| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.2, | |
| "eval_steps": 500, | |
| "global_step": 200, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.8932701796293259, | |
| "epoch": 0.01, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 17.141883850097656, | |
| "learning_rate": 9.649999999999999e-07, | |
| "loss": -0.0, | |
| "num_tokens": 49344.0, | |
| "reward": -459.925, | |
| "reward_std": 173.0284324645996, | |
| "rewards/reward_func/mean": -459.925, | |
| "rewards/reward_func/std": 173.02844009399413, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.5676407814025879, | |
| "epoch": 0.02, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 12.152915000915527, | |
| "learning_rate": 9.15e-07, | |
| "loss": 0.0, | |
| "num_tokens": 104592.0, | |
| "reward": -394.4625, | |
| "reward_std": 221.84682922363282, | |
| "rewards/reward_func/mean": -394.4625, | |
| "rewards/reward_func/std": 221.84682998657226, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.20221070423722268, | |
| "epoch": 0.03, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 6.0900750160217285, | |
| "learning_rate": 8.65e-07, | |
| "loss": 0.0, | |
| "num_tokens": 152208.0, | |
| "reward": -79.8125, | |
| "reward_std": 144.8038761138916, | |
| "rewards/reward_func/mean": -79.8125, | |
| "rewards/reward_func/std": 144.80387077331542, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.09748994875699282, | |
| "epoch": 0.04, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.9179723262786865, | |
| "learning_rate": 8.149999999999999e-07, | |
| "loss": -0.0, | |
| "num_tokens": 201216.0, | |
| "reward": 9.3625, | |
| "reward_std": 81.3893858909607, | |
| "rewards/reward_func/mean": 9.3625, | |
| "rewards/reward_func/std": 81.38938302993775, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.08117130994796753, | |
| "epoch": 0.05, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.0272433757781982, | |
| "learning_rate": 7.65e-07, | |
| "loss": -0.0, | |
| "num_tokens": 252344.0, | |
| "reward": 60.8, | |
| "reward_std": 22.109106731414794, | |
| "rewards/reward_func/mean": 60.8, | |
| "rewards/reward_func/std": 22.109107208251952, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.0658429590985179, | |
| "epoch": 0.06, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.603912115097046, | |
| "learning_rate": 7.149999999999999e-07, | |
| "loss": 0.0, | |
| "num_tokens": 303800.0, | |
| "reward": 65.1875, | |
| "reward_std": 14.921637725830077, | |
| "rewards/reward_func/mean": 65.1875, | |
| "rewards/reward_func/std": 14.921638202667236, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.08412722386419773, | |
| "epoch": 0.07, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.701263189315796, | |
| "learning_rate": 6.65e-07, | |
| "loss": 0.0, | |
| "num_tokens": 355936.0, | |
| "reward": 74.2625, | |
| "reward_std": 19.148268938064575, | |
| "rewards/reward_func/mean": 74.2625, | |
| "rewards/reward_func/std": 19.14826898574829, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.08717511333525181, | |
| "epoch": 0.08, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.643070936203003, | |
| "learning_rate": 6.149999999999999e-07, | |
| "loss": -0.0, | |
| "num_tokens": 409632.0, | |
| "reward": 71.3, | |
| "reward_std": 19.00277919769287, | |
| "rewards/reward_func/mean": 71.3, | |
| "rewards/reward_func/std": 19.0027795791626, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.10131141170859337, | |
| "epoch": 0.09, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 4.0365071296691895, | |
| "learning_rate": 5.649999999999999e-07, | |
| "loss": 0.0, | |
| "num_tokens": 457048.0, | |
| "reward": 81.2875, | |
| "reward_std": 12.263180470466613, | |
| "rewards/reward_func/mean": 81.2875, | |
| "rewards/reward_func/std": 12.263180875778199, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.07999873682856559, | |
| "epoch": 0.1, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 4.262997150421143, | |
| "learning_rate": 5.149999999999999e-07, | |
| "loss": 0.0, | |
| "num_tokens": 506584.0, | |
| "reward": 73.05, | |
| "reward_std": 16.826302528381348, | |
| "rewards/reward_func/mean": 73.05, | |
| "rewards/reward_func/std": 16.82630310058594, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.07205904349684715, | |
| "epoch": 0.11, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.98404598236084, | |
| "learning_rate": 4.65e-07, | |
| "loss": 0.0, | |
| "num_tokens": 556800.0, | |
| "reward": 80.7875, | |
| "reward_std": 11.600197219848633, | |
| "rewards/reward_func/mean": 80.7875, | |
| "rewards/reward_func/std": 11.600197505950927, | |
| "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": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.08184156678617001, | |
| "epoch": 0.12, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 2.9209563732147217, | |
| "learning_rate": 4.1499999999999994e-07, | |
| "loss": -0.0, | |
| "num_tokens": 605280.0, | |
| "reward": 82.5625, | |
| "reward_std": 13.004734754562378, | |
| "rewards/reward_func/mean": 82.5625, | |
| "rewards/reward_func/std": 13.004735040664674, | |
| "step": 120 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.07969092782586813, | |
| "epoch": 0.13, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 4.116878509521484, | |
| "learning_rate": 3.65e-07, | |
| "loss": 0.0, | |
| "num_tokens": 657160.0, | |
| "reward": 55.6875, | |
| "reward_std": 24.997920417785643, | |
| "rewards/reward_func/mean": 55.6875, | |
| "rewards/reward_func/std": 24.99792127609253, | |
| "step": 130 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.08586088940501213, | |
| "epoch": 0.14, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.09967041015625, | |
| "learning_rate": 3.15e-07, | |
| "loss": 0.0, | |
| "num_tokens": 707272.0, | |
| "reward": 78.0375, | |
| "reward_std": 13.64857816696167, | |
| "rewards/reward_func/mean": 78.0375, | |
| "rewards/reward_func/std": 13.64857850074768, | |
| "step": 140 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.07407695688307285, | |
| "epoch": 0.15, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.6434996128082275, | |
| "learning_rate": 2.65e-07, | |
| "loss": 0.0, | |
| "num_tokens": 755712.0, | |
| "reward": 62.6875, | |
| "reward_std": 29.81237063407898, | |
| "rewards/reward_func/mean": 62.6875, | |
| "rewards/reward_func/std": 29.812370777130127, | |
| "step": 150 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.08506322149187326, | |
| "epoch": 0.16, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.3035037517547607, | |
| "learning_rate": 2.1499999999999998e-07, | |
| "loss": 0.0, | |
| "num_tokens": 806232.0, | |
| "reward": 80.3875, | |
| "reward_std": 12.722473907470704, | |
| "rewards/reward_func/mean": 80.3875, | |
| "rewards/reward_func/std": 12.722473907470704, | |
| "step": 160 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.07220943029969931, | |
| "epoch": 0.17, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 4.2111358642578125, | |
| "learning_rate": 1.65e-07, | |
| "loss": 0.0, | |
| "num_tokens": 858016.0, | |
| "reward": 81.125, | |
| "reward_std": 11.847685623168946, | |
| "rewards/reward_func/mean": 81.125, | |
| "rewards/reward_func/std": 11.84768624305725, | |
| "step": 170 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.07000144328922034, | |
| "epoch": 0.18, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.138835906982422, | |
| "learning_rate": 1.15e-07, | |
| "loss": 0.0, | |
| "num_tokens": 907688.0, | |
| "reward": 72.55, | |
| "reward_std": 17.49347562789917, | |
| "rewards/reward_func/mean": 72.55, | |
| "rewards/reward_func/std": 17.493476104736327, | |
| "step": 180 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.06578830443322659, | |
| "epoch": 0.19, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 4.317309379577637, | |
| "learning_rate": 6.5e-08, | |
| "loss": -0.0, | |
| "num_tokens": 960144.0, | |
| "reward": 78.3875, | |
| "reward_std": 13.952195310592652, | |
| "rewards/reward_func/mean": 78.3875, | |
| "rewards/reward_func/std": 13.952195501327514, | |
| "step": 190 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 1.0, | |
| "completions/max_length": 256.0, | |
| "completions/max_terminated_length": 0.0, | |
| "completions/mean_length": 256.0, | |
| "completions/mean_terminated_length": 0.0, | |
| "completions/min_length": 256.0, | |
| "completions/min_terminated_length": 0.0, | |
| "entropy": 0.06528270887210966, | |
| "epoch": 0.2, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 2.896949291229248, | |
| "learning_rate": 1.5e-08, | |
| "loss": -0.0, | |
| "num_tokens": 1006496.0, | |
| "reward": 75.8875, | |
| "reward_std": 12.274623441696168, | |
| "rewards/reward_func/mean": 75.8875, | |
| "rewards/reward_func/std": 12.274623727798462, | |
| "step": 200 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 1006496, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |