| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.997867803837953, | |
| "eval_steps": 100, | |
| "global_step": 117, | |
| "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": 0.259765625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1016.0, | |
| "completions/mean_length": 622.638671875, | |
| "completions/mean_terminated_length": 481.79156494140625, | |
| "completions/min_length": 6.0, | |
| "completions/min_terminated_length": 6.0, | |
| "epoch": 0.008528784648187633, | |
| "grad_norm": 0.8145178930391833, | |
| "kl": 0.00037097930908203125, | |
| "learning_rate": 0.0, | |
| "loss": 0.0196, | |
| "num_tokens": 392023.0, | |
| "reward": 0.634765625, | |
| "reward_std": 0.33141425251960754, | |
| "rewards/accuracy_reward/mean": 0.634765625, | |
| "rewards/accuracy_reward/std": 0.4819667339324951, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "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": 0.2314453125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1019.75, | |
| "completions/mean_length": 595.36865234375, | |
| "completions/mean_terminated_length": 466.6201705932617, | |
| "completions/min_length": 7.25, | |
| "completions/min_terminated_length": 7.25, | |
| "epoch": 0.042643923240938165, | |
| "grad_norm": 0.4773668247144155, | |
| "kl": 0.00041562318801879883, | |
| "learning_rate": 1e-06, | |
| "loss": 0.0477, | |
| "num_tokens": 1910946.0, | |
| "reward": 0.64501953125, | |
| "reward_std": 0.37129127234220505, | |
| "rewards/accuracy_reward/mean": 0.64501953125, | |
| "rewards/accuracy_reward/std": 0.47847116738557816, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 5 | |
| }, | |
| { | |
| "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": 0.2375, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1019.2, | |
| "completions/mean_length": 618.060546875, | |
| "completions/mean_terminated_length": 491.431787109375, | |
| "completions/min_length": 10.8, | |
| "completions/min_terminated_length": 10.8, | |
| "epoch": 0.08528784648187633, | |
| "grad_norm": 0.3298661973983867, | |
| "kl": 0.0007388591766357422, | |
| "learning_rate": 2.25e-06, | |
| "loss": 0.0181, | |
| "num_tokens": 3860773.0, | |
| "reward": 0.6421875, | |
| "reward_std": 0.33047744631767273, | |
| "rewards/accuracy_reward/mean": 0.6421875, | |
| "rewards/accuracy_reward/std": 0.4783483386039734, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 10 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.203125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1017.6, | |
| "completions/mean_length": 603.215625, | |
| "completions/mean_terminated_length": 495.9770202636719, | |
| "completions/min_length": 45.2, | |
| "completions/min_terminated_length": 45.2, | |
| "epoch": 0.1279317697228145, | |
| "grad_norm": 0.25911245136387523, | |
| "kl": 0.0020538330078125, | |
| "learning_rate": 2.9973151946516025e-06, | |
| "loss": 0.0687, | |
| "num_tokens": 5749941.0, | |
| "reward": 0.734375, | |
| "reward_std": 0.2392146944999695, | |
| "rewards/accuracy_reward/mean": 0.734375, | |
| "rewards/accuracy_reward/std": 0.44130025506019593, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 15 | |
| }, | |
| { | |
| "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": 0.15078125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1018.2, | |
| "completions/mean_length": 585.84765625, | |
| "completions/mean_terminated_length": 508.5999755859375, | |
| "completions/min_length": 148.8, | |
| "completions/min_terminated_length": 148.8, | |
| "epoch": 0.17057569296375266, | |
| "grad_norm": 0.19464089548857677, | |
| "kl": 0.0031925201416015624, | |
| "learning_rate": 2.9672214011007086e-06, | |
| "loss": 0.0434, | |
| "num_tokens": 7604799.0, | |
| "reward": 0.76953125, | |
| "reward_std": 0.181648451089859, | |
| "rewards/accuracy_reward/mean": 0.76953125, | |
| "rewards/accuracy_reward/std": 0.41696075201034544, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 20 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.1265625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1015.4, | |
| "completions/mean_length": 569.831640625, | |
| "completions/mean_terminated_length": 505.2327087402344, | |
| "completions/min_length": 126.2, | |
| "completions/min_terminated_length": 126.2, | |
| "epoch": 0.21321961620469082, | |
| "grad_norm": 0.11280922307838523, | |
| "kl": 0.003424835205078125, | |
| "learning_rate": 2.904352305959606e-06, | |
| "loss": 0.0414, | |
| "num_tokens": 9406560.0, | |
| "reward": 0.800390625, | |
| "reward_std": 0.1668152093887329, | |
| "rewards/accuracy_reward/mean": 0.800390625, | |
| "rewards/accuracy_reward/std": 0.3962217092514038, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 25 | |
| }, | |
| { | |
| "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": 0.123828125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1019.4, | |
| "completions/mean_length": 572.0875, | |
| "completions/mean_terminated_length": 508.3062744140625, | |
| "completions/min_length": 119.0, | |
| "completions/min_terminated_length": 119.0, | |
| "epoch": 0.255863539445629, | |
| "grad_norm": 0.17292784666143843, | |
| "kl": 0.0038557052612304688, | |
| "learning_rate": 2.8101123009256945e-06, | |
| "loss": 0.0444, | |
| "num_tokens": 11235280.0, | |
| "reward": 0.779296875, | |
| "reward_std": 0.18240008652210235, | |
| "rewards/accuracy_reward/mean": 0.779296875, | |
| "rewards/accuracy_reward/std": 0.4136933445930481, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 30 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.10546875, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1020.2, | |
| "completions/mean_length": 563.6625, | |
| "completions/mean_terminated_length": 509.4048156738281, | |
| "completions/min_length": 91.2, | |
| "completions/min_terminated_length": 91.2, | |
| "epoch": 0.29850746268656714, | |
| "grad_norm": 0.1417198452884755, | |
| "kl": 0.004206085205078125, | |
| "learning_rate": 2.6866065519845123e-06, | |
| "loss": 0.0393, | |
| "num_tokens": 13029192.0, | |
| "reward": 0.796875, | |
| "reward_std": 0.1687217354774475, | |
| "rewards/accuracy_reward/mean": 0.796875, | |
| "rewards/accuracy_reward/std": 0.40032246708869934, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 35 | |
| }, | |
| { | |
| "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": 0.11015625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1019.8, | |
| "completions/mean_length": 555.61328125, | |
| "completions/mean_terminated_length": 498.0840698242188, | |
| "completions/min_length": 122.6, | |
| "completions/min_terminated_length": 122.6, | |
| "epoch": 0.3411513859275053, | |
| "grad_norm": 0.2464658893666007, | |
| "kl": 0.003921127319335938, | |
| "learning_rate": 2.5365939734802974e-06, | |
| "loss": 0.0384, | |
| "num_tokens": 14806018.0, | |
| "reward": 0.796484375, | |
| "reward_std": 0.16222528517246246, | |
| "rewards/accuracy_reward/mean": 0.796484375, | |
| "rewards/accuracy_reward/std": 0.40181103348731995, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 40 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.121875, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1019.4, | |
| "completions/mean_length": 568.50703125, | |
| "completions/mean_terminated_length": 505.3303649902344, | |
| "completions/min_length": 146.8, | |
| "completions/min_terminated_length": 146.8, | |
| "epoch": 0.3837953091684435, | |
| "grad_norm": 0.229924275356646, | |
| "kl": 0.004242706298828125, | |
| "learning_rate": 2.3634255985285104e-06, | |
| "loss": 0.0377, | |
| "num_tokens": 16620316.0, | |
| "reward": 0.81171875, | |
| "reward_std": 0.16615791916847228, | |
| "rewards/accuracy_reward/mean": 0.81171875, | |
| "rewards/accuracy_reward/std": 0.39103145599365235, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 45 | |
| }, | |
| { | |
| "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": 0.122265625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1014.2, | |
| "completions/mean_length": 565.183984375, | |
| "completions/mean_terminated_length": 501.5477783203125, | |
| "completions/min_length": 119.6, | |
| "completions/min_terminated_length": 119.6, | |
| "epoch": 0.42643923240938164, | |
| "grad_norm": 0.2797579550491506, | |
| "kl": 0.00439910888671875, | |
| "learning_rate": 2.1709697224734488e-06, | |
| "loss": 0.0355, | |
| "num_tokens": 18421059.0, | |
| "reward": 0.766015625, | |
| "reward_std": 0.1700347661972046, | |
| "rewards/accuracy_reward/mean": 0.766015625, | |
| "rewards/accuracy_reward/std": 0.41948946118354796, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 50 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.111328125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1014.0, | |
| "completions/mean_length": 558.638671875, | |
| "completions/mean_terminated_length": 500.31190795898436, | |
| "completions/min_length": 114.4, | |
| "completions/min_terminated_length": 114.4, | |
| "epoch": 0.4690831556503198, | |
| "grad_norm": 0.2632961965405764, | |
| "kl": 0.0050487518310546875, | |
| "learning_rate": 1.963525491562421e-06, | |
| "loss": 0.0222, | |
| "num_tokens": 20196126.0, | |
| "reward": 0.7828125, | |
| "reward_std": 0.1633252739906311, | |
| "rewards/accuracy_reward/mean": 0.7828125, | |
| "rewards/accuracy_reward/std": 0.41108490228652955, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 55 | |
| }, | |
| { | |
| "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": 0.117578125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1020.4, | |
| "completions/mean_length": 581.921875, | |
| "completions/mean_terminated_length": 522.9397216796875, | |
| "completions/min_length": 131.6, | |
| "completions/min_terminated_length": 131.6, | |
| "epoch": 0.511727078891258, | |
| "grad_norm": 0.0965979552879068, | |
| "kl": 0.004494857788085937, | |
| "learning_rate": 1.7457268671227065e-06, | |
| "loss": 0.0396, | |
| "num_tokens": 22047438.0, | |
| "reward": 0.77734375, | |
| "reward_std": 0.16624618470668792, | |
| "rewards/accuracy_reward/mean": 0.77734375, | |
| "rewards/accuracy_reward/std": 0.41559609174728396, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 60 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.116015625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1014.8, | |
| "completions/mean_length": 552.442578125, | |
| "completions/mean_terminated_length": 490.71544189453124, | |
| "completions/min_length": 101.6, | |
| "completions/min_terminated_length": 101.6, | |
| "epoch": 0.5543710021321961, | |
| "grad_norm": 0.17017618787674732, | |
| "kl": 0.004914474487304687, | |
| "learning_rate": 1.5224391105228955e-06, | |
| "loss": 0.0323, | |
| "num_tokens": 23838739.0, | |
| "reward": 0.774609375, | |
| "reward_std": 0.1651999369263649, | |
| "rewards/accuracy_reward/mean": 0.774609375, | |
| "rewards/accuracy_reward/std": 0.4162830650806427, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 65 | |
| }, | |
| { | |
| "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": 0.122265625, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1019.2, | |
| "completions/mean_length": 572.6484375, | |
| "completions/mean_terminated_length": 510.603759765625, | |
| "completions/min_length": 98.0, | |
| "completions/min_terminated_length": 98.0, | |
| "epoch": 0.5970149253731343, | |
| "grad_norm": 0.12570880852383362, | |
| "kl": 0.004351425170898438, | |
| "learning_rate": 1.2986501012735172e-06, | |
| "loss": 0.0361, | |
| "num_tokens": 25673351.0, | |
| "reward": 0.75234375, | |
| "reward_std": 0.18619680404663086, | |
| "rewards/accuracy_reward/mean": 0.75234375, | |
| "rewards/accuracy_reward/std": 0.42559565901756286, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 70 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.105078125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1013.8, | |
| "completions/mean_length": 539.875, | |
| "completions/mean_terminated_length": 483.08538818359375, | |
| "completions/min_length": 105.2, | |
| "completions/min_terminated_length": 105.2, | |
| "epoch": 0.6396588486140725, | |
| "grad_norm": 0.19371389016243035, | |
| "kl": 0.005275344848632813, | |
| "learning_rate": 1.079358916040996e-06, | |
| "loss": 0.0372, | |
| "num_tokens": 27420743.0, | |
| "reward": 0.78046875, | |
| "reward_std": 0.17587369978427886, | |
| "rewards/accuracy_reward/mean": 0.78046875, | |
| "rewards/accuracy_reward/std": 0.41278024911880495, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 75 | |
| }, | |
| { | |
| "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": 0.107421875, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1014.6, | |
| "completions/mean_length": 551.752734375, | |
| "completions/mean_terminated_length": 495.41151733398436, | |
| "completions/min_length": 92.4, | |
| "completions/min_terminated_length": 92.4, | |
| "epoch": 0.6823027718550106, | |
| "grad_norm": 0.12351401523997702, | |
| "kl": 0.004492950439453125, | |
| "learning_rate": 8.69464157535652e-07, | |
| "loss": 0.0355, | |
| "num_tokens": 29190430.0, | |
| "reward": 0.775390625, | |
| "reward_std": 0.16248664855957032, | |
| "rewards/accuracy_reward/mean": 0.775390625, | |
| "rewards/accuracy_reward/std": 0.4170423984527588, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 80 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.10546875, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1017.8, | |
| "completions/mean_length": 564.80859375, | |
| "completions/mean_terminated_length": 510.5605102539063, | |
| "completions/min_length": 99.2, | |
| "completions/min_terminated_length": 99.2, | |
| "epoch": 0.7249466950959488, | |
| "grad_norm": 0.11800011260860456, | |
| "kl": 0.004195022583007813, | |
| "learning_rate": 6.736545278218464e-07, | |
| "loss": 0.0232, | |
| "num_tokens": 30987452.0, | |
| "reward": 0.792578125, | |
| "reward_std": 0.17066446840763091, | |
| "rewards/accuracy_reward/mean": 0.7951016902923584, | |
| "rewards/accuracy_reward/std": 0.40229706168174745, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 85 | |
| }, | |
| { | |
| "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": 0.108203125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1020.4, | |
| "completions/mean_length": 555.931640625, | |
| "completions/mean_terminated_length": 499.55921630859376, | |
| "completions/min_length": 100.0, | |
| "completions/min_terminated_length": 100.0, | |
| "epoch": 0.767590618336887, | |
| "grad_norm": 0.1766286378559812, | |
| "kl": 0.004351806640625, | |
| "learning_rate": 4.963040904617131e-07, | |
| "loss": 0.0305, | |
| "num_tokens": 32791613.0, | |
| "reward": 0.762890625, | |
| "reward_std": 0.16622219383716583, | |
| "rewards/accuracy_reward/mean": 0.762890625, | |
| "rewards/accuracy_reward/std": 0.42507564425468447, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 90 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.110546875, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1020.4, | |
| "completions/mean_length": 567.742578125, | |
| "completions/mean_terminated_length": 511.2064636230469, | |
| "completions/min_length": 89.6, | |
| "completions/min_terminated_length": 89.6, | |
| "epoch": 0.8102345415778252, | |
| "grad_norm": 0.13097760910471143, | |
| "kl": 0.004387664794921875, | |
| "learning_rate": 3.4137456116310087e-07, | |
| "loss": 0.0181, | |
| "num_tokens": 34613034.0, | |
| "reward": 0.76953125, | |
| "reward_std": 0.16518225371837617, | |
| "rewards/accuracy_reward/mean": 0.76953125, | |
| "rewards/accuracy_reward/std": 0.42095457315444945, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 95 | |
| }, | |
| { | |
| "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": 0.10234375, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1016.8, | |
| "completions/mean_length": 570.152734375, | |
| "completions/mean_terminated_length": 518.7150634765625, | |
| "completions/min_length": 120.8, | |
| "completions/min_terminated_length": 120.8, | |
| "epoch": 0.8528784648187633, | |
| "grad_norm": 0.1265224718572913, | |
| "kl": 0.004515457153320313, | |
| "learning_rate": 2.1232680959720086e-07, | |
| "loss": 0.0369, | |
| "num_tokens": 36434305.0, | |
| "reward": 0.77890625, | |
| "reward_std": 0.16256613433361053, | |
| "rewards/accuracy_reward/mean": 0.7815476179122924, | |
| "rewards/accuracy_reward/std": 0.41042966246604917, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.8528784648187633, | |
| "eval_clip_ratio/high_max": 0.0, | |
| "eval_clip_ratio/high_mean": 0.0, | |
| "eval_clip_ratio/low_mean": 0.0, | |
| "eval_clip_ratio/low_min": 0.0, | |
| "eval_clip_ratio/region_mean": 0.0, | |
| "eval_completions/clipped_ratio": 0.093025, | |
| "eval_completions/max_length": 1017.8096, | |
| "eval_completions/max_terminated_length": 956.9056, | |
| "eval_completions/mean_length": 544.63845, | |
| "eval_completions/mean_terminated_length": 497.0462825683594, | |
| "eval_completions/min_length": 184.3936, | |
| "eval_completions/min_terminated_length": 184.3936, | |
| "eval_kl": 0.00679013671875, | |
| "eval_loss": 0.01577703095972538, | |
| "eval_num_tokens": 36434305.0, | |
| "eval_reward": 0.694925, | |
| "eval_reward_std": 0.20438497549295426, | |
| "eval_rewards/accuracy_reward/mean": 0.694925, | |
| "eval_rewards/accuracy_reward/std": 0.44015464553833006, | |
| "eval_rewards/format_reward/mean": 0.0, | |
| "eval_rewards/format_reward/std": 0.0, | |
| "eval_runtime": 4105.356, | |
| "eval_samples_per_second": 1.218, | |
| "eval_steps_per_second": 0.019, | |
| "step": 100 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.093359375, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1006.6, | |
| "completions/mean_length": 551.971875, | |
| "completions/mean_terminated_length": 503.73955688476565, | |
| "completions/min_length": 124.4, | |
| "completions/min_terminated_length": 124.4, | |
| "epoch": 0.8955223880597015, | |
| "grad_norm": 0.11809578314174589, | |
| "kl": 0.004218292236328125, | |
| "learning_rate": 1.1204354928900495e-07, | |
| "loss": 0.0265, | |
| "num_tokens": 38198833.0, | |
| "reward": 0.803125, | |
| "reward_std": 0.1608460694551468, | |
| "rewards/accuracy_reward/mean": 0.803125, | |
| "rewards/accuracy_reward/std": 0.39661717414855957, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 105 | |
| }, | |
| { | |
| "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": 0.08984375, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1018.0, | |
| "completions/mean_length": 559.37890625, | |
| "completions/mean_terminated_length": 514.0422424316406, | |
| "completions/min_length": 139.2, | |
| "completions/min_terminated_length": 139.2, | |
| "epoch": 0.9381663113006397, | |
| "grad_norm": 0.1543236952292962, | |
| "kl": 0.0042022705078125, | |
| "learning_rate": 4.276494256069874e-08, | |
| "loss": 0.0293, | |
| "num_tokens": 39998331.0, | |
| "reward": 0.817578125, | |
| "reward_std": 0.17398287057876588, | |
| "rewards/accuracy_reward/mean": 0.817578125, | |
| "rewards/accuracy_reward/std": 0.3809166610240936, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 110 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.093359375, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1012.8, | |
| "completions/mean_length": 548.9046875, | |
| "completions/mean_terminated_length": 499.97105102539064, | |
| "completions/min_length": 94.4, | |
| "completions/min_terminated_length": 94.4, | |
| "epoch": 0.9808102345415778, | |
| "grad_norm": 0.15606339132567215, | |
| "kl": 0.00447540283203125, | |
| "learning_rate": 6.038559007141398e-09, | |
| "loss": 0.0288, | |
| "num_tokens": 41769847.0, | |
| "reward": 0.787890625, | |
| "reward_std": 0.1641687572002411, | |
| "rewards/accuracy_reward/mean": 0.787890625, | |
| "rewards/accuracy_reward/std": 0.4066555380821228, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 115 | |
| }, | |
| { | |
| "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": 0.08664344879518071, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 1010.0, | |
| "completions/mean_length": 558.7720031738281, | |
| "completions/mean_terminated_length": 514.6327362060547, | |
| "completions/min_length": 124.0, | |
| "completions/min_terminated_length": 124.0, | |
| "epoch": 0.997867803837953, | |
| "kl": 0.003989219665527344, | |
| "num_tokens": 42497058.0, | |
| "reward": 0.7802734375, | |
| "reward_std": 0.19359246641397476, | |
| "rewards/accuracy_reward/mean": 0.7802734375, | |
| "rewards/accuracy_reward/std": 0.41445116698741913, | |
| "rewards/format_reward/mean": 0.0, | |
| "rewards/format_reward/std": 0.0, | |
| "step": 117, | |
| "total_flos": 0.0, | |
| "train_loss": 0.03524629572708892, | |
| "train_runtime": 9378.8625, | |
| "train_samples_per_second": 0.8, | |
| "train_steps_per_second": 0.012 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 117, | |
| "num_input_tokens_seen": 42497058, | |
| "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 | |
| } | |