whodisidk's picture
End of training
8860844 verified
Raw
History Blame Contribute Delete
38.2 kB
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 9.097560975609756,
"eval_steps": 500,
"global_step": 100,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"completion_length": 170.46875,
"epoch": 0.1951219512195122,
"grad_norm": 6.099827289581299,
"kl": 0.5973675912246108,
"learning_rate": 1.6666666666666665e-07,
"loss": 0.0006,
"reward": 2.4682722240686417,
"reward_std": 0.4695738647133112,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.7388660460710526,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.3125,
"rewards/xmlcount_reward_func": 1.166906252503395,
"step": 2
},
{
"completion_length": 114.125,
"epoch": 0.3902439024390244,
"grad_norm": 5.554270267486572,
"kl": 1.5984905734658241,
"learning_rate": 5e-07,
"loss": 0.0016,
"reward": 2.225109323859215,
"reward_std": 0.6870723395841196,
"rewards/concensus_correctness_reward_func": 0.06012500077486038,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.5827342979609966,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.390625,
"rewards/xmlcount_reward_func": 1.1291249990463257,
"step": 4
},
{
"completion_length": 148.15625,
"epoch": 0.5853658536585366,
"grad_norm": 5.069469451904297,
"kl": 3.3319540731608868,
"learning_rate": 4.994757065594279e-07,
"loss": 0.0033,
"reward": 4.637000426650047,
"reward_std": 2.4823005218058825,
"rewards/concensus_correctness_reward_func": 1.4928749948740005,
"rewards/consensus_reward_func": 0.1875,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.875,
"rewards/question_recreation_reward_func": 0.7499066442251205,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.328125,
"rewards/xmlcount_reward_func": 1.0035937614738941,
"step": 6
},
{
"completion_length": 159.59375,
"epoch": 0.7804878048780488,
"grad_norm": 22.011062622070312,
"kl": 5.900226164609194,
"learning_rate": 4.979050253066063e-07,
"loss": 0.0059,
"reward": 2.358607694506645,
"reward_std": 0.5496477753622457,
"rewards/concensus_correctness_reward_func": 0.06012500077486038,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.7734514437615871,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.296875,
"rewards/xmlcount_reward_func": 1.1656562685966492,
"step": 8
},
{
"completion_length": 175.625,
"epoch": 0.975609756097561,
"grad_norm": 4.487865447998047,
"kl": 0.8899677954614162,
"learning_rate": 4.952945442245597e-07,
"loss": 0.0009,
"reward": 2.114151880145073,
"reward_std": 0.5495768673717976,
"rewards/concensus_correctness_reward_func": 0.0598750002682209,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.5913393385708332,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.25,
"rewards/xmlcount_reward_func": 1.1504375040531158,
"step": 10
},
{
"completion_length": 137.7,
"epoch": 1.0975609756097562,
"grad_norm": 5.117649078369141,
"kl": 0.46071633845567705,
"learning_rate": 4.916552125781528e-07,
"loss": 0.0003,
"reward": 3.7879202365875244,
"reward_std": 1.7626623034477233,
"rewards/concensus_correctness_reward_func": 0.4884000062942505,
"rewards/consensus_reward_func": 0.3,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.8,
"rewards/question_recreation_reward_func": 0.7498702496290207,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.3,
"rewards/xmlcount_reward_func": 1.1496500015258788,
"step": 12
},
{
"completion_length": 174.71875,
"epoch": 1.2926829268292683,
"grad_norm": 4.460564136505127,
"kl": 0.9902181550860405,
"learning_rate": 4.870022949890676e-07,
"loss": 0.001,
"reward": 2.211259737610817,
"reward_std": 0.41652211552718654,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.7058847844600677,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.328125,
"rewards/xmlcount_reward_func": 1.1772499978542328,
"step": 14
},
{
"completion_length": 121.15625,
"epoch": 1.4878048780487805,
"grad_norm": 5.077511787414551,
"kl": 2.23406695201993,
"learning_rate": 4.81355307410676e-07,
"loss": 0.0022,
"reward": 4.244883641600609,
"reward_std": 2.1680759941227734,
"rewards/concensus_correctness_reward_func": 1.4302500039339066,
"rewards/consensus_reward_func": 0.1875,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.9261960908770561,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.28125,
"rewards/xmlcount_reward_func": 1.1696875020861626,
"step": 16
},
{
"completion_length": 183.53125,
"epoch": 1.6829268292682928,
"grad_norm": 5.903010845184326,
"kl": 2.295085714198649,
"learning_rate": 4.747379352713488e-07,
"loss": 0.0023,
"reward": 2.392599403858185,
"reward_std": 0.9273251239210367,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.5,
"rewards/question_recreation_reward_func": 0.792318120598793,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.1875,
"rewards/xmlcount_reward_func": 0.9127812534570694,
"step": 18
},
{
"completion_length": 142.8125,
"epoch": 1.8780487804878048,
"grad_norm": 1.9898862838745117,
"kl": 0.6429034024477005,
"learning_rate": 4.6717793412953776e-07,
"loss": 0.0006,
"reward": 2.9371677935123444,
"reward_std": 0.8608561623841524,
"rewards/concensus_correctness_reward_func": 0.12025000154972076,
"rewards/consensus_reward_func": 0.125,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.5,
"rewards/question_recreation_reward_func": 0.6914178244769573,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.34375,
"rewards/xmlcount_reward_func": 1.1567500084638596,
"step": 20
},
{
"completion_length": 217.05,
"epoch": 2.0,
"grad_norm": 1.1733369827270508,
"kl": 4.470386900007725,
"learning_rate": 4.5870701325731773e-07,
"loss": 0.0028,
"reward": 2.0891710042953493,
"reward_std": 0.5484436511993408,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.7873210072517395,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.25,
"rewards/xmlcount_reward_func": 1.0518500089645386,
"step": 22
},
{
"completion_length": 139.375,
"epoch": 2.1951219512195124,
"grad_norm": 4.537793159484863,
"kl": 0.816438952460885,
"learning_rate": 4.4936070264068016e-07,
"loss": 0.0008,
"reward": 2.3938701897859573,
"reward_std": 0.34673458884935826,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.6716827414929867,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.296875,
"rewards/xmlcount_reward_func": 1.1753125041723251,
"step": 24
},
{
"completion_length": 245.0,
"epoch": 2.3902439024390243,
"grad_norm": 3.754828929901123,
"kl": 2.14161329343915,
"learning_rate": 4.391782039544238e-07,
"loss": 0.0021,
"reward": 4.017070218920708,
"reward_std": 2.1269510723650455,
"rewards/concensus_correctness_reward_func": 1.9351250007748604,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.5311014428734779,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.28125,
"rewards/xmlcount_reward_func": 0.9570937529206276,
"step": 26
},
{
"completion_length": 188.6875,
"epoch": 2.5853658536585367,
"grad_norm": 4.370049476623535,
"kl": 41.1987736299634,
"learning_rate": 4.282022261367073e-07,
"loss": 0.0412,
"reward": 3.2786895185709,
"reward_std": 2.1890580281615257,
"rewards/concensus_correctness_reward_func": 0.8053750023245811,
"rewards/consensus_reward_func": 0.1875,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.1875,
"rewards/question_recreation_reward_func": 0.7307832725346088,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.265625,
"rewards/xmlcount_reward_func": 1.1019062548875809,
"step": 28
},
{
"completion_length": 201.25,
"epoch": 2.7804878048780486,
"grad_norm": 3.327321767807007,
"kl": 0.41819524113088846,
"learning_rate": 4.1647880625292027e-07,
"loss": 0.0004,
"reward": 2.4246954917907715,
"reward_std": 0.5871567726135254,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.6943517699837685,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.34375,
"rewards/xmlcount_reward_func": 1.136593759059906,
"step": 30
},
{
"completion_length": 130.5625,
"epoch": 2.975609756097561,
"grad_norm": 4.2605109214782715,
"kl": 1.8532675383612514,
"learning_rate": 4.040571164002318e-07,
"loss": 0.0019,
"reward": 3.6226596236228943,
"reward_std": 2.205760184675455,
"rewards/concensus_correctness_reward_func": 0.8050000034272671,
"rewards/consensus_reward_func": 0.1875,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.4375,
"rewards/question_recreation_reward_func": 0.7430970743298531,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.3125,
"rewards/xmlcount_reward_func": 1.1370624974370003,
"step": 32
},
{
"completion_length": 104.65,
"epoch": 3.097560975609756,
"grad_norm": 5.712991237640381,
"kl": 0.6544200539588928,
"learning_rate": 3.909892574627266e-07,
"loss": 0.0004,
"reward": 1.9994364500045776,
"reward_std": 0.34640331044793127,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.6102864652872085,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.225,
"rewards/xmlcount_reward_func": 1.1641499996185303,
"step": 34
},
{
"completion_length": 164.625,
"epoch": 3.292682926829268,
"grad_norm": 3.5782923698425293,
"kl": 0.47456978261470795,
"learning_rate": 3.773300405821908e-07,
"loss": 0.0005,
"reward": 2.4073937088251114,
"reward_std": 0.8083425159566104,
"rewards/concensus_correctness_reward_func": 0.12012499943375587,
"rewards/consensus_reward_func": 0.125,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.7384249614551663,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.265625,
"rewards/xmlcount_reward_func": 1.1582187488675117,
"step": 36
},
{
"completion_length": 145.4375,
"epoch": 3.4878048780487805,
"grad_norm": 3.6999077796936035,
"kl": 0.7185129821300507,
"learning_rate": 3.6313675726113475e-07,
"loss": 0.0007,
"reward": 3.2990617603063583,
"reward_std": 1.678582351654768,
"rewards/concensus_correctness_reward_func": 0.625,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.5,
"rewards/question_recreation_reward_func": 0.6626242846250534,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.359375,
"rewards/xmlcount_reward_func": 1.1520625054836273,
"step": 38
},
{
"completion_length": 219.40625,
"epoch": 3.682926829268293,
"grad_norm": 3.1566965579986572,
"kl": 0.7571137994527817,
"learning_rate": 3.484689390623218e-07,
"loss": 0.0008,
"reward": 4.262388929724693,
"reward_std": 1.6960864844731987,
"rewards/concensus_correctness_reward_func": 1.875,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.7855764981359243,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.296875,
"rewards/xmlcount_reward_func": 1.0549375042319298,
"step": 40
},
{
"completion_length": 124.125,
"epoch": 3.8780487804878048,
"grad_norm": 5.992835998535156,
"kl": 4.541106943972409,
"learning_rate": 3.3338810791270517e-07,
"loss": 0.0045,
"reward": 3.2296720147132874,
"reward_std": 1.5638214345090091,
"rewards/concensus_correctness_reward_func": 0.625,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.7207032516598701,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.40625,
"rewards/xmlcount_reward_func": 1.2277187556028366,
"step": 42
},
{
"completion_length": 143.7,
"epoch": 4.0,
"grad_norm": 1.4582439661026,
"kl": 13.396405771374702,
"learning_rate": 3.179575180590857e-07,
"loss": 0.0084,
"reward": 5.374096298217774,
"reward_std": 3.2581148982048034,
"rewards/concensus_correctness_reward_func": 2.1923000037670137,
"rewards/consensus_reward_func": 0.2,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.8,
"rewards/question_recreation_reward_func": 0.8091962337493896,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.25,
"rewards/xmlcount_reward_func": 1.122599995136261,
"step": 44
},
{
"completion_length": 160.03125,
"epoch": 4.195121951219512,
"grad_norm": 4.867384910583496,
"kl": 1.008437397889793,
"learning_rate": 3.022418907578188e-07,
"loss": 0.001,
"reward": 3.2849575728178024,
"reward_std": 1.9591746106743813,
"rewards/concensus_correctness_reward_func": 0.6851250007748604,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.746645025908947,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.40625,
"rewards/xmlcount_reward_func": 1.1344375163316727,
"step": 46
},
{
"completion_length": 168.625,
"epoch": 4.390243902439025,
"grad_norm": 5.173135280609131,
"kl": 0.954070495441556,
"learning_rate": 2.863071428113726e-07,
"loss": 0.001,
"reward": 3.638337269425392,
"reward_std": 1.8555740788578987,
"rewards/concensus_correctness_reward_func": 1.25,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.6886498406529427,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.328125,
"rewards/xmlcount_reward_func": 1.1215624958276749,
"step": 48
},
{
"completion_length": 147.625,
"epoch": 4.585365853658536,
"grad_norm": 3.766300678253174,
"kl": 1.248833647929132,
"learning_rate": 2.7022011009035107e-07,
"loss": 0.0012,
"reward": 2.902125522494316,
"reward_std": 0.7541704792529345,
"rewards/concensus_correctness_reward_func": 0.12250000238418579,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.5,
"rewards/question_recreation_reward_func": 0.8029379919171333,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.265625,
"rewards/xmlcount_reward_func": 1.1485625207424164,
"step": 50
},
{
"completion_length": 221.5625,
"epoch": 4.780487804878049,
"grad_norm": 42.74195098876953,
"kl": 19.366400617174804,
"learning_rate": 2.540482672006254e-07,
"loss": 0.0194,
"reward": 2.4607008695602417,
"reward_std": 0.582477293908596,
"rewards/concensus_correctness_reward_func": 0.06012500077486038,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.7755759283900261,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.234375,
"rewards/xmlcount_reward_func": 1.0781250149011612,
"step": 52
},
{
"completion_length": 121.65625,
"epoch": 4.975609756097561,
"grad_norm": 4.044496059417725,
"kl": 1.2229662016034126,
"learning_rate": 2.37859444471388e-07,
"loss": 0.0012,
"reward": 3.0761840641498566,
"reward_std": 1.7169672884047031,
"rewards/concensus_correctness_reward_func": 0.6851250007748604,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.6576527766883373,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.296875,
"rewards/xmlcount_reward_func": 1.1240312531590462,
"step": 54
},
{
"completion_length": 190.15,
"epoch": 5.097560975609756,
"grad_norm": 2.5054287910461426,
"kl": 0.40689231753349303,
"learning_rate": 2.2172154345117894e-07,
"loss": 0.0003,
"reward": 2.3544807910919188,
"reward_std": 0.4992034614086151,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.4,
"rewards/question_recreation_reward_func": 0.6489308118820191,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.175,
"rewards/xmlcount_reward_func": 1.130550003051758,
"step": 56
},
{
"completion_length": 167.46875,
"epoch": 5.2926829268292686,
"grad_norm": 4.939743518829346,
"kl": 1.0670354459434748,
"learning_rate": 2.0570225210519433e-07,
"loss": 0.0011,
"reward": 2.940119907259941,
"reward_std": 0.5183741468936205,
"rewards/concensus_correctness_reward_func": 0.06012500077486038,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.5,
"rewards/question_recreation_reward_func": 0.7578699514269829,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.375,
"rewards/xmlcount_reward_func": 1.1846249997615814,
"step": 58
},
{
"completion_length": 129.78125,
"epoch": 5.487804878048781,
"grad_norm": 14.262602806091309,
"kl": 6.911861918866634,
"learning_rate": 1.8986876090843664e-07,
"loss": 0.0069,
"reward": 2.512383133172989,
"reward_std": 0.6689571645110846,
"rewards/concensus_correctness_reward_func": 0.06012500077486038,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.7206018827855587,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.25,
"rewards/xmlcount_reward_func": 1.1691562682390213,
"step": 60
},
{
"completion_length": 153.84375,
"epoch": 5.682926829268292,
"grad_norm": 4.040970802307129,
"kl": 1.8997416272759438,
"learning_rate": 1.7428748102551234e-07,
"loss": 0.0019,
"reward": 4.094300732016563,
"reward_std": 3.2450541853904724,
"rewards/concensus_correctness_reward_func": 1.3101250007748604,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.5,
"rewards/question_recreation_reward_func": 0.7321133241057396,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.359375,
"rewards/xmlcount_reward_func": 1.130187500268221,
"step": 62
},
{
"completion_length": 149.0,
"epoch": 5.878048780487805,
"grad_norm": 4.407651424407959,
"kl": 1.1698362613096833,
"learning_rate": 1.5902376575912814e-07,
"loss": 0.0012,
"reward": 2.2420406192541122,
"reward_std": 0.5333225438371301,
"rewards/concensus_correctness_reward_func": 0.0598750002682209,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.792165607213974,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.21875,
"rewards/xmlcount_reward_func": 1.1087500154972076,
"step": 64
},
{
"completion_length": 112.15,
"epoch": 6.0,
"grad_norm": 3.1488583087921143,
"kl": 0.7777493536472321,
"learning_rate": 1.4414163643562753e-07,
"loss": 0.0005,
"reward": 2.53222017288208,
"reward_std": 0.6459437549114228,
"rewards/concensus_correctness_reward_func": 0.09580000042915345,
"rewards/consensus_reward_func": 0.1,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.7860701441764831,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.35,
"rewards/xmlcount_reward_func": 1.2003499746322632,
"step": 66
},
{
"completion_length": 136.46875,
"epoch": 6.195121951219512,
"grad_norm": 5.8898539543151855,
"kl": 1.8731475602835417,
"learning_rate": 1.2970351387729872e-07,
"loss": 0.0019,
"reward": 2.372854396700859,
"reward_std": 0.46924145889352076,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.6067919060587883,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.34375,
"rewards/xmlcount_reward_func": 1.1723125129938126,
"step": 68
},
{
"completion_length": 162.78125,
"epoch": 6.390243902439025,
"grad_norm": 6.582674026489258,
"kl": 1.30052063241601,
"learning_rate": 1.1576995658775404e-07,
"loss": 0.0013,
"reward": 4.535748735070229,
"reward_std": 1.4339275071397424,
"rewards/concensus_correctness_reward_func": 2.05774999409914,
"rewards/consensus_reward_func": 0.125,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.6988736353814602,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.25,
"rewards/xmlcount_reward_func": 1.1541250050067902,
"step": 70
},
{
"completion_length": 108.125,
"epoch": 6.585365853658536,
"grad_norm": 6.37650728225708,
"kl": 0.5360646117478609,
"learning_rate": 1.0239940674851941e-07,
"loss": 0.0005,
"reward": 3.5589722096920013,
"reward_std": 1.8011039877310395,
"rewards/concensus_correctness_reward_func": 0.6851250007748604,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.5,
"rewards/question_recreation_reward_func": 0.7731909388676286,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.3125,
"rewards/xmlcount_reward_func": 1.2256562411785126,
"step": 72
},
{
"completion_length": 138.21875,
"epoch": 6.780487804878049,
"grad_norm": 4.734973430633545,
"kl": 3.8659601393155754,
"learning_rate": 8.964794509221507e-08,
"loss": 0.0039,
"reward": 2.692448616027832,
"reward_std": 0.35131036676466465,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.8613236248493195,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.359375,
"rewards/xmlcount_reward_func": 1.221750020980835,
"step": 74
},
{
"completion_length": 151.0625,
"epoch": 6.975609756097561,
"grad_norm": 5.129486560821533,
"kl": 1.5537091912701726,
"learning_rate": 7.756905568047392e-08,
"loss": 0.0016,
"reward": 2.6046862453222275,
"reward_std": 0.7406165972352028,
"rewards/concensus_correctness_reward_func": 0.12250000238418579,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.6871863044798374,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.3125,
"rewards/xmlcount_reward_func": 1.1699999868869781,
"step": 76
},
{
"completion_length": 131.7,
"epoch": 7.097560975609756,
"grad_norm": 3.7079029083251953,
"kl": 1.5399453818798066,
"learning_rate": 6.621340157319996e-08,
"loss": 0.001,
"reward": 2.7136191368103026,
"reward_std": 0.9075609683990479,
"rewards/concensus_correctness_reward_func": 0.09620000123977661,
"rewards/consensus_reward_func": 0.1,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.3,
"rewards/question_recreation_reward_func": 0.7861691594123841,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.25,
"rewards/xmlcount_reward_func": 1.18125,
"step": 78
},
{
"completion_length": 181.09375,
"epoch": 7.2926829268292686,
"grad_norm": 6.1535162925720215,
"kl": 0.573518211953342,
"learning_rate": 5.5628612330087724e-08,
"loss": 0.0006,
"reward": 3.8459194749593735,
"reward_std": 1.7336784086655825,
"rewards/concensus_correctness_reward_func": 1.25,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.8142632357776165,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.34375,
"rewards/xmlcount_reward_func": 1.1879062503576279,
"step": 80
},
{
"completion_length": 152.875,
"epoch": 7.487804878048781,
"grad_norm": 6.165317535400391,
"kl": 2.319924622774124,
"learning_rate": 4.5859084235697235e-08,
"loss": 0.0023,
"reward": 2.0710987001657486,
"reward_std": 0.751331745646894,
"rewards/concensus_correctness_reward_func": 0.06012500077486038,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.55609873495996,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.265625,
"rewards/xmlcount_reward_func": 1.1267500072717667,
"step": 82
},
{
"completion_length": 174.9375,
"epoch": 7.682926829268292,
"grad_norm": 5.046776294708252,
"kl": 2.245447961613536,
"learning_rate": 3.6945794086007705e-08,
"loss": 0.0022,
"reward": 2.5292265713214874,
"reward_std": 1.143988709896803,
"rewards/concensus_correctness_reward_func": 0.24050000309944153,
"rewards/consensus_reward_func": 0.25,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.6357265987899154,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.3125,
"rewards/xmlcount_reward_func": 1.0905000045895576,
"step": 84
},
{
"completion_length": 164.0,
"epoch": 7.878048780487805,
"grad_norm": 3.5354819297790527,
"kl": 0.5519012669101357,
"learning_rate": 2.892612731749414e-08,
"loss": 0.0006,
"reward": 3.282635271549225,
"reward_std": 1.7596494741737843,
"rewards/concensus_correctness_reward_func": 0.625,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.375,
"rewards/question_recreation_reward_func": 0.807010255753994,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.328125,
"rewards/xmlcount_reward_func": 1.1475000083446503,
"step": 86
},
{
"completion_length": 108.4,
"epoch": 8.0,
"grad_norm": 2.0044972896575928,
"kl": 0.35757209062576295,
"learning_rate": 2.183372119961499e-08,
"loss": 0.0002,
"reward": 6.687426805496216,
"reward_std": 4.642478483915329,
"rewards/concensus_correctness_reward_func": 3.388399910926819,
"rewards/consensus_reward_func": 0.3,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.8,
"rewards/question_recreation_reward_func": 0.6740266382694244,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.3,
"rewards/xmlcount_reward_func": 1.225,
"step": 88
},
{
"completion_length": 120.59375,
"epoch": 8.195121951219512,
"grad_norm": 3.868396520614624,
"kl": 2.700576603412628,
"learning_rate": 1.5698323748414122e-08,
"loss": 0.0027,
"reward": 3.760663390159607,
"reward_std": 2.0940339118242264,
"rewards/concensus_correctness_reward_func": 0.9279374964535236,
"rewards/consensus_reward_func": 0.25,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.8017571456730366,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.328125,
"rewards/xmlcount_reward_func": 1.202843725681305,
"step": 90
},
{
"completion_length": 203.96875,
"epoch": 8.390243902439025,
"grad_norm": 4.034110069274902,
"kl": 3.7330283680930734,
"learning_rate": 1.054566895300324e-08,
"loss": 0.0037,
"reward": 3.24531951546669,
"reward_std": 1.75903601013124,
"rewards/concensus_correctness_reward_func": 0.8073750026524067,
"rewards/consensus_reward_func": 0.125,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.6574131920933723,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.28125,
"rewards/xmlcount_reward_func": 1.1242812499403954,
"step": 92
},
{
"completion_length": 113.84375,
"epoch": 8.585365853658537,
"grad_norm": 3.601546287536621,
"kl": 2.1267421934753656,
"learning_rate": 6.397368838268496e-09,
"loss": 0.0021,
"reward": 2.7754287719726562,
"reward_std": 0.6228232402354479,
"rewards/concensus_correctness_reward_func": 0.05999999865889549,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.8581162914633751,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.359375,
"rewards/xmlcount_reward_func": 1.1854375004768372,
"step": 94
},
{
"completion_length": 198.03125,
"epoch": 8.78048780487805,
"grad_norm": 3.0875630378723145,
"kl": 0.6049554403871298,
"learning_rate": 3.2708228165273244e-09,
"loss": 0.0006,
"reward": 4.5220272690057755,
"reward_std": 1.8342011701315641,
"rewards/concensus_correctness_reward_func": 1.9952500015497208,
"rewards/consensus_reward_func": 0.125,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.7617147043347359,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.28125,
"rewards/xmlcount_reward_func": 1.1088124960660934,
"step": 96
},
{
"completion_length": 147.875,
"epoch": 8.975609756097562,
"grad_norm": 6.384603977203369,
"kl": 2.2161214109510183,
"learning_rate": 1.1791447083465133e-09,
"loss": 0.0022,
"reward": 2.5484085381031036,
"reward_std": 0.6816056333482265,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.5,
"rewards/question_recreation_reward_func": 0.5684084966778755,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.375,
"rewards/xmlcount_reward_func": 1.1050000041723251,
"step": 98
},
{
"completion_length": 105.9,
"epoch": 9.097560975609756,
"grad_norm": 3.4388906955718994,
"kl": 0.5531518459320068,
"learning_rate": 1.3110773862126667e-10,
"loss": 0.0003,
"reward": 2.561092567443848,
"reward_std": 0.5468424066901207,
"rewards/concensus_correctness_reward_func": 0.09610000252723694,
"rewards/consensus_reward_func": 0.1,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.6962425768375397,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.45,
"rewards/xmlcount_reward_func": 1.21875,
"step": 100
},
{
"epoch": 9.097560975609756,
"step": 100,
"total_flos": 0.0,
"train_loss": 0.0029315173596842214,
"train_runtime": 554.1656,
"train_samples_per_second": 2.887,
"train_steps_per_second": 0.18
}
],
"logging_steps": 2,
"max_steps": 100,
"num_input_tokens_seen": 0,
"num_train_epochs": 10,
"save_steps": 100,
"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": 4,
"trial_name": null,
"trial_params": null
}