| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9989071038251366, | |
| "eval_steps": 400, | |
| "global_step": 457, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01092896174863388, | |
| "grad_norm": 315.2451683922351, | |
| "learning_rate": 1.0869565217391303e-07, | |
| "logits/chosen": -1.0065257549285889, | |
| "logits/rejected": -1.0008176565170288, | |
| "logps/chosen": -0.28065255284309387, | |
| "logps/rejected": -0.28539329767227173, | |
| "loss": 3.4114, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -2.806525468826294, | |
| "rewards/margins": 0.04740738496184349, | |
| "rewards/rejected": -2.8539328575134277, | |
| "semantic_entropy": 0.7513969540596008, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.02185792349726776, | |
| "grad_norm": 181.1484663719842, | |
| "learning_rate": 2.1739130434782607e-07, | |
| "logits/chosen": -1.0534369945526123, | |
| "logits/rejected": -1.0029994249343872, | |
| "logps/chosen": -0.2570807933807373, | |
| "logps/rejected": -0.27113229036331177, | |
| "loss": 3.3911, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -2.570807933807373, | |
| "rewards/margins": 0.14051488041877747, | |
| "rewards/rejected": -2.711322784423828, | |
| "semantic_entropy": 0.710273802280426, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03278688524590164, | |
| "grad_norm": 178.65067568018998, | |
| "learning_rate": 3.260869565217391e-07, | |
| "logits/chosen": -1.0082308053970337, | |
| "logits/rejected": -0.9609392285346985, | |
| "logps/chosen": -0.26744094491004944, | |
| "logps/rejected": -0.27332359552383423, | |
| "loss": 3.3533, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -2.6744089126586914, | |
| "rewards/margins": 0.05882669612765312, | |
| "rewards/rejected": -2.7332358360290527, | |
| "semantic_entropy": 0.7273439168930054, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.04371584699453552, | |
| "grad_norm": 258.46792679447157, | |
| "learning_rate": 4.3478260869565214e-07, | |
| "logits/chosen": -0.9462105631828308, | |
| "logits/rejected": -0.8957524299621582, | |
| "logps/chosen": -0.27257752418518066, | |
| "logps/rejected": -0.2848864197731018, | |
| "loss": 3.3976, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -2.7257750034332275, | |
| "rewards/margins": 0.12308906018733978, | |
| "rewards/rejected": -2.8488640785217285, | |
| "semantic_entropy": 0.7455072999000549, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.0546448087431694, | |
| "grad_norm": 273.0706988791117, | |
| "learning_rate": 5.434782608695652e-07, | |
| "logits/chosen": -0.9422909617424011, | |
| "logits/rejected": -0.8697713613510132, | |
| "logps/chosen": -0.2761459946632385, | |
| "logps/rejected": -0.2941877543926239, | |
| "loss": 3.346, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -2.761460065841675, | |
| "rewards/margins": 0.18041765689849854, | |
| "rewards/rejected": -2.9418773651123047, | |
| "semantic_entropy": 0.7553174495697021, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.06557377049180328, | |
| "grad_norm": 260.6158887162735, | |
| "learning_rate": 6.521739130434782e-07, | |
| "logits/chosen": -1.0548616647720337, | |
| "logits/rejected": -0.9892600774765015, | |
| "logps/chosen": -0.2682558596134186, | |
| "logps/rejected": -0.284037709236145, | |
| "loss": 3.4058, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": -2.682558536529541, | |
| "rewards/margins": 0.1578185111284256, | |
| "rewards/rejected": -2.8403773307800293, | |
| "semantic_entropy": 0.7250551581382751, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.07650273224043716, | |
| "grad_norm": 124.4121289092538, | |
| "learning_rate": 7.608695652173913e-07, | |
| "logits/chosen": -1.0096337795257568, | |
| "logits/rejected": -0.9423675537109375, | |
| "logps/chosen": -0.2600244879722595, | |
| "logps/rejected": -0.27900081872940063, | |
| "loss": 3.313, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -2.6002449989318848, | |
| "rewards/margins": 0.18976299464702606, | |
| "rewards/rejected": -2.790008068084717, | |
| "semantic_entropy": 0.7209498882293701, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.08743169398907104, | |
| "grad_norm": 147.2873491379567, | |
| "learning_rate": 8.695652173913043e-07, | |
| "logits/chosen": -0.9600120782852173, | |
| "logits/rejected": -0.8983286619186401, | |
| "logps/chosen": -0.2835314869880676, | |
| "logps/rejected": -0.2980334460735321, | |
| "loss": 3.454, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -2.835314989089966, | |
| "rewards/margins": 0.1450195610523224, | |
| "rewards/rejected": -2.980334520339966, | |
| "semantic_entropy": 0.7609063386917114, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.09836065573770492, | |
| "grad_norm": 93.94573860721647, | |
| "learning_rate": 9.782608695652173e-07, | |
| "logits/chosen": -1.0171349048614502, | |
| "logits/rejected": -0.9333709478378296, | |
| "logps/chosen": -0.28658169507980347, | |
| "logps/rejected": -0.3058907389640808, | |
| "loss": 3.1741, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": -2.865816831588745, | |
| "rewards/margins": 0.1930905282497406, | |
| "rewards/rejected": -3.0589072704315186, | |
| "semantic_entropy": 0.7610034346580505, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.1092896174863388, | |
| "grad_norm": 286.4609615603791, | |
| "learning_rate": 9.997663088532014e-07, | |
| "logits/chosen": -0.9543835520744324, | |
| "logits/rejected": -0.8730956315994263, | |
| "logps/chosen": -0.2823755145072937, | |
| "logps/rejected": -0.2902544140815735, | |
| "loss": 3.3278, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": -2.8237555027008057, | |
| "rewards/margins": 0.07878823578357697, | |
| "rewards/rejected": -2.902543544769287, | |
| "semantic_entropy": 0.7529318332672119, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12021857923497267, | |
| "grad_norm": 122.74091082377944, | |
| "learning_rate": 9.98817312944725e-07, | |
| "logits/chosen": -0.9809161424636841, | |
| "logits/rejected": -0.8649328947067261, | |
| "logps/chosen": -0.2803560495376587, | |
| "logps/rejected": -0.3149644732475281, | |
| "loss": 3.1758, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -2.803560256958008, | |
| "rewards/margins": 0.3460845947265625, | |
| "rewards/rejected": -3.1496450901031494, | |
| "semantic_entropy": 0.7631497979164124, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.13114754098360656, | |
| "grad_norm": 87.3444735452583, | |
| "learning_rate": 9.971397915250336e-07, | |
| "logits/chosen": -1.033070683479309, | |
| "logits/rejected": -0.9856836199760437, | |
| "logps/chosen": -0.2708079218864441, | |
| "logps/rejected": -0.31113672256469727, | |
| "loss": 3.0283, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -2.7080790996551514, | |
| "rewards/margins": 0.40328770875930786, | |
| "rewards/rejected": -3.1113669872283936, | |
| "semantic_entropy": 0.7593907713890076, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14207650273224043, | |
| "grad_norm": 162.53486819318002, | |
| "learning_rate": 9.94736194623663e-07, | |
| "logits/chosen": -1.0021493434906006, | |
| "logits/rejected": -0.9318512082099915, | |
| "logps/chosen": -0.31085288524627686, | |
| "logps/rejected": -0.3405633866786957, | |
| "loss": 3.2559, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -3.1085288524627686, | |
| "rewards/margins": 0.29710477590560913, | |
| "rewards/rejected": -3.4056334495544434, | |
| "semantic_entropy": 0.8082467317581177, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.15300546448087432, | |
| "grad_norm": 348.65729079716573, | |
| "learning_rate": 9.916100327075037e-07, | |
| "logits/chosen": -0.9440506100654602, | |
| "logits/rejected": -0.9250672459602356, | |
| "logps/chosen": -0.29765281081199646, | |
| "logps/rejected": -0.3226909935474396, | |
| "loss": 3.0703, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -2.9765281677246094, | |
| "rewards/margins": 0.2503815293312073, | |
| "rewards/rejected": -3.22691011428833, | |
| "semantic_entropy": 0.7739163041114807, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16393442622950818, | |
| "grad_norm": 78.90080101900976, | |
| "learning_rate": 9.877658715537428e-07, | |
| "logits/chosen": -0.9282974004745483, | |
| "logits/rejected": -0.912223219871521, | |
| "logps/chosen": -0.3213742971420288, | |
| "logps/rejected": -0.351571649312973, | |
| "loss": 3.1657, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -3.213742733001709, | |
| "rewards/margins": 0.30197376012802124, | |
| "rewards/rejected": -3.515717029571533, | |
| "semantic_entropy": 0.8054312467575073, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.17486338797814208, | |
| "grad_norm": 86.95229439852206, | |
| "learning_rate": 9.832093255815216e-07, | |
| "logits/chosen": -0.9330040812492371, | |
| "logits/rejected": -0.8699474334716797, | |
| "logps/chosen": -0.3141597807407379, | |
| "logps/rejected": -0.3350343406200409, | |
| "loss": 3.1277, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -3.1415977478027344, | |
| "rewards/margins": 0.20874571800231934, | |
| "rewards/rejected": -3.350343704223633, | |
| "semantic_entropy": 0.7814024686813354, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18579234972677597, | |
| "grad_norm": 129.31831333568232, | |
| "learning_rate": 9.779470496520441e-07, | |
| "logits/chosen": -0.9329907298088074, | |
| "logits/rejected": -0.8826324343681335, | |
| "logps/chosen": -0.31158381700515747, | |
| "logps/rejected": -0.3667066693305969, | |
| "loss": 2.9946, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -3.1158382892608643, | |
| "rewards/margins": 0.5512284636497498, | |
| "rewards/rejected": -3.667067050933838, | |
| "semantic_entropy": 0.802832305431366, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.19672131147540983, | |
| "grad_norm": 101.69393552819945, | |
| "learning_rate": 9.719867293491144e-07, | |
| "logits/chosen": -1.0058772563934326, | |
| "logits/rejected": -0.9232236742973328, | |
| "logps/chosen": -0.33966144919395447, | |
| "logps/rejected": -0.37590381503105164, | |
| "loss": 3.0893, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -3.3966145515441895, | |
| "rewards/margins": 0.3624236285686493, | |
| "rewards/rejected": -3.759038209915161, | |
| "semantic_entropy": 0.8507563471794128, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20765027322404372, | |
| "grad_norm": 91.30230852825771, | |
| "learning_rate": 9.653370697542987e-07, | |
| "logits/chosen": -0.9487398862838745, | |
| "logits/rejected": -0.9483828544616699, | |
| "logps/chosen": -0.3381520211696625, | |
| "logps/rejected": -0.362968385219574, | |
| "loss": 2.9005, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -3.3815205097198486, | |
| "rewards/margins": 0.24816343188285828, | |
| "rewards/rejected": -3.62968373298645, | |
| "semantic_entropy": 0.8359481692314148, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.2185792349726776, | |
| "grad_norm": 108.89056387914384, | |
| "learning_rate": 9.580077827331037e-07, | |
| "logits/chosen": -0.9571771621704102, | |
| "logits/rejected": -0.9095252752304077, | |
| "logps/chosen": -0.3725859522819519, | |
| "logps/rejected": -0.4279399514198303, | |
| "loss": 2.8907, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -3.7258598804473877, | |
| "rewards/margins": 0.553540050983429, | |
| "rewards/rejected": -4.2793989181518555, | |
| "semantic_entropy": 0.8787097930908203, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.22950819672131148, | |
| "grad_norm": 127.65723295420393, | |
| "learning_rate": 9.500095727507419e-07, | |
| "logits/chosen": -1.008998155593872, | |
| "logits/rejected": -0.9757212400436401, | |
| "logps/chosen": -0.3544849753379822, | |
| "logps/rejected": -0.3977915942668915, | |
| "loss": 2.9862, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -3.5448498725891113, | |
| "rewards/margins": 0.4330664277076721, | |
| "rewards/rejected": -3.9779160022735596, | |
| "semantic_entropy": 0.8548823595046997, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.24043715846994534, | |
| "grad_norm": 114.56272891566377, | |
| "learning_rate": 9.413541212382004e-07, | |
| "logits/chosen": -0.9939772486686707, | |
| "logits/rejected": -0.9762369990348816, | |
| "logps/chosen": -0.3617566227912903, | |
| "logps/rejected": -0.4416491985321045, | |
| "loss": 2.8457, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -3.617565870285034, | |
| "rewards/margins": 0.7989261746406555, | |
| "rewards/rejected": -4.416492462158203, | |
| "semantic_entropy": 0.8938226699829102, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25136612021857924, | |
| "grad_norm": 93.3647181422965, | |
| "learning_rate": 9.320540695314438e-07, | |
| "logits/chosen": -1.007943868637085, | |
| "logits/rejected": -0.9657400846481323, | |
| "logps/chosen": -0.3646220564842224, | |
| "logps/rejected": -0.46302324533462524, | |
| "loss": 2.7542, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -3.6462206840515137, | |
| "rewards/margins": 0.9840116500854492, | |
| "rewards/rejected": -4.630232334136963, | |
| "semantic_entropy": 0.878681480884552, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.26229508196721313, | |
| "grad_norm": 83.92874813803249, | |
| "learning_rate": 9.221230004086721e-07, | |
| "logits/chosen": -1.0344518423080444, | |
| "logits/rejected": -0.9601195454597473, | |
| "logps/chosen": -0.37165606021881104, | |
| "logps/rejected": -0.42726248502731323, | |
| "loss": 2.8688, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -3.7165608406066895, | |
| "rewards/margins": 0.5560643076896667, | |
| "rewards/rejected": -4.272624969482422, | |
| "semantic_entropy": 0.8952873945236206, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.273224043715847, | |
| "grad_norm": 190.904470995411, | |
| "learning_rate": 9.11575418252596e-07, | |
| "logits/chosen": -0.9347244501113892, | |
| "logits/rejected": -0.8975458145141602, | |
| "logps/chosen": -0.38535335659980774, | |
| "logps/rejected": -0.4541945457458496, | |
| "loss": 2.7207, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -3.853533983230591, | |
| "rewards/margins": 0.68841153383255, | |
| "rewards/rejected": -4.541945457458496, | |
| "semantic_entropy": 0.9004859924316406, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.28415300546448086, | |
| "grad_norm": 92.76045922249655, | |
| "learning_rate": 9.004267278667031e-07, | |
| "logits/chosen": -0.9624107480049133, | |
| "logits/rejected": -0.9534618258476257, | |
| "logps/chosen": -0.4056780934333801, | |
| "logps/rejected": -0.5245551466941833, | |
| "loss": 2.7139, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -4.056780815124512, | |
| "rewards/margins": 1.1887714862823486, | |
| "rewards/rejected": -5.245552062988281, | |
| "semantic_entropy": 0.8837997317314148, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29508196721311475, | |
| "grad_norm": 73.9902335755549, | |
| "learning_rate": 8.886932119764565e-07, | |
| "logits/chosen": -1.0003821849822998, | |
| "logits/rejected": -0.9125338792800903, | |
| "logps/chosen": -0.3864729106426239, | |
| "logps/rejected": -0.4857531189918518, | |
| "loss": 2.6137, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -3.8647284507751465, | |
| "rewards/margins": 0.9928020238876343, | |
| "rewards/rejected": -4.8575310707092285, | |
| "semantic_entropy": 0.9038194417953491, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.30601092896174864, | |
| "grad_norm": 86.17327659708472, | |
| "learning_rate": 8.763920074482809e-07, | |
| "logits/chosen": -0.9963301420211792, | |
| "logits/rejected": -0.9396141767501831, | |
| "logps/chosen": -0.4139133095741272, | |
| "logps/rejected": -0.5436104536056519, | |
| "loss": 2.4158, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -4.139132976531982, | |
| "rewards/margins": 1.296971082687378, | |
| "rewards/rejected": -5.4361042976379395, | |
| "semantic_entropy": 0.9314233064651489, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31693989071038253, | |
| "grad_norm": 95.91548705026597, | |
| "learning_rate": 8.635410802610723e-07, | |
| "logits/chosen": -0.9837471842765808, | |
| "logits/rejected": -0.96197909116745, | |
| "logps/chosen": -0.3892672657966614, | |
| "logps/rejected": -0.4463191032409668, | |
| "loss": 2.5469, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -3.8926727771759033, | |
| "rewards/margins": 0.5705188512802124, | |
| "rewards/rejected": -4.463191986083984, | |
| "semantic_entropy": 0.9131715893745422, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.32786885245901637, | |
| "grad_norm": 62.831854072245996, | |
| "learning_rate": 8.501591992667849e-07, | |
| "logits/chosen": -1.0432965755462646, | |
| "logits/rejected": -1.0063092708587646, | |
| "logps/chosen": -0.42201298475265503, | |
| "logps/rejected": -0.5824503302574158, | |
| "loss": 2.4081, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -4.22012996673584, | |
| "rewards/margins": 1.6043736934661865, | |
| "rewards/rejected": -5.8245038986206055, | |
| "semantic_entropy": 0.9204033613204956, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33879781420765026, | |
| "grad_norm": 71.34642946039108, | |
| "learning_rate": 8.362659087784152e-07, | |
| "logits/chosen": -1.0033342838287354, | |
| "logits/rejected": -0.943057656288147, | |
| "logps/chosen": -0.4163185656070709, | |
| "logps/rejected": -0.5151209831237793, | |
| "loss": 2.5079, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -4.163184642791748, | |
| "rewards/margins": 0.9880247116088867, | |
| "rewards/rejected": -5.151209831237793, | |
| "semantic_entropy": 0.9188691973686218, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.34972677595628415, | |
| "grad_norm": 95.49736374278223, | |
| "learning_rate": 8.218815000254231e-07, | |
| "logits/chosen": -1.036727786064148, | |
| "logits/rejected": -0.9749704599380493, | |
| "logps/chosen": -0.46870869398117065, | |
| "logps/rejected": -0.5485578775405884, | |
| "loss": 2.5679, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -4.68708610534668, | |
| "rewards/margins": 0.7984916567802429, | |
| "rewards/rejected": -5.4855780601501465, | |
| "semantic_entropy": 0.9369996786117554, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36065573770491804, | |
| "grad_norm": 90.06304805222751, | |
| "learning_rate": 8.07026981518276e-07, | |
| "logits/chosen": -1.0219743251800537, | |
| "logits/rejected": -0.9637954831123352, | |
| "logps/chosen": -0.4483868181705475, | |
| "logps/rejected": -0.5347827076911926, | |
| "loss": 2.5189, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -4.48386812210083, | |
| "rewards/margins": 0.8639583587646484, | |
| "rewards/rejected": -5.3478264808654785, | |
| "semantic_entropy": 0.9511811137199402, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.37158469945355194, | |
| "grad_norm": 99.59126068255084, | |
| "learning_rate": 7.917240483654e-07, | |
| "logits/chosen": -1.01731276512146, | |
| "logits/rejected": -0.9495924115180969, | |
| "logps/chosen": -0.44470348954200745, | |
| "logps/rejected": -0.5198506712913513, | |
| "loss": 2.6168, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -4.447035312652588, | |
| "rewards/margins": 0.7514716386795044, | |
| "rewards/rejected": -5.198506832122803, | |
| "semantic_entropy": 0.9501636624336243, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.3825136612021858, | |
| "grad_norm": 91.44599232569668, | |
| "learning_rate": 7.759950505873521e-07, | |
| "logits/chosen": -1.067455768585205, | |
| "logits/rejected": -1.031198263168335, | |
| "logps/chosen": -0.4639251232147217, | |
| "logps/rejected": -0.5280762910842896, | |
| "loss": 2.4665, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -4.639250755310059, | |
| "rewards/margins": 0.6415112614631653, | |
| "rewards/rejected": -5.280762672424316, | |
| "semantic_entropy": 0.9261935949325562, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.39344262295081966, | |
| "grad_norm": 90.82885753738844, | |
| "learning_rate": 7.598629604744872e-07, | |
| "logits/chosen": -1.0707954168319702, | |
| "logits/rejected": -1.0595567226409912, | |
| "logps/chosen": -0.43981847167015076, | |
| "logps/rejected": -0.5758017897605896, | |
| "loss": 2.3437, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -4.398184776306152, | |
| "rewards/margins": 1.3598332405090332, | |
| "rewards/rejected": -5.7580180168151855, | |
| "semantic_entropy": 0.966151237487793, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.40437158469945356, | |
| "grad_norm": 67.92403731887116, | |
| "learning_rate": 7.433513390357989e-07, | |
| "logits/chosen": -1.108884572982788, | |
| "logits/rejected": -1.1143901348114014, | |
| "logps/chosen": -0.46474918723106384, | |
| "logps/rejected": -0.5912537574768066, | |
| "loss": 2.3628, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -4.647492408752441, | |
| "rewards/margins": 1.265044927597046, | |
| "rewards/rejected": -5.912537097930908, | |
| "semantic_entropy": 0.9750612378120422, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 0.41530054644808745, | |
| "grad_norm": 108.44092763175198, | |
| "learning_rate": 7.264843015879321e-07, | |
| "logits/chosen": -1.1020928621292114, | |
| "logits/rejected": -1.0545780658721924, | |
| "logps/chosen": -0.4519892632961273, | |
| "logps/rejected": -0.6003154516220093, | |
| "loss": 2.4166, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -4.51989221572876, | |
| "rewards/margins": 1.483262062072754, | |
| "rewards/rejected": -6.003154754638672, | |
| "semantic_entropy": 0.9666361808776855, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4262295081967213, | |
| "grad_norm": 83.4733675057699, | |
| "learning_rate": 7.092864825346266e-07, | |
| "logits/chosen": -1.129482626914978, | |
| "logits/rejected": -1.0993843078613281, | |
| "logps/chosen": -0.5358282327651978, | |
| "logps/rejected": -0.7053772211074829, | |
| "loss": 2.467, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -5.358282566070557, | |
| "rewards/margins": 1.6954904794692993, | |
| "rewards/rejected": -7.053772926330566, | |
| "semantic_entropy": 0.9725033044815063, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.4371584699453552, | |
| "grad_norm": 84.36629950168863, | |
| "learning_rate": 6.917829993880302e-07, | |
| "logits/chosen": -1.1204617023468018, | |
| "logits/rejected": -1.0279228687286377, | |
| "logps/chosen": -0.5025330185890198, | |
| "logps/rejected": -0.6288330554962158, | |
| "loss": 2.3371, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -5.025330543518066, | |
| "rewards/margins": 1.263000249862671, | |
| "rewards/rejected": -6.288330078125, | |
| "semantic_entropy": 0.9786258935928345, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44808743169398907, | |
| "grad_norm": 69.08292746057649, | |
| "learning_rate": 6.739994160844309e-07, | |
| "logits/chosen": -1.0733792781829834, | |
| "logits/rejected": -1.0833173990249634, | |
| "logps/chosen": -0.4674602448940277, | |
| "logps/rejected": -0.6061697006225586, | |
| "loss": 2.2298, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -4.67460298538208, | |
| "rewards/margins": 1.3870941400527954, | |
| "rewards/rejected": -6.061697006225586, | |
| "semantic_entropy": 0.9808717966079712, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 0.45901639344262296, | |
| "grad_norm": 173.88772431903314, | |
| "learning_rate": 6.559617056479827e-07, | |
| "logits/chosen": -1.1001962423324585, | |
| "logits/rejected": -1.0926573276519775, | |
| "logps/chosen": -0.5027323961257935, | |
| "logps/rejected": -0.6716328263282776, | |
| "loss": 2.2974, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -5.027324676513672, | |
| "rewards/margins": 1.6890045404434204, | |
| "rewards/rejected": -6.7163286209106445, | |
| "semantic_entropy": 0.9590319395065308, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46994535519125685, | |
| "grad_norm": 94.73792447265116, | |
| "learning_rate": 6.376962122569567e-07, | |
| "logits/chosen": -1.1140978336334229, | |
| "logits/rejected": -1.0545861721038818, | |
| "logps/chosen": -0.5114679336547852, | |
| "logps/rejected": -0.6812509298324585, | |
| "loss": 2.4425, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -5.114679336547852, | |
| "rewards/margins": 1.6978304386138916, | |
| "rewards/rejected": -6.812510013580322, | |
| "semantic_entropy": 1.0018761157989502, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 0.4808743169398907, | |
| "grad_norm": 92.5570818396132, | |
| "learning_rate": 6.192296127679192e-07, | |
| "logits/chosen": -1.1659886837005615, | |
| "logits/rejected": -1.1348073482513428, | |
| "logps/chosen": -0.5338795781135559, | |
| "logps/rejected": -0.6664601564407349, | |
| "loss": 2.2908, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -5.338796138763428, | |
| "rewards/margins": 1.325805425643921, | |
| "rewards/rejected": -6.664601802825928, | |
| "semantic_entropy": 0.9806681871414185, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.4918032786885246, | |
| "grad_norm": 69.72669762595787, | |
| "learning_rate": 6.005888777540319e-07, | |
| "logits/chosen": -1.1951611042022705, | |
| "logits/rejected": -1.1498881578445435, | |
| "logps/chosen": -0.5304981470108032, | |
| "logps/rejected": -0.6832343935966492, | |
| "loss": 2.335, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -5.304981708526611, | |
| "rewards/margins": 1.5273630619049072, | |
| "rewards/rejected": -6.832344055175781, | |
| "semantic_entropy": 1.0007470846176147, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.5027322404371585, | |
| "grad_norm": 75.21829123168051, | |
| "learning_rate": 5.818012321143773e-07, | |
| "logits/chosen": -1.0969598293304443, | |
| "logits/rejected": -1.0912028551101685, | |
| "logps/chosen": -0.5201154351234436, | |
| "logps/rejected": -0.7016697525978088, | |
| "loss": 2.2528, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -5.2011542320251465, | |
| "rewards/margins": 1.8155431747436523, | |
| "rewards/rejected": -7.016697883605957, | |
| "semantic_entropy": 0.9929972887039185, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5136612021857924, | |
| "grad_norm": 83.71823549926938, | |
| "learning_rate": 5.628941153118388e-07, | |
| "logits/chosen": -1.1018104553222656, | |
| "logits/rejected": -1.0575555562973022, | |
| "logps/chosen": -0.5212110280990601, | |
| "logps/rejected": -0.664139986038208, | |
| "loss": 2.2375, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -5.21211051940918, | |
| "rewards/margins": 1.4292891025543213, | |
| "rewards/rejected": -6.641399383544922, | |
| "semantic_entropy": 0.9852234125137329, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 0.5245901639344263, | |
| "grad_norm": 69.35272919599275, | |
| "learning_rate": 5.438951412976098e-07, | |
| "logits/chosen": -1.1364176273345947, | |
| "logits/rejected": -1.141788125038147, | |
| "logps/chosen": -0.49681615829467773, | |
| "logps/rejected": -0.6832265853881836, | |
| "loss": 2.0805, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -4.9681620597839355, | |
| "rewards/margins": 1.8641045093536377, | |
| "rewards/rejected": -6.832266330718994, | |
| "semantic_entropy": 1.0116462707519531, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5355191256830601, | |
| "grad_norm": 71.19254935234711, | |
| "learning_rate": 5.248320581808619e-07, | |
| "logits/chosen": -1.0613957643508911, | |
| "logits/rejected": -1.0091025829315186, | |
| "logps/chosen": -0.5141640901565552, | |
| "logps/rejected": -0.6968377828598022, | |
| "loss": 2.1843, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -5.141640663146973, | |
| "rewards/margins": 1.8267381191253662, | |
| "rewards/rejected": -6.96837854385376, | |
| "semantic_entropy": 0.9718330502510071, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 0.546448087431694, | |
| "grad_norm": 70.83927823916859, | |
| "learning_rate": 5.057327077024744e-07, | |
| "logits/chosen": -1.146533727645874, | |
| "logits/rejected": -1.1047497987747192, | |
| "logps/chosen": -0.5028788447380066, | |
| "logps/rejected": -0.6331702470779419, | |
| "loss": 2.2718, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -5.028789043426514, | |
| "rewards/margins": 1.3029136657714844, | |
| "rewards/rejected": -6.331702709197998, | |
| "semantic_entropy": 0.9849420785903931, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5573770491803278, | |
| "grad_norm": 78.29453714448445, | |
| "learning_rate": 4.866249845720132e-07, | |
| "logits/chosen": -1.1301579475402832, | |
| "logits/rejected": -1.091973900794983, | |
| "logps/chosen": -0.555388331413269, | |
| "logps/rejected": -0.7187477946281433, | |
| "loss": 2.1692, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -5.553883075714111, | |
| "rewards/margins": 1.633594274520874, | |
| "rewards/rejected": -7.187478065490723, | |
| "semantic_entropy": 0.9999436140060425, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 0.5683060109289617, | |
| "grad_norm": 74.66530559540921, | |
| "learning_rate": 4.675367957273505e-07, | |
| "logits/chosen": -1.096861720085144, | |
| "logits/rejected": -1.0846450328826904, | |
| "logps/chosen": -0.5131552815437317, | |
| "logps/rejected": -0.671288251876831, | |
| "loss": 2.1911, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -5.131552219390869, | |
| "rewards/margins": 1.5813300609588623, | |
| "rewards/rejected": -6.712882041931152, | |
| "semantic_entropy": 0.9927156567573547, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5792349726775956, | |
| "grad_norm": 122.60471760919329, | |
| "learning_rate": 4.4849601957642285e-07, | |
| "logits/chosen": -1.124089002609253, | |
| "logits/rejected": -1.0828189849853516, | |
| "logps/chosen": -0.5186060070991516, | |
| "logps/rejected": -0.6825847625732422, | |
| "loss": 2.1964, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -5.186059474945068, | |
| "rewards/margins": 1.6397874355316162, | |
| "rewards/rejected": -6.8258466720581055, | |
| "semantic_entropy": 0.9880490303039551, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 0.5901639344262295, | |
| "grad_norm": 86.34568213442027, | |
| "learning_rate": 4.295304652806592e-07, | |
| "logits/chosen": -1.1392979621887207, | |
| "logits/rejected": -1.1078673601150513, | |
| "logps/chosen": -0.5172940492630005, | |
| "logps/rejected": -0.699386477470398, | |
| "loss": 2.0791, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -5.172940254211426, | |
| "rewards/margins": 1.8209247589111328, | |
| "rewards/rejected": -6.993865013122559, | |
| "semantic_entropy": 0.9866276979446411, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.6010928961748634, | |
| "grad_norm": 58.20062582439532, | |
| "learning_rate": 4.106678321395433e-07, | |
| "logits/chosen": -1.1032135486602783, | |
| "logits/rejected": -1.0302824974060059, | |
| "logps/chosen": -0.5297619104385376, | |
| "logps/rejected": -0.627161979675293, | |
| "loss": 2.1916, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -5.2976179122924805, | |
| "rewards/margins": 0.9740018844604492, | |
| "rewards/rejected": -6.271620750427246, | |
| "semantic_entropy": 0.9879854917526245, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.6120218579234973, | |
| "grad_norm": 84.89751844734043, | |
| "learning_rate": 3.9193566913562915e-07, | |
| "logits/chosen": -1.0617036819458008, | |
| "logits/rejected": -1.0624239444732666, | |
| "logps/chosen": -0.5223734378814697, | |
| "logps/rejected": -0.7275804281234741, | |
| "loss": 2.1763, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -5.223733901977539, | |
| "rewards/margins": 2.0520694255828857, | |
| "rewards/rejected": -7.275804042816162, | |
| "semantic_entropy": 0.986528754234314, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6229508196721312, | |
| "grad_norm": 72.19485427611927, | |
| "learning_rate": 3.7336133469909623e-07, | |
| "logits/chosen": -1.1958709955215454, | |
| "logits/rejected": -1.1594369411468506, | |
| "logps/chosen": -0.5087668895721436, | |
| "logps/rejected": -0.702479898929596, | |
| "loss": 2.1358, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -5.0876688957214355, | |
| "rewards/margins": 1.9371296167373657, | |
| "rewards/rejected": -7.024799346923828, | |
| "semantic_entropy": 0.9978361129760742, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 0.6338797814207651, | |
| "grad_norm": 81.98995443073827, | |
| "learning_rate": 3.549719567506076e-07, | |
| "logits/chosen": -1.1317315101623535, | |
| "logits/rejected": -1.0870287418365479, | |
| "logps/chosen": -0.5346897840499878, | |
| "logps/rejected": -0.6969183087348938, | |
| "loss": 2.1523, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -5.346898078918457, | |
| "rewards/margins": 1.62228524684906, | |
| "rewards/rejected": -6.969183444976807, | |
| "semantic_entropy": 1.0123668909072876, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.644808743169399, | |
| "grad_norm": 67.90034025619751, | |
| "learning_rate": 3.3679439308082774e-07, | |
| "logits/chosen": -1.115994930267334, | |
| "logits/rejected": -1.1152690649032593, | |
| "logps/chosen": -0.5274439454078674, | |
| "logps/rejected": -0.7239035367965698, | |
| "loss": 1.9278, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -5.274438381195068, | |
| "rewards/margins": 1.9645967483520508, | |
| "rewards/rejected": -7.239035606384277, | |
| "semantic_entropy": 1.0061827898025513, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 0.6557377049180327, | |
| "grad_norm": 70.85439265034472, | |
| "learning_rate": 3.1885519212446716e-07, | |
| "logits/chosen": -1.144639253616333, | |
| "logits/rejected": -1.1228580474853516, | |
| "logps/chosen": -0.542576253414154, | |
| "logps/rejected": -0.7291213274002075, | |
| "loss": 2.0159, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -5.42576265335083, | |
| "rewards/margins": 1.865450143814087, | |
| "rewards/rejected": -7.291213035583496, | |
| "semantic_entropy": 0.9855409860610962, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6666666666666666, | |
| "grad_norm": 76.04305212768787, | |
| "learning_rate": 3.0118055418614295e-07, | |
| "logits/chosen": -1.1450592279434204, | |
| "logits/rejected": -1.0869606733322144, | |
| "logps/chosen": -0.5319762229919434, | |
| "logps/rejected": -0.7148723006248474, | |
| "loss": 2.1436, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -5.319762706756592, | |
| "rewards/margins": 1.8289600610733032, | |
| "rewards/rejected": -7.1487226486206055, | |
| "semantic_entropy": 1.007246732711792, | |
| "step": 305 | |
| }, | |
| { | |
| "epoch": 0.6775956284153005, | |
| "grad_norm": 108.42660481521786, | |
| "learning_rate": 2.83796293174686e-07, | |
| "logits/chosen": -1.0885827541351318, | |
| "logits/rejected": -1.092543363571167, | |
| "logps/chosen": -0.5401273369789124, | |
| "logps/rejected": -0.748576283454895, | |
| "loss": 2.2592, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -5.401273250579834, | |
| "rewards/margins": 2.084489107131958, | |
| "rewards/rejected": -7.485762119293213, | |
| "semantic_entropy": 0.9948571920394897, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6885245901639344, | |
| "grad_norm": 70.86094179855972, | |
| "learning_rate": 2.6672779890178046e-07, | |
| "logits/chosen": -1.1491663455963135, | |
| "logits/rejected": -1.1490873098373413, | |
| "logps/chosen": -0.5631974935531616, | |
| "logps/rejected": -0.6903260946273804, | |
| "loss": 2.1308, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -5.631974697113037, | |
| "rewards/margins": 1.2712849378585815, | |
| "rewards/rejected": -6.903260231018066, | |
| "semantic_entropy": 0.9926439523696899, | |
| "step": 315 | |
| }, | |
| { | |
| "epoch": 0.6994535519125683, | |
| "grad_norm": 57.13767602235213, | |
| "learning_rate": 2.500000000000001e-07, | |
| "logits/chosen": -1.2022249698638916, | |
| "logits/rejected": -1.1517468690872192, | |
| "logps/chosen": -0.5420448184013367, | |
| "logps/rejected": -0.747357964515686, | |
| "loss": 2.0999, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -5.42044734954834, | |
| "rewards/margins": 2.053131580352783, | |
| "rewards/rejected": -7.473579406738281, | |
| "semantic_entropy": 1.00899338722229, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.7103825136612022, | |
| "grad_norm": 82.09122654285451, | |
| "learning_rate": 2.3363732751439923e-07, | |
| "logits/chosen": -1.1618945598602295, | |
| "logits/rejected": -1.143754243850708, | |
| "logps/chosen": -0.5291402339935303, | |
| "logps/rejected": -0.7228410243988037, | |
| "loss": 2.0999, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -5.291402339935303, | |
| "rewards/margins": 1.9370079040527344, | |
| "rewards/rejected": -7.228410243988037, | |
| "semantic_entropy": 1.0088245868682861, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.7213114754098361, | |
| "grad_norm": 62.18032792736575, | |
| "learning_rate": 2.1766367922083283e-07, | |
| "logits/chosen": -1.112157940864563, | |
| "logits/rejected": -1.0798307657241821, | |
| "logps/chosen": -0.4986172318458557, | |
| "logps/rejected": -0.7466678023338318, | |
| "loss": 2.0623, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -4.986172676086426, | |
| "rewards/margins": 2.4805047512054443, | |
| "rewards/rejected": -7.466677188873291, | |
| "semantic_entropy": 0.98463374376297, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.73224043715847, | |
| "grad_norm": 77.93323322539872, | |
| "learning_rate": 2.021023847231202e-07, | |
| "logits/chosen": -1.1002051830291748, | |
| "logits/rejected": -1.0612647533416748, | |
| "logps/chosen": -0.5647180080413818, | |
| "logps/rejected": -0.758574366569519, | |
| "loss": 2.0578, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -5.647180557250977, | |
| "rewards/margins": 1.938563346862793, | |
| "rewards/rejected": -7.5857439041137695, | |
| "semantic_entropy": 0.9851423501968384, | |
| "step": 335 | |
| }, | |
| { | |
| "epoch": 0.7431693989071039, | |
| "grad_norm": 86.17026425537334, | |
| "learning_rate": 1.869761713800254e-07, | |
| "logits/chosen": -1.1088799238204956, | |
| "logits/rejected": -1.064263939857483, | |
| "logps/chosen": -0.5516521334648132, | |
| "logps/rejected": -0.7183089256286621, | |
| "loss": 2.1191, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -5.516521453857422, | |
| "rewards/margins": 1.6665668487548828, | |
| "rewards/rejected": -7.183088779449463, | |
| "semantic_entropy": 0.9953676462173462, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7540983606557377, | |
| "grad_norm": 80.58633018790611, | |
| "learning_rate": 1.7230713111182164e-07, | |
| "logits/chosen": -1.156589150428772, | |
| "logits/rejected": -1.1543285846710205, | |
| "logps/chosen": -0.5463498830795288, | |
| "logps/rejected": -0.7534428238868713, | |
| "loss": 2.1553, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -5.463499546051025, | |
| "rewards/margins": 2.0709292888641357, | |
| "rewards/rejected": -7.534428596496582, | |
| "semantic_entropy": 0.9916456341743469, | |
| "step": 345 | |
| }, | |
| { | |
| "epoch": 0.7650273224043715, | |
| "grad_norm": 89.6142113445473, | |
| "learning_rate": 1.5811668813491696e-07, | |
| "logits/chosen": -1.1436890363693237, | |
| "logits/rejected": -1.124874234199524, | |
| "logps/chosen": -0.5143482685089111, | |
| "logps/rejected": -0.6786841154098511, | |
| "loss": 2.0898, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -5.143482685089111, | |
| "rewards/margins": 1.643358588218689, | |
| "rewards/rejected": -6.786840915679932, | |
| "semantic_entropy": 0.9892457127571106, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7759562841530054, | |
| "grad_norm": 66.31246023221092, | |
| "learning_rate": 1.4442556767166369e-07, | |
| "logits/chosen": -1.1231715679168701, | |
| "logits/rejected": -1.0935585498809814, | |
| "logps/chosen": -0.5266932845115662, | |
| "logps/rejected": -0.6977185010910034, | |
| "loss": 2.0776, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -5.266932487487793, | |
| "rewards/margins": 1.7102525234222412, | |
| "rewards/rejected": -6.9771857261657715, | |
| "semantic_entropy": 1.0020959377288818, | |
| "step": 355 | |
| }, | |
| { | |
| "epoch": 0.7868852459016393, | |
| "grad_norm": 83.110534751007, | |
| "learning_rate": 1.312537656810549e-07, | |
| "logits/chosen": -1.0739078521728516, | |
| "logits/rejected": -1.0743262767791748, | |
| "logps/chosen": -0.5362976789474487, | |
| "logps/rejected": -0.714411735534668, | |
| "loss": 2.1365, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -5.362977504730225, | |
| "rewards/margins": 1.7811400890350342, | |
| "rewards/rejected": -7.144117832183838, | |
| "semantic_entropy": 0.9827717542648315, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7978142076502732, | |
| "grad_norm": 141.91877356203636, | |
| "learning_rate": 1.1862051965451214e-07, | |
| "logits/chosen": -1.1579176187515259, | |
| "logits/rejected": -1.1566094160079956, | |
| "logps/chosen": -0.5423863530158997, | |
| "logps/rejected": -0.7343495488166809, | |
| "loss": 2.054, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -5.423863410949707, | |
| "rewards/margins": 1.9196319580078125, | |
| "rewards/rejected": -7.3434953689575195, | |
| "semantic_entropy": 1.011482834815979, | |
| "step": 365 | |
| }, | |
| { | |
| "epoch": 0.8087431693989071, | |
| "grad_norm": 78.59703524378126, | |
| "learning_rate": 1.0654428051942138e-07, | |
| "logits/chosen": -1.165038824081421, | |
| "logits/rejected": -1.1290335655212402, | |
| "logps/chosen": -0.5577388405799866, | |
| "logps/rejected": -0.7825115919113159, | |
| "loss": 2.1829, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -5.577388286590576, | |
| "rewards/margins": 2.247727632522583, | |
| "rewards/rejected": -7.825116157531738, | |
| "semantic_entropy": 1.0018393993377686, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.819672131147541, | |
| "grad_norm": 63.6353735816688, | |
| "learning_rate": 9.504268569144763e-08, | |
| "logits/chosen": -1.2013657093048096, | |
| "logits/rejected": -1.1338837146759033, | |
| "logps/chosen": -0.5316141843795776, | |
| "logps/rejected": -0.7204681038856506, | |
| "loss": 2.0911, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -5.316141605377197, | |
| "rewards/margins": 1.8885393142700195, | |
| "rewards/rejected": -7.204681396484375, | |
| "semantic_entropy": 1.0054622888565063, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.8306010928961749, | |
| "grad_norm": 64.71607446988344, | |
| "learning_rate": 8.413253331499049e-08, | |
| "logits/chosen": -1.0807088613510132, | |
| "logits/rejected": -1.102399230003357, | |
| "logps/chosen": -0.549530029296875, | |
| "logps/rejected": -0.7361005544662476, | |
| "loss": 2.0374, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -5.49530029296875, | |
| "rewards/margins": 1.8657052516937256, | |
| "rewards/rejected": -7.361004829406738, | |
| "semantic_entropy": 1.0044524669647217, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8415300546448088, | |
| "grad_norm": 67.03443291040514, | |
| "learning_rate": 7.382975772939865e-08, | |
| "logits/chosen": -1.1790930032730103, | |
| "logits/rejected": -1.1615774631500244, | |
| "logps/chosen": -0.590388834476471, | |
| "logps/rejected": -0.7754439115524292, | |
| "loss": 2.1706, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -5.903887748718262, | |
| "rewards/margins": 1.850551962852478, | |
| "rewards/rejected": -7.754439353942871, | |
| "semantic_entropy": 1.0115418434143066, | |
| "step": 385 | |
| }, | |
| { | |
| "epoch": 0.8524590163934426, | |
| "grad_norm": 103.88519596336293, | |
| "learning_rate": 6.414940619677734e-08, | |
| "logits/chosen": -1.166526436805725, | |
| "logits/rejected": -1.1425046920776367, | |
| "logps/chosen": -0.5350316762924194, | |
| "logps/rejected": -0.7514439821243286, | |
| "loss": 2.1188, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -5.350315570831299, | |
| "rewards/margins": 2.164124011993408, | |
| "rewards/rejected": -7.514439582824707, | |
| "semantic_entropy": 1.0120559930801392, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8633879781420765, | |
| "grad_norm": 71.31872002268815, | |
| "learning_rate": 5.5105616925376296e-08, | |
| "logits/chosen": -1.1460245847702026, | |
| "logits/rejected": -1.1267726421356201, | |
| "logps/chosen": -0.5442631244659424, | |
| "logps/rejected": -0.6925864815711975, | |
| "loss": 2.0494, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -5.442631721496582, | |
| "rewards/margins": 1.4832336902618408, | |
| "rewards/rejected": -6.925864219665527, | |
| "semantic_entropy": 1.0200846195220947, | |
| "step": 395 | |
| }, | |
| { | |
| "epoch": 0.8743169398907104, | |
| "grad_norm": 75.02341041681119, | |
| "learning_rate": 4.6711598420656976e-08, | |
| "logits/chosen": -1.0774163007736206, | |
| "logits/rejected": -1.0491969585418701, | |
| "logps/chosen": -0.5711244344711304, | |
| "logps/rejected": -0.7721945643424988, | |
| "loss": 2.0424, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -5.711243629455566, | |
| "rewards/margins": 2.010702133178711, | |
| "rewards/rejected": -7.721946716308594, | |
| "semantic_entropy": 0.9843025207519531, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8743169398907104, | |
| "eval_logits/chosen": -1.373081088066101, | |
| "eval_logits/rejected": -1.332649827003479, | |
| "eval_logps/chosen": -0.5314387083053589, | |
| "eval_logps/rejected": -0.7141891121864319, | |
| "eval_loss": 2.0935795307159424, | |
| "eval_rewards/accuracies": 0.7771084308624268, | |
| "eval_rewards/chosen": -5.314386367797852, | |
| "eval_rewards/margins": 1.8275047540664673, | |
| "eval_rewards/rejected": -7.1418914794921875, | |
| "eval_runtime": 37.9946, | |
| "eval_samples_per_second": 34.689, | |
| "eval_semantic_entropy": 0.9976296424865723, | |
| "eval_steps_per_second": 2.185, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8852459016393442, | |
| "grad_norm": 68.55432141696805, | |
| "learning_rate": 3.897961019419516e-08, | |
| "logits/chosen": -1.1141546964645386, | |
| "logits/rejected": -1.046690583229065, | |
| "logps/chosen": -0.5127943754196167, | |
| "logps/rejected": -0.656581699848175, | |
| "loss": 1.9895, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -5.127943992614746, | |
| "rewards/margins": 1.4378730058670044, | |
| "rewards/rejected": -6.565816402435303, | |
| "semantic_entropy": 1.010331630706787, | |
| "step": 405 | |
| }, | |
| { | |
| "epoch": 0.8961748633879781, | |
| "grad_norm": 78.80129315841073, | |
| "learning_rate": 3.192094485859526e-08, | |
| "logits/chosen": -1.1211316585540771, | |
| "logits/rejected": -1.1407296657562256, | |
| "logps/chosen": -0.5510164499282837, | |
| "logps/rejected": -0.7787143588066101, | |
| "loss": 2.0771, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -5.5101637840271, | |
| "rewards/margins": 2.2769789695739746, | |
| "rewards/rejected": -7.787143707275391, | |
| "semantic_entropy": 0.9897591471672058, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.907103825136612, | |
| "grad_norm": 63.07603955603986, | |
| "learning_rate": 2.5545911634565265e-08, | |
| "logits/chosen": -1.1598929166793823, | |
| "logits/rejected": -1.1571664810180664, | |
| "logps/chosen": -0.5616727471351624, | |
| "logps/rejected": -0.7835390567779541, | |
| "loss": 2.1028, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -5.616726875305176, | |
| "rewards/margins": 2.2186641693115234, | |
| "rewards/rejected": -7.835390567779541, | |
| "semantic_entropy": 0.9976503252983093, | |
| "step": 415 | |
| }, | |
| { | |
| "epoch": 0.9180327868852459, | |
| "grad_norm": 89.74277499235377, | |
| "learning_rate": 1.9863821294241522e-08, | |
| "logits/chosen": -1.1581684350967407, | |
| "logits/rejected": -1.1270772218704224, | |
| "logps/chosen": -0.5167144536972046, | |
| "logps/rejected": -0.7354345321655273, | |
| "loss": 1.9853, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -5.167144298553467, | |
| "rewards/margins": 2.1872007846832275, | |
| "rewards/rejected": -7.354344844818115, | |
| "semantic_entropy": 1.002636194229126, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9289617486338798, | |
| "grad_norm": 71.82508706026904, | |
| "learning_rate": 1.4882972562753615e-08, | |
| "logits/chosen": -1.1459519863128662, | |
| "logits/rejected": -1.1357475519180298, | |
| "logps/chosen": -0.5768141746520996, | |
| "logps/rejected": -0.7887662053108215, | |
| "loss": 2.1228, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -5.768141746520996, | |
| "rewards/margins": 2.1195199489593506, | |
| "rewards/rejected": -7.887660980224609, | |
| "semantic_entropy": 0.9999387860298157, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.9398907103825137, | |
| "grad_norm": 91.24129023810345, | |
| "learning_rate": 1.0610639997888915e-08, | |
| "logits/chosen": -1.0857610702514648, | |
| "logits/rejected": -1.0863049030303955, | |
| "logps/chosen": -0.520858883857727, | |
| "logps/rejected": -0.7221606969833374, | |
| "loss": 1.9229, | |
| "rewards/accuracies": 0.8687499761581421, | |
| "rewards/chosen": -5.208588600158691, | |
| "rewards/margins": 2.0130181312561035, | |
| "rewards/rejected": -7.221606254577637, | |
| "semantic_entropy": 1.0139762163162231, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9508196721311475, | |
| "grad_norm": 88.53510645254667, | |
| "learning_rate": 7.053063365559997e-09, | |
| "logits/chosen": -1.147918939590454, | |
| "logits/rejected": -1.1737738847732544, | |
| "logps/chosen": -0.5408393740653992, | |
| "logps/rejected": -0.7648274898529053, | |
| "loss": 2.0597, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -5.408394813537598, | |
| "rewards/margins": 2.2398805618286133, | |
| "rewards/rejected": -7.6482744216918945, | |
| "semantic_entropy": 1.0074278116226196, | |
| "step": 435 | |
| }, | |
| { | |
| "epoch": 0.9617486338797814, | |
| "grad_norm": 77.47275941246323, | |
| "learning_rate": 4.215438526591064e-09, | |
| "logits/chosen": -1.08914053440094, | |
| "logits/rejected": -1.0488555431365967, | |
| "logps/chosen": -0.5592411160469055, | |
| "logps/rejected": -0.7156537175178528, | |
| "loss": 2.0346, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -5.592411518096924, | |
| "rewards/margins": 1.5641257762908936, | |
| "rewards/rejected": -7.156537055969238, | |
| "semantic_entropy": 0.9892482757568359, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9726775956284153, | |
| "grad_norm": 85.37542604306078, | |
| "learning_rate": 2.1019098481337426e-09, | |
| "logits/chosen": -1.1617168188095093, | |
| "logits/rejected": -1.1342122554779053, | |
| "logps/chosen": -0.5354763865470886, | |
| "logps/rejected": -0.7130267024040222, | |
| "loss": 2.0073, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -5.354763984680176, | |
| "rewards/margins": 1.7755035161972046, | |
| "rewards/rejected": -7.130267143249512, | |
| "semantic_entropy": 1.0182139873504639, | |
| "step": 445 | |
| }, | |
| { | |
| "epoch": 0.9836065573770492, | |
| "grad_norm": 92.55545564226749, | |
| "learning_rate": 7.155641507955445e-10, | |
| "logits/chosen": -1.0736119747161865, | |
| "logits/rejected": -1.0667097568511963, | |
| "logps/chosen": -0.5805756449699402, | |
| "logps/rejected": -0.7572126388549805, | |
| "loss": 2.146, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -5.805756568908691, | |
| "rewards/margins": 1.7663694620132446, | |
| "rewards/rejected": -7.5721259117126465, | |
| "semantic_entropy": 0.9835384488105774, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.994535519125683, | |
| "grad_norm": 79.92510322372067, | |
| "learning_rate": 5.842620032053824e-11, | |
| "logits/chosen": -1.0938892364501953, | |
| "logits/rejected": -1.0882636308670044, | |
| "logps/chosen": -0.5764094591140747, | |
| "logps/rejected": -0.7290435433387756, | |
| "loss": 2.187, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -5.764094829559326, | |
| "rewards/margins": 1.5263407230377197, | |
| "rewards/rejected": -7.290434837341309, | |
| "semantic_entropy": 1.0177193880081177, | |
| "step": 455 | |
| }, | |
| { | |
| "epoch": 0.9989071038251366, | |
| "step": 457, | |
| "total_flos": 0.0, | |
| "train_loss": 2.4655840506438875, | |
| "train_runtime": 5955.1851, | |
| "train_samples_per_second": 9.833, | |
| "train_steps_per_second": 0.077 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 457, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |