| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9992520568436799, | |
| "eval_steps": 300, | |
| "global_step": 334, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0029917726252804786, | |
| "grad_norm": 0.515625, | |
| "learning_rate": 1.4705882352941176e-08, | |
| "logits/chosen": -11.9029541015625, | |
| "logits/rejected": -11.867537498474121, | |
| "logps/chosen": -446.77239990234375, | |
| "logps/rejected": -451.92742919921875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.029917726252804786, | |
| "grad_norm": 2.84375, | |
| "learning_rate": 1.4705882352941175e-07, | |
| "logits/chosen": -11.970257759094238, | |
| "logits/rejected": -11.80525016784668, | |
| "logps/chosen": -548.0371704101562, | |
| "logps/rejected": -528.268798828125, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4184027910232544, | |
| "rewards/chosen": 0.0011305785737931728, | |
| "rewards/margins": 0.0008499751565977931, | |
| "rewards/rejected": 0.0002806035045068711, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05983545250560957, | |
| "grad_norm": 1.484375, | |
| "learning_rate": 2.941176470588235e-07, | |
| "logits/chosen": -12.091507911682129, | |
| "logits/rejected": -11.931904792785645, | |
| "logps/chosen": -605.2603149414062, | |
| "logps/rejected": -594.6846923828125, | |
| "loss": 0.6946, | |
| "rewards/accuracies": 0.4781250059604645, | |
| "rewards/chosen": -0.0016426166985183954, | |
| "rewards/margins": -0.001760849030688405, | |
| "rewards/rejected": 0.00011823121167253703, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08975317875841436, | |
| "grad_norm": 0.8125, | |
| "learning_rate": 4.4117647058823526e-07, | |
| "logits/chosen": -12.251832962036133, | |
| "logits/rejected": -11.93907356262207, | |
| "logps/chosen": -627.1741333007812, | |
| "logps/rejected": -592.0576782226562, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.4765625, | |
| "rewards/chosen": -0.013599636033177376, | |
| "rewards/margins": 0.001697429222986102, | |
| "rewards/rejected": -0.015297065488994122, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.11967090501121914, | |
| "grad_norm": 0.7734375, | |
| "learning_rate": 4.995066821070679e-07, | |
| "logits/chosen": -12.038887023925781, | |
| "logits/rejected": -11.891900062561035, | |
| "logps/chosen": -573.0090942382812, | |
| "logps/rejected": -555.4964599609375, | |
| "loss": 0.6966, | |
| "rewards/accuracies": 0.4390625059604645, | |
| "rewards/chosen": -0.009484687820076942, | |
| "rewards/margins": -0.005814659409224987, | |
| "rewards/rejected": -0.0036700288765132427, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.14958863126402394, | |
| "grad_norm": 1.1796875, | |
| "learning_rate": 4.964990092676262e-07, | |
| "logits/chosen": -12.046560287475586, | |
| "logits/rejected": -11.745036125183105, | |
| "logps/chosen": -487.1591796875, | |
| "logps/rejected": -467.3463439941406, | |
| "loss": 0.6961, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": -0.014206953346729279, | |
| "rewards/margins": -0.0039894962683320045, | |
| "rewards/rejected": -0.010217458009719849, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.17950635751682872, | |
| "grad_norm": 0.65234375, | |
| "learning_rate": 4.907906416994145e-07, | |
| "logits/chosen": -11.951323509216309, | |
| "logits/rejected": -11.81787109375, | |
| "logps/chosen": -524.3377685546875, | |
| "logps/rejected": -518.5364990234375, | |
| "loss": 0.6951, | |
| "rewards/accuracies": 0.4906249940395355, | |
| "rewards/chosen": -0.007045179605484009, | |
| "rewards/margins": -0.0024539525620639324, | |
| "rewards/rejected": -0.004591226577758789, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.2094240837696335, | |
| "grad_norm": 0.7109375, | |
| "learning_rate": 4.824441214720628e-07, | |
| "logits/chosen": -12.155089378356934, | |
| "logits/rejected": -11.935697555541992, | |
| "logps/chosen": -614.5479736328125, | |
| "logps/rejected": -598.7986450195312, | |
| "loss": 0.694, | |
| "rewards/accuracies": 0.4921875, | |
| "rewards/chosen": -0.004119172692298889, | |
| "rewards/margins": -0.0007487249677069485, | |
| "rewards/rejected": -0.003370448248460889, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.2393418100224383, | |
| "grad_norm": 1.125, | |
| "learning_rate": 4.7155089480780365e-07, | |
| "logits/chosen": -12.310976028442383, | |
| "logits/rejected": -11.926130294799805, | |
| "logps/chosen": -590.0772705078125, | |
| "logps/rejected": -552.1112060546875, | |
| "loss": 0.694, | |
| "rewards/accuracies": 0.4703125059604645, | |
| "rewards/chosen": -0.014592505991458893, | |
| "rewards/margins": -0.000543795176781714, | |
| "rewards/rejected": -0.014048713259398937, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.26925953627524307, | |
| "grad_norm": 1.7890625, | |
| "learning_rate": 4.582303101775248e-07, | |
| "logits/chosen": -12.02568244934082, | |
| "logits/rejected": -11.863547325134277, | |
| "logps/chosen": -596.1324462890625, | |
| "logps/rejected": -583.9989624023438, | |
| "loss": 0.6946, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": -0.00669059157371521, | |
| "rewards/margins": -0.0019042941275984049, | |
| "rewards/rejected": -0.004786298610270023, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.2991772625280479, | |
| "grad_norm": 1.1015625, | |
| "learning_rate": 4.426283106939473e-07, | |
| "logits/chosen": -12.22739028930664, | |
| "logits/rejected": -11.815451622009277, | |
| "logps/chosen": -604.3762817382812, | |
| "logps/rejected": -574.3903198242188, | |
| "loss": 0.6949, | |
| "rewards/accuracies": 0.4828124940395355, | |
| "rewards/chosen": 0.0030080166179686785, | |
| "rewards/margins": -0.0021901517175137997, | |
| "rewards/rejected": 0.0051981681026518345, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.32909498878085264, | |
| "grad_norm": 1.453125, | |
| "learning_rate": 4.249158351283413e-07, | |
| "logits/chosen": -12.154890060424805, | |
| "logits/rejected": -11.859006881713867, | |
| "logps/chosen": -524.4495849609375, | |
| "logps/rejected": -509.00946044921875, | |
| "loss": 0.6954, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": 0.002577757928520441, | |
| "rewards/margins": -0.002684831153601408, | |
| "rewards/rejected": 0.005262589547783136, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.35901271503365745, | |
| "grad_norm": 3.0625, | |
| "learning_rate": 4.0528694506957754e-07, | |
| "logits/chosen": -12.192368507385254, | |
| "logits/rejected": -12.028105735778809, | |
| "logps/chosen": -528.8504028320312, | |
| "logps/rejected": -517.0180053710938, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.5015624761581421, | |
| "rewards/chosen": 0.007784114684909582, | |
| "rewards/margins": 0.0019344612956047058, | |
| "rewards/rejected": 0.005849653389304876, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.3889304412864622, | |
| "grad_norm": 1.0546875, | |
| "learning_rate": 3.839566987447491e-07, | |
| "logits/chosen": -12.14574146270752, | |
| "logits/rejected": -12.06743335723877, | |
| "logps/chosen": -646.7266845703125, | |
| "logps/rejected": -643.1847534179688, | |
| "loss": 0.6923, | |
| "rewards/accuracies": 0.4906249940395355, | |
| "rewards/chosen": -0.00342937046661973, | |
| "rewards/margins": 0.0035477778874337673, | |
| "rewards/rejected": -0.006977148354053497, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.418848167539267, | |
| "grad_norm": 0.7421875, | |
| "learning_rate": 3.6115879479623183e-07, | |
| "logits/chosen": -11.896639823913574, | |
| "logits/rejected": -11.759775161743164, | |
| "logps/chosen": -548.5411376953125, | |
| "logps/rejected": -533.6222534179688, | |
| "loss": 0.6943, | |
| "rewards/accuracies": 0.4984374940395355, | |
| "rewards/chosen": -0.006624075584113598, | |
| "rewards/margins": -0.00133815489243716, | |
| "rewards/rejected": -0.005285919643938541, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4487658937920718, | |
| "grad_norm": 0.87109375, | |
| "learning_rate": 3.371430118304538e-07, | |
| "logits/chosen": -11.987521171569824, | |
| "logits/rejected": -11.82154655456543, | |
| "logps/chosen": -599.1915283203125, | |
| "logps/rejected": -586.5440673828125, | |
| "loss": 0.6945, | |
| "rewards/accuracies": 0.4593749940395355, | |
| "rewards/chosen": -0.0005397430504672229, | |
| "rewards/margins": -0.001651047496125102, | |
| "rewards/rejected": 0.0011113042710348964, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.4786836200448766, | |
| "grad_norm": 2.078125, | |
| "learning_rate": 3.121724717912138e-07, | |
| "logits/chosen": -12.010518074035645, | |
| "logits/rejected": -11.801678657531738, | |
| "logps/chosen": -477.989013671875, | |
| "logps/rejected": -461.06695556640625, | |
| "loss": 0.6923, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.002107062842696905, | |
| "rewards/margins": 0.002797911176458001, | |
| "rewards/rejected": -0.00490497425198555, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5086013462976814, | |
| "grad_norm": 1.046875, | |
| "learning_rate": 2.865207571406029e-07, | |
| "logits/chosen": -12.087748527526855, | |
| "logits/rejected": -11.823697090148926, | |
| "logps/chosen": -565.2094116210938, | |
| "logps/rejected": -535.0115966796875, | |
| "loss": 0.6945, | |
| "rewards/accuracies": 0.45625001192092896, | |
| "rewards/chosen": -0.010714416392147541, | |
| "rewards/margins": -0.0014880959643051028, | |
| "rewards/rejected": -0.009226320311427116, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5385190725504861, | |
| "grad_norm": 0.69140625, | |
| "learning_rate": 2.6046891343229986e-07, | |
| "logits/chosen": -12.05897331237793, | |
| "logits/rejected": -11.826770782470703, | |
| "logps/chosen": -639.4141845703125, | |
| "logps/rejected": -625.544189453125, | |
| "loss": 0.694, | |
| "rewards/accuracies": 0.4703125059604645, | |
| "rewards/chosen": -0.008839543908834457, | |
| "rewards/margins": -0.0007707075565122068, | |
| "rewards/rejected": -0.008068837225437164, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.5684367988032909, | |
| "grad_norm": 2.15625, | |
| "learning_rate": 2.3430237011767164e-07, | |
| "logits/chosen": -12.140950202941895, | |
| "logits/rejected": -11.830968856811523, | |
| "logps/chosen": -564.7281494140625, | |
| "logps/rejected": -540.7762451171875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.484375, | |
| "rewards/chosen": -0.00785023346543312, | |
| "rewards/margins": 0.0012117780279368162, | |
| "rewards/rejected": -0.00906201172620058, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.5983545250560958, | |
| "grad_norm": 0.64453125, | |
| "learning_rate": 2.0830781332097445e-07, | |
| "logits/chosen": -12.064250946044922, | |
| "logits/rejected": -11.847702026367188, | |
| "logps/chosen": -601.1182861328125, | |
| "logps/rejected": -584.1577758789062, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.46562498807907104, | |
| "rewards/chosen": -0.0038043882232159376, | |
| "rewards/margins": 0.0006765492144040763, | |
| "rewards/rejected": -0.0044809365645051, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.6282722513089005, | |
| "grad_norm": 0.6796875, | |
| "learning_rate": 1.8277004484618357e-07, | |
| "logits/chosen": -12.159658432006836, | |
| "logits/rejected": -11.93867015838623, | |
| "logps/chosen": -570.8473510742188, | |
| "logps/rejected": -558.4066162109375, | |
| "loss": 0.6956, | |
| "rewards/accuracies": 0.46562498807907104, | |
| "rewards/chosen": -0.01011698879301548, | |
| "rewards/margins": -0.003837681608274579, | |
| "rewards/rejected": -0.006279306020587683, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.6581899775617053, | |
| "grad_norm": 0.90625, | |
| "learning_rate": 1.579688618288305e-07, | |
| "logits/chosen": -12.136404991149902, | |
| "logits/rejected": -11.933469772338867, | |
| "logps/chosen": -653.62841796875, | |
| "logps/rejected": -629.8015747070312, | |
| "loss": 0.6917, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.0027369544841349125, | |
| "rewards/margins": 0.003812599228695035, | |
| "rewards/rejected": -0.0010756452102214098, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.6881077038145101, | |
| "grad_norm": 1.046875, | |
| "learning_rate": 1.341759912200346e-07, | |
| "logits/chosen": -12.061933517456055, | |
| "logits/rejected": -11.955540657043457, | |
| "logps/chosen": -541.6209716796875, | |
| "logps/rejected": -530.1668701171875, | |
| "loss": 0.69, | |
| "rewards/accuracies": 0.526562511920929, | |
| "rewards/chosen": 0.006943908985704184, | |
| "rewards/margins": 0.007641922682523727, | |
| "rewards/rejected": -0.0006980125908739865, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7180254300673149, | |
| "grad_norm": 2.078125, | |
| "learning_rate": 1.11652112689164e-07, | |
| "logits/chosen": -11.996115684509277, | |
| "logits/rejected": -11.906721115112305, | |
| "logps/chosen": -514.420166015625, | |
| "logps/rejected": -510.46307373046875, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.004082207567989826, | |
| "rewards/margins": 0.0010683867149055004, | |
| "rewards/rejected": 0.003013820853084326, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7479431563201197, | |
| "grad_norm": 0.60546875, | |
| "learning_rate": 9.064400256282755e-08, | |
| "logits/chosen": -11.868894577026367, | |
| "logits/rejected": -11.739983558654785, | |
| "logps/chosen": -636.687744140625, | |
| "logps/rejected": -622.6748657226562, | |
| "loss": 0.6937, | |
| "rewards/accuracies": 0.4984374940395355, | |
| "rewards/chosen": 0.007722427137196064, | |
| "rewards/margins": -8.60728334828309e-07, | |
| "rewards/rejected": 0.007723286747932434, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.7778608825729244, | |
| "grad_norm": 0.625, | |
| "learning_rate": 7.138183009179921e-08, | |
| "logits/chosen": -11.95314884185791, | |
| "logits/rejected": -11.800286293029785, | |
| "logps/chosen": -516.5999755859375, | |
| "logps/rejected": -502.53424072265625, | |
| "loss": 0.6941, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 0.003324865596368909, | |
| "rewards/margins": -0.0007609135354869068, | |
| "rewards/rejected": 0.004085779655724764, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8077786088257293, | |
| "grad_norm": 2.03125, | |
| "learning_rate": 5.4076635668540065e-08, | |
| "logits/chosen": -11.870991706848145, | |
| "logits/rejected": -11.772178649902344, | |
| "logps/chosen": -537.8231201171875, | |
| "logps/rejected": -536.79296875, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": -0.0003947736695408821, | |
| "rewards/margins": 0.0007606123690493405, | |
| "rewards/rejected": -0.0011553869117051363, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.837696335078534, | |
| "grad_norm": 0.59765625, | |
| "learning_rate": 3.8918018624496286e-08, | |
| "logits/chosen": -12.036161422729492, | |
| "logits/rejected": -11.860921859741211, | |
| "logps/chosen": -554.2711791992188, | |
| "logps/rejected": -541.3438720703125, | |
| "loss": 0.6953, | |
| "rewards/accuracies": 0.4781250059604645, | |
| "rewards/chosen": 0.00031470804242417216, | |
| "rewards/margins": -0.0031545311212539673, | |
| "rewards/rejected": 0.0034692403860390186, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.8676140613313388, | |
| "grad_norm": 2.578125, | |
| "learning_rate": 2.6072059940146772e-08, | |
| "logits/chosen": -12.018091201782227, | |
| "logits/rejected": -11.894245147705078, | |
| "logps/chosen": -569.8294677734375, | |
| "logps/rejected": -563.2034912109375, | |
| "loss": 0.6936, | |
| "rewards/accuracies": 0.4921875, | |
| "rewards/chosen": 0.004260566085577011, | |
| "rewards/margins": 0.0009015941177494824, | |
| "rewards/rejected": 0.003358972491696477, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.8975317875841436, | |
| "grad_norm": 1.015625, | |
| "learning_rate": 1.5679502627027136e-08, | |
| "logits/chosen": -12.114303588867188, | |
| "logits/rejected": -11.98996639251709, | |
| "logps/chosen": -567.474853515625, | |
| "logps/rejected": -552.4931640625, | |
| "loss": 0.6943, | |
| "rewards/accuracies": 0.46406251192092896, | |
| "rewards/chosen": 0.0020927421282976866, | |
| "rewards/margins": -0.0002106330794049427, | |
| "rewards/rejected": 0.0023033744655549526, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.8975317875841436, | |
| "eval_logits/chosen": -12.036696434020996, | |
| "eval_logits/rejected": -11.858353614807129, | |
| "eval_logps/chosen": -579.9698486328125, | |
| "eval_logps/rejected": -567.8892822265625, | |
| "eval_loss": 0.6939929127693176, | |
| "eval_rewards/accuracies": 0.4747757911682129, | |
| "eval_rewards/chosen": 0.0061609940603375435, | |
| "eval_rewards/margins": -0.00024100362497847527, | |
| "eval_rewards/rejected": 0.006401997059583664, | |
| "eval_runtime": 2311.0773, | |
| "eval_samples_per_second": 2.314, | |
| "eval_steps_per_second": 0.289, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9274495138369484, | |
| "grad_norm": 0.7890625, | |
| "learning_rate": 7.85420971784223e-09, | |
| "logits/chosen": -12.027294158935547, | |
| "logits/rejected": -11.775075912475586, | |
| "logps/chosen": -637.2244262695312, | |
| "logps/rejected": -614.4982299804688, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.49687498807907104, | |
| "rewards/chosen": 0.008133328519761562, | |
| "rewards/margins": 0.0024830640759319067, | |
| "rewards/rejected": 0.005650263279676437, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.9573672400897532, | |
| "grad_norm": 0.99609375, | |
| "learning_rate": 2.6819167592529168e-09, | |
| "logits/chosen": -11.920928955078125, | |
| "logits/rejected": -11.692464828491211, | |
| "logps/chosen": -622.5579223632812, | |
| "logps/rejected": -609.9359130859375, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.48906248807907104, | |
| "rewards/chosen": 0.012452816590666771, | |
| "rewards/margins": 0.0013520055217668414, | |
| "rewards/rejected": 0.011100810021162033, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.9872849663425579, | |
| "grad_norm": 1.734375, | |
| "learning_rate": 2.1929247528540418e-10, | |
| "logits/chosen": -12.089780807495117, | |
| "logits/rejected": -11.94688606262207, | |
| "logps/chosen": -568.0908203125, | |
| "logps/rejected": -560.9391479492188, | |
| "loss": 0.695, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.013248354196548462, | |
| "rewards/margins": -0.0018901375588029623, | |
| "rewards/rejected": 0.015138491988182068, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.9992520568436799, | |
| "step": 334, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6938572500994106, | |
| "train_runtime": 20825.8975, | |
| "train_samples_per_second": 1.027, | |
| "train_steps_per_second": 0.016 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 334, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 300, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |