522055892e-07, | |
| "loss": 0.15, | |
| "num_tokens": 32533547.0, | |
| "reward": 0.2140625, | |
| "reward_std": 0.26049785539507864, | |
| "rewards/accuracy_reward/mean": 0.2140625, | |
| "rewards/accuracy_reward/std": 0.3133357249200344, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 350 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1640625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 822.4, | |
| "completions/mean_length": 535.765625, | |
| "completions/mean_terminated_length": 441.6015251159668, | |
| "completions/min_length": 126.925, | |
| "completions/min_terminated_length": 126.925, | |
| "epoch": 0.7573333333333333, | |
| "grad_norm": 1.9814797295208133, | |
| "kl": 3.35029296875, | |
| "learning_rate": 5.108059645979812e-07, | |
| "loss": 0.1853, | |
| "num_tokens": 32969357.0, | |
| "reward": 0.21875, | |
| "reward_std": 0.28468980863690374, | |
| "rewards/accuracy_reward/mean": 0.21875, | |
| "rewards/accuracy_reward/std": 0.35949567407369615, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 355 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1578125, | |
| "completions/max_length": 972.05, | |
| "completions/max_terminated_length": 812.65, | |
| "completions/mean_length": 522.775, | |
| "completions/mean_terminated_length": 431.6031944274902, | |
| "completions/min_length": 99.15, | |
| "completions/min_terminated_length": 99.15, | |
| "epoch": 0.768, | |
| "grad_norm": 1.213077412434873, | |
| "kl": 2.9021240234375, | |
| "learning_rate": 4.694319905234322e-07, | |
| "loss": 0.1545, | |
| "num_tokens": 33399765.0, | |
| "reward": 0.2203125, | |
| "reward_std": 0.26060167923569677, | |
| "rewards/accuracy_reward/mean": 0.2203125, | |
| "rewards/accuracy_reward/std": 0.33386615812778475, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 360 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.171875, | |
| "completions/max_length": 1012.75, | |
| "completions/max_terminated_length": 848.125, | |
| "completions/mean_length": 540.746875, | |
| "completions/mean_terminated_length": 442.72081604003904, | |
| "completions/min_length": 106.125, | |
| "completions/min_terminated_length": 106.125, | |
| "epoch": 0.7786666666666666, | |
| "grad_norm": 2.7577857686089398, | |
| "kl": 3.180859375, | |
| "learning_rate": 4.2949252068991326e-07, | |
| "loss": 0.1558, | |
| "num_tokens": 33836803.0, | |
| "reward": 0.1703125, | |
| "reward_std": 0.24860655926167965, | |
| "rewards/accuracy_reward/mean": 0.1703125, | |
| "rewards/accuracy_reward/std": 0.3174226224422455, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 365 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1390625, | |
| "completions/max_length": 1014.45, | |
| "completions/max_terminated_length": 857.475, | |
| "completions/mean_length": 531.0265625, | |
| "completions/mean_terminated_length": 451.6034767150879, | |
| "completions/min_length": 87.1, | |
| "completions/min_terminated_length": 87.1, | |
| "epoch": 0.7893333333333333, | |
| "grad_norm": 1.9163359591102496, | |
| "kl": 2.898291015625, | |
| "learning_rate": 3.91043149040118e-07, | |
| "loss": 0.1431, | |
| "num_tokens": 34267244.0, | |
| "reward": 0.15625, | |
| "reward_std": 0.22862185277044772, | |
| "rewards/accuracy_reward/mean": 0.15625, | |
| "rewards/accuracy_reward/std": 0.3019771084189415, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 370 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1375, | |
| "completions/max_length": 999.2, | |
| "completions/max_terminated_length": 855.575, | |
| "completions/mean_length": 511.95, | |
| "completions/mean_terminated_length": 432.5935943603516, | |
| "completions/min_length": 115.575, | |
| "completions/min_terminated_length": 115.575, | |
| "epoch": 0.8, | |
| "grad_norm": 7.856282117594133, | |
| "kl": 2.960986328125, | |
| "learning_rate": 3.541373953671986e-07, | |
| "loss": 0.1759, | |
| "num_tokens": 34681772.0, | |
| "reward": 0.2546875, | |
| "reward_std": 0.2616026271134615, | |
| "rewards/accuracy_reward/mean": 0.2546875, | |
| "rewards/accuracy_reward/std": 0.32934130281209945, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 375 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1234375, | |
| "completions/max_length": 999.7, | |
| "completions/max_terminated_length": 814.5, | |
| "completions/mean_length": 519.0625, | |
| "completions/mean_terminated_length": 451.72740325927737, | |
| "completions/min_length": 147.0, | |
| "completions/min_terminated_length": 147.0, | |
| "epoch": 0.8106666666666666, | |
| "grad_norm": 1.5846595981949476, | |
| "kl": 1.841064453125, | |
| "learning_rate": 3.18826630817626e-07, | |
| "loss": 0.123, | |
| "num_tokens": 35113548.0, | |
| "reward": 0.3171875, | |
| "reward_std": 0.3077378235757351, | |
| "rewards/accuracy_reward/mean": 0.3171875, | |
| "rewards/accuracy_reward/std": 0.39782319590449333, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 380 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.165625, | |
| "completions/max_length": 984.15, | |
| "completions/max_terminated_length": 822.15, | |
| "completions/mean_length": 540.75625, | |
| "completions/mean_terminated_length": 446.8059135437012, | |
| "completions/min_length": 113.4, | |
| "completions/min_terminated_length": 113.4, | |
| "epoch": 0.8213333333333334, | |
| "grad_norm": 2.121434797315098, | |
| "kl": 2.864453125, | |
| "learning_rate": 2.851600063848651e-07, | |
| "loss": 0.1622, | |
| "num_tokens": 35554176.0, | |
| "reward": 0.2546875, | |
| "reward_std": 0.2852169893682003, | |
| "rewards/accuracy_reward/mean": 0.25625, | |
| "rewards/accuracy_reward/std": 0.36226451173424723, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 385 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.125, | |
| "completions/max_length": 994.7, | |
| "completions/max_terminated_length": 881.975, | |
| "completions/mean_length": 542.5953125, | |
| "completions/mean_terminated_length": 474.35446014404295, | |
| "completions/min_length": 147.075, | |
| "completions/min_terminated_length": 147.075, | |
| "epoch": 0.832, | |
| "grad_norm": 2.11442210090208, | |
| "kl": 3.627099609375, | |
| "learning_rate": 2.5318438449341047e-07, | |
| "loss": 0.1706, | |
| "num_tokens": 35991597.0, | |
| "reward": 0.2, | |
| "reward_std": 0.26254682019352915, | |
| "rewards/accuracy_reward/mean": 0.2, | |
| "rewards/accuracy_reward/std": 0.3453776828944683, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 390 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.18125, | |
| "completions/max_length": 1022.825, | |
| "completions/max_terminated_length": 830.85, | |
| "completions/mean_length": 552.690625, | |
| "completions/mean_terminated_length": 449.2019546508789, | |
| "completions/min_length": 128.5, | |
| "completions/min_terminated_length": 128.5, | |
| "epoch": 0.8426666666666667, | |
| "grad_norm": 1.5077720298338484, | |
| "kl": 3.9841552734375, | |
| "learning_rate": 2.2294427376840055e-07, | |
| "loss": 0.212, | |
| "num_tokens": 36435279.0, | |
| "reward": 0.2421875, | |
| "reward_std": 0.264066056907177, | |
| "rewards/accuracy_reward/mean": 0.2421875, | |
| "rewards/accuracy_reward/std": 0.3508764863014221, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 395 | |
| }, | |
| { | |
| "epoch": 0.8533333333333334, | |
| "grad_norm": 1.2192246871952803, | |
| "learning_rate": 1.9448176708161513e-07, | |
| "loss": 0.2023, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8533333333333334, | |
| "eval_clip_ratio": 0.0, | |
| "eval_completions/clipped_ratio": 0.150125, | |
| "eval_completions/max_length": 1005.3156, | |
| "eval_completions/max_terminated_length": 839.3408, | |
| "eval_completions/mean_length": 531.305675, | |
| "eval_completions/mean_terminated_length": 446.57037117919924, | |
| "eval_completions/min_length": 132.6728, | |
| "eval_completions/min_terminated_length": 132.6728, | |
| "eval_kl": 3.71179091796875, | |
| "eval_loss": 0.19600652158260345, | |
| "eval_num_tokens": 36868550.0, | |
| "eval_reward": 0.20215, | |
| "eval_reward_std": 0.25261726505756377, | |
| "eval_rewards/accuracy_reward/mean": 0.20215, | |
| "eval_rewards/accuracy_reward/std": 0.3302568965673447, | |
| "eval_rewards/format_reward/mean": 0.0, | |
| "eval_rewards/format_reward/std": 0.0, | |
| "eval_runtime": 17634.5575, | |
| "eval_samples_per_second": 0.284, | |
| "eval_steps_per_second": 0.018, | |
| "step": 400 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.15, | |
| "completions/max_length": 1005.6875, | |
| "completions/max_terminated_length": 833.475, | |
| "completions/mean_length": 535.62890625, | |
| "completions/mean_terminated_length": 452.4154224395752, | |
| "completions/min_length": 145.45, | |
| "completions/min_terminated_length": 145.45, | |
| "epoch": 0.864, | |
| "grad_norm": 4.507654850853612, | |
| "kl": 3.559375, | |
| "learning_rate": 1.6783648296009728e-07, | |
| "loss": 0.1897, | |
| "num_tokens": 37296524.0, | |
| "reward": 0.23984375, | |
| "reward_std": 0.28821724094450474, | |
| "rewards/accuracy_reward/mean": 0.23984375, | |
| "rewards/accuracy_reward/std": 0.36311594024300575, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 405 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1515625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 885.625, | |
| "completions/mean_length": 544.4328125, | |
| "completions/mean_terminated_length": 461.6532211303711, | |
| "completions/min_length": 137.025, | |
| "completions/min_terminated_length": 137.025, | |
| "epoch": 0.8746666666666667, | |
| "grad_norm": 0.9085864417795622, | |
| "kl": 3.36923828125, | |
| "learning_rate": 1.430455104389463e-07, | |
| "loss": 0.1799, | |
| "num_tokens": 37734025.0, | |
| "reward": 0.215625, | |
| "reward_std": 0.2822155307978392, | |
| "rewards/accuracy_reward/mean": 0.215625, | |
| "rewards/accuracy_reward/std": 0.350068199634552, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 410 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1359375, | |
| "completions/max_length": 998.525, | |
| "completions/max_terminated_length": 864.525, | |
| "completions/mean_length": 531.0125, | |
| "completions/mean_terminated_length": 456.2215087890625, | |
| "completions/min_length": 119.025, | |
| "completions/min_terminated_length": 119.025, | |
| "epoch": 0.8853333333333333, | |
| "grad_norm": 1.4991122052195613, | |
| "kl": 3.038134765625, | |
| "learning_rate": 1.2014335743505234e-07, | |
| "loss": 0.132, | |
| "num_tokens": 38162537.0, | |
| "reward": 0.2125, | |
| "reward_std": 0.26438754238188267, | |
| "rewards/accuracy_reward/mean": 0.2125, | |
| "rewards/accuracy_reward/std": 0.35445903837680814, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 415 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1328125, | |
| "completions/max_length": 1004.3, | |
| "completions/max_terminated_length": 819.325, | |
| "completions/mean_length": 519.06875, | |
| "completions/mean_terminated_length": 442.86925506591797, | |
| "completions/min_length": 138.95, | |
| "completions/min_terminated_length": 138.95, | |
| "epoch": 0.896, | |
| "grad_norm": 2.3887742458853194, | |
| "kl": 2.613427734375, | |
| "learning_rate": 9.916190271363112e-08, | |
| "loss": 0.1644, | |
| "num_tokens": 38579869.0, | |
| "reward": 0.2765625, | |
| "reward_std": 0.3201499585062265, | |
| "rewards/accuracy_reward/mean": 0.2765625, | |
| "rewards/accuracy_reward/std": 0.3956493496894836, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 420 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.15625, | |
| "completions/max_length": 985.525, | |
| "completions/max_terminated_length": 810.875, | |
| "completions/mean_length": 539.99375, | |
| "completions/mean_terminated_length": 451.690470123291, | |
| "completions/min_length": 120.5, | |
| "completions/min_terminated_length": 120.5, | |
| "epoch": 0.9066666666666666, | |
| "grad_norm": 1.3401864604257274, | |
| "kl": 2.73095703125, | |
| "learning_rate": 8.013035151441828e-08, | |
| "loss": 0.171, | |
| "num_tokens": 39024553.0, | |
| "reward": 0.2234375, | |
| "reward_std": 0.25571046136319636, | |
| "rewards/accuracy_reward/mean": 0.2234375, | |
| "rewards/accuracy_reward/std": 0.31687583178281786, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 425 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1171875, | |
| "completions/max_length": 1005.1, | |
| "completions/max_terminated_length": 851.7, | |
| "completions/mean_length": 518.528125, | |
| "completions/mean_terminated_length": 452.1327270507812, | |
| "completions/min_length": 130.3, | |
| "completions/min_terminated_length": 130.3, | |
| "epoch": 0.9173333333333333, | |
| "grad_norm": 2.167470698605123, | |
| "kl": 3.103076171875, | |
| "learning_rate": 6.307519489929209e-08, | |
| "loss": 0.1795, | |
| "num_tokens": 39445811.0, | |
| "reward": 0.271875, | |
| "reward_std": 0.33798977397382257, | |
| "rewards/accuracy_reward/mean": 0.271875, | |
| "rewards/accuracy_reward/std": 0.3924663387238979, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 430 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1578125, | |
| "completions/max_length": 999.3, | |
| "completions/max_terminated_length": 834.65, | |
| "completions/mean_length": 530.0015625, | |
| "completions/mean_terminated_length": 438.1447872161865, | |
| "completions/min_length": 111.7, | |
| "completions/min_terminated_length": 111.7, | |
| "epoch": 0.928, | |
| "grad_norm": 1.846618428525097, | |
| "kl": 3.21512451171875, | |
| "learning_rate": 4.8020172877908494e-08, | |
| "loss": 0.1451, | |
| "num_tokens": 39879828.0, | |
| "reward": 0.19375, | |
| "reward_std": 0.26397562474012376, | |
| "rewards/accuracy_reward/mean": 0.19375, | |
| "rewards/accuracy_reward/std": 0.34237141236662866, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 435 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1546875, | |
| "completions/max_length": 985.5, | |
| "completions/max_terminated_length": 826.025, | |
| "completions/mean_length": 539.29375, | |
| "completions/mean_terminated_length": 453.41730575561525, | |
| "completions/min_length": 141.975, | |
| "completions/min_terminated_length": 141.975, | |
| "epoch": 0.9386666666666666, | |
| "grad_norm": 1.616441959368508, | |
| "kl": 2.7537109375, | |
| "learning_rate": 3.498624136267653e-08, | |
| "loss": 0.1327, | |
| "num_tokens": 40309160.0, | |
| "reward": 0.2390625, | |
| "reward_std": 0.28233275562524796, | |
| "rewards/accuracy_reward/mean": 0.2390625, | |
| "rewards/accuracy_reward/std": 0.3621255189180374, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 440 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1234375, | |
| "completions/max_length": 1013.775, | |
| "completions/max_terminated_length": 843.275, | |
| "completions/mean_length": 515.215625, | |
| "completions/mean_terminated_length": 444.84048385620116, | |
| "completions/min_length": 131.525, | |
| "completions/min_terminated_length": 131.525, | |
| "epoch": 0.9493333333333334, | |
| "grad_norm": 1.9458890335400199, | |
| "kl": 2.9161376953125, | |
| "learning_rate": 2.39915429990733e-08, | |
| "loss": 0.166, | |
| "num_tokens": 40735346.0, | |
| "reward": 0.1859375, | |
| "reward_std": 0.26107093654572966, | |
| "rewards/accuracy_reward/mean": 0.1859375, | |
| "rewards/accuracy_reward/std": 0.3356678240001202, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 445 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.14375, | |
| "completions/max_length": 994.675, | |
| "completions/max_terminated_length": 836.625, | |
| "completions/mean_length": 542.1921875, | |
| "completions/mean_terminated_length": 463.8310737609863, | |
| "completions/min_length": 143.95, | |
| "completions/min_terminated_length": 143.95, | |
| "epoch": 0.96, | |
| "grad_norm": 1.801995464485617, | |
| "kl": 3.32373046875, | |
| "learning_rate": 1.5051381911898253e-08, | |
| "loss": 0.1805, | |
| "num_tokens": 41171661.0, | |
| "reward": 0.190625, | |
| "reward_std": 0.24303234927356243, | |
| "rewards/accuracy_reward/mean": 0.190625, | |
| "rewards/accuracy_reward/std": 0.30271864533424375, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 450 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.15625, | |
| "completions/max_length": 1004.325, | |
| "completions/max_terminated_length": 831.425, | |
| "completions/mean_length": 526.859375, | |
| "completions/mean_terminated_length": 436.33713760375974, | |
| "completions/min_length": 115.4, | |
| "completions/min_terminated_length": 115.4, | |
| "epoch": 0.9706666666666667, | |
| "grad_norm": 3.175555647761817, | |
| "kl": 3.665087890625, | |
| "learning_rate": 8.178202402621349e-09, | |
| "loss": 0.2182, | |
| "num_tokens": 41601355.0, | |
| "reward": 0.2296875, | |
| "reward_std": 0.2530842948704958, | |
| "rewards/accuracy_reward/mean": 0.2296875, | |
| "rewards/accuracy_reward/std": 0.3312860429286957, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 455 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1375, | |
| "completions/max_length": 1002.925, | |
| "completions/max_terminated_length": 839.05, | |
| "completions/mean_length": 538.9109375, | |
| "completions/mean_terminated_length": 461.0768562316895, | |
| "completions/min_length": 141.975, | |
| "completions/min_terminated_length": 141.975, | |
| "epoch": 0.9813333333333333, | |
| "grad_norm": 1.693673832434065, | |
| "kl": 2.9865234375, | |
| "learning_rate": 3.381571627475488e-09, | |
| "loss": 0.1702, | |
| "num_tokens": 42034314.0, | |
| "reward": 0.221875, | |
| "reward_std": 0.2828045580536127, | |
| "rewards/accuracy_reward/mean": 0.221875, | |
| "rewards/accuracy_reward/std": 0.3636058814823627, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 460 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.1609375, | |
| "completions/max_length": 996.125, | |
| "completions/max_terminated_length": 834.275, | |
| "completions/mean_length": 526.0546875, | |
| "completions/mean_terminated_length": 432.3683372497559, | |
| "completions/min_length": 118.925, | |
| "completions/min_terminated_length": 118.925, | |
| "epoch": 0.992, | |
| "grad_norm": 1.2046637410754644, | |
| "kl": 3.300732421875, | |
| "learning_rate": 6.681662804065569e-10, | |
| "loss": 0.1793, | |
| "num_tokens": 42460397.0, | |
| "reward": 0.2484375, | |
| "reward_std": 0.2802219405770302, | |
| "rewards/accuracy_reward/mean": 0.2484375, | |
| "rewards/accuracy_reward/std": 0.36698284819722177, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 465 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completions/clipped_ratio": 0.15104166666666666, | |
| "completions/max_length": 965.4166666666666, | |
| "completions/max_terminated_length": 821.9166666666666, | |
| "completions/mean_length": 551.5199699401855, | |
| "completions/mean_terminated_length": 467.1291364034017, | |
| "completions/min_length": 121.79166666666667, | |
| "completions/min_terminated_length": 121.79166666666667, | |
| "epoch": 0.9984, | |
| "kl": 3.1354166666666665, | |
| "num_tokens": 42724152.0, | |
| "reward": 0.2421875, | |
| "reward_std": 0.28976076406737167, | |
| "rewards/accuracy_reward/mean": 0.2421875, | |
| "rewards/accuracy_reward/std": 0.35445400203267735, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 468, | |
| "total_flos": 0.0, | |
| "train_loss": 0.12545046692467335, | |
| "train_runtime": 98008.0298, | |
| "train_samples_per_second": 0.077, | |
| "train_steps_per_second": 0.005 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 468, | |
| "num_input_tokens_seen": 42724152, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |