| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.4605543710021323, | |
| "eval_steps": 10, | |
| "global_step": 201, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "completion_length": 605.8180999755859, | |
| "epoch": 0.017057569296375266, | |
| "grad_norm": 11.068035125732422, | |
| "kl": 0.0, | |
| "learning_rate": 6.000000000000001e-07, | |
| "loss": 0.0, | |
| "reward": 0.6227678880095482, | |
| "reward_std": 0.34988817013800144, | |
| "rewards/accuracy_reward": 0.6227678880095482, | |
| "rewards/format_reward": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "completion_length": 606.5859651565552, | |
| "epoch": 0.08528784648187633, | |
| "grad_norm": 10.528549194335938, | |
| "kl": 0.0009853541851043701, | |
| "learning_rate": 3e-06, | |
| "loss": 0.0, | |
| "reward": 0.5948660997673869, | |
| "reward_std": 0.3529963130131364, | |
| "rewards/accuracy_reward": 0.5948660997673869, | |
| "rewards/format_reward": 0.0, | |
| "step": 5 | |
| }, | |
| { | |
| "completion_length": 623.1681053161622, | |
| "epoch": 0.17057569296375266, | |
| "grad_norm": 0.528905987739563, | |
| "kl": 0.008502006530761719, | |
| "learning_rate": 2.9095389311788626e-06, | |
| "loss": 0.0003, | |
| "reward": 0.689285746216774, | |
| "reward_std": 0.29115155190229414, | |
| "rewards/accuracy_reward": 0.689285746216774, | |
| "rewards/format_reward": 0.0, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.17057569296375266, | |
| "eval_completion_length": 573.9749799455915, | |
| "eval_kl": 1.3873639787946428, | |
| "eval_loss": 0.056397709995508194, | |
| "eval_reward": 0.7653061492102486, | |
| "eval_reward_std": 0.2849572343485696, | |
| "eval_rewards/accuracy_reward": 0.7653061492102486, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 229.3061, | |
| "eval_samples_per_second": 0.436, | |
| "eval_steps_per_second": 0.004, | |
| "step": 10 | |
| }, | |
| { | |
| "completion_length": 624.6886390686035, | |
| "epoch": 0.255863539445629, | |
| "grad_norm": 0.8065505623817444, | |
| "kl": 0.014313316345214844, | |
| "learning_rate": 2.649066664678467e-06, | |
| "loss": 0.0006, | |
| "reward": 0.7506696745753288, | |
| "reward_std": 0.23205664344131946, | |
| "rewards/accuracy_reward": 0.7506696745753288, | |
| "rewards/format_reward": 0.0, | |
| "step": 15 | |
| }, | |
| { | |
| "completion_length": 613.3145324707032, | |
| "epoch": 0.3411513859275053, | |
| "grad_norm": 0.18300795555114746, | |
| "kl": 0.38666706085205077, | |
| "learning_rate": 2.25e-06, | |
| "loss": 0.0154, | |
| "reward": 0.7591518208384513, | |
| "reward_std": 0.21772289499640465, | |
| "rewards/accuracy_reward": 0.7591518208384513, | |
| "rewards/format_reward": 0.0, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.3411513859275053, | |
| "eval_completion_length": 533.1733834402902, | |
| "eval_kl": 0.005168369838169643, | |
| "eval_loss": 0.0009401756688021123, | |
| "eval_reward": 0.812500034059797, | |
| "eval_reward_std": 0.22666969469615392, | |
| "eval_rewards/accuracy_reward": 0.812500034059797, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 225.084, | |
| "eval_samples_per_second": 0.444, | |
| "eval_steps_per_second": 0.004, | |
| "step": 20 | |
| }, | |
| { | |
| "completion_length": 611.4100685119629, | |
| "epoch": 0.42643923240938164, | |
| "grad_norm": 0.2118988335132599, | |
| "kl": 0.003771209716796875, | |
| "learning_rate": 1.7604722665003958e-06, | |
| "loss": 0.0002, | |
| "reward": 0.7513393223285675, | |
| "reward_std": 0.19753552060574292, | |
| "rewards/accuracy_reward": 0.7513393223285675, | |
| "rewards/format_reward": 0.0, | |
| "step": 25 | |
| }, | |
| { | |
| "completion_length": 612.1899833679199, | |
| "epoch": 0.511727078891258, | |
| "grad_norm": 0.2857199013233185, | |
| "kl": 0.003532218933105469, | |
| "learning_rate": 1.2395277334996047e-06, | |
| "loss": 0.0001, | |
| "reward": 0.7535714626312255, | |
| "reward_std": 0.17617593705654144, | |
| "rewards/accuracy_reward": 0.7535714626312255, | |
| "rewards/format_reward": 0.0, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.511727078891258, | |
| "eval_completion_length": 524.0120588030134, | |
| "eval_kl": 0.004640851702008929, | |
| "eval_loss": 0.000920308637432754, | |
| "eval_reward": 0.8431122728756496, | |
| "eval_reward_std": 0.16344326840979712, | |
| "eval_rewards/accuracy_reward": 0.8431122728756496, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 223.9002, | |
| "eval_samples_per_second": 0.447, | |
| "eval_steps_per_second": 0.004, | |
| "step": 30 | |
| }, | |
| { | |
| "completion_length": 600.6205627441407, | |
| "epoch": 0.5970149253731343, | |
| "grad_norm": 0.10232926905155182, | |
| "kl": 0.00328369140625, | |
| "learning_rate": 7.500000000000003e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7500000350177288, | |
| "reward_std": 0.18227138370275497, | |
| "rewards/accuracy_reward": 0.7500000350177288, | |
| "rewards/format_reward": 0.0, | |
| "step": 35 | |
| }, | |
| { | |
| "completion_length": 588.3980171203614, | |
| "epoch": 0.6823027718550106, | |
| "grad_norm": 0.12149345874786377, | |
| "kl": 0.013450050354003906, | |
| "learning_rate": 3.5093333532153313e-07, | |
| "loss": 0.0005, | |
| "reward": 0.7700893267989158, | |
| "reward_std": 0.18259718772023917, | |
| "rewards/accuracy_reward": 0.7700893267989158, | |
| "rewards/format_reward": 0.0, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.6823027718550106, | |
| "eval_completion_length": 515.6179722377232, | |
| "eval_kl": 0.005268641880580357, | |
| "eval_loss": 0.00021133186237420887, | |
| "eval_reward": 0.8303571854318891, | |
| "eval_reward_std": 0.1784549355506897, | |
| "eval_rewards/accuracy_reward": 0.8303571854318891, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 223.7869, | |
| "eval_samples_per_second": 0.447, | |
| "eval_steps_per_second": 0.004, | |
| "step": 40 | |
| }, | |
| { | |
| "completion_length": 608.0940032958985, | |
| "epoch": 0.767590618336887, | |
| "grad_norm": 0.11995360255241394, | |
| "kl": 0.0032606124877929688, | |
| "learning_rate": 9.046106882113752e-08, | |
| "loss": 0.0001, | |
| "reward": 0.7495535999536515, | |
| "reward_std": 0.19272922482341528, | |
| "rewards/accuracy_reward": 0.7495535999536515, | |
| "rewards/format_reward": 0.0, | |
| "step": 45 | |
| }, | |
| { | |
| "completion_length": 614.151146697998, | |
| "epoch": 0.8528784648187633, | |
| "grad_norm": 2.694963216781616, | |
| "kl": 0.004784202575683594, | |
| "learning_rate": 0.0, | |
| "loss": 0.0002, | |
| "reward": 0.7437500387430191, | |
| "reward_std": 0.19476121049374343, | |
| "rewards/accuracy_reward": 0.7437500387430191, | |
| "rewards/format_reward": 0.0, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.8528784648187633, | |
| "eval_completion_length": 521.0712454659598, | |
| "eval_kl": 0.005817958286830357, | |
| "eval_loss": 0.00023309080279432237, | |
| "eval_reward": 0.8392857568604606, | |
| "eval_reward_std": 0.2051741225378854, | |
| "eval_rewards/accuracy_reward": 0.8392857568604606, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 225.9, | |
| "eval_samples_per_second": 0.443, | |
| "eval_steps_per_second": 0.004, | |
| "step": 50 | |
| }, | |
| { | |
| "completion_length": 588.8913177490234, | |
| "epoch": 0.9381663113006397, | |
| "grad_norm": 0.08098618686199188, | |
| "kl": 0.0032186508178710938, | |
| "learning_rate": 1.5e-06, | |
| "loss": 0.0001, | |
| "reward": 0.7845982491970063, | |
| "reward_std": 0.1832274880260229, | |
| "rewards/accuracy_reward": 0.7845982491970063, | |
| "rewards/format_reward": 0.0, | |
| "step": 55 | |
| }, | |
| { | |
| "completion_length": 579.3964538574219, | |
| "epoch": 1.0341151385927505, | |
| "grad_norm": 0.12469803541898727, | |
| "kl": 0.003044976128472222, | |
| "learning_rate": 1.2395277334996047e-06, | |
| "loss": 0.0001, | |
| "reward": 0.7807540045844183, | |
| "reward_std": 0.16871399796671338, | |
| "rewards/accuracy_reward": 0.7807540045844183, | |
| "rewards/format_reward": 0.0, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0341151385927505, | |
| "eval_completion_length": 504.27440970284596, | |
| "eval_kl": 0.005137852260044643, | |
| "eval_loss": 0.00020576055976562202, | |
| "eval_reward": 0.8545918720109122, | |
| "eval_reward_std": 0.15936485332037723, | |
| "eval_rewards/accuracy_reward": 0.8545918720109122, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 218.3536, | |
| "eval_samples_per_second": 0.458, | |
| "eval_steps_per_second": 0.005, | |
| "step": 60 | |
| }, | |
| { | |
| "completion_length": 566.6837295532226, | |
| "epoch": 1.1194029850746268, | |
| "grad_norm": 0.08429302275180817, | |
| "kl": 0.0031642913818359375, | |
| "learning_rate": 9.86969785011497e-07, | |
| "loss": 0.0001, | |
| "reward": 0.783928607404232, | |
| "reward_std": 0.16070863213390113, | |
| "rewards/accuracy_reward": 0.783928607404232, | |
| "rewards/format_reward": 0.0, | |
| "step": 65 | |
| }, | |
| { | |
| "completion_length": 565.8973487854004, | |
| "epoch": 1.2046908315565032, | |
| "grad_norm": 0.1427806168794632, | |
| "kl": 0.0034132003784179688, | |
| "learning_rate": 7.500000000000003e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7837053954601287, | |
| "reward_std": 0.14835985810495914, | |
| "rewards/accuracy_reward": 0.7837053954601287, | |
| "rewards/format_reward": 0.0, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2046908315565032, | |
| "eval_completion_length": 487.16007341657365, | |
| "eval_kl": 0.006696428571428571, | |
| "eval_loss": 0.0002668871602509171, | |
| "eval_reward": 0.8686224818229675, | |
| "eval_reward_std": 0.14786859175988606, | |
| "eval_rewards/accuracy_reward": 0.8686224818229675, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 220.7164, | |
| "eval_samples_per_second": 0.453, | |
| "eval_steps_per_second": 0.005, | |
| "step": 70 | |
| }, | |
| { | |
| "completion_length": 578.0143165588379, | |
| "epoch": 1.2899786780383795, | |
| "grad_norm": 0.16407515108585358, | |
| "kl": 0.0034801483154296873, | |
| "learning_rate": 5.358185854701909e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7758928909897804, | |
| "reward_std": 0.17325164508074523, | |
| "rewards/accuracy_reward": 0.7758928909897804, | |
| "rewards/format_reward": 0.0, | |
| "step": 75 | |
| }, | |
| { | |
| "completion_length": 567.1640846252442, | |
| "epoch": 1.375266524520256, | |
| "grad_norm": 0.06801605224609375, | |
| "kl": 0.0032863616943359375, | |
| "learning_rate": 3.5093333532153313e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7915178939700127, | |
| "reward_std": 0.14882502863183616, | |
| "rewards/accuracy_reward": 0.7915178939700127, | |
| "rewards/format_reward": 0.0, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.375266524520256, | |
| "eval_completion_length": 490.3266034807478, | |
| "eval_kl": 0.004799979073660714, | |
| "eval_loss": 0.0001912058360176161, | |
| "eval_reward": 0.8494898336274284, | |
| "eval_reward_std": 0.165963847722326, | |
| "eval_rewards/accuracy_reward": 0.8494898336274284, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 221.7278, | |
| "eval_samples_per_second": 0.451, | |
| "eval_steps_per_second": 0.005, | |
| "step": 80 | |
| }, | |
| { | |
| "completion_length": 568.004940032959, | |
| "epoch": 1.4605543710021323, | |
| "grad_norm": 0.11520688235759735, | |
| "kl": 0.00350799560546875, | |
| "learning_rate": 2.0096189432334195e-07, | |
| "loss": 0.0001, | |
| "reward": 0.791741105914116, | |
| "reward_std": 0.15448834607377648, | |
| "rewards/accuracy_reward": 0.791741105914116, | |
| "rewards/format_reward": 0.0, | |
| "step": 85 | |
| }, | |
| { | |
| "completion_length": 579.3893096923828, | |
| "epoch": 1.5458422174840085, | |
| "grad_norm": 0.096395343542099, | |
| "kl": 0.0031429290771484374, | |
| "learning_rate": 9.046106882113752e-08, | |
| "loss": 0.0001, | |
| "reward": 0.7698661059141159, | |
| "reward_std": 0.16837037783116102, | |
| "rewards/accuracy_reward": 0.7698661059141159, | |
| "rewards/format_reward": 0.0, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5458422174840085, | |
| "eval_completion_length": 489.6377737862723, | |
| "eval_kl": 0.005275181361607143, | |
| "eval_loss": 0.00021086516790091991, | |
| "eval_reward": 0.8596939103943961, | |
| "eval_reward_std": 0.16273953127009527, | |
| "eval_rewards/accuracy_reward": 0.8596939103943961, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 220.5157, | |
| "eval_samples_per_second": 0.453, | |
| "eval_steps_per_second": 0.005, | |
| "step": 90 | |
| }, | |
| { | |
| "completion_length": 579.5056037902832, | |
| "epoch": 1.6311300639658848, | |
| "grad_norm": 0.08153486996889114, | |
| "kl": 0.003292083740234375, | |
| "learning_rate": 2.278837048168797e-08, | |
| "loss": 0.0001, | |
| "reward": 0.7850446745753288, | |
| "reward_std": 0.17225408796221017, | |
| "rewards/accuracy_reward": 0.7850446745753288, | |
| "rewards/format_reward": 0.0, | |
| "step": 95 | |
| }, | |
| { | |
| "completion_length": 569.5910942077637, | |
| "epoch": 1.716417910447761, | |
| "grad_norm": 0.07987989485263824, | |
| "kl": 0.0033542633056640623, | |
| "learning_rate": 0.0, | |
| "loss": 0.0001, | |
| "reward": 0.797544676065445, | |
| "reward_std": 0.15488467887043952, | |
| "rewards/accuracy_reward": 0.797544676065445, | |
| "rewards/format_reward": 0.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.716417910447761, | |
| "eval_completion_length": 488.6938912527902, | |
| "eval_kl": 0.005329677036830357, | |
| "eval_loss": 0.0009469491196796298, | |
| "eval_reward": 0.8507653389658246, | |
| "eval_reward_std": 0.1611656959035567, | |
| "eval_rewards/accuracy_reward": 0.8507653389658246, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 224.3175, | |
| "eval_samples_per_second": 0.446, | |
| "eval_steps_per_second": 0.004, | |
| "step": 100 | |
| }, | |
| { | |
| "completion_length": 583.1759201049805, | |
| "epoch": 1.8017057569296375, | |
| "grad_norm": 0.10411060601472855, | |
| "kl": 0.003289031982421875, | |
| "learning_rate": 7.500000000000003e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7611607491970063, | |
| "reward_std": 0.16414006650447846, | |
| "rewards/accuracy_reward": 0.7611607491970063, | |
| "rewards/format_reward": 0.0, | |
| "step": 105 | |
| }, | |
| { | |
| "completion_length": 582.3049369812012, | |
| "epoch": 1.886993603411514, | |
| "grad_norm": 0.15299023687839508, | |
| "kl": 0.003238677978515625, | |
| "learning_rate": 6.04262112445821e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7812500387430191, | |
| "reward_std": 0.16288615837693216, | |
| "rewards/accuracy_reward": 0.7812500387430191, | |
| "rewards/format_reward": 0.0, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.886993603411514, | |
| "eval_completion_length": 482.53951154436385, | |
| "eval_kl": 0.005589076450892857, | |
| "eval_loss": 0.00022271877969615161, | |
| "eval_reward": 0.8698980041912624, | |
| "eval_reward_std": 0.148396335542202, | |
| "eval_rewards/accuracy_reward": 0.8698980041912624, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 218.9173, | |
| "eval_samples_per_second": 0.457, | |
| "eval_steps_per_second": 0.005, | |
| "step": 110 | |
| }, | |
| { | |
| "completion_length": 569.8573867797852, | |
| "epoch": 1.9722814498933903, | |
| "grad_norm": 0.06015090271830559, | |
| "kl": 0.0032745361328125, | |
| "learning_rate": 4.7063754319689976e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7970982477068901, | |
| "reward_std": 0.16446643033996225, | |
| "rewards/accuracy_reward": 0.7970982477068901, | |
| "rewards/format_reward": 0.0, | |
| "step": 115 | |
| }, | |
| { | |
| "completion_length": 567.0639133029514, | |
| "epoch": 2.068230277185501, | |
| "grad_norm": 0.17577879130840302, | |
| "kl": 0.003118896484375, | |
| "learning_rate": 3.5093333532153313e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7855159123738606, | |
| "reward_std": 0.16644797714220153, | |
| "rewards/accuracy_reward": 0.7855159123738606, | |
| "rewards/format_reward": 0.0, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.068230277185501, | |
| "eval_completion_length": 480.4123011997768, | |
| "eval_kl": 0.006166730608258929, | |
| "eval_loss": 0.00024634136934764683, | |
| "eval_reward": 0.848214328289032, | |
| "eval_reward_std": 0.1697287346635546, | |
| "eval_rewards/accuracy_reward": 0.848214328289032, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 217.3304, | |
| "eval_samples_per_second": 0.46, | |
| "eval_steps_per_second": 0.005, | |
| "step": 120 | |
| }, | |
| { | |
| "completion_length": 545.7962287902832, | |
| "epoch": 2.1535181236673773, | |
| "grad_norm": 0.15464283525943756, | |
| "kl": 0.00330963134765625, | |
| "learning_rate": 2.467682828805956e-07, | |
| "loss": 0.0001, | |
| "reward": 0.796428607404232, | |
| "reward_std": 0.15349247655831277, | |
| "rewards/accuracy_reward": 0.796428607404232, | |
| "rewards/format_reward": 0.0, | |
| "step": 125 | |
| }, | |
| { | |
| "completion_length": 574.9627479553222, | |
| "epoch": 2.2388059701492535, | |
| "grad_norm": 0.11914752423763275, | |
| "kl": 0.0033355712890625, | |
| "learning_rate": 1.5955103951488177e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7667411088943481, | |
| "reward_std": 0.15913840509019792, | |
| "rewards/accuracy_reward": 0.7667411088943481, | |
| "rewards/format_reward": 0.0, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.2388059701492535, | |
| "eval_completion_length": 475.7303728376116, | |
| "eval_kl": 0.005048479352678571, | |
| "eval_loss": 0.00020054024935234338, | |
| "eval_reward": 0.8596939189093453, | |
| "eval_reward_std": 0.14824596978724003, | |
| "eval_rewards/accuracy_reward": 0.8596939189093453, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 220.4076, | |
| "eval_samples_per_second": 0.454, | |
| "eval_steps_per_second": 0.005, | |
| "step": 130 | |
| }, | |
| { | |
| "completion_length": 557.3685493469238, | |
| "epoch": 2.3240938166311302, | |
| "grad_norm": 0.07406862825155258, | |
| "kl": 0.0036487579345703125, | |
| "learning_rate": 9.046106882113752e-08, | |
| "loss": 0.0001, | |
| "reward": 0.7955357536673546, | |
| "reward_std": 0.15597790591418742, | |
| "rewards/accuracy_reward": 0.7955357536673546, | |
| "rewards/format_reward": 0.0, | |
| "step": 135 | |
| }, | |
| { | |
| "completion_length": 564.0457908630372, | |
| "epoch": 2.4093816631130065, | |
| "grad_norm": 0.14250978827476501, | |
| "kl": 0.003148651123046875, | |
| "learning_rate": 4.0432694130264294e-08, | |
| "loss": 0.0001, | |
| "reward": 0.8013393253087997, | |
| "reward_std": 0.1507680752314627, | |
| "rewards/accuracy_reward": 0.8013393253087997, | |
| "rewards/format_reward": 0.0, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.4093816631130065, | |
| "eval_completion_length": 482.3264944893973, | |
| "eval_kl": 0.005327497209821429, | |
| "eval_loss": 0.00020964255963917822, | |
| "eval_reward": 0.8584184050559998, | |
| "eval_reward_std": 0.15163347657237733, | |
| "eval_rewards/accuracy_reward": 0.8584184050559998, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 216.7121, | |
| "eval_samples_per_second": 0.461, | |
| "eval_steps_per_second": 0.005, | |
| "step": 140 | |
| }, | |
| { | |
| "completion_length": 573.7694519042968, | |
| "epoch": 2.4946695095948828, | |
| "grad_norm": 0.11774393916130066, | |
| "kl": 0.0032398223876953123, | |
| "learning_rate": 1.0142463387085465e-08, | |
| "loss": 0.0001, | |
| "reward": 0.7705357506871223, | |
| "reward_std": 0.16194519642740487, | |
| "rewards/accuracy_reward": 0.7705357506871223, | |
| "rewards/format_reward": 0.0, | |
| "step": 145 | |
| }, | |
| { | |
| "completion_length": 554.9973442077637, | |
| "epoch": 2.579957356076759, | |
| "grad_norm": 0.10069162398576736, | |
| "kl": 0.003186798095703125, | |
| "learning_rate": 0.0, | |
| "loss": 0.0001, | |
| "reward": 0.7941964656114578, | |
| "reward_std": 0.15693624839186668, | |
| "rewards/accuracy_reward": 0.7941964656114578, | |
| "rewards/format_reward": 0.0, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.579957356076759, | |
| "eval_completion_length": 478.3360900878906, | |
| "eval_kl": 0.005617414202008929, | |
| "eval_loss": 0.00022248993627727032, | |
| "eval_reward": 0.8380102430071149, | |
| "eval_reward_std": 0.17689391651323863, | |
| "eval_rewards/accuracy_reward": 0.8380102430071149, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 216.3428, | |
| "eval_samples_per_second": 0.462, | |
| "eval_steps_per_second": 0.005, | |
| "step": 150 | |
| }, | |
| { | |
| "completion_length": 560.3859672546387, | |
| "epoch": 2.6652452025586353, | |
| "grad_norm": 0.08840309083461761, | |
| "kl": 0.0033924102783203123, | |
| "learning_rate": 4.3933982822017883e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7861607491970062, | |
| "reward_std": 0.15671849967911838, | |
| "rewards/accuracy_reward": 0.7861607491970062, | |
| "rewards/format_reward": 0.0, | |
| "step": 155 | |
| }, | |
| { | |
| "completion_length": 585.1167678833008, | |
| "epoch": 2.750533049040512, | |
| "grad_norm": 0.07753895968198776, | |
| "kl": 0.003240966796875, | |
| "learning_rate": 3.5093333532153313e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7700893253087997, | |
| "reward_std": 0.17175142988562583, | |
| "rewards/accuracy_reward": 0.7700893253087997, | |
| "rewards/format_reward": 0.0, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.750533049040512, | |
| "eval_completion_length": 480.5063912527902, | |
| "eval_kl": 0.005467006138392857, | |
| "eval_loss": 0.00021828452008776367, | |
| "eval_reward": 0.8647959572928292, | |
| "eval_reward_std": 0.14701131811099394, | |
| "eval_rewards/accuracy_reward": 0.8647959572928292, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 220.7103, | |
| "eval_samples_per_second": 0.453, | |
| "eval_steps_per_second": 0.005, | |
| "step": 160 | |
| }, | |
| { | |
| "completion_length": 578.4064949035644, | |
| "epoch": 2.835820895522388, | |
| "grad_norm": 0.09659363329410553, | |
| "kl": 0.00327301025390625, | |
| "learning_rate": 2.7127193356651214e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7720982506871223, | |
| "reward_std": 0.1718197108246386, | |
| "rewards/accuracy_reward": 0.7720982506871223, | |
| "rewards/format_reward": 0.0, | |
| "step": 165 | |
| }, | |
| { | |
| "completion_length": 568.9587310791015, | |
| "epoch": 2.9211087420042645, | |
| "grad_norm": 0.13311569392681122, | |
| "kl": 0.0035663604736328124, | |
| "learning_rate": 2.0096189432334195e-07, | |
| "loss": 0.0001, | |
| "reward": 0.7866071820259094, | |
| "reward_std": 0.16021770192310214, | |
| "rewards/accuracy_reward": 0.7866071820259094, | |
| "rewards/format_reward": 0.0, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.9211087420042645, | |
| "eval_completion_length": 485.0064479282924, | |
| "eval_kl": 0.005305698939732143, | |
| "eval_loss": 0.000213223320315592, | |
| "eval_reward": 0.8584183965410505, | |
| "eval_reward_std": 0.156809002161026, | |
| "eval_rewards/accuracy_reward": 0.8584183965410505, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 215.4697, | |
| "eval_samples_per_second": 0.464, | |
| "eval_steps_per_second": 0.005, | |
| "step": 170 | |
| }, | |
| { | |
| "completion_length": 568.291493733724, | |
| "epoch": 3.0170575692963753, | |
| "grad_norm": 0.1568661630153656, | |
| "kl": 0.0034328884548611113, | |
| "learning_rate": 1.405383194450251e-07, | |
| "loss": 0.0002, | |
| "reward": 0.798611146873898, | |
| "reward_std": 0.1725065532657835, | |
| "rewards/accuracy_reward": 0.798611146873898, | |
| "rewards/format_reward": 0.0, | |
| "step": 175 | |
| }, | |
| { | |
| "completion_length": 563.8743576049804, | |
| "epoch": 3.1023454157782515, | |
| "grad_norm": 0.09700077027082443, | |
| "kl": 0.00376434326171875, | |
| "learning_rate": 9.046106882113752e-08, | |
| "loss": 0.0002, | |
| "reward": 0.7857143208384514, | |
| "reward_std": 0.15767460195347666, | |
| "rewards/accuracy_reward": 0.7857143208384514, | |
| "rewards/format_reward": 0.0, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 3.1023454157782515, | |
| "eval_completion_length": 484.9557364327567, | |
| "eval_kl": 0.005445207868303571, | |
| "eval_loss": 0.00021743674005847424, | |
| "eval_reward": 0.8622449381010873, | |
| "eval_reward_std": 0.14529993225421226, | |
| "eval_rewards/accuracy_reward": 0.8622449381010873, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 222.0031, | |
| "eval_samples_per_second": 0.45, | |
| "eval_steps_per_second": 0.005, | |
| "step": 180 | |
| }, | |
| { | |
| "completion_length": 537.6498031616211, | |
| "epoch": 3.1876332622601278, | |
| "grad_norm": 0.09764347225427628, | |
| "kl": 0.003476715087890625, | |
| "learning_rate": 5.11112605663977e-08, | |
| "loss": 0.0001, | |
| "reward": 0.8029018208384514, | |
| "reward_std": 0.15667821522802114, | |
| "rewards/accuracy_reward": 0.8029018208384514, | |
| "rewards/format_reward": 0.0, | |
| "step": 185 | |
| }, | |
| { | |
| "completion_length": 580.5053825378418, | |
| "epoch": 3.272921108742004, | |
| "grad_norm": 0.10233303159475327, | |
| "kl": 0.0035114288330078125, | |
| "learning_rate": 2.278837048168797e-08, | |
| "loss": 0.0001, | |
| "reward": 0.7638393223285675, | |
| "reward_std": 0.18133262265473604, | |
| "rewards/accuracy_reward": 0.7638393223285675, | |
| "rewards/format_reward": 0.0, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.272921108742004, | |
| "eval_completion_length": 475.9809875488281, | |
| "eval_kl": 0.0054931640625, | |
| "eval_loss": 0.00022025364160072058, | |
| "eval_reward": 0.8609694242477417, | |
| "eval_reward_std": 0.14643230502094542, | |
| "eval_rewards/accuracy_reward": 0.8609694242477417, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 217.2878, | |
| "eval_samples_per_second": 0.46, | |
| "eval_steps_per_second": 0.005, | |
| "step": 190 | |
| }, | |
| { | |
| "completion_length": 554.5051574707031, | |
| "epoch": 3.3582089552238807, | |
| "grad_norm": 0.11225981265306473, | |
| "kl": 0.003415489196777344, | |
| "learning_rate": 5.707952862381682e-09, | |
| "loss": 0.0001, | |
| "reward": 0.802678607404232, | |
| "reward_std": 0.1451864595990628, | |
| "rewards/accuracy_reward": 0.802678607404232, | |
| "rewards/format_reward": 0.0, | |
| "step": 195 | |
| }, | |
| { | |
| "completion_length": 560.3868598937988, | |
| "epoch": 3.443496801705757, | |
| "grad_norm": 0.1300434023141861, | |
| "kl": 0.0034725189208984373, | |
| "learning_rate": 0.0, | |
| "loss": 0.0001, | |
| "reward": 0.799553605914116, | |
| "reward_std": 0.14306436143815518, | |
| "rewards/accuracy_reward": 0.799553605914116, | |
| "rewards/format_reward": 0.0, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.443496801705757, | |
| "eval_completion_length": 486.1337367466518, | |
| "eval_kl": 0.005728585379464286, | |
| "eval_loss": 0.00022739818086847663, | |
| "eval_reward": 0.871173518044608, | |
| "eval_reward_std": 0.13813474109130247, | |
| "eval_rewards/accuracy_reward": 0.871173518044608, | |
| "eval_rewards/format_reward": 0.0, | |
| "eval_runtime": 216.9308, | |
| "eval_samples_per_second": 0.461, | |
| "eval_steps_per_second": 0.005, | |
| "step": 200 | |
| }, | |
| { | |
| "completion_length": 571.4754753112793, | |
| "epoch": 3.4605543710021323, | |
| "kl": 0.0033359527587890625, | |
| "reward": 0.7745536044239998, | |
| "reward_std": 0.1426069471053779, | |
| "rewards/accuracy_reward": 0.7745536044239998, | |
| "rewards/format_reward": 0.0, | |
| "step": 201, | |
| "total_flos": 0.0, | |
| "train_loss": 3.630263656748468e-05, | |
| "train_runtime": 619.0643, | |
| "train_samples_per_second": 289.469, | |
| "train_steps_per_second": 0.323 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "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": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |