| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 4.0, | |
| "eval_steps": 100, | |
| "global_step": 1540, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.282051282051282e-07, | |
| "logits/chosen": -1.7278180122375488, | |
| "logits/rejected": -1.7377450466156006, | |
| "logps/chosen": -29.553977966308594, | |
| "logps/rejected": -42.813133239746094, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.282051282051282e-06, | |
| "logits/chosen": -1.8666962385177612, | |
| "logits/rejected": -1.8709977865219116, | |
| "logps/chosen": -36.98939514160156, | |
| "logps/rejected": -33.66963195800781, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.5694444179534912, | |
| "rewards/chosen": 0.00017197892884723842, | |
| "rewards/margins": 0.0005675320862792432, | |
| "rewards/rejected": -0.0003955531574320048, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.564102564102564e-06, | |
| "logits/chosen": -1.9978935718536377, | |
| "logits/rejected": -2.000532627105713, | |
| "logps/chosen": -29.66562843322754, | |
| "logps/rejected": -29.045883178710938, | |
| "loss": 0.6934, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": -0.00023434234026353806, | |
| "rewards/margins": -0.0004099405778106302, | |
| "rewards/rejected": 0.0001755982666509226, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 3.846153846153847e-06, | |
| "logits/chosen": -1.9211324453353882, | |
| "logits/rejected": -1.9184545278549194, | |
| "logps/chosen": -31.41294288635254, | |
| "logps/rejected": -33.23053741455078, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 3.152530553052202e-05, | |
| "rewards/margins": 0.000152341352077201, | |
| "rewards/rejected": -0.00012081606837455183, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.999896948438434e-06, | |
| "logits/chosen": -2.017341375350952, | |
| "logits/rejected": -2.0086092948913574, | |
| "logps/chosen": -32.60146713256836, | |
| "logps/rejected": -32.49399185180664, | |
| "loss": 0.6934, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": -0.0002466029836796224, | |
| "rewards/margins": -0.0004333632532507181, | |
| "rewards/rejected": 0.00018676018225960433, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.987541037542187e-06, | |
| "logits/chosen": -1.862633466720581, | |
| "logits/rejected": -1.8518692255020142, | |
| "logps/chosen": -33.55931091308594, | |
| "logps/rejected": -35.44870376586914, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -1.831089502957184e-05, | |
| "rewards/margins": -5.47249146620743e-05, | |
| "rewards/rejected": 3.641402145149186e-05, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.954691471941119e-06, | |
| "logits/chosen": -1.9409154653549194, | |
| "logits/rejected": -1.9428699016571045, | |
| "logps/chosen": -32.53916549682617, | |
| "logps/rejected": -33.24130630493164, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.0006101715262047946, | |
| "rewards/margins": 0.0013094183523207903, | |
| "rewards/rejected": -0.0006992466514930129, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.901618883413549e-06, | |
| "logits/chosen": -2.070591926574707, | |
| "logits/rejected": -2.075544834136963, | |
| "logps/chosen": -34.023067474365234, | |
| "logps/rejected": -36.647151947021484, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.00034581663203425705, | |
| "rewards/margins": 0.0004369783273432404, | |
| "rewards/rejected": -0.0007827949011698365, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.828760511501322e-06, | |
| "logits/chosen": -1.9308092594146729, | |
| "logits/rejected": -1.933943748474121, | |
| "logps/chosen": -34.318023681640625, | |
| "logps/rejected": -34.67802429199219, | |
| "loss": 0.6922, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": 0.0010978971840813756, | |
| "rewards/margins": 0.0019540609791874886, | |
| "rewards/rejected": -0.000856163795106113, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.7367166013034295e-06, | |
| "logits/chosen": -1.9380912780761719, | |
| "logits/rejected": -1.9425855875015259, | |
| "logps/chosen": -32.38385009765625, | |
| "logps/rejected": -32.35346603393555, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.0008357145707122982, | |
| "rewards/margins": 0.0007813175907358527, | |
| "rewards/rejected": 5.439693995867856e-05, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.626245458345211e-06, | |
| "logits/chosen": -2.035137891769409, | |
| "logits/rejected": -2.0331528186798096, | |
| "logps/chosen": -32.112831115722656, | |
| "logps/rejected": -31.29166030883789, | |
| "loss": 0.6924, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.0012800416443496943, | |
| "rewards/margins": 0.0015345367137342691, | |
| "rewards/rejected": -0.0002544948656577617, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_logits/chosen": -2.2300801277160645, | |
| "eval_logits/rejected": -2.225238084793091, | |
| "eval_logps/chosen": -34.04683303833008, | |
| "eval_logps/rejected": -37.53927230834961, | |
| "eval_loss": 0.6930972337722778, | |
| "eval_rewards/accuracies": 0.5186877250671387, | |
| "eval_rewards/chosen": -0.00012280470400583, | |
| "eval_rewards/margins": 0.00010372586984885857, | |
| "eval_rewards/rejected": -0.0002265305956825614, | |
| "eval_runtime": 145.7259, | |
| "eval_samples_per_second": 2.354, | |
| "eval_steps_per_second": 0.295, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.498257201263691e-06, | |
| "logits/chosen": -1.989782691001892, | |
| "logits/rejected": -1.9873950481414795, | |
| "logps/chosen": -33.12385559082031, | |
| "logps/rejected": -34.011810302734375, | |
| "loss": 0.6926, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.0011996207758784294, | |
| "rewards/margins": 0.001024017808958888, | |
| "rewards/rejected": 0.00017560287960805, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.353806263777678e-06, | |
| "logits/chosen": -2.0008151531219482, | |
| "logits/rejected": -1.992500901222229, | |
| "logps/chosen": -32.320838928222656, | |
| "logps/rejected": -32.128170013427734, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.001240686746314168, | |
| "rewards/margins": 0.0009073130786418915, | |
| "rewards/rejected": 0.0003333735803607851, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.1940827077152755e-06, | |
| "logits/chosen": -2.0289230346679688, | |
| "logits/rejected": -2.020946502685547, | |
| "logps/chosen": -30.313907623291016, | |
| "logps/rejected": -32.086116790771484, | |
| "loss": 0.6922, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.0016180993989109993, | |
| "rewards/margins": 0.0019491963321343064, | |
| "rewards/rejected": -0.00033109664218500257, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.0204024186666215e-06, | |
| "logits/chosen": -1.9592479467391968, | |
| "logits/rejected": -1.9694607257843018, | |
| "logps/chosen": -31.223953247070312, | |
| "logps/rejected": -32.547454833984375, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.0019570994190871716, | |
| "rewards/margins": 0.002082846825942397, | |
| "rewards/rejected": -0.00012574761058203876, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 3.834196265035119e-06, | |
| "logits/chosen": -1.8708124160766602, | |
| "logits/rejected": -1.8719879388809204, | |
| "logps/chosen": -33.877174377441406, | |
| "logps/rejected": -34.78774642944336, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.003155181184411049, | |
| "rewards/margins": 0.0034600873477756977, | |
| "rewards/rejected": -0.00030490627977997065, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.636998309800573e-06, | |
| "logits/chosen": -1.9215673208236694, | |
| "logits/rejected": -1.9181665182113647, | |
| "logps/chosen": -36.011531829833984, | |
| "logps/rejected": -32.685707092285156, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.001601455733180046, | |
| "rewards/margins": 0.001231002388522029, | |
| "rewards/rejected": 0.00037045328645035625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.4304331721118078e-06, | |
| "logits/chosen": -2.021604061126709, | |
| "logits/rejected": -2.014291524887085, | |
| "logps/chosen": -33.482086181640625, | |
| "logps/rejected": -31.404422760009766, | |
| "loss": 0.6913, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.0031577465124428272, | |
| "rewards/margins": 0.003683448536321521, | |
| "rewards/rejected": -0.0005257020820863545, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.2162026428305436e-06, | |
| "logits/chosen": -2.027444839477539, | |
| "logits/rejected": -2.032665729522705, | |
| "logps/chosen": -32.183101654052734, | |
| "logps/rejected": -32.39936065673828, | |
| "loss": 0.6918, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.0036168727092444897, | |
| "rewards/margins": 0.0027590212412178516, | |
| "rewards/rejected": 0.0008578516426496208, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 2.996071664294641e-06, | |
| "logits/chosen": -2.027879238128662, | |
| "logits/rejected": -2.025132656097412, | |
| "logps/chosen": -31.258464813232422, | |
| "logps/rejected": -31.348388671875, | |
| "loss": 0.6919, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.0023048892617225647, | |
| "rewards/margins": 0.0026066480204463005, | |
| "rewards/rejected": -0.000301758642308414, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.7718537898066833e-06, | |
| "logits/chosen": -1.8983129262924194, | |
| "logits/rejected": -1.902967095375061, | |
| "logps/chosen": -31.276391983032227, | |
| "logps/rejected": -32.81935119628906, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.0032989257015287876, | |
| "rewards/margins": 0.00358308176510036, | |
| "rewards/rejected": -0.00028415597626008093, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "eval_logits/chosen": -2.2249655723571777, | |
| "eval_logits/rejected": -2.220139503479004, | |
| "eval_logps/chosen": -34.04255294799805, | |
| "eval_logps/rejected": -37.55300521850586, | |
| "eval_loss": 0.6930080056190491, | |
| "eval_rewards/accuracies": 0.5245016813278198, | |
| "eval_rewards/chosen": -8.006239659152925e-05, | |
| "eval_rewards/margins": 0.00028380370349623263, | |
| "eval_rewards/rejected": -0.00036386612919159234, | |
| "eval_runtime": 145.5269, | |
| "eval_samples_per_second": 2.357, | |
| "eval_steps_per_second": 0.295, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2.5453962426402006e-06, | |
| "logits/chosen": -2.010593891143799, | |
| "logits/rejected": -2.021207332611084, | |
| "logps/chosen": -31.7437801361084, | |
| "logps/rejected": -33.93886947631836, | |
| "loss": 0.6916, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.0022752191871404648, | |
| "rewards/margins": 0.003036911366507411, | |
| "rewards/rejected": -0.000761692295782268, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.3185646976551794e-06, | |
| "logits/chosen": -1.9029136896133423, | |
| "logits/rejected": -1.9176632165908813, | |
| "logps/chosen": -29.78145408630371, | |
| "logps/rejected": -31.63638687133789, | |
| "loss": 0.6911, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.003300876123830676, | |
| "rewards/margins": 0.0040829661302268505, | |
| "rewards/rejected": -0.000782089657150209, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.0932279108998323e-06, | |
| "logits/chosen": -1.9591538906097412, | |
| "logits/rejected": -1.9631026983261108, | |
| "logps/chosen": -33.05189895629883, | |
| "logps/rejected": -31.594707489013672, | |
| "loss": 0.6911, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.003655704203993082, | |
| "rewards/margins": 0.004109731875360012, | |
| "rewards/rejected": -0.00045402703108265996, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.8712423238279358e-06, | |
| "logits/chosen": -1.9572566747665405, | |
| "logits/rejected": -1.9354870319366455, | |
| "logps/chosen": -33.83857727050781, | |
| "logps/rejected": -35.12303924560547, | |
| "loss": 0.6907, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.0031517534516751766, | |
| "rewards/margins": 0.004874187987297773, | |
| "rewards/rejected": -0.0017224351176992059, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.6544367689701824e-06, | |
| "logits/chosen": -1.998875379562378, | |
| "logits/rejected": -1.9955555200576782, | |
| "logps/chosen": -32.72559356689453, | |
| "logps/rejected": -36.2435417175293, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.0019947518594563007, | |
| "rewards/margins": 0.0021809376776218414, | |
| "rewards/rejected": -0.0001861859782366082, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.4445974030621963e-06, | |
| "logits/chosen": -1.8661177158355713, | |
| "logits/rejected": -1.8636993169784546, | |
| "logps/chosen": -33.959014892578125, | |
| "logps/rejected": -35.526344299316406, | |
| "loss": 0.6919, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.0022938635665923357, | |
| "rewards/margins": 0.0025111136492341757, | |
| "rewards/rejected": -0.0002172500389860943, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.243452991757889e-06, | |
| "logits/chosen": -1.8511241674423218, | |
| "logits/rejected": -1.8487510681152344, | |
| "logps/chosen": -34.16337585449219, | |
| "logps/rejected": -31.830408096313477, | |
| "loss": 0.6917, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.002342230873182416, | |
| "rewards/margins": 0.0029330006800591946, | |
| "rewards/rejected": -0.0005907699232921004, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0526606671603523e-06, | |
| "logits/chosen": -1.9535648822784424, | |
| "logits/rejected": -1.943101167678833, | |
| "logps/chosen": -35.01304244995117, | |
| "logps/rejected": -31.87521743774414, | |
| "loss": 0.6913, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.0034359837882220745, | |
| "rewards/margins": 0.0037782168947160244, | |
| "rewards/rejected": -0.0003422332229092717, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.737922755071455e-07, | |
| "logits/chosen": -2.048783302307129, | |
| "logits/rejected": -2.0339112281799316, | |
| "logps/chosen": -30.716812133789062, | |
| "logps/rejected": -32.62614059448242, | |
| "loss": 0.6922, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.0020731096155941486, | |
| "rewards/margins": 0.001816184027120471, | |
| "rewards/rejected": 0.00025692558847367764, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 7.08321427484816e-07, | |
| "logits/chosen": -1.9198474884033203, | |
| "logits/rejected": -1.917340636253357, | |
| "logps/chosen": -32.29683303833008, | |
| "logps/rejected": -30.91409683227539, | |
| "loss": 0.6895, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.0063027567230165005, | |
| "rewards/margins": 0.007275627460330725, | |
| "rewards/rejected": -0.0009728703880682588, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "eval_logits/chosen": -2.221111297607422, | |
| "eval_logits/rejected": -2.2162926197052, | |
| "eval_logps/chosen": -34.0648307800293, | |
| "eval_logps/rejected": -37.58684158325195, | |
| "eval_loss": 0.6929495930671692, | |
| "eval_rewards/accuracies": 0.5419435501098633, | |
| "eval_rewards/chosen": -0.0003027978236787021, | |
| "eval_rewards/margins": 0.0003993964346591383, | |
| "eval_rewards/rejected": -0.0007021942874416709, | |
| "eval_runtime": 145.7415, | |
| "eval_samples_per_second": 2.353, | |
| "eval_steps_per_second": 0.295, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 4.84533120650964e-06, | |
| "logits/chosen": -1.9055675268173218, | |
| "logits/rejected": -1.902345895767212, | |
| "logps/chosen": -31.301956176757812, | |
| "logps/rejected": -33.823036193847656, | |
| "loss": 0.6912, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.0030139132868498564, | |
| "rewards/margins": 0.0038227462209761143, | |
| "rewards/rejected": -0.0008088329923339188, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.825108134172131e-06, | |
| "logits/chosen": -1.951906442642212, | |
| "logits/rejected": -1.939772605895996, | |
| "logps/chosen": -34.27196502685547, | |
| "logps/rejected": -33.685001373291016, | |
| "loss": 0.6908, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.003241057973355055, | |
| "rewards/margins": 0.004702677950263023, | |
| "rewards/rejected": -0.0014616195112466812, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 4.80369052967602e-06, | |
| "logits/chosen": -1.9854780435562134, | |
| "logits/rejected": -1.9840580224990845, | |
| "logps/chosen": -33.0145378112793, | |
| "logps/rejected": -32.56486511230469, | |
| "loss": 0.6904, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.004712558351457119, | |
| "rewards/margins": 0.005565387196838856, | |
| "rewards/rejected": -0.0008528297767043114, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 4.781089396387968e-06, | |
| "logits/chosen": -2.070883274078369, | |
| "logits/rejected": -2.055272102355957, | |
| "logps/chosen": -33.69978713989258, | |
| "logps/rejected": -33.0802001953125, | |
| "loss": 0.6909, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.004796043038368225, | |
| "rewards/margins": 0.004417680203914642, | |
| "rewards/rejected": 0.0003783629508689046, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 4.757316345716554e-06, | |
| "logits/chosen": -1.944435477256775, | |
| "logits/rejected": -1.943645715713501, | |
| "logps/chosen": -32.76495361328125, | |
| "logps/rejected": -32.4921760559082, | |
| "loss": 0.6905, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.0051714470610022545, | |
| "rewards/margins": 0.005441715009510517, | |
| "rewards/rejected": -0.0002702682395465672, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 4.73238359114687e-06, | |
| "logits/chosen": -1.8958152532577515, | |
| "logits/rejected": -1.9060084819793701, | |
| "logps/chosen": -31.695724487304688, | |
| "logps/rejected": -35.41404342651367, | |
| "loss": 0.69, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.005089647602289915, | |
| "rewards/margins": 0.0063424864783883095, | |
| "rewards/rejected": -0.0012528380611911416, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 4.706303941965804e-06, | |
| "logits/chosen": -2.029942035675049, | |
| "logits/rejected": -2.0236124992370605, | |
| "logps/chosen": -33.23334884643555, | |
| "logps/rejected": -29.281543731689453, | |
| "loss": 0.6908, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.004186153877526522, | |
| "rewards/margins": 0.004737343639135361, | |
| "rewards/rejected": -0.0005511896451935172, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 4.679090796681225e-06, | |
| "logits/chosen": -1.8858661651611328, | |
| "logits/rejected": -1.8880888223648071, | |
| "logps/chosen": -33.61238098144531, | |
| "logps/rejected": -30.986286163330078, | |
| "loss": 0.6895, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.0062666991725564, | |
| "rewards/margins": 0.007403238210827112, | |
| "rewards/rejected": -0.001136539620347321, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "learning_rate": 4.650758136138454e-06, | |
| "logits/chosen": -1.9126472473144531, | |
| "logits/rejected": -1.9113785028457642, | |
| "logps/chosen": -33.73168182373047, | |
| "logps/rejected": -36.05659484863281, | |
| "loss": 0.6879, | |
| "rewards/accuracies": 0.7458333373069763, | |
| "rewards/chosen": 0.006531029939651489, | |
| "rewards/margins": 0.010536923073232174, | |
| "rewards/rejected": -0.004005893599241972, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 4.621320516337559e-06, | |
| "logits/chosen": -1.8457567691802979, | |
| "logits/rejected": -1.8373829126358032, | |
| "logps/chosen": -30.92877197265625, | |
| "logps/rejected": -36.478904724121094, | |
| "loss": 0.6863, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.008452029898762703, | |
| "rewards/margins": 0.013835062272846699, | |
| "rewards/rejected": -0.005383032839745283, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "eval_logits/chosen": -2.193706512451172, | |
| "eval_logits/rejected": -2.1888742446899414, | |
| "eval_logps/chosen": -34.14311218261719, | |
| "eval_logps/rejected": -37.68904113769531, | |
| "eval_loss": 0.6928316950798035, | |
| "eval_rewards/accuracies": 0.5681062936782837, | |
| "eval_rewards/chosen": -0.0010856210719794035, | |
| "eval_rewards/margins": 0.0006385648157447577, | |
| "eval_rewards/rejected": -0.001724186004139483, | |
| "eval_runtime": 146.0208, | |
| "eval_samples_per_second": 2.349, | |
| "eval_steps_per_second": 0.294, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "learning_rate": 4.590793060955158e-06, | |
| "logits/chosen": -2.0138370990753174, | |
| "logits/rejected": -2.0166878700256348, | |
| "logps/chosen": -32.178985595703125, | |
| "logps/rejected": -35.35575485229492, | |
| "loss": 0.686, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.008749553933739662, | |
| "rewards/margins": 0.014378642663359642, | |
| "rewards/rejected": -0.005629089195281267, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 4.559191453574582e-06, | |
| "logits/chosen": -1.8486782312393188, | |
| "logits/rejected": -1.8472837209701538, | |
| "logps/chosen": -28.309524536132812, | |
| "logps/rejected": -32.836753845214844, | |
| "loss": 0.6868, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.0075803459621965885, | |
| "rewards/margins": 0.012771248817443848, | |
| "rewards/rejected": -0.005190903786569834, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 4.52653192962838e-06, | |
| "logits/chosen": -1.802756905555725, | |
| "logits/rejected": -1.7958400249481201, | |
| "logps/chosen": -33.09931182861328, | |
| "logps/rejected": -34.53899002075195, | |
| "loss": 0.6868, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 0.009747742675244808, | |
| "rewards/margins": 0.012684956192970276, | |
| "rewards/rejected": -0.002937213983386755, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "learning_rate": 4.492831268057307e-06, | |
| "logits/chosen": -1.9703264236450195, | |
| "logits/rejected": -1.9651902914047241, | |
| "logps/chosen": -30.736658096313477, | |
| "logps/rejected": -32.6190071105957, | |
| "loss": 0.6847, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 0.010466375388205051, | |
| "rewards/margins": 0.01698034629225731, | |
| "rewards/rejected": -0.006513969041407108, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 4.458106782690094e-06, | |
| "logits/chosen": -1.8493196964263916, | |
| "logits/rejected": -1.8536157608032227, | |
| "logps/chosen": -33.46088409423828, | |
| "logps/rejected": -33.30448532104492, | |
| "loss": 0.6844, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 0.010549607686698437, | |
| "rewards/margins": 0.01770811341702938, | |
| "rewards/rejected": -0.007158507592976093, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 4.422376313348405e-06, | |
| "logits/chosen": -1.8494908809661865, | |
| "logits/rejected": -1.843927025794983, | |
| "logps/chosen": -34.2591552734375, | |
| "logps/rejected": -35.904815673828125, | |
| "loss": 0.6827, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 0.011603695340454578, | |
| "rewards/margins": 0.02116088569164276, | |
| "rewards/rejected": -0.009557187557220459, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 4.3856582166815696e-06, | |
| "logits/chosen": -1.868131399154663, | |
| "logits/rejected": -1.8679981231689453, | |
| "logps/chosen": -33.08659362792969, | |
| "logps/rejected": -34.75391387939453, | |
| "loss": 0.685, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.01081737782806158, | |
| "rewards/margins": 0.0165016558021307, | |
| "rewards/rejected": -0.005684278905391693, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 4.347971356735789e-06, | |
| "logits/chosen": -1.9114658832550049, | |
| "logits/rejected": -1.8928560018539429, | |
| "logps/chosen": -32.96870040893555, | |
| "logps/rejected": -33.964908599853516, | |
| "loss": 0.6828, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.011920640245079994, | |
| "rewards/margins": 0.020869914442300797, | |
| "rewards/rejected": -0.008949270471930504, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "learning_rate": 4.309335095262675e-06, | |
| "logits/chosen": -1.8733381032943726, | |
| "logits/rejected": -1.8726457357406616, | |
| "logps/chosen": -30.497507095336914, | |
| "logps/rejected": -31.803579330444336, | |
| "loss": 0.6849, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 0.011137938126921654, | |
| "rewards/margins": 0.016713283956050873, | |
| "rewards/rejected": -0.005575346294790506, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 4.269769281772082e-06, | |
| "logits/chosen": -1.8297357559204102, | |
| "logits/rejected": -1.8228442668914795, | |
| "logps/chosen": -31.466567993164062, | |
| "logps/rejected": -35.563499450683594, | |
| "loss": 0.6826, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.012035631574690342, | |
| "rewards/margins": 0.021403178572654724, | |
| "rewards/rejected": -0.009367546997964382, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "eval_logits/chosen": -2.122058868408203, | |
| "eval_logits/rejected": -2.117284059524536, | |
| "eval_logps/chosen": -34.34208297729492, | |
| "eval_logps/rejected": -37.94715118408203, | |
| "eval_loss": 0.6925419569015503, | |
| "eval_rewards/accuracies": 0.5651993155479431, | |
| "eval_rewards/chosen": -0.003075304673984647, | |
| "eval_rewards/margins": 0.001229992602020502, | |
| "eval_rewards/rejected": -0.0043052975088357925, | |
| "eval_runtime": 145.8949, | |
| "eval_samples_per_second": 2.351, | |
| "eval_steps_per_second": 0.295, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 4.22929424333435e-06, | |
| "logits/chosen": -1.8198668956756592, | |
| "logits/rejected": -1.8234672546386719, | |
| "logps/chosen": -28.312463760375977, | |
| "logps/rejected": -33.89719772338867, | |
| "loss": 0.6837, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.008623605594038963, | |
| "rewards/margins": 0.019002709537744522, | |
| "rewards/rejected": -0.010379104875028133, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "learning_rate": 4.1879307741372085e-06, | |
| "logits/chosen": -1.8108766078948975, | |
| "logits/rejected": -1.8216520547866821, | |
| "logps/chosen": -32.165672302246094, | |
| "logps/rejected": -31.733028411865234, | |
| "loss": 0.6824, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.010966637171804905, | |
| "rewards/margins": 0.021780062466859818, | |
| "rewards/rejected": -0.010813427157700062, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "learning_rate": 4.145700124802693e-06, | |
| "logits/chosen": -1.74923837184906, | |
| "logits/rejected": -1.7469356060028076, | |
| "logps/chosen": -30.605663299560547, | |
| "logps/rejected": -31.276514053344727, | |
| "loss": 0.6825, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.010614162310957909, | |
| "rewards/margins": 0.021611668169498444, | |
| "rewards/rejected": -0.010997505858540535, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 4.102623991469562e-06, | |
| "logits/chosen": -1.816229224205017, | |
| "logits/rejected": -1.8094854354858398, | |
| "logps/chosen": -33.24816131591797, | |
| "logps/rejected": -34.189598083496094, | |
| "loss": 0.6825, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.009846633300185204, | |
| "rewards/margins": 0.021567735821008682, | |
| "rewards/rejected": -0.011721103452146053, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "learning_rate": 4.058724504646834e-06, | |
| "logits/chosen": -1.7789214849472046, | |
| "logits/rejected": -1.7853628396987915, | |
| "logps/chosen": -30.978107452392578, | |
| "logps/rejected": -33.693607330322266, | |
| "loss": 0.6847, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.007750834338366985, | |
| "rewards/margins": 0.017099570482969284, | |
| "rewards/rejected": -0.009348735213279724, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "learning_rate": 4.014024217844167e-06, | |
| "logits/chosen": -1.8461157083511353, | |
| "logits/rejected": -1.8232545852661133, | |
| "logps/chosen": -30.5151424407959, | |
| "logps/rejected": -33.84736633300781, | |
| "loss": 0.6844, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": 0.009224263951182365, | |
| "rewards/margins": 0.01773734949529171, | |
| "rewards/rejected": -0.008513087406754494, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "learning_rate": 3.968546095984911e-06, | |
| "logits/chosen": -1.7744262218475342, | |
| "logits/rejected": -1.769487738609314, | |
| "logps/chosen": -31.48854637145996, | |
| "logps/rejected": -33.041587829589844, | |
| "loss": 0.684, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.009281843900680542, | |
| "rewards/margins": 0.018488582223653793, | |
| "rewards/rejected": -0.009206734597682953, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 3.922313503607806e-06, | |
| "logits/chosen": -1.8039462566375732, | |
| "logits/rejected": -1.8057708740234375, | |
| "logps/chosen": -33.585567474365234, | |
| "logps/rejected": -36.357948303222656, | |
| "loss": 0.6816, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.007902255281805992, | |
| "rewards/margins": 0.023472566157579422, | |
| "rewards/rejected": -0.015570309944450855, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "learning_rate": 3.875350192863368e-06, | |
| "logits/chosen": -1.779675841331482, | |
| "logits/rejected": -1.7792049646377563, | |
| "logps/chosen": -29.52834129333496, | |
| "logps/rejected": -32.76404571533203, | |
| "loss": 0.6815, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 0.010637165978550911, | |
| "rewards/margins": 0.023661229759454727, | |
| "rewards/rejected": -0.013024063780903816, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 3.8276802913111436e-06, | |
| "logits/chosen": -1.7808748483657837, | |
| "logits/rejected": -1.778590440750122, | |
| "logps/chosen": -32.0461540222168, | |
| "logps/rejected": -33.55706024169922, | |
| "loss": 0.6823, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.009521286003291607, | |
| "rewards/margins": 0.02209232933819294, | |
| "rewards/rejected": -0.012571041472256184, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "eval_logits/chosen": -2.0287230014801025, | |
| "eval_logits/rejected": -2.024071455001831, | |
| "eval_logps/chosen": -34.73115539550781, | |
| "eval_logps/rejected": -38.49046325683594, | |
| "eval_loss": 0.6917924880981445, | |
| "eval_rewards/accuracies": 0.5830564498901367, | |
| "eval_rewards/chosen": -0.006966045591980219, | |
| "eval_rewards/margins": 0.0027723864186555147, | |
| "eval_rewards/rejected": -0.009738431312143803, | |
| "eval_runtime": 145.7839, | |
| "eval_samples_per_second": 2.353, | |
| "eval_steps_per_second": 0.295, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "learning_rate": 3.7793282895240927e-06, | |
| "logits/chosen": -1.8076483011245728, | |
| "logits/rejected": -1.8139461278915405, | |
| "logps/chosen": -31.64394187927246, | |
| "logps/rejected": -33.57398986816406, | |
| "loss": 0.6828, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.0059343790635466576, | |
| "rewards/margins": 0.02104238048195839, | |
| "rewards/rejected": -0.015108002349734306, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 3.730319028506478e-06, | |
| "logits/chosen": -1.753603219985962, | |
| "logits/rejected": -1.75141179561615, | |
| "logps/chosen": -33.820560455322266, | |
| "logps/rejected": -32.37050247192383, | |
| "loss": 0.6817, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.009353202767670155, | |
| "rewards/margins": 0.0232031662017107, | |
| "rewards/rejected": -0.013849964365363121, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "learning_rate": 3.6806776869317074e-06, | |
| "logits/chosen": -1.6926358938217163, | |
| "logits/rejected": -1.686195731163025, | |
| "logps/chosen": -34.429847717285156, | |
| "logps/rejected": -33.97523880004883, | |
| "loss": 0.6801, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": 0.01034373790025711, | |
| "rewards/margins": 0.026524048298597336, | |
| "rewards/rejected": -0.016180310398340225, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "learning_rate": 3.6304297682067146e-06, | |
| "logits/chosen": -1.7083446979522705, | |
| "logits/rejected": -1.7146565914154053, | |
| "logps/chosen": -33.29853820800781, | |
| "logps/rejected": -34.668426513671875, | |
| "loss": 0.6827, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.006807624362409115, | |
| "rewards/margins": 0.021089300513267517, | |
| "rewards/rejected": -0.014281675219535828, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "learning_rate": 3.579601087369492e-06, | |
| "logits/chosen": -1.7786967754364014, | |
| "logits/rejected": -1.792654037475586, | |
| "logps/chosen": -31.198848724365234, | |
| "logps/rejected": -33.51192855834961, | |
| "loss": 0.6825, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.006536136381328106, | |
| "rewards/margins": 0.021678542718291283, | |
| "rewards/rejected": -0.015142406336963177, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 3.5282177578265295e-06, | |
| "logits/chosen": -1.6418495178222656, | |
| "logits/rejected": -1.6386057138442993, | |
| "logps/chosen": -32.84505081176758, | |
| "logps/rejected": -36.883094787597656, | |
| "loss": 0.6772, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.010779094882309437, | |
| "rewards/margins": 0.03260749578475952, | |
| "rewards/rejected": -0.02182840369641781, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 3.476306177936961e-06, | |
| "logits/chosen": -1.7246978282928467, | |
| "logits/rejected": -1.7246736288070679, | |
| "logps/chosen": -30.864843368530273, | |
| "logps/rejected": -36.09869384765625, | |
| "loss": 0.6799, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.004207477904856205, | |
| "rewards/margins": 0.026930591091513634, | |
| "rewards/rejected": -0.022723112255334854, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "learning_rate": 3.423893017450324e-06, | |
| "logits/chosen": -1.6627376079559326, | |
| "logits/rejected": -1.6593656539916992, | |
| "logps/chosen": -30.347408294677734, | |
| "logps/rejected": -34.78777313232422, | |
| "loss": 0.6809, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.005249389447271824, | |
| "rewards/margins": 0.0249490849673748, | |
| "rewards/rejected": -0.0196996983140707, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "learning_rate": 3.3710052038048794e-06, | |
| "logits/chosen": -1.676200270652771, | |
| "logits/rejected": -1.6763780117034912, | |
| "logps/chosen": -29.362756729125977, | |
| "logps/rejected": -32.716041564941406, | |
| "loss": 0.6792, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": 0.008003375492990017, | |
| "rewards/margins": 0.028458837419748306, | |
| "rewards/rejected": -0.020455462858080864, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 3.3176699082935546e-06, | |
| "logits/chosen": -1.5964815616607666, | |
| "logits/rejected": -1.599886417388916, | |
| "logps/chosen": -33.50843048095703, | |
| "logps/rejected": -33.53223419189453, | |
| "loss": 0.6784, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.009700378403067589, | |
| "rewards/margins": 0.03037584200501442, | |
| "rewards/rejected": -0.02067546173930168, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "eval_logits/chosen": -1.9464259147644043, | |
| "eval_logits/rejected": -1.9419163465499878, | |
| "eval_logps/chosen": -35.35507583618164, | |
| "eval_logps/rejected": -39.15093994140625, | |
| "eval_loss": 0.6916440725326538, | |
| "eval_rewards/accuracies": 0.565614640712738, | |
| "eval_rewards/chosen": -0.013205258175730705, | |
| "eval_rewards/margins": 0.0031379179563373327, | |
| "eval_rewards/rejected": -0.01634317822754383, | |
| "eval_runtime": 145.8665, | |
| "eval_samples_per_second": 2.351, | |
| "eval_steps_per_second": 0.295, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "learning_rate": 3.2639145321045933e-06, | |
| "logits/chosen": -1.667773962020874, | |
| "logits/rejected": -1.6596691608428955, | |
| "logps/chosen": -36.030296325683594, | |
| "logps/rejected": -33.893470764160156, | |
| "loss": 0.681, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.004817272536456585, | |
| "rewards/margins": 0.024693841114640236, | |
| "rewards/rejected": -0.019876569509506226, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "learning_rate": 3.2097666922441107e-06, | |
| "logits/chosen": -1.6702191829681396, | |
| "logits/rejected": -1.6717958450317383, | |
| "logps/chosen": -36.07275390625, | |
| "logps/rejected": -35.63324737548828, | |
| "loss": 0.6788, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 0.0036263135261833668, | |
| "rewards/margins": 0.02919856831431389, | |
| "rewards/rejected": -0.025572258979082108, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "learning_rate": 3.1552542073477554e-06, | |
| "logits/chosen": -1.6882798671722412, | |
| "logits/rejected": -1.6859245300292969, | |
| "logps/chosen": -31.580810546875, | |
| "logps/rejected": -34.97660446166992, | |
| "loss": 0.6787, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.009825185872614384, | |
| "rewards/margins": 0.02962224744260311, | |
| "rewards/rejected": -0.019797060638666153, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 3.100405083388799e-06, | |
| "logits/chosen": -1.6548511981964111, | |
| "logits/rejected": -1.6600011587142944, | |
| "logps/chosen": -30.910289764404297, | |
| "logps/rejected": -35.500179290771484, | |
| "loss": 0.6774, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.008754345588386059, | |
| "rewards/margins": 0.03220795840024948, | |
| "rewards/rejected": -0.023453611880540848, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "learning_rate": 3.0452474992899645e-06, | |
| "logits/chosen": -1.609222412109375, | |
| "logits/rejected": -1.6077518463134766, | |
| "logps/chosen": -32.664878845214844, | |
| "logps/rejected": -37.466697692871094, | |
| "loss": 0.6774, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.004600999411195517, | |
| "rewards/margins": 0.032408393919467926, | |
| "rewards/rejected": -0.02780739590525627, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 2.989809792446417e-06, | |
| "logits/chosen": -1.4765026569366455, | |
| "logits/rejected": -1.472049355506897, | |
| "logps/chosen": -35.379676818847656, | |
| "logps/rejected": -38.33124542236328, | |
| "loss": 0.6749, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.006187664810568094, | |
| "rewards/margins": 0.03741595149040222, | |
| "rewards/rejected": -0.031228289008140564, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 2.9341204441673267e-06, | |
| "logits/chosen": -1.5943939685821533, | |
| "logits/rejected": -1.5989573001861572, | |
| "logps/chosen": -34.98912811279297, | |
| "logps/rejected": -36.04502487182617, | |
| "loss": 0.6801, | |
| "rewards/accuracies": 0.7208333611488342, | |
| "rewards/chosen": 0.00224525248631835, | |
| "rewards/margins": 0.026747092604637146, | |
| "rewards/rejected": -0.024501841515302658, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "learning_rate": 2.878208065043501e-06, | |
| "logits/chosen": -1.542252779006958, | |
| "logits/rejected": -1.5407251119613647, | |
| "logps/chosen": -32.663124084472656, | |
| "logps/rejected": -38.704864501953125, | |
| "loss": 0.6661, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 0.012384667061269283, | |
| "rewards/margins": 0.05538671463727951, | |
| "rewards/rejected": -0.04300205036997795, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "learning_rate": 2.8221013802485974e-06, | |
| "logits/chosen": -1.5762343406677246, | |
| "logits/rejected": -1.5737056732177734, | |
| "logps/chosen": -32.12613296508789, | |
| "logps/rejected": -36.56070327758789, | |
| "loss": 0.6701, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 0.012496042996644974, | |
| "rewards/margins": 0.04723736643791199, | |
| "rewards/rejected": -0.03474132716655731, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 2.76582921478147e-06, | |
| "logits/chosen": -1.4931247234344482, | |
| "logits/rejected": -1.487870454788208, | |
| "logps/chosen": -33.71710968017578, | |
| "logps/rejected": -34.95537567138672, | |
| "loss": 0.6729, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.006659658160060644, | |
| "rewards/margins": 0.04157082363963127, | |
| "rewards/rejected": -0.034911174327135086, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "eval_logits/chosen": -1.8692306280136108, | |
| "eval_logits/rejected": -1.8648308515548706, | |
| "eval_logps/chosen": -35.941200256347656, | |
| "eval_logps/rejected": -39.86221694946289, | |
| "eval_loss": 0.6910557746887207, | |
| "eval_rewards/accuracies": 0.5539867281913757, | |
| "eval_rewards/chosen": -0.019066473469138145, | |
| "eval_rewards/margins": 0.004389475099742413, | |
| "eval_rewards/rejected": -0.023455949500203133, | |
| "eval_runtime": 145.7021, | |
| "eval_samples_per_second": 2.354, | |
| "eval_steps_per_second": 0.295, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "learning_rate": 2.7094204786572254e-06, | |
| "logits/chosen": -1.579530954360962, | |
| "logits/rejected": -1.58658766746521, | |
| "logps/chosen": -31.1917724609375, | |
| "logps/rejected": -38.586029052734375, | |
| "loss": 0.6677, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 0.009364848956465721, | |
| "rewards/margins": 0.05266670510172844, | |
| "rewards/rejected": -0.04330185800790787, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 2.6529041520546072e-06, | |
| "logits/chosen": -1.5474834442138672, | |
| "logits/rejected": -1.5493825674057007, | |
| "logps/chosen": -31.922176361083984, | |
| "logps/rejected": -36.21441650390625, | |
| "loss": 0.6755, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.006489004008471966, | |
| "rewards/margins": 0.03633515536785126, | |
| "rewards/rejected": -0.029846150428056717, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "learning_rate": 2.5963092704273302e-06, | |
| "logits/chosen": -1.4332886934280396, | |
| "logits/rejected": -1.4374314546585083, | |
| "logps/chosen": -32.152000427246094, | |
| "logps/rejected": -39.53594207763672, | |
| "loss": 0.6685, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 0.004394562914967537, | |
| "rewards/margins": 0.05065219849348068, | |
| "rewards/rejected": -0.046257637441158295, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "learning_rate": 2.53966490958702e-06, | |
| "logits/chosen": -1.4998013973236084, | |
| "logits/rejected": -1.496098518371582, | |
| "logps/chosen": -32.54491424560547, | |
| "logps/rejected": -36.80445098876953, | |
| "loss": 0.6725, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.004237356595695019, | |
| "rewards/margins": 0.04233536496758461, | |
| "rewards/rejected": -0.03809800371527672, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "learning_rate": 2.4830001707654135e-06, | |
| "logits/chosen": -1.5474607944488525, | |
| "logits/rejected": -1.549788236618042, | |
| "logps/chosen": -31.952754974365234, | |
| "logps/rejected": -40.619407653808594, | |
| "loss": 0.6651, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 0.00822580885142088, | |
| "rewards/margins": 0.057800523936748505, | |
| "rewards/rejected": -0.04957471415400505, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "learning_rate": 2.4263441656635054e-06, | |
| "logits/chosen": -1.3754708766937256, | |
| "logits/rejected": -1.3716084957122803, | |
| "logps/chosen": -36.57474136352539, | |
| "logps/rejected": -37.647613525390625, | |
| "loss": 0.6716, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.002930630696937442, | |
| "rewards/margins": 0.04462647810578346, | |
| "rewards/rejected": -0.04755710810422897, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "learning_rate": 2.3697260014953107e-06, | |
| "logits/chosen": -1.4011175632476807, | |
| "logits/rejected": -1.4010181427001953, | |
| "logps/chosen": -35.5493049621582, | |
| "logps/rejected": -40.10515213012695, | |
| "loss": 0.6662, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.0029350135009735823, | |
| "rewards/margins": 0.055523864924907684, | |
| "rewards/rejected": -0.052588850259780884, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "learning_rate": 2.3131747660339396e-06, | |
| "logits/chosen": -1.4290226697921753, | |
| "logits/rejected": -1.4171994924545288, | |
| "logps/chosen": -33.74538040161133, | |
| "logps/rejected": -38.27408981323242, | |
| "loss": 0.6654, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.0036120389122515917, | |
| "rewards/margins": 0.057132624089717865, | |
| "rewards/rejected": -0.05352058261632919, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.31, | |
| "learning_rate": 2.256719512667651e-06, | |
| "logits/chosen": -1.5206860303878784, | |
| "logits/rejected": -1.5256131887435913, | |
| "logps/chosen": -33.839393615722656, | |
| "logps/rejected": -38.63503646850586, | |
| "loss": 0.6659, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.005094523541629314, | |
| "rewards/margins": 0.056891001760959625, | |
| "rewards/rejected": -0.061985522508621216, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 2.2003892454735786e-06, | |
| "logits/chosen": -1.4437249898910522, | |
| "logits/rejected": -1.4366403818130493, | |
| "logps/chosen": -34.63188552856445, | |
| "logps/rejected": -38.323524475097656, | |
| "loss": 0.6635, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 0.0022219305392354727, | |
| "rewards/margins": 0.06143581122159958, | |
| "rewards/rejected": -0.05921388417482376, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "eval_logits/chosen": -1.7487633228302002, | |
| "eval_logits/rejected": -1.7445435523986816, | |
| "eval_logps/chosen": -37.12141799926758, | |
| "eval_logps/rejected": -41.19174575805664, | |
| "eval_loss": 0.6904172301292419, | |
| "eval_rewards/accuracies": 0.5365448594093323, | |
| "eval_rewards/chosen": -0.03086867742240429, | |
| "eval_rewards/margins": 0.00588257284834981, | |
| "eval_rewards/rejected": -0.03675125539302826, | |
| "eval_runtime": 145.8716, | |
| "eval_samples_per_second": 2.351, | |
| "eval_steps_per_second": 0.295, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "learning_rate": 2.1442129043167877e-06, | |
| "logits/chosen": -1.433040738105774, | |
| "logits/rejected": -1.4330635070800781, | |
| "logps/chosen": -31.19219970703125, | |
| "logps/rejected": -41.520694732666016, | |
| "loss": 0.6604, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.00025037964223884046, | |
| "rewards/margins": 0.06805966049432755, | |
| "rewards/rejected": -0.06831003725528717, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "learning_rate": 2.088219349982323e-06, | |
| "logits/chosen": -1.3752285242080688, | |
| "logits/rejected": -1.3669588565826416, | |
| "logps/chosen": -32.70459747314453, | |
| "logps/rejected": -40.24443817138672, | |
| "loss": 0.6653, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.008656435646116734, | |
| "rewards/margins": 0.058074213564395905, | |
| "rewards/rejected": -0.06673064827919006, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "learning_rate": 2.0324373493478803e-06, | |
| "logits/chosen": -1.5194597244262695, | |
| "logits/rejected": -1.517913818359375, | |
| "logps/chosen": -30.247411727905273, | |
| "logps/rejected": -39.21205520629883, | |
| "loss": 0.664, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.002503753872588277, | |
| "rewards/margins": 0.06096818298101425, | |
| "rewards/rejected": -0.06347193568944931, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "learning_rate": 1.976895560604729e-06, | |
| "logits/chosen": -1.4003164768218994, | |
| "logits/rejected": -1.4109015464782715, | |
| "logps/chosen": -35.36278533935547, | |
| "logps/rejected": -40.1971549987793, | |
| "loss": 0.6632, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.009377234615385532, | |
| "rewards/margins": 0.06269785016775131, | |
| "rewards/rejected": -0.07207508385181427, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 1.921622518534466e-06, | |
| "logits/chosen": -1.4363105297088623, | |
| "logits/rejected": -1.4391801357269287, | |
| "logps/chosen": -31.704153060913086, | |
| "logps/rejected": -38.39413070678711, | |
| "loss": 0.6663, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.010564305819571018, | |
| "rewards/margins": 0.056016188114881516, | |
| "rewards/rejected": -0.06658048927783966, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.49, | |
| "learning_rate": 1.8666466198491794e-06, | |
| "logits/chosen": -1.411747932434082, | |
| "logits/rejected": -1.4068377017974854, | |
| "logps/chosen": -34.955177307128906, | |
| "logps/rejected": -41.352115631103516, | |
| "loss": 0.6619, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.008799608796834946, | |
| "rewards/margins": 0.06565765291452408, | |
| "rewards/rejected": -0.07445726543664932, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 2.52, | |
| "learning_rate": 1.8119961086025376e-06, | |
| "logits/chosen": -1.3314030170440674, | |
| "logits/rejected": -1.3333518505096436, | |
| "logps/chosen": -33.6180534362793, | |
| "logps/rejected": -42.525047302246094, | |
| "loss": 0.6617, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.008053036406636238, | |
| "rewards/margins": 0.06538151204586029, | |
| "rewards/rejected": -0.07343455404043198, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 1.7576990616793139e-06, | |
| "logits/chosen": -1.3536522388458252, | |
| "logits/rejected": -1.3478691577911377, | |
| "logps/chosen": -37.943336486816406, | |
| "logps/rejected": -44.42793655395508, | |
| "loss": 0.6663, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.025382736697793007, | |
| "rewards/margins": 0.0562770739197731, | |
| "rewards/rejected": -0.08165980130434036, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 1.7037833743707892e-06, | |
| "logits/chosen": -1.3406635522842407, | |
| "logits/rejected": -1.3342511653900146, | |
| "logps/chosen": -32.205875396728516, | |
| "logps/rejected": -44.02067565917969, | |
| "loss": 0.6603, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.012029221281409264, | |
| "rewards/margins": 0.0682787075638771, | |
| "rewards/rejected": -0.08030791580677032, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 1.6502767460434588e-06, | |
| "logits/chosen": -1.3228440284729004, | |
| "logits/rejected": -1.3131605386734009, | |
| "logps/chosen": -33.538795471191406, | |
| "logps/rejected": -35.858123779296875, | |
| "loss": 0.6719, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.01790105737745762, | |
| "rewards/margins": 0.04468285292387009, | |
| "rewards/rejected": -0.06258390843868256, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "eval_logits/chosen": -1.660041332244873, | |
| "eval_logits/rejected": -1.6560044288635254, | |
| "eval_logps/chosen": -38.48649597167969, | |
| "eval_logps/rejected": -42.736080169677734, | |
| "eval_loss": 0.6896607875823975, | |
| "eval_rewards/accuracies": 0.5485880374908447, | |
| "eval_rewards/chosen": -0.04451945051550865, | |
| "eval_rewards/margins": 0.007675125263631344, | |
| "eval_rewards/rejected": -0.05219458416104317, | |
| "eval_runtime": 145.8291, | |
| "eval_samples_per_second": 2.352, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "learning_rate": 1.5972066659083796e-06, | |
| "logits/chosen": -1.3907979726791382, | |
| "logits/rejected": -1.3904699087142944, | |
| "logps/chosen": -33.401824951171875, | |
| "logps/rejected": -37.548805236816406, | |
| "loss": 0.6661, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.015776285901665688, | |
| "rewards/margins": 0.05712694674730301, | |
| "rewards/rejected": -0.07290322333574295, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "learning_rate": 1.5446003988985041e-06, | |
| "logits/chosen": -1.4413875341415405, | |
| "logits/rejected": -1.4416849613189697, | |
| "logps/chosen": -33.36183547973633, | |
| "logps/rejected": -38.4256477355957, | |
| "loss": 0.665, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.014182562939822674, | |
| "rewards/margins": 0.05841582268476486, | |
| "rewards/rejected": -0.07259838283061981, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 2.68, | |
| "learning_rate": 1.4924849716612211e-06, | |
| "logits/chosen": -1.3972914218902588, | |
| "logits/rejected": -1.4021806716918945, | |
| "logps/chosen": -34.44452667236328, | |
| "logps/rejected": -34.45269775390625, | |
| "loss": 0.6727, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.021644581109285355, | |
| "rewards/margins": 0.043045226484537125, | |
| "rewards/rejected": -0.06468981504440308, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 2.7, | |
| "learning_rate": 1.440887158673332e-06, | |
| "logits/chosen": -1.3861340284347534, | |
| "logits/rejected": -1.377633810043335, | |
| "logps/chosen": -32.912872314453125, | |
| "logps/rejected": -42.408958435058594, | |
| "loss": 0.6588, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.01734977774322033, | |
| "rewards/margins": 0.07173751294612885, | |
| "rewards/rejected": -0.08908729255199432, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "learning_rate": 1.3898334684855647e-06, | |
| "logits/chosen": -1.3528214693069458, | |
| "logits/rejected": -1.3631798028945923, | |
| "logps/chosen": -35.93256378173828, | |
| "logps/rejected": -40.24216079711914, | |
| "loss": 0.666, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.023794613778591156, | |
| "rewards/margins": 0.05617685988545418, | |
| "rewards/rejected": -0.07997147738933563, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 1.3393501301037245e-06, | |
| "logits/chosen": -1.4136943817138672, | |
| "logits/rejected": -1.405368447303772, | |
| "logps/chosen": -35.31805419921875, | |
| "logps/rejected": -45.923988342285156, | |
| "loss": 0.6566, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.018172740936279297, | |
| "rewards/margins": 0.07775326073169708, | |
| "rewards/rejected": -0.09592600166797638, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 1.2894630795134454e-06, | |
| "logits/chosen": -1.320759892463684, | |
| "logits/rejected": -1.3234620094299316, | |
| "logps/chosen": -37.43547439575195, | |
| "logps/rejected": -40.544471740722656, | |
| "loss": 0.6625, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.014461873099207878, | |
| "rewards/margins": 0.06400416046380997, | |
| "rewards/rejected": -0.078466035425663, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "learning_rate": 1.2401979463554984e-06, | |
| "logits/chosen": -1.4195467233657837, | |
| "logits/rejected": -1.4183709621429443, | |
| "logps/chosen": -34.790035247802734, | |
| "logps/rejected": -43.464012145996094, | |
| "loss": 0.6569, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.017323976382613182, | |
| "rewards/margins": 0.07615931332111359, | |
| "rewards/rejected": -0.09348328411579132, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "learning_rate": 1.1915800407584705e-06, | |
| "logits/chosen": -1.4136641025543213, | |
| "logits/rejected": -1.4168442487716675, | |
| "logps/chosen": -32.618568420410156, | |
| "logps/rejected": -41.96255111694336, | |
| "loss": 0.6611, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.017252514138817787, | |
| "rewards/margins": 0.06689468771219254, | |
| "rewards/rejected": -0.08414719998836517, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "learning_rate": 1.1436343403356019e-06, | |
| "logits/chosen": -1.3993356227874756, | |
| "logits/rejected": -1.4037957191467285, | |
| "logps/chosen": -35.935604095458984, | |
| "logps/rejected": -36.58147430419922, | |
| "loss": 0.6749, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.023666027933359146, | |
| "rewards/margins": 0.038131967186927795, | |
| "rewards/rejected": -0.06179799512028694, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "eval_logits/chosen": -1.6285730600357056, | |
| "eval_logits/rejected": -1.6245777606964111, | |
| "eval_logps/chosen": -38.960731506347656, | |
| "eval_logps/rejected": -43.267208099365234, | |
| "eval_loss": 0.6894330978393555, | |
| "eval_rewards/accuracies": 0.5365448594093323, | |
| "eval_rewards/chosen": -0.04926181212067604, | |
| "eval_rewards/margins": 0.008244064636528492, | |
| "eval_rewards/rejected": -0.057505879551172256, | |
| "eval_runtime": 145.7626, | |
| "eval_samples_per_second": 2.353, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "learning_rate": 1.0963854773524548e-06, | |
| "logits/chosen": -1.3872336149215698, | |
| "logits/rejected": -1.3884273767471313, | |
| "logps/chosen": -34.366111755371094, | |
| "logps/rejected": -38.44298553466797, | |
| "loss": 0.6646, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.014259574934840202, | |
| "rewards/margins": 0.06001114100217819, | |
| "rewards/rejected": -0.07427072525024414, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "learning_rate": 1.049857726072005e-06, | |
| "logits/chosen": -1.2578824758529663, | |
| "logits/rejected": -1.2607439756393433, | |
| "logps/chosen": -36.28580856323242, | |
| "logps/rejected": -40.985992431640625, | |
| "loss": 0.6636, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.017208317294716835, | |
| "rewards/margins": 0.06223895400762558, | |
| "rewards/rejected": -0.07944727689027786, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "learning_rate": 1.0040749902836508e-06, | |
| "logits/chosen": -1.2802751064300537, | |
| "logits/rejected": -1.2788641452789307, | |
| "logps/chosen": -33.49232864379883, | |
| "logps/rejected": -38.528602600097656, | |
| "loss": 0.67, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.024806050583720207, | |
| "rewards/margins": 0.04882260411977768, | |
| "rewards/rejected": -0.07362865656614304, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 9.59060791022566e-07, | |
| "logits/chosen": -1.3983075618743896, | |
| "logits/rejected": -1.3943830728530884, | |
| "logps/chosen": -34.20863723754883, | |
| "logps/rejected": -41.15024185180664, | |
| "loss": 0.6594, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.011688429862260818, | |
| "rewards/margins": 0.07027387619018555, | |
| "rewards/rejected": -0.08196230232715607, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "learning_rate": 9.148382544856885e-07, | |
| "logits/chosen": -1.3060632944107056, | |
| "logits/rejected": -1.2978880405426025, | |
| "logps/chosen": -35.88400650024414, | |
| "logps/rejected": -39.35108184814453, | |
| "loss": 0.6663, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.024322666227817535, | |
| "rewards/margins": 0.05626412108540535, | |
| "rewards/rejected": -0.08058679848909378, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 3.01, | |
| "learning_rate": 8.714301001505568e-07, | |
| "logits/chosen": -1.332467794418335, | |
| "logits/rejected": -1.3338046073913574, | |
| "logps/chosen": -35.79069900512695, | |
| "logps/rejected": -38.749183654785156, | |
| "loss": 0.6671, | |
| "rewards/accuracies": 0.7541667222976685, | |
| "rewards/chosen": -0.021089451387524605, | |
| "rewards/margins": 0.05438787862658501, | |
| "rewards/rejected": -0.07547733187675476, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 3.04, | |
| "learning_rate": 8.288586291031025e-07, | |
| "logits/chosen": -1.4120080471038818, | |
| "logits/rejected": -1.4067761898040771, | |
| "logps/chosen": -35.47734069824219, | |
| "logps/rejected": -40.39026641845703, | |
| "loss": 0.667, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.019513698294758797, | |
| "rewards/margins": 0.05497678369283676, | |
| "rewards/rejected": -0.074490487575531, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "learning_rate": 7.871457125803897e-07, | |
| "logits/chosen": -1.3105064630508423, | |
| "logits/rejected": -1.3196675777435303, | |
| "logps/chosen": -35.979042053222656, | |
| "logps/rejected": -40.29875946044922, | |
| "loss": 0.6678, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.026460427790880203, | |
| "rewards/margins": 0.05291280895471573, | |
| "rewards/rejected": -0.07937324047088623, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "learning_rate": 7.463127807341966e-07, | |
| "logits/chosen": -1.3309608697891235, | |
| "logits/rejected": -1.325539231300354, | |
| "logps/chosen": -33.38233947753906, | |
| "logps/rejected": -41.38855743408203, | |
| "loss": 0.66, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.012087189592421055, | |
| "rewards/margins": 0.06896394491195679, | |
| "rewards/rejected": -0.08105112612247467, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "learning_rate": 7.063808116212021e-07, | |
| "logits/chosen": -1.2911185026168823, | |
| "logits/rejected": -1.2924482822418213, | |
| "logps/chosen": -35.253963470458984, | |
| "logps/rejected": -42.57808303833008, | |
| "loss": 0.6551, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.017635192722082138, | |
| "rewards/margins": 0.08054514229297638, | |
| "rewards/rejected": -0.09818033874034882, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "eval_logits/chosen": -1.6238328218460083, | |
| "eval_logits/rejected": -1.6198344230651855, | |
| "eval_logps/chosen": -39.01145935058594, | |
| "eval_logps/rejected": -43.33791732788086, | |
| "eval_loss": 0.6893402338027954, | |
| "eval_rewards/accuracies": 0.5365448594093323, | |
| "eval_rewards/chosen": -0.049769096076488495, | |
| "eval_rewards/margins": 0.008443917147815228, | |
| "eval_rewards/rejected": -0.058213010430336, | |
| "eval_runtime": 145.8737, | |
| "eval_samples_per_second": 2.351, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.14, | |
| "learning_rate": 6.673703204254348e-07, | |
| "logits/chosen": -1.2456345558166504, | |
| "logits/rejected": -1.2452775239944458, | |
| "logps/chosen": -37.221336364746094, | |
| "logps/rejected": -42.06071853637695, | |
| "loss": 0.6544, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.01168847642838955, | |
| "rewards/margins": 0.08182507008314133, | |
| "rewards/rejected": -0.09351354837417603, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "learning_rate": 6.293013489185315e-07, | |
| "logits/chosen": -1.3760040998458862, | |
| "logits/rejected": -1.3689346313476562, | |
| "logps/chosen": -33.46622848510742, | |
| "logps/rejected": -42.4327392578125, | |
| "loss": 0.6562, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.017625439912080765, | |
| "rewards/margins": 0.07791656255722046, | |
| "rewards/rejected": -0.09554200619459152, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "learning_rate": 5.921934551632086e-07, | |
| "logits/chosen": -1.2549601793289185, | |
| "logits/rejected": -1.2447240352630615, | |
| "logps/chosen": -35.567508697509766, | |
| "logps/rejected": -42.02611541748047, | |
| "loss": 0.6551, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.011818263679742813, | |
| "rewards/margins": 0.07950461655855179, | |
| "rewards/rejected": -0.0913228839635849, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 3.22, | |
| "learning_rate": 5.560657034652405e-07, | |
| "logits/chosen": -1.3439350128173828, | |
| "logits/rejected": -1.338648796081543, | |
| "logps/chosen": -33.191280364990234, | |
| "logps/rejected": -36.837867736816406, | |
| "loss": 0.6678, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.024910490959882736, | |
| "rewards/margins": 0.05479263514280319, | |
| "rewards/rejected": -0.07970312982797623, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "learning_rate": 5.2093665457911e-07, | |
| "logits/chosen": -1.3508336544036865, | |
| "logits/rejected": -1.3587679862976074, | |
| "logps/chosen": -37.35521697998047, | |
| "logps/rejected": -39.65736770629883, | |
| "loss": 0.6635, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.019598282873630524, | |
| "rewards/margins": 0.06179226562380791, | |
| "rewards/rejected": -0.08139055222272873, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 3.27, | |
| "learning_rate": 4.868243561723535e-07, | |
| "logits/chosen": -1.3516252040863037, | |
| "logits/rejected": -1.3514872789382935, | |
| "logps/chosen": -35.94284439086914, | |
| "logps/rejected": -42.4056510925293, | |
| "loss": 0.6611, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.029142867773771286, | |
| "rewards/margins": 0.0667649507522583, | |
| "rewards/rejected": -0.09590782225131989, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 3.3, | |
| "learning_rate": 4.537463335535161e-07, | |
| "logits/chosen": -1.265148401260376, | |
| "logits/rejected": -1.2630140781402588, | |
| "logps/chosen": -34.26659393310547, | |
| "logps/rejected": -43.09412384033203, | |
| "loss": 0.6533, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.01175761315971613, | |
| "rewards/margins": 0.08397753536701202, | |
| "rewards/rejected": -0.09573514014482498, | |
| "step": 1270 | |
| }, | |
| { | |
| "epoch": 3.32, | |
| "learning_rate": 4.217195806684629e-07, | |
| "logits/chosen": -1.1799885034561157, | |
| "logits/rejected": -1.1763312816619873, | |
| "logps/chosen": -36.85099411010742, | |
| "logps/rejected": -38.88633346557617, | |
| "loss": 0.6628, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.016135532408952713, | |
| "rewards/margins": 0.06346292048692703, | |
| "rewards/rejected": -0.07959844172000885, | |
| "step": 1280 | |
| }, | |
| { | |
| "epoch": 3.35, | |
| "learning_rate": 3.907605513696808e-07, | |
| "logits/chosen": -1.353476881980896, | |
| "logits/rejected": -1.339864730834961, | |
| "logps/chosen": -36.990108489990234, | |
| "logps/rejected": -45.27104568481445, | |
| "loss": 0.6558, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.0264253169298172, | |
| "rewards/margins": 0.0786319449543953, | |
| "rewards/rejected": -0.1050572618842125, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 3.38, | |
| "learning_rate": 3.6088515096305675e-07, | |
| "logits/chosen": -1.304051160812378, | |
| "logits/rejected": -1.3077205419540405, | |
| "logps/chosen": -35.44999313354492, | |
| "logps/rejected": -47.28888702392578, | |
| "loss": 0.6489, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.019323688000440598, | |
| "rewards/margins": 0.09253297001123428, | |
| "rewards/rejected": -0.11185667663812637, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.38, | |
| "eval_logits/chosen": -1.6211615800857544, | |
| "eval_logits/rejected": -1.6171820163726807, | |
| "eval_logps/chosen": -39.072174072265625, | |
| "eval_logps/rejected": -43.41142654418945, | |
| "eval_loss": 0.6892833113670349, | |
| "eval_rewards/accuracies": 0.5394518375396729, | |
| "eval_rewards/chosen": -0.050376225262880325, | |
| "eval_rewards/margins": 0.008571851067245007, | |
| "eval_rewards/rejected": -0.058948077261447906, | |
| "eval_runtime": 145.8858, | |
| "eval_samples_per_second": 2.351, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "learning_rate": 3.321087280364757e-07, | |
| "logits/chosen": -1.2905550003051758, | |
| "logits/rejected": -1.2912893295288086, | |
| "logps/chosen": -38.21614074707031, | |
| "logps/rejected": -47.26013946533203, | |
| "loss": 0.6547, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.01972118392586708, | |
| "rewards/margins": 0.08203905820846558, | |
| "rewards/rejected": -0.10176024585962296, | |
| "step": 1310 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "learning_rate": 3.044460665744284e-07, | |
| "logits/chosen": -1.3596677780151367, | |
| "logits/rejected": -1.3584003448486328, | |
| "logps/chosen": -33.941978454589844, | |
| "logps/rejected": -39.85774230957031, | |
| "loss": 0.6596, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.01848122850060463, | |
| "rewards/margins": 0.07056263089179993, | |
| "rewards/rejected": -0.08904386311769485, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 3.45, | |
| "learning_rate": 2.779113783626916e-07, | |
| "logits/chosen": -1.2977750301361084, | |
| "logits/rejected": -1.2993113994598389, | |
| "logps/chosen": -35.68281936645508, | |
| "logps/rejected": -42.771202087402344, | |
| "loss": 0.6558, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.013085673563182354, | |
| "rewards/margins": 0.07813958078622818, | |
| "rewards/rejected": -0.09122525155544281, | |
| "step": 1330 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 2.5251829568697204e-07, | |
| "logits/chosen": -1.3435966968536377, | |
| "logits/rejected": -1.3425482511520386, | |
| "logps/chosen": -32.46406555175781, | |
| "logps/rejected": -40.374244689941406, | |
| "loss": 0.659, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.012653985992074013, | |
| "rewards/margins": 0.07116873562335968, | |
| "rewards/rejected": -0.08382271975278854, | |
| "step": 1340 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "learning_rate": 2.2827986432927774e-07, | |
| "logits/chosen": -1.362319827079773, | |
| "logits/rejected": -1.3474690914154053, | |
| "logps/chosen": -36.95580291748047, | |
| "logps/rejected": -47.64240264892578, | |
| "loss": 0.6519, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.026391323655843735, | |
| "rewards/margins": 0.08738837391138077, | |
| "rewards/rejected": -0.11377968639135361, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 3.53, | |
| "learning_rate": 2.0520853686560177e-07, | |
| "logits/chosen": -1.3275715112686157, | |
| "logits/rejected": -1.3374977111816406, | |
| "logps/chosen": -33.35503387451172, | |
| "logps/rejected": -41.160377502441406, | |
| "loss": 0.6572, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.012865647673606873, | |
| "rewards/margins": 0.07634075731039047, | |
| "rewards/rejected": -0.08920640498399734, | |
| "step": 1360 | |
| }, | |
| { | |
| "epoch": 3.56, | |
| "learning_rate": 1.833161662683672e-07, | |
| "logits/chosen": -1.4463578462600708, | |
| "logits/rejected": -1.4461679458618164, | |
| "logps/chosen": -32.90170669555664, | |
| "logps/rejected": -47.23381423950195, | |
| "loss": 0.6432, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.008712759241461754, | |
| "rewards/margins": 0.10531653463840485, | |
| "rewards/rejected": -0.11402928829193115, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 3.58, | |
| "learning_rate": 1.626139998169246e-07, | |
| "logits/chosen": -1.3187510967254639, | |
| "logits/rejected": -1.3259624242782593, | |
| "logps/chosen": -35.57271957397461, | |
| "logps/rejected": -47.857994079589844, | |
| "loss": 0.6506, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.013922369107604027, | |
| "rewards/margins": 0.09016112238168716, | |
| "rewards/rejected": -0.10408350080251694, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 3.61, | |
| "learning_rate": 1.4311267331922535e-07, | |
| "logits/chosen": -1.2786071300506592, | |
| "logits/rejected": -1.2746905088424683, | |
| "logps/chosen": -35.84669876098633, | |
| "logps/rejected": -39.81802749633789, | |
| "loss": 0.659, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.010794862173497677, | |
| "rewards/margins": 0.07141149789094925, | |
| "rewards/rejected": -0.0822063684463501, | |
| "step": 1390 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "learning_rate": 1.2482220564763669e-07, | |
| "logits/chosen": -1.401760458946228, | |
| "logits/rejected": -1.4005050659179688, | |
| "logps/chosen": -32.612770080566406, | |
| "logps/rejected": -40.44251251220703, | |
| "loss": 0.6597, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.013162161223590374, | |
| "rewards/margins": 0.06981828063726425, | |
| "rewards/rejected": -0.0829804539680481, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "eval_logits/chosen": -1.6204686164855957, | |
| "eval_logits/rejected": -1.6164851188659668, | |
| "eval_logps/chosen": -39.084259033203125, | |
| "eval_logps/rejected": -43.422786712646484, | |
| "eval_loss": 0.6892901659011841, | |
| "eval_rewards/accuracies": 0.5423588156700134, | |
| "eval_rewards/chosen": -0.050497058779001236, | |
| "eval_rewards/margins": 0.008564572781324387, | |
| "eval_rewards/rejected": -0.05906163901090622, | |
| "eval_runtime": 145.7418, | |
| "eval_samples_per_second": 2.353, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.66, | |
| "learning_rate": 1.0775199359171346e-07, | |
| "logits/chosen": -1.372238278388977, | |
| "logits/rejected": -1.3673722743988037, | |
| "logps/chosen": -35.539161682128906, | |
| "logps/rejected": -37.22252655029297, | |
| "loss": 0.6659, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.018966300413012505, | |
| "rewards/margins": 0.05771438404917717, | |
| "rewards/rejected": -0.07668069750070572, | |
| "step": 1410 | |
| }, | |
| { | |
| "epoch": 3.69, | |
| "learning_rate": 9.191080703056604e-08, | |
| "logits/chosen": -1.321447730064392, | |
| "logits/rejected": -1.3224408626556396, | |
| "logps/chosen": -34.87453079223633, | |
| "logps/rejected": -43.16680145263672, | |
| "loss": 0.6608, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.015096550807356834, | |
| "rewards/margins": 0.06733135879039764, | |
| "rewards/rejected": -0.08242791891098022, | |
| "step": 1420 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "learning_rate": 7.730678442730539e-08, | |
| "logits/chosen": -1.271436095237732, | |
| "logits/rejected": -1.265836477279663, | |
| "logps/chosen": -35.28139114379883, | |
| "logps/rejected": -47.02886199951172, | |
| "loss": 0.6503, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.01139620877802372, | |
| "rewards/margins": 0.09199832379817963, | |
| "rewards/rejected": -0.10339454561471939, | |
| "step": 1430 | |
| }, | |
| { | |
| "epoch": 3.74, | |
| "learning_rate": 6.394742864787806e-08, | |
| "logits/chosen": -1.285681962966919, | |
| "logits/rejected": -1.2799713611602783, | |
| "logps/chosen": -30.83676528930664, | |
| "logps/rejected": -40.77880096435547, | |
| "loss": 0.6574, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.015235202386975288, | |
| "rewards/margins": 0.07524626702070236, | |
| "rewards/rejected": -0.0904814749956131, | |
| "step": 1440 | |
| }, | |
| { | |
| "epoch": 3.77, | |
| "learning_rate": 5.183960310644748e-08, | |
| "logits/chosen": -1.337096929550171, | |
| "logits/rejected": -1.3268693685531616, | |
| "logps/chosen": -34.95880889892578, | |
| "logps/rejected": -44.96342086791992, | |
| "loss": 0.6587, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.02477274276316166, | |
| "rewards/margins": 0.07200786471366882, | |
| "rewards/rejected": -0.09678061306476593, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "learning_rate": 4.098952823928693e-08, | |
| "logits/chosen": -1.2949423789978027, | |
| "logits/rejected": -1.2914998531341553, | |
| "logps/chosen": -35.32928466796875, | |
| "logps/rejected": -39.03660583496094, | |
| "loss": 0.664, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.02019507996737957, | |
| "rewards/margins": 0.06126277893781662, | |
| "rewards/rejected": -0.08145786076784134, | |
| "step": 1460 | |
| }, | |
| { | |
| "epoch": 3.82, | |
| "learning_rate": 3.1402778309014284e-08, | |
| "logits/chosen": -1.3512227535247803, | |
| "logits/rejected": -1.3558355569839478, | |
| "logps/chosen": -33.2025260925293, | |
| "logps/rejected": -41.91522979736328, | |
| "loss": 0.656, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.011333522386848927, | |
| "rewards/margins": 0.07819454371929169, | |
| "rewards/rejected": -0.08952806890010834, | |
| "step": 1470 | |
| }, | |
| { | |
| "epoch": 3.84, | |
| "learning_rate": 2.3084278540791427e-08, | |
| "logits/chosen": -1.3535398244857788, | |
| "logits/rejected": -1.3626043796539307, | |
| "logps/chosen": -32.97187042236328, | |
| "logps/rejected": -37.688446044921875, | |
| "loss": 0.6605, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.012766157276928425, | |
| "rewards/margins": 0.06783264130353928, | |
| "rewards/rejected": -0.08059880137443542, | |
| "step": 1480 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "learning_rate": 1.6038302591975807e-08, | |
| "logits/chosen": -1.2888884544372559, | |
| "logits/rejected": -1.2834962606430054, | |
| "logps/chosen": -35.54216003417969, | |
| "logps/rejected": -40.272682189941406, | |
| "loss": 0.6634, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.018856767565011978, | |
| "rewards/margins": 0.061765290796756744, | |
| "rewards/rejected": -0.08062206208705902, | |
| "step": 1490 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "learning_rate": 1.0268470356514237e-08, | |
| "logits/chosen": -1.3495625257492065, | |
| "logits/rejected": -1.346825122833252, | |
| "logps/chosen": -35.6667366027832, | |
| "logps/rejected": -43.17388153076172, | |
| "loss": 0.6557, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.020411662757396698, | |
| "rewards/margins": 0.0787249356508255, | |
| "rewards/rejected": -0.0991365909576416, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "eval_logits/chosen": -1.6205651760101318, | |
| "eval_logits/rejected": -1.616579294204712, | |
| "eval_logps/chosen": -39.0870246887207, | |
| "eval_logps/rejected": -43.4185791015625, | |
| "eval_loss": 0.6893215179443359, | |
| "eval_rewards/accuracies": 0.5423588156700134, | |
| "eval_rewards/chosen": -0.05052470788359642, | |
| "eval_rewards/margins": 0.008494864217936993, | |
| "eval_rewards/rejected": -0.05901956930756569, | |
| "eval_runtime": 145.7021, | |
| "eval_samples_per_second": 2.354, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.92, | |
| "learning_rate": 5.777746105209147e-09, | |
| "logits/chosen": -1.4113116264343262, | |
| "logits/rejected": -1.411259651184082, | |
| "logps/chosen": -30.929424285888672, | |
| "logps/rejected": -41.88774871826172, | |
| "loss": 0.6548, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.011131499893963337, | |
| "rewards/margins": 0.08034153282642365, | |
| "rewards/rejected": -0.09147302061319351, | |
| "step": 1510 | |
| }, | |
| { | |
| "epoch": 3.95, | |
| "learning_rate": 2.5684369628148352e-09, | |
| "logits/chosen": -1.276719331741333, | |
| "logits/rejected": -1.2754055261611938, | |
| "logps/chosen": -34.34500503540039, | |
| "logps/rejected": -42.410675048828125, | |
| "loss": 0.6587, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.013031329028308392, | |
| "rewards/margins": 0.07233807444572449, | |
| "rewards/rejected": -0.08536941558122635, | |
| "step": 1520 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "learning_rate": 6.421917227455999e-10, | |
| "logits/chosen": -1.4115439653396606, | |
| "logits/rejected": -1.4091360569000244, | |
| "logps/chosen": -33.317054748535156, | |
| "logps/rejected": -40.02583694458008, | |
| "loss": 0.6612, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.018429961055517197, | |
| "rewards/margins": 0.0669919028878212, | |
| "rewards/rejected": -0.0854218453168869, | |
| "step": 1530 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -1.398667573928833, | |
| "logits/rejected": -1.3992483615875244, | |
| "logps/chosen": -33.088409423828125, | |
| "logps/rejected": -36.9452018737793, | |
| "loss": 0.6707, | |
| "rewards/accuracies": 0.7791666388511658, | |
| "rewards/chosen": -0.024684693664312363, | |
| "rewards/margins": 0.0468655489385128, | |
| "rewards/rejected": -0.07155025750398636, | |
| "step": 1540 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 1540, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5396727961379212, | |
| "train_runtime": 10793.7948, | |
| "train_samples_per_second": 1.141, | |
| "train_steps_per_second": 0.143 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1540, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |