3a1e3b3b4b
Model: jackf857/llama-3-8b-base-robust-dpo-ultrafeedback-8xh200 Source: Original Platform
796 lines
27 KiB
JSON
796 lines
27 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 0.9989528795811519,
|
|
"eval_steps": 200,
|
|
"global_step": 477,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"epoch": 0.0020942408376963353,
|
|
"grad_norm": 178.72232055664062,
|
|
"learning_rate": 0.0,
|
|
"logits/chosen": -0.615048885345459,
|
|
"logits/rejected": -0.6184952855110168,
|
|
"logps/chosen": -340.7542724609375,
|
|
"logps/rejected": -224.71102905273438,
|
|
"loss": 2.7815,
|
|
"rewards/accuracies": 0.5625,
|
|
"rewards/chosen": -0.0005076047964394093,
|
|
"rewards/margins": 0.0013701794669032097,
|
|
"rewards/rejected": -0.0018777847290039062,
|
|
"step": 1
|
|
},
|
|
{
|
|
"epoch": 0.020942408376963352,
|
|
"grad_norm": 165.93289184570312,
|
|
"learning_rate": 9.375e-08,
|
|
"logits/chosen": -0.6438462734222412,
|
|
"logits/rejected": -0.6377246379852295,
|
|
"logps/chosen": -259.61663818359375,
|
|
"logps/rejected": -258.05615234375,
|
|
"loss": 2.7689,
|
|
"rewards/accuracies": 0.5208333134651184,
|
|
"rewards/chosen": 0.0028021351899951696,
|
|
"rewards/margins": 0.0066328952088952065,
|
|
"rewards/rejected": -0.0038307607173919678,
|
|
"step": 10
|
|
},
|
|
{
|
|
"epoch": 0.041884816753926704,
|
|
"grad_norm": 185.72096252441406,
|
|
"learning_rate": 1.9791666666666664e-07,
|
|
"logits/chosen": -0.6060315370559692,
|
|
"logits/rejected": -0.6294665336608887,
|
|
"logps/chosen": -292.277587890625,
|
|
"logps/rejected": -243.3096923828125,
|
|
"loss": 2.7606,
|
|
"rewards/accuracies": 0.5562499761581421,
|
|
"rewards/chosen": 0.007751139812171459,
|
|
"rewards/margins": 0.006523211486637592,
|
|
"rewards/rejected": 0.001227927627041936,
|
|
"step": 20
|
|
},
|
|
{
|
|
"epoch": 0.06282722513089005,
|
|
"grad_norm": 177.17030334472656,
|
|
"learning_rate": 3.020833333333333e-07,
|
|
"logits/chosen": -0.5830520987510681,
|
|
"logits/rejected": -0.5622227191925049,
|
|
"logps/chosen": -266.82086181640625,
|
|
"logps/rejected": -252.55810546875,
|
|
"loss": 2.7334,
|
|
"rewards/accuracies": 0.5874999761581421,
|
|
"rewards/chosen": 0.02364853397011757,
|
|
"rewards/margins": 0.014692515134811401,
|
|
"rewards/rejected": 0.008956017903983593,
|
|
"step": 30
|
|
},
|
|
{
|
|
"epoch": 0.08376963350785341,
|
|
"grad_norm": 161.6173095703125,
|
|
"learning_rate": 4.0625e-07,
|
|
"logits/chosen": -0.610925555229187,
|
|
"logits/rejected": -0.6422015428543091,
|
|
"logps/chosen": -283.6108093261719,
|
|
"logps/rejected": -263.230712890625,
|
|
"loss": 2.6428,
|
|
"rewards/accuracies": 0.65625,
|
|
"rewards/chosen": 0.07614050805568695,
|
|
"rewards/margins": 0.0455891415476799,
|
|
"rewards/rejected": 0.030551373958587646,
|
|
"step": 40
|
|
},
|
|
{
|
|
"epoch": 0.10471204188481675,
|
|
"grad_norm": 168.7168426513672,
|
|
"learning_rate": 4.999932966293553e-07,
|
|
"logits/chosen": -0.6716187596321106,
|
|
"logits/rejected": -0.6886256337165833,
|
|
"logps/chosen": -264.2910461425781,
|
|
"logps/rejected": -256.81341552734375,
|
|
"loss": 2.5166,
|
|
"rewards/accuracies": 0.6875,
|
|
"rewards/chosen": 0.14609374105930328,
|
|
"rewards/margins": 0.11362485587596893,
|
|
"rewards/rejected": 0.032468877732753754,
|
|
"step": 50
|
|
},
|
|
{
|
|
"epoch": 0.1256544502617801,
|
|
"grad_norm": 161.653076171875,
|
|
"learning_rate": 4.991893270335525e-07,
|
|
"logits/chosen": -0.6373323798179626,
|
|
"logits/rejected": -0.6400020718574524,
|
|
"logps/chosen": -292.5552062988281,
|
|
"logps/rejected": -283.8487243652344,
|
|
"loss": 2.3448,
|
|
"rewards/accuracies": 0.731249988079071,
|
|
"rewards/chosen": 0.2640538215637207,
|
|
"rewards/margins": 0.24639348685741425,
|
|
"rewards/rejected": 0.01766030862927437,
|
|
"step": 60
|
|
},
|
|
{
|
|
"epoch": 0.14659685863874344,
|
|
"grad_norm": 150.28529357910156,
|
|
"learning_rate": 4.970496218214204e-07,
|
|
"logits/chosen": -0.7395513653755188,
|
|
"logits/rejected": -0.73731929063797,
|
|
"logps/chosen": -294.8056945800781,
|
|
"logps/rejected": -290.65618896484375,
|
|
"loss": 2.2036,
|
|
"rewards/accuracies": 0.706250011920929,
|
|
"rewards/chosen": 0.38670551776885986,
|
|
"rewards/margins": 0.33332276344299316,
|
|
"rewards/rejected": 0.05338277295231819,
|
|
"step": 70
|
|
},
|
|
{
|
|
"epoch": 0.16753926701570682,
|
|
"grad_norm": 150.14097595214844,
|
|
"learning_rate": 4.935856505068998e-07,
|
|
"logits/chosen": -0.6947239637374878,
|
|
"logits/rejected": -0.7096772193908691,
|
|
"logps/chosen": -265.2959289550781,
|
|
"logps/rejected": -247.0114288330078,
|
|
"loss": 2.0865,
|
|
"rewards/accuracies": 0.675000011920929,
|
|
"rewards/chosen": 0.44251909852027893,
|
|
"rewards/margins": 0.39249858260154724,
|
|
"rewards/rejected": 0.0500204861164093,
|
|
"step": 80
|
|
},
|
|
{
|
|
"epoch": 0.18848167539267016,
|
|
"grad_norm": 158.30775451660156,
|
|
"learning_rate": 4.8881598109976e-07,
|
|
"logits/chosen": -0.6669536828994751,
|
|
"logits/rejected": -0.699147641658783,
|
|
"logps/chosen": -281.6817932128906,
|
|
"logps/rejected": -246.4118194580078,
|
|
"loss": 2.024,
|
|
"rewards/accuracies": 0.7124999761581421,
|
|
"rewards/chosen": 0.4167642593383789,
|
|
"rewards/margins": 0.561462938785553,
|
|
"rewards/rejected": -0.14469867944717407,
|
|
"step": 90
|
|
},
|
|
{
|
|
"epoch": 0.2094240837696335,
|
|
"grad_norm": 159.54420471191406,
|
|
"learning_rate": 4.827661805750437e-07,
|
|
"logits/chosen": -0.6790419220924377,
|
|
"logits/rejected": -0.6878204345703125,
|
|
"logps/chosen": -258.7181701660156,
|
|
"logps/rejected": -271.3045349121094,
|
|
"loss": 2.0356,
|
|
"rewards/accuracies": 0.6625000238418579,
|
|
"rewards/chosen": 0.389530211687088,
|
|
"rewards/margins": 0.5577359199523926,
|
|
"rewards/rejected": -0.16820567846298218,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.23036649214659685,
|
|
"grad_norm": 148.32960510253906,
|
|
"learning_rate": 4.75468677825789e-07,
|
|
"logits/chosen": -0.649848461151123,
|
|
"logits/rejected": -0.7013173699378967,
|
|
"logps/chosen": -278.6050109863281,
|
|
"logps/rejected": -287.51776123046875,
|
|
"loss": 1.7628,
|
|
"rewards/accuracies": 0.731249988079071,
|
|
"rewards/chosen": 0.2824258506298065,
|
|
"rewards/margins": 0.6535213589668274,
|
|
"rewards/rejected": -0.3710955083370209,
|
|
"step": 110
|
|
},
|
|
{
|
|
"epoch": 0.2513089005235602,
|
|
"grad_norm": 159.53060913085938,
|
|
"learning_rate": 4.669625898336438e-07,
|
|
"logits/chosen": -0.7270513772964478,
|
|
"logits/rejected": -0.7322048544883728,
|
|
"logps/chosen": -268.2157287597656,
|
|
"logps/rejected": -264.9087829589844,
|
|
"loss": 1.9148,
|
|
"rewards/accuracies": 0.6937500238418579,
|
|
"rewards/chosen": 0.10032379627227783,
|
|
"rewards/margins": 0.6243712902069092,
|
|
"rewards/rejected": -0.5240474343299866,
|
|
"step": 120
|
|
},
|
|
{
|
|
"epoch": 0.27225130890052357,
|
|
"grad_norm": 180.24705505371094,
|
|
"learning_rate": 4.5729351198915705e-07,
|
|
"logits/chosen": -0.7340003848075867,
|
|
"logits/rejected": -0.718673825263977,
|
|
"logps/chosen": -288.23260498046875,
|
|
"logps/rejected": -291.12969970703125,
|
|
"loss": 1.8436,
|
|
"rewards/accuracies": 0.6312500238418579,
|
|
"rewards/chosen": 0.08194615691900253,
|
|
"rewards/margins": 0.6275433897972107,
|
|
"rewards/rejected": -0.5455971956253052,
|
|
"step": 130
|
|
},
|
|
{
|
|
"epoch": 0.2931937172774869,
|
|
"grad_norm": 163.4424285888672,
|
|
"learning_rate": 4.4651327368569684e-07,
|
|
"logits/chosen": -0.7740471363067627,
|
|
"logits/rejected": -0.7457703351974487,
|
|
"logps/chosen": -292.3294982910156,
|
|
"logps/rejected": -274.18505859375,
|
|
"loss": 1.8836,
|
|
"rewards/accuracies": 0.6499999761581421,
|
|
"rewards/chosen": 0.17513397336006165,
|
|
"rewards/margins": 0.7150871157646179,
|
|
"rewards/rejected": -0.5399531126022339,
|
|
"step": 140
|
|
},
|
|
{
|
|
"epoch": 0.31413612565445026,
|
|
"grad_norm": 183.37110900878906,
|
|
"learning_rate": 4.346796604970912e-07,
|
|
"logits/chosen": -0.7887325882911682,
|
|
"logits/rejected": -0.7826852202415466,
|
|
"logps/chosen": -306.27752685546875,
|
|
"logps/rejected": -263.0813903808594,
|
|
"loss": 1.6791,
|
|
"rewards/accuracies": 0.7250000238418579,
|
|
"rewards/chosen": 0.32484668493270874,
|
|
"rewards/margins": 1.0725610256195068,
|
|
"rewards/rejected": -0.7477144002914429,
|
|
"step": 150
|
|
},
|
|
{
|
|
"epoch": 0.33507853403141363,
|
|
"grad_norm": 154.94068908691406,
|
|
"learning_rate": 4.218561044282098e-07,
|
|
"logits/chosen": -0.7592242956161499,
|
|
"logits/rejected": -0.7694119215011597,
|
|
"logps/chosen": -269.1500244140625,
|
|
"logps/rejected": -244.9143829345703,
|
|
"loss": 1.4581,
|
|
"rewards/accuracies": 0.793749988079071,
|
|
"rewards/chosen": -0.11840873956680298,
|
|
"rewards/margins": 1.2886579036712646,
|
|
"rewards/rejected": -1.4070665836334229,
|
|
"step": 160
|
|
},
|
|
{
|
|
"epoch": 0.35602094240837695,
|
|
"grad_norm": 188.69107055664062,
|
|
"learning_rate": 4.081113438988443e-07,
|
|
"logits/chosen": -0.7660834193229675,
|
|
"logits/rejected": -0.7688754796981812,
|
|
"logps/chosen": -288.95806884765625,
|
|
"logps/rejected": -266.93170166015625,
|
|
"loss": 1.5308,
|
|
"rewards/accuracies": 0.75,
|
|
"rewards/chosen": 0.014453152194619179,
|
|
"rewards/margins": 1.154816746711731,
|
|
"rewards/rejected": -1.1403634548187256,
|
|
"step": 170
|
|
},
|
|
{
|
|
"epoch": 0.3769633507853403,
|
|
"grad_norm": 145.07247924804688,
|
|
"learning_rate": 3.935190552834828e-07,
|
|
"logits/chosen": -0.7244433760643005,
|
|
"logits/rejected": -0.7473747134208679,
|
|
"logps/chosen": -271.08758544921875,
|
|
"logps/rejected": -252.1222686767578,
|
|
"loss": 1.6745,
|
|
"rewards/accuracies": 0.731249988079071,
|
|
"rewards/chosen": -0.007142066955566406,
|
|
"rewards/margins": 1.086653470993042,
|
|
"rewards/rejected": -1.093795657157898,
|
|
"step": 180
|
|
},
|
|
{
|
|
"epoch": 0.39790575916230364,
|
|
"grad_norm": 160.0677490234375,
|
|
"learning_rate": 3.781574579820464e-07,
|
|
"logits/chosen": -0.7775353193283081,
|
|
"logits/rejected": -0.7835763692855835,
|
|
"logps/chosen": -284.01068115234375,
|
|
"logps/rejected": -260.91009521484375,
|
|
"loss": 1.4346,
|
|
"rewards/accuracies": 0.7562500238418579,
|
|
"rewards/chosen": 0.1680121123790741,
|
|
"rewards/margins": 1.4267839193344116,
|
|
"rewards/rejected": -1.2587717771530151,
|
|
"step": 190
|
|
},
|
|
{
|
|
"epoch": 0.418848167539267,
|
|
"grad_norm": 195.3789520263672,
|
|
"learning_rate": 3.621088951385353e-07,
|
|
"logits/chosen": -0.7822957634925842,
|
|
"logits/rejected": -0.8039867281913757,
|
|
"logps/chosen": -267.6298522949219,
|
|
"logps/rejected": -259.41973876953125,
|
|
"loss": 1.5682,
|
|
"rewards/accuracies": 0.6875,
|
|
"rewards/chosen": 0.25724172592163086,
|
|
"rewards/margins": 1.3147997856140137,
|
|
"rewards/rejected": -1.0575581789016724,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.418848167539267,
|
|
"eval_logits/chosen": -0.800130307674408,
|
|
"eval_logits/rejected": -0.7831319570541382,
|
|
"eval_logps/chosen": -300.2963562011719,
|
|
"eval_logps/rejected": -282.6593322753906,
|
|
"eval_loss": 0.4018401503562927,
|
|
"eval_rewards/accuracies": 0.7540322542190552,
|
|
"eval_rewards/chosen": -0.24218061566352844,
|
|
"eval_rewards/margins": 1.2888014316558838,
|
|
"eval_rewards/rejected": -1.5309820175170898,
|
|
"eval_runtime": 44.8688,
|
|
"eval_samples_per_second": 44.574,
|
|
"eval_steps_per_second": 1.404,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.4397905759162304,
|
|
"grad_norm": 197.96990966796875,
|
|
"learning_rate": 3.454593922550693e-07,
|
|
"logits/chosen": -0.7248128056526184,
|
|
"logits/rejected": -0.7440636157989502,
|
|
"logps/chosen": -288.216552734375,
|
|
"logps/rejected": -284.04571533203125,
|
|
"loss": 1.5993,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": -0.09156915545463562,
|
|
"rewards/margins": 1.5361555814743042,
|
|
"rewards/rejected": -1.6277246475219727,
|
|
"step": 210
|
|
},
|
|
{
|
|
"epoch": 0.4607329842931937,
|
|
"grad_norm": 184.7930145263672,
|
|
"learning_rate": 3.2829819606729477e-07,
|
|
"logits/chosen": -0.7869287133216858,
|
|
"logits/rejected": -0.770258367061615,
|
|
"logps/chosen": -275.52459716796875,
|
|
"logps/rejected": -306.2828063964844,
|
|
"loss": 1.3904,
|
|
"rewards/accuracies": 0.75,
|
|
"rewards/chosen": -0.46098509430885315,
|
|
"rewards/margins": 1.4248541593551636,
|
|
"rewards/rejected": -1.8858392238616943,
|
|
"step": 220
|
|
},
|
|
{
|
|
"epoch": 0.4816753926701571,
|
|
"grad_norm": 209.72683715820312,
|
|
"learning_rate": 3.1071729615293424e-07,
|
|
"logits/chosen": -0.7838747501373291,
|
|
"logits/rejected": -0.7623480558395386,
|
|
"logps/chosen": -318.6150207519531,
|
|
"logps/rejected": -280.46673583984375,
|
|
"loss": 1.2821,
|
|
"rewards/accuracies": 0.793749988079071,
|
|
"rewards/chosen": -0.25955599546432495,
|
|
"rewards/margins": 1.878294587135315,
|
|
"rewards/rejected": -2.137850761413574,
|
|
"step": 230
|
|
},
|
|
{
|
|
"epoch": 0.5026178010471204,
|
|
"grad_norm": 189.98550415039062,
|
|
"learning_rate": 2.9281093183781403e-07,
|
|
"logits/chosen": -0.7183490991592407,
|
|
"logits/rejected": -0.7486238479614258,
|
|
"logps/chosen": -254.42953491210938,
|
|
"logps/rejected": -258.20489501953125,
|
|
"loss": 1.492,
|
|
"rewards/accuracies": 0.7250000238418579,
|
|
"rewards/chosen": -0.5049279928207397,
|
|
"rewards/margins": 1.325303077697754,
|
|
"rewards/rejected": -1.8302310705184937,
|
|
"step": 240
|
|
},
|
|
{
|
|
"epoch": 0.5235602094240838,
|
|
"grad_norm": 206.22911071777344,
|
|
"learning_rate": 2.7467508704251135e-07,
|
|
"logits/chosen": -0.7619749903678894,
|
|
"logits/rejected": -0.7551571726799011,
|
|
"logps/chosen": -296.3054504394531,
|
|
"logps/rejected": -302.1717224121094,
|
|
"loss": 1.5437,
|
|
"rewards/accuracies": 0.7562500238418579,
|
|
"rewards/chosen": -0.023396695032715797,
|
|
"rewards/margins": 1.536604404449463,
|
|
"rewards/rejected": -1.5600011348724365,
|
|
"step": 250
|
|
},
|
|
{
|
|
"epoch": 0.5445026178010471,
|
|
"grad_norm": 225.17259216308594,
|
|
"learning_rate": 2.5640697577740815e-07,
|
|
"logits/chosen": -0.7599934935569763,
|
|
"logits/rejected": -0.760217547416687,
|
|
"logps/chosen": -283.0222473144531,
|
|
"logps/rejected": -279.87725830078125,
|
|
"loss": 1.4859,
|
|
"rewards/accuracies": 0.7749999761581421,
|
|
"rewards/chosen": -0.262519896030426,
|
|
"rewards/margins": 1.3884146213531494,
|
|
"rewards/rejected": -1.6509345769882202,
|
|
"step": 260
|
|
},
|
|
{
|
|
"epoch": 0.5654450261780105,
|
|
"grad_norm": 178.20997619628906,
|
|
"learning_rate": 2.381045210440644e-07,
|
|
"logits/chosen": -0.751593828201294,
|
|
"logits/rejected": -0.740311861038208,
|
|
"logps/chosen": -262.66802978515625,
|
|
"logps/rejected": -278.4461975097656,
|
|
"loss": 1.3796,
|
|
"rewards/accuracies": 0.75,
|
|
"rewards/chosen": -0.4121341109275818,
|
|
"rewards/margins": 1.4869859218597412,
|
|
"rewards/rejected": -1.8991199731826782,
|
|
"step": 270
|
|
},
|
|
{
|
|
"epoch": 0.5863874345549738,
|
|
"grad_norm": 166.0771026611328,
|
|
"learning_rate": 2.1986582993616925e-07,
|
|
"logits/chosen": -0.7442865967750549,
|
|
"logits/rejected": -0.7617285847663879,
|
|
"logps/chosen": -291.0567321777344,
|
|
"logps/rejected": -286.50579833984375,
|
|
"loss": 1.3057,
|
|
"rewards/accuracies": 0.737500011920929,
|
|
"rewards/chosen": -0.2755025029182434,
|
|
"rewards/margins": 1.6462242603302002,
|
|
"rewards/rejected": -1.9217264652252197,
|
|
"step": 280
|
|
},
|
|
{
|
|
"epoch": 0.6073298429319371,
|
|
"grad_norm": 179.26734924316406,
|
|
"learning_rate": 2.0178866775369774e-07,
|
|
"logits/chosen": -0.7708907723426819,
|
|
"logits/rejected": -0.7562252879142761,
|
|
"logps/chosen": -316.54803466796875,
|
|
"logps/rejected": -304.83404541015625,
|
|
"loss": 1.5853,
|
|
"rewards/accuracies": 0.7124999761581421,
|
|
"rewards/chosen": -0.6070529818534851,
|
|
"rewards/margins": 1.3637443780899048,
|
|
"rewards/rejected": -1.9707973003387451,
|
|
"step": 290
|
|
},
|
|
{
|
|
"epoch": 0.6282722513089005,
|
|
"grad_norm": 188.23605346679688,
|
|
"learning_rate": 1.839699339491937e-07,
|
|
"logits/chosen": -0.7966352105140686,
|
|
"logits/rejected": -0.7720645666122437,
|
|
"logps/chosen": -269.08026123046875,
|
|
"logps/rejected": -288.74798583984375,
|
|
"loss": 1.5072,
|
|
"rewards/accuracies": 0.731249988079071,
|
|
"rewards/chosen": 0.051731567829847336,
|
|
"rewards/margins": 1.4731868505477905,
|
|
"rewards/rejected": -1.4214551448822021,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.6492146596858639,
|
|
"grad_norm": 193.12405395507812,
|
|
"learning_rate": 1.6650514271527465e-07,
|
|
"logits/chosen": -0.7465989589691162,
|
|
"logits/rejected": -0.732526421546936,
|
|
"logps/chosen": -315.72998046875,
|
|
"logps/rejected": -311.20135498046875,
|
|
"loss": 1.3191,
|
|
"rewards/accuracies": 0.7749999761581421,
|
|
"rewards/chosen": 0.06771044433116913,
|
|
"rewards/margins": 1.4948314428329468,
|
|
"rewards/rejected": -1.4271211624145508,
|
|
"step": 310
|
|
},
|
|
{
|
|
"epoch": 0.6701570680628273,
|
|
"grad_norm": 208.85194396972656,
|
|
"learning_rate": 1.4948791099758052e-07,
|
|
"logits/chosen": -0.7754351496696472,
|
|
"logits/rejected": -0.7965367436408997,
|
|
"logps/chosen": -287.69244384765625,
|
|
"logps/rejected": -258.49078369140625,
|
|
"loss": 1.3408,
|
|
"rewards/accuracies": 0.7250000238418579,
|
|
"rewards/chosen": -0.2836495339870453,
|
|
"rewards/margins": 1.5221548080444336,
|
|
"rewards/rejected": -1.8058042526245117,
|
|
"step": 320
|
|
},
|
|
{
|
|
"epoch": 0.6910994764397905,
|
|
"grad_norm": 196.4761962890625,
|
|
"learning_rate": 1.3300945667758012e-07,
|
|
"logits/chosen": -0.7553143501281738,
|
|
"logits/rejected": -0.7459635138511658,
|
|
"logps/chosen": -284.33258056640625,
|
|
"logps/rejected": -292.96905517578125,
|
|
"loss": 1.3833,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": -0.4384119510650635,
|
|
"rewards/margins": 1.677755355834961,
|
|
"rewards/rejected": -2.1161673069000244,
|
|
"step": 330
|
|
},
|
|
{
|
|
"epoch": 0.7120418848167539,
|
|
"grad_norm": 196.71824645996094,
|
|
"learning_rate": 1.1715810961514072e-07,
|
|
"logits/chosen": -0.7940613031387329,
|
|
"logits/rejected": -0.7792786955833435,
|
|
"logps/chosen": -260.6127624511719,
|
|
"logps/rejected": -280.0231018066406,
|
|
"loss": 1.3467,
|
|
"rewards/accuracies": 0.7124999761581421,
|
|
"rewards/chosen": -0.8941448330879211,
|
|
"rewards/margins": 1.461021065711975,
|
|
"rewards/rejected": -2.355165958404541,
|
|
"step": 340
|
|
},
|
|
{
|
|
"epoch": 0.7329842931937173,
|
|
"grad_norm": 206.063232421875,
|
|
"learning_rate": 1.0201883817182949e-07,
|
|
"logits/chosen": -0.8100167512893677,
|
|
"logits/rejected": -0.8022691011428833,
|
|
"logps/chosen": -272.3794250488281,
|
|
"logps/rejected": -279.54290771484375,
|
|
"loss": 1.4061,
|
|
"rewards/accuracies": 0.706250011920929,
|
|
"rewards/chosen": -0.7607845664024353,
|
|
"rewards/margins": 1.3687453269958496,
|
|
"rewards/rejected": -2.1295297145843506,
|
|
"step": 350
|
|
},
|
|
{
|
|
"epoch": 0.7539267015706806,
|
|
"grad_norm": 180.1078338623047,
|
|
"learning_rate": 8.76727937529367e-08,
|
|
"logits/chosen": -0.7735427021980286,
|
|
"logits/rejected": -0.7678244709968567,
|
|
"logps/chosen": -281.62591552734375,
|
|
"logps/rejected": -278.86700439453125,
|
|
"loss": 1.5415,
|
|
"rewards/accuracies": 0.706250011920929,
|
|
"rewards/chosen": -0.6388282775878906,
|
|
"rewards/margins": 1.561563491821289,
|
|
"rewards/rejected": -2.2003917694091797,
|
|
"step": 360
|
|
},
|
|
{
|
|
"epoch": 0.774869109947644,
|
|
"grad_norm": 159.87562561035156,
|
|
"learning_rate": 7.419687580962222e-08,
|
|
"logits/chosen": -0.7762534618377686,
|
|
"logits/rejected": -0.7863459587097168,
|
|
"logps/chosen": -306.4834289550781,
|
|
"logps/rejected": -304.45135498046875,
|
|
"loss": 1.1556,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": -0.3031012713909149,
|
|
"rewards/margins": 1.8383687734603882,
|
|
"rewards/rejected": -2.141469955444336,
|
|
"step": 370
|
|
},
|
|
{
|
|
"epoch": 0.7958115183246073,
|
|
"grad_norm": 225.36546325683594,
|
|
"learning_rate": 6.166331963291519e-08,
|
|
"logits/chosen": -0.7447551488876343,
|
|
"logits/rejected": -0.7403131723403931,
|
|
"logps/chosen": -300.09930419921875,
|
|
"logps/rejected": -274.04351806640625,
|
|
"loss": 1.4064,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": -0.3262189030647278,
|
|
"rewards/margins": 1.6647300720214844,
|
|
"rewards/rejected": -1.9909489154815674,
|
|
"step": 380
|
|
},
|
|
{
|
|
"epoch": 0.8167539267015707,
|
|
"grad_norm": 182.28985595703125,
|
|
"learning_rate": 5.013930914912476e-08,
|
|
"logits/chosen": -0.8010516166687012,
|
|
"logits/rejected": -0.7907552719116211,
|
|
"logps/chosen": -266.9259338378906,
|
|
"logps/rejected": -266.32281494140625,
|
|
"loss": 1.257,
|
|
"rewards/accuracies": 0.7437499761581421,
|
|
"rewards/chosen": -0.2791803777217865,
|
|
"rewards/margins": 1.4971604347229004,
|
|
"rewards/rejected": -1.7763407230377197,
|
|
"step": 390
|
|
},
|
|
{
|
|
"epoch": 0.837696335078534,
|
|
"grad_norm": 180.9009246826172,
|
|
"learning_rate": 3.968661679220467e-08,
|
|
"logits/chosen": -0.7620838284492493,
|
|
"logits/rejected": -0.7844998240470886,
|
|
"logps/chosen": -293.094482421875,
|
|
"logps/rejected": -261.71636962890625,
|
|
"loss": 1.3853,
|
|
"rewards/accuracies": 0.731249988079071,
|
|
"rewards/chosen": -0.32791513204574585,
|
|
"rewards/margins": 1.4900782108306885,
|
|
"rewards/rejected": -1.8179931640625,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.837696335078534,
|
|
"eval_logits/chosen": -0.7842384576797485,
|
|
"eval_logits/rejected": -0.7659946084022522,
|
|
"eval_logps/chosen": -300.5436706542969,
|
|
"eval_logps/rejected": -286.3951416015625,
|
|
"eval_loss": 0.3504386246204376,
|
|
"eval_rewards/accuracies": 0.7459677457809448,
|
|
"eval_rewards/chosen": -0.2669134736061096,
|
|
"eval_rewards/margins": 1.6376467943191528,
|
|
"eval_rewards/rejected": -1.9045602083206177,
|
|
"eval_runtime": 44.6161,
|
|
"eval_samples_per_second": 44.827,
|
|
"eval_steps_per_second": 1.412,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.8586387434554974,
|
|
"grad_norm": 201.3262176513672,
|
|
"learning_rate": 3.036127238347164e-08,
|
|
"logits/chosen": -0.7589632272720337,
|
|
"logits/rejected": -0.7365776896476746,
|
|
"logps/chosen": -272.2435302734375,
|
|
"logps/rejected": -303.0001525878906,
|
|
"loss": 1.3037,
|
|
"rewards/accuracies": 0.6937500238418579,
|
|
"rewards/chosen": -0.48334985971450806,
|
|
"rewards/margins": 1.4791133403778076,
|
|
"rewards/rejected": -1.9624630212783813,
|
|
"step": 410
|
|
},
|
|
{
|
|
"epoch": 0.8795811518324608,
|
|
"grad_norm": 210.60845947265625,
|
|
"learning_rate": 2.2213262793589482e-08,
|
|
"logits/chosen": -0.7427655458450317,
|
|
"logits/rejected": -0.733137309551239,
|
|
"logps/chosen": -323.9688415527344,
|
|
"logps/rejected": -285.21478271484375,
|
|
"loss": 1.236,
|
|
"rewards/accuracies": 0.6812499761581421,
|
|
"rewards/chosen": -0.3501654267311096,
|
|
"rewards/margins": 1.6594756841659546,
|
|
"rewards/rejected": -2.009641170501709,
|
|
"step": 420
|
|
},
|
|
{
|
|
"epoch": 0.900523560209424,
|
|
"grad_norm": 193.39053344726562,
|
|
"learning_rate": 1.5286263996730026e-08,
|
|
"logits/chosen": -0.7872425317764282,
|
|
"logits/rejected": -0.7826786041259766,
|
|
"logps/chosen": -306.154052734375,
|
|
"logps/rejected": -300.08172607421875,
|
|
"loss": 1.2994,
|
|
"rewards/accuracies": 0.71875,
|
|
"rewards/chosen": -0.4893164038658142,
|
|
"rewards/margins": 1.622433066368103,
|
|
"rewards/rejected": -2.1117494106292725,
|
|
"step": 430
|
|
},
|
|
{
|
|
"epoch": 0.9214659685863874,
|
|
"grad_norm": 236.22662353515625,
|
|
"learning_rate": 9.617406953185136e-09,
|
|
"logits/chosen": -0.7628771662712097,
|
|
"logits/rejected": -0.7661261558532715,
|
|
"logps/chosen": -288.91278076171875,
|
|
"logps/rejected": -315.32373046875,
|
|
"loss": 1.5465,
|
|
"rewards/accuracies": 0.6625000238418579,
|
|
"rewards/chosen": -0.37518563866615295,
|
|
"rewards/margins": 1.1774741411209106,
|
|
"rewards/rejected": -1.5526599884033203,
|
|
"step": 440
|
|
},
|
|
{
|
|
"epoch": 0.9424083769633508,
|
|
"grad_norm": 187.93885803222656,
|
|
"learning_rate": 5.2370785753763356e-09,
|
|
"logits/chosen": -0.7725846767425537,
|
|
"logits/rejected": -0.7734043002128601,
|
|
"logps/chosen": -284.969970703125,
|
|
"logps/rejected": -260.22076416015625,
|
|
"loss": 1.1488,
|
|
"rewards/accuracies": 0.7562500238418579,
|
|
"rewards/chosen": -0.275113046169281,
|
|
"rewards/margins": 1.5526530742645264,
|
|
"rewards/rejected": -1.8277660608291626,
|
|
"step": 450
|
|
},
|
|
{
|
|
"epoch": 0.9633507853403142,
|
|
"grad_norm": 191.54318237304688,
|
|
"learning_rate": 2.168758844148272e-09,
|
|
"logits/chosen": -0.7661501169204712,
|
|
"logits/rejected": -0.7916361689567566,
|
|
"logps/chosen": -326.5988464355469,
|
|
"logps/rejected": -314.1842041015625,
|
|
"loss": 1.485,
|
|
"rewards/accuracies": 0.71875,
|
|
"rewards/chosen": -0.322204053401947,
|
|
"rewards/margins": 1.6917155981063843,
|
|
"rewards/rejected": -2.0139195919036865,
|
|
"step": 460
|
|
},
|
|
{
|
|
"epoch": 0.9842931937172775,
|
|
"grad_norm": 176.15220642089844,
|
|
"learning_rate": 4.288949484559934e-10,
|
|
"logits/chosen": -0.7752399444580078,
|
|
"logits/rejected": -0.7753234505653381,
|
|
"logps/chosen": -258.8638916015625,
|
|
"logps/rejected": -276.8975524902344,
|
|
"loss": 1.2765,
|
|
"rewards/accuracies": 0.7562500238418579,
|
|
"rewards/chosen": -0.06555650383234024,
|
|
"rewards/margins": 1.781582236289978,
|
|
"rewards/rejected": -1.8471390008926392,
|
|
"step": 470
|
|
},
|
|
{
|
|
"epoch": 0.9989528795811519,
|
|
"step": 477,
|
|
"total_flos": 0.0,
|
|
"train_loss": 1.6700679651096433,
|
|
"train_runtime": 3395.1223,
|
|
"train_samples_per_second": 18.007,
|
|
"train_steps_per_second": 0.14
|
|
}
|
|
],
|
|
"logging_steps": 10,
|
|
"max_steps": 477,
|
|
"num_input_tokens_seen": 0,
|
|
"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": 4,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|