| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 1.9979429914781075, |
| "eval_steps": 200, |
| "global_step": 5100, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.019590557351356647, |
| "grad_norm": 140.09786987304688, |
| "learning_rate": 3.2666666666666663e-07, |
| "logits/chosen": -0.9822076559066772, |
| "logits/rejected": -0.6774621605873108, |
| "logps/chosen": -352.589111328125, |
| "logps/rejected": -284.4425354003906, |
| "loss": 0.6901, |
| "rewards/accuracies": 0.47999998927116394, |
| "rewards/chosen": 0.0031729438342154026, |
| "rewards/margins": 0.008149052038788795, |
| "rewards/rejected": -0.004976108204573393, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.03918111470271329, |
| "grad_norm": 128.5372314453125, |
| "learning_rate": 6.6e-07, |
| "logits/chosen": -0.9334858059883118, |
| "logits/rejected": -0.6409775018692017, |
| "logps/chosen": -366.2322692871094, |
| "logps/rejected": -287.4626770019531, |
| "loss": 0.6758, |
| "rewards/accuracies": 0.6150000095367432, |
| "rewards/chosen": 0.02407059632241726, |
| "rewards/margins": 0.04141266271471977, |
| "rewards/rejected": -0.017342066392302513, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.05877167205406994, |
| "grad_norm": 103.95256805419922, |
| "learning_rate": 9.933333333333333e-07, |
| "logits/chosen": -0.8153188228607178, |
| "logits/rejected": -0.5176287293434143, |
| "logps/chosen": -373.8503112792969, |
| "logps/rejected": -307.275634765625, |
| "loss": 0.6225, |
| "rewards/accuracies": 0.6775000095367432, |
| "rewards/chosen": 0.07690317183732986, |
| "rewards/margins": 0.18566985428333282, |
| "rewards/rejected": -0.10876669734716415, |
| "step": 150 |
| }, |
| { |
| "epoch": 0.07836222940542659, |
| "grad_norm": 130.03802490234375, |
| "learning_rate": 9.99894936347371e-07, |
| "logits/chosen": -0.8271468877792358, |
| "logits/rejected": -0.5104550719261169, |
| "logps/chosen": -350.3367614746094, |
| "logps/rejected": -296.0983581542969, |
| "loss": 0.6152, |
| "rewards/accuracies": 0.6775000095367432, |
| "rewards/chosen": -0.0007632786291651428, |
| "rewards/margins": 0.25347909331321716, |
| "rewards/rejected": -0.2542423605918884, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.07836222940542659, |
| "eval_logits/chosen": -0.8247441053390503, |
| "eval_logits/rejected": -0.5691510438919067, |
| "eval_logps/chosen": -352.2530212402344, |
| "eval_logps/rejected": -288.40521240234375, |
| "eval_loss": 0.6011638641357422, |
| "eval_rewards/accuracies": 0.684013307094574, |
| "eval_rewards/chosen": 0.006044471170753241, |
| "eval_rewards/margins": 0.30330440402030945, |
| "eval_rewards/rejected": -0.2972599267959595, |
| "eval_runtime": 180.5524, |
| "eval_samples_per_second": 13.304, |
| "eval_steps_per_second": 6.652, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.09795278675678323, |
| "grad_norm": 127.07317352294922, |
| "learning_rate": 9.995711712979657e-07, |
| "logits/chosen": -0.7803874015808105, |
| "logits/rejected": -0.5503089427947998, |
| "logps/chosen": -354.3157043457031, |
| "logps/rejected": -271.8356018066406, |
| "loss": 0.5935, |
| "rewards/accuracies": 0.6825000047683716, |
| "rewards/chosen": -0.0023533145431429148, |
| "rewards/margins": 0.3317827880382538, |
| "rewards/rejected": -0.33413612842559814, |
| "step": 250 |
| }, |
| { |
| "epoch": 0.11754334410813988, |
| "grad_norm": 107.00403594970703, |
| "learning_rate": 9.990288027658056e-07, |
| "logits/chosen": -0.8059212565422058, |
| "logits/rejected": -0.5893051028251648, |
| "logps/chosen": -349.48760986328125, |
| "logps/rejected": -285.4340515136719, |
| "loss": 0.5528, |
| "rewards/accuracies": 0.7599999904632568, |
| "rewards/chosen": 0.04797028377652168, |
| "rewards/margins": 0.4950464367866516, |
| "rewards/rejected": -0.44707614183425903, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.13713390145949653, |
| "grad_norm": 118.62214660644531, |
| "learning_rate": 9.982680680817391e-07, |
| "logits/chosen": -0.9657462239265442, |
| "logits/rejected": -0.647233784198761, |
| "logps/chosen": -355.55096435546875, |
| "logps/rejected": -308.36053466796875, |
| "loss": 0.56, |
| "rewards/accuracies": 0.7074999809265137, |
| "rewards/chosen": -0.11357284337282181, |
| "rewards/margins": 0.5093904137611389, |
| "rewards/rejected": -0.6229632496833801, |
| "step": 350 |
| }, |
| { |
| "epoch": 0.15672445881085317, |
| "grad_norm": 80.19744110107422, |
| "learning_rate": 9.972893001297698e-07, |
| "logits/chosen": -1.0753679275512695, |
| "logits/rejected": -0.731191873550415, |
| "logps/chosen": -374.8988952636719, |
| "logps/rejected": -319.61199951171875, |
| "loss": 0.5683, |
| "rewards/accuracies": 0.7049999833106995, |
| "rewards/chosen": -0.20595918595790863, |
| "rewards/margins": 0.5294285416603088, |
| "rewards/rejected": -0.7353877425193787, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.15672445881085317, |
| "eval_logits/chosen": -1.0395917892456055, |
| "eval_logits/rejected": -0.7957448363304138, |
| "eval_logps/chosen": -354.0768127441406, |
| "eval_logps/rejected": -293.1625061035156, |
| "eval_loss": 0.5456792116165161, |
| "eval_rewards/accuracies": 0.7264779210090637, |
| "eval_rewards/chosen": -0.17633533477783203, |
| "eval_rewards/margins": 0.5966517329216003, |
| "eval_rewards/rejected": -0.7729870080947876, |
| "eval_runtime": 180.5278, |
| "eval_samples_per_second": 13.305, |
| "eval_steps_per_second": 6.653, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.17631501616220982, |
| "grad_norm": 97.5062026977539, |
| "learning_rate": 9.96092927201391e-07, |
| "logits/chosen": -1.0175271034240723, |
| "logits/rejected": -0.752033531665802, |
| "logps/chosen": -355.3705749511719, |
| "logps/rejected": -302.32373046875, |
| "loss": 0.5547, |
| "rewards/accuracies": 0.7024999856948853, |
| "rewards/chosen": -0.2221955806016922, |
| "rewards/margins": 0.6058707237243652, |
| "rewards/rejected": -0.828066349029541, |
| "step": 450 |
| }, |
| { |
| "epoch": 0.19590557351356647, |
| "grad_norm": 109.52022552490234, |
| "learning_rate": 9.946794728081737e-07, |
| "logits/chosen": -1.2173391580581665, |
| "logits/rejected": -0.8966682553291321, |
| "logps/chosen": -345.02984619140625, |
| "logps/rejected": -300.69268798828125, |
| "loss": 0.5145, |
| "rewards/accuracies": 0.7250000238418579, |
| "rewards/chosen": -0.26011258363723755, |
| "rewards/margins": 0.7154355049133301, |
| "rewards/rejected": -0.9755480885505676, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.2154961308649231, |
| "grad_norm": 65.85977172851562, |
| "learning_rate": 9.930495554526878e-07, |
| "logits/chosen": -1.1911191940307617, |
| "logits/rejected": -0.9048889875411987, |
| "logps/chosen": -360.3658142089844, |
| "logps/rejected": -297.0860900878906, |
| "loss": 0.5025, |
| "rewards/accuracies": 0.7275000214576721, |
| "rewards/chosen": -0.28734278678894043, |
| "rewards/margins": 0.8380499482154846, |
| "rewards/rejected": -1.1253927946090698, |
| "step": 550 |
| }, |
| { |
| "epoch": 0.23508668821627976, |
| "grad_norm": 63.237606048583984, |
| "learning_rate": 9.912038883578552e-07, |
| "logits/chosen": -1.0660635232925415, |
| "logits/rejected": -0.7923431992530823, |
| "logps/chosen": -363.8102722167969, |
| "logps/rejected": -295.3084716796875, |
| "loss": 0.5179, |
| "rewards/accuracies": 0.7300000190734863, |
| "rewards/chosen": -0.4248141050338745, |
| "rewards/margins": 0.8246171474456787, |
| "rewards/rejected": -1.2494312524795532, |
| "step": 600 |
| }, |
| { |
| "epoch": 0.23508668821627976, |
| "eval_logits/chosen": -1.0540363788604736, |
| "eval_logits/rejected": -0.8268531560897827, |
| "eval_logps/chosen": -355.8159484863281, |
| "eval_logps/rejected": -297.4617919921875, |
| "eval_loss": 0.5122258067131042, |
| "eval_rewards/accuracies": 0.7373022437095642, |
| "eval_rewards/chosen": -0.35024747252464294, |
| "eval_rewards/margins": 0.8526709675788879, |
| "eval_rewards/rejected": -1.2029184103012085, |
| "eval_runtime": 181.1919, |
| "eval_samples_per_second": 13.257, |
| "eval_steps_per_second": 6.628, |
| "step": 600 |
| }, |
| { |
| "epoch": 0.2546772455676364, |
| "grad_norm": 194.335693359375, |
| "learning_rate": 9.891432791548562e-07, |
| "logits/chosen": -1.0396273136138916, |
| "logits/rejected": -0.8581287860870361, |
| "logps/chosen": -375.4624938964844, |
| "logps/rejected": -302.572998046875, |
| "loss": 0.4782, |
| "rewards/accuracies": 0.7799999713897705, |
| "rewards/chosen": -0.3973628580570221, |
| "rewards/margins": 0.9747829437255859, |
| "rewards/rejected": -1.3721458911895752, |
| "step": 650 |
| }, |
| { |
| "epoch": 0.27426780291899305, |
| "grad_norm": 65.69367980957031, |
| "learning_rate": 9.868686295297258e-07, |
| "logits/chosen": -1.070594072341919, |
| "logits/rejected": -0.848624050617218, |
| "logps/chosen": -358.2480773925781, |
| "logps/rejected": -311.5008850097656, |
| "loss": 0.5109, |
| "rewards/accuracies": 0.7475000023841858, |
| "rewards/chosen": -0.465026319026947, |
| "rewards/margins": 0.9383654594421387, |
| "rewards/rejected": -1.4033918380737305, |
| "step": 700 |
| }, |
| { |
| "epoch": 0.29385836027034967, |
| "grad_norm": 116.7066650390625, |
| "learning_rate": 9.8438093482879e-07, |
| "logits/chosen": -1.0724475383758545, |
| "logits/rejected": -0.8224292993545532, |
| "logps/chosen": -373.66754150390625, |
| "logps/rejected": -297.4761047363281, |
| "loss": 0.483, |
| "rewards/accuracies": 0.7699999809265137, |
| "rewards/chosen": -0.39762890338897705, |
| "rewards/margins": 1.163187026977539, |
| "rewards/rejected": -1.5608159303665161, |
| "step": 750 |
| }, |
| { |
| "epoch": 0.31344891762170635, |
| "grad_norm": 124.66069793701172, |
| "learning_rate": 9.81681283623121e-07, |
| "logits/chosen": -1.0539674758911133, |
| "logits/rejected": -0.779601514339447, |
| "logps/chosen": -347.1493225097656, |
| "logps/rejected": -300.728515625, |
| "loss": 0.5427, |
| "rewards/accuracies": 0.7250000238418579, |
| "rewards/chosen": -0.48545289039611816, |
| "rewards/margins": 0.8515650033950806, |
| "rewards/rejected": -1.3370177745819092, |
| "step": 800 |
| }, |
| { |
| "epoch": 0.31344891762170635, |
| "eval_logits/chosen": -0.9799102544784546, |
| "eval_logits/rejected": -0.7651399970054626, |
| "eval_logps/chosen": -356.1210021972656, |
| "eval_logps/rejected": -299.1585388183594, |
| "eval_loss": 0.4938255250453949, |
| "eval_rewards/accuracies": 0.7493755221366882, |
| "eval_rewards/chosen": -0.38075584173202515, |
| "eval_rewards/margins": 0.9918379783630371, |
| "eval_rewards/rejected": -1.372593641281128, |
| "eval_runtime": 181.2826, |
| "eval_samples_per_second": 13.25, |
| "eval_steps_per_second": 6.625, |
| "step": 800 |
| }, |
| { |
| "epoch": 0.33303947497306297, |
| "grad_norm": 120.1912612915039, |
| "learning_rate": 9.787708572321983e-07, |
| "logits/chosen": -1.034041404724121, |
| "logits/rejected": -0.810564398765564, |
| "logps/chosen": -367.707763671875, |
| "logps/rejected": -309.762939453125, |
| "loss": 0.4326, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": -0.3861949145793915, |
| "rewards/margins": 1.226701021194458, |
| "rewards/rejected": -1.6128960847854614, |
| "step": 850 |
| }, |
| { |
| "epoch": 0.35263003232441964, |
| "grad_norm": 79.74503326416016, |
| "learning_rate": 9.756509292069825e-07, |
| "logits/chosen": -1.125342845916748, |
| "logits/rejected": -0.875031590461731, |
| "logps/chosen": -357.8549499511719, |
| "logps/rejected": -289.9258117675781, |
| "loss": 0.4864, |
| "rewards/accuracies": 0.762499988079071, |
| "rewards/chosen": -0.5919533967971802, |
| "rewards/margins": 1.2158567905426025, |
| "rewards/rejected": -1.8078101873397827, |
| "step": 900 |
| }, |
| { |
| "epoch": 0.37222058967577626, |
| "grad_norm": 136.9542236328125, |
| "learning_rate": 9.72322864772634e-07, |
| "logits/chosen": -1.1560072898864746, |
| "logits/rejected": -0.886911928653717, |
| "logps/chosen": -364.5694274902344, |
| "logps/rejected": -293.49981689453125, |
| "loss": 0.51, |
| "rewards/accuracies": 0.7400000095367432, |
| "rewards/chosen": -0.5705673098564148, |
| "rewards/margins": 1.0413060188293457, |
| "rewards/rejected": -1.6118732690811157, |
| "step": 950 |
| }, |
| { |
| "epoch": 0.39181114702713293, |
| "grad_norm": 116.3714370727539, |
| "learning_rate": 9.687881202311132e-07, |
| "logits/chosen": -1.0079857110977173, |
| "logits/rejected": -0.7752091884613037, |
| "logps/chosen": -349.3457946777344, |
| "logps/rejected": -286.99993896484375, |
| "loss": 0.4722, |
| "rewards/accuracies": 0.7749999761581421, |
| "rewards/chosen": -0.4944779872894287, |
| "rewards/margins": 1.1931304931640625, |
| "rewards/rejected": -1.6876084804534912, |
| "step": 1000 |
| }, |
| { |
| "epoch": 0.39181114702713293, |
| "eval_logits/chosen": -1.0761741399765015, |
| "eval_logits/rejected": -0.8803514838218689, |
| "eval_logps/chosen": -357.8018798828125, |
| "eval_logps/rejected": -302.2410888671875, |
| "eval_loss": 0.4866333603858948, |
| "eval_rewards/accuracies": 0.7510408163070679, |
| "eval_rewards/chosen": -0.5488451719284058, |
| "eval_rewards/margins": 1.1320006847381592, |
| "eval_rewards/rejected": -1.680845856666565, |
| "eval_runtime": 181.5986, |
| "eval_samples_per_second": 13.227, |
| "eval_steps_per_second": 6.613, |
| "step": 1000 |
| }, |
| { |
| "epoch": 0.41140170437848955, |
| "grad_norm": 169.1030731201172, |
| "learning_rate": 9.65048242323929e-07, |
| "logits/chosen": -1.0233768224716187, |
| "logits/rejected": -0.740478515625, |
| "logps/chosen": -380.3099365234375, |
| "logps/rejected": -323.1108703613281, |
| "loss": 0.4805, |
| "rewards/accuracies": 0.7774999737739563, |
| "rewards/chosen": -0.6764897704124451, |
| "rewards/margins": 1.2643505334854126, |
| "rewards/rejected": -1.940840482711792, |
| "step": 1050 |
| }, |
| { |
| "epoch": 0.4309922617298462, |
| "grad_norm": 107.12033081054688, |
| "learning_rate": 9.611048675553127e-07, |
| "logits/chosen": -1.0799275636672974, |
| "logits/rejected": -0.8886963129043579, |
| "logps/chosen": -387.2933349609375, |
| "logps/rejected": -321.2319030761719, |
| "loss": 0.4268, |
| "rewards/accuracies": 0.7774999737739563, |
| "rewards/chosen": -0.5030026435852051, |
| "rewards/margins": 1.4225863218307495, |
| "rewards/rejected": -1.925588846206665, |
| "step": 1100 |
| }, |
| { |
| "epoch": 0.45058281908120285, |
| "grad_norm": 124.23180389404297, |
| "learning_rate": 9.569597214761124e-07, |
| "logits/chosen": -1.0482264757156372, |
| "logits/rejected": -0.8745830059051514, |
| "logps/chosen": -352.1131591796875, |
| "logps/rejected": -303.13592529296875, |
| "loss": 0.4846, |
| "rewards/accuracies": 0.7674999833106995, |
| "rewards/chosen": -0.689760684967041, |
| "rewards/margins": 1.3204187154769897, |
| "rewards/rejected": -2.0101795196533203, |
| "step": 1150 |
| }, |
| { |
| "epoch": 0.4701733764325595, |
| "grad_norm": 68.6223373413086, |
| "learning_rate": 9.526146179287222e-07, |
| "logits/chosen": -1.097909927368164, |
| "logits/rejected": -0.8678469657897949, |
| "logps/chosen": -358.7846984863281, |
| "logps/rejected": -314.0942077636719, |
| "loss": 0.4829, |
| "rewards/accuracies": 0.7649999856948853, |
| "rewards/chosen": -0.6658291816711426, |
| "rewards/margins": 1.3254380226135254, |
| "rewards/rejected": -1.991267204284668, |
| "step": 1200 |
| }, |
| { |
| "epoch": 0.4701733764325595, |
| "eval_logits/chosen": -1.1058708429336548, |
| "eval_logits/rejected": -0.9207807183265686, |
| "eval_logps/chosen": -358.45928955078125, |
| "eval_logps/rejected": -304.12744140625, |
| "eval_loss": 0.4782937467098236, |
| "eval_rewards/accuracies": 0.7585345506668091, |
| "eval_rewards/chosen": -0.6145861148834229, |
| "eval_rewards/margins": 1.2548983097076416, |
| "eval_rewards/rejected": -1.8694841861724854, |
| "eval_runtime": 181.5576, |
| "eval_samples_per_second": 13.23, |
| "eval_steps_per_second": 6.615, |
| "step": 1200 |
| }, |
| { |
| "epoch": 0.48976393378391614, |
| "grad_norm": 182.0519561767578, |
| "learning_rate": 9.480714582533773e-07, |
| "logits/chosen": -1.0906615257263184, |
| "logits/rejected": -0.8988884091377258, |
| "logps/chosen": -363.1963195800781, |
| "logps/rejected": -310.1281433105469, |
| "loss": 0.5144, |
| "rewards/accuracies": 0.7200000286102295, |
| "rewards/chosen": -0.7843419909477234, |
| "rewards/margins": 1.230003833770752, |
| "rewards/rejected": -2.014345645904541, |
| "step": 1250 |
| }, |
| { |
| "epoch": 0.5093544911352728, |
| "grad_norm": 124.72360229492188, |
| "learning_rate": 9.433322304561614e-07, |
| "logits/chosen": -1.069496750831604, |
| "logits/rejected": -0.8835853338241577, |
| "logps/chosen": -361.09356689453125, |
| "logps/rejected": -313.1136779785156, |
| "loss": 0.463, |
| "rewards/accuracies": 0.7825000286102295, |
| "rewards/chosen": -0.662202000617981, |
| "rewards/margins": 1.3414545059204102, |
| "rewards/rejected": -2.0036566257476807, |
| "step": 1300 |
| }, |
| { |
| "epoch": 0.5289450484866295, |
| "grad_norm": 124.38642120361328, |
| "learning_rate": 9.383990083390903e-07, |
| "logits/chosen": -1.080662488937378, |
| "logits/rejected": -0.8675041794776917, |
| "logps/chosen": -355.84405517578125, |
| "logps/rejected": -322.5790710449219, |
| "loss": 0.4625, |
| "rewards/accuracies": 0.7674999833106995, |
| "rewards/chosen": -0.5591977834701538, |
| "rewards/margins": 1.3175528049468994, |
| "rewards/rejected": -1.8767504692077637, |
| "step": 1350 |
| }, |
| { |
| "epoch": 0.5485356058379861, |
| "grad_norm": 110.81146240234375, |
| "learning_rate": 9.332739505926528e-07, |
| "logits/chosen": -1.1147152185440063, |
| "logits/rejected": -0.9014944434165955, |
| "logps/chosen": -360.80523681640625, |
| "logps/rejected": -295.3592834472656, |
| "loss": 0.4819, |
| "rewards/accuracies": 0.737500011920929, |
| "rewards/chosen": -0.6334275603294373, |
| "rewards/margins": 1.295997142791748, |
| "rewards/rejected": -1.92942476272583, |
| "step": 1400 |
| }, |
| { |
| "epoch": 0.5485356058379861, |
| "eval_logits/chosen": -1.0615795850753784, |
| "eval_logits/rejected": -0.8742749094963074, |
| "eval_logps/chosen": -358.2158508300781, |
| "eval_logps/rejected": -304.3434753417969, |
| "eval_loss": 0.4777006208896637, |
| "eval_rewards/accuracies": 0.7635303735733032, |
| "eval_rewards/chosen": -0.590242862701416, |
| "eval_rewards/margins": 1.3008415699005127, |
| "eval_rewards/rejected": -1.8910844326019287, |
| "eval_runtime": 182.5694, |
| "eval_samples_per_second": 13.157, |
| "eval_steps_per_second": 6.578, |
| "step": 1400 |
| }, |
| { |
| "epoch": 0.5681261631893427, |
| "grad_norm": 99.77544403076172, |
| "learning_rate": 9.279592998512064e-07, |
| "logits/chosen": -1.0060768127441406, |
| "logits/rejected": -0.828403115272522, |
| "logps/chosen": -386.8955078125, |
| "logps/rejected": -324.961669921875, |
| "loss": 0.4593, |
| "rewards/accuracies": 0.7749999761581421, |
| "rewards/chosen": -0.6105374693870544, |
| "rewards/margins": 1.4274816513061523, |
| "rewards/rejected": -2.0380191802978516, |
| "step": 1450 |
| }, |
| { |
| "epoch": 0.5877167205406993, |
| "grad_norm": 113.63347625732422, |
| "learning_rate": 9.224573817116396e-07, |
| "logits/chosen": -1.1257575750350952, |
| "logits/rejected": -0.8088225722312927, |
| "logps/chosen": -356.369140625, |
| "logps/rejected": -299.9069519042969, |
| "loss": 0.4751, |
| "rewards/accuracies": 0.7674999833106995, |
| "rewards/chosen": -0.7516566514968872, |
| "rewards/margins": 1.3119940757751465, |
| "rewards/rejected": -2.063650608062744, |
| "step": 1500 |
| }, |
| { |
| "epoch": 0.6073072778920561, |
| "grad_norm": 71.0758285522461, |
| "learning_rate": 9.16770603715733e-07, |
| "logits/chosen": -1.0219570398330688, |
| "logits/rejected": -0.8258044719696045, |
| "logps/chosen": -365.3489074707031, |
| "logps/rejected": -317.51776123046875, |
| "loss": 0.4755, |
| "rewards/accuracies": 0.7724999785423279, |
| "rewards/chosen": -0.722920298576355, |
| "rewards/margins": 1.3333240747451782, |
| "rewards/rejected": -2.0562446117401123, |
| "step": 1550 |
| }, |
| { |
| "epoch": 0.6268978352434127, |
| "grad_norm": 63.271759033203125, |
| "learning_rate": 9.109014542966609e-07, |
| "logits/chosen": -1.0383367538452148, |
| "logits/rejected": -0.8199602365493774, |
| "logps/chosen": -361.697509765625, |
| "logps/rejected": -319.4376525878906, |
| "loss": 0.4271, |
| "rewards/accuracies": 0.8100000023841858, |
| "rewards/chosen": -0.8095757961273193, |
| "rewards/margins": 1.4505212306976318, |
| "rewards/rejected": -2.260097026824951, |
| "step": 1600 |
| }, |
| { |
| "epoch": 0.6268978352434127, |
| "eval_logits/chosen": -1.0107686519622803, |
| "eval_logits/rejected": -0.826732873916626, |
| "eval_logps/chosen": -360.0285339355469, |
| "eval_logps/rejected": -307.601318359375, |
| "eval_loss": 0.4775010347366333, |
| "eval_rewards/accuracies": 0.768109917640686, |
| "eval_rewards/chosen": -0.7715086936950684, |
| "eval_rewards/margins": 1.445361852645874, |
| "eval_rewards/rejected": -2.2168707847595215, |
| "eval_runtime": 181.1951, |
| "eval_samples_per_second": 13.256, |
| "eval_steps_per_second": 6.628, |
| "step": 1600 |
| }, |
| { |
| "epoch": 0.6464883925947693, |
| "grad_norm": 69.98014831542969, |
| "learning_rate": 9.04852501690097e-07, |
| "logits/chosen": -1.0106360912322998, |
| "logits/rejected": -0.769228458404541, |
| "logps/chosen": -361.59783935546875, |
| "logps/rejected": -298.4144287109375, |
| "loss": 0.4807, |
| "rewards/accuracies": 0.7749999761581421, |
| "rewards/chosen": -0.6844636797904968, |
| "rewards/margins": 1.4386271238327026, |
| "rewards/rejected": -2.1230905055999756, |
| "step": 1650 |
| }, |
| { |
| "epoch": 0.6660789499461259, |
| "grad_norm": 73.16764831542969, |
| "learning_rate": 8.986263928104007e-07, |
| "logits/chosen": -1.0747100114822388, |
| "logits/rejected": -0.8522970676422119, |
| "logps/chosen": -377.2361755371094, |
| "logps/rejected": -314.54180908203125, |
| "loss": 0.475, |
| "rewards/accuracies": 0.7925000190734863, |
| "rewards/chosen": -0.6853989958763123, |
| "rewards/margins": 1.3294869661331177, |
| "rewards/rejected": -2.014885902404785, |
| "step": 1700 |
| }, |
| { |
| "epoch": 0.6856695072974827, |
| "grad_norm": 159.627685546875, |
| "learning_rate": 8.922258520923739e-07, |
| "logits/chosen": -1.1866579055786133, |
| "logits/rejected": -0.9568431973457336, |
| "logps/chosen": -363.7584228515625, |
| "logps/rejected": -335.536376953125, |
| "loss": 0.4701, |
| "rewards/accuracies": 0.7549999952316284, |
| "rewards/chosen": -0.6905403733253479, |
| "rewards/margins": 1.4639062881469727, |
| "rewards/rejected": -2.1544463634490967, |
| "step": 1750 |
| }, |
| { |
| "epoch": 0.7052600646488393, |
| "grad_norm": 59.52585983276367, |
| "learning_rate": 8.856536802990969e-07, |
| "logits/chosen": -1.2136218547821045, |
| "logits/rejected": -0.915833055973053, |
| "logps/chosen": -368.9427795410156, |
| "logps/rejected": -318.7424621582031, |
| "loss": 0.4355, |
| "rewards/accuracies": 0.7925000190734863, |
| "rewards/chosen": -0.7404943704605103, |
| "rewards/margins": 1.615696907043457, |
| "rewards/rejected": -2.3561911582946777, |
| "step": 1800 |
| }, |
| { |
| "epoch": 0.7052600646488393, |
| "eval_logits/chosen": -1.137018084526062, |
| "eval_logits/rejected": -0.9567646384239197, |
| "eval_logps/chosen": -360.75030517578125, |
| "eval_logps/rejected": -307.82464599609375, |
| "eval_loss": 0.4773857891559601, |
| "eval_rewards/accuracies": 0.770191490650177, |
| "eval_rewards/chosen": -0.8436892032623291, |
| "eval_rewards/margins": 1.3955140113830566, |
| "eval_rewards/rejected": -2.2392032146453857, |
| "eval_runtime": 181.4174, |
| "eval_samples_per_second": 13.24, |
| "eval_steps_per_second": 6.62, |
| "step": 1800 |
| }, |
| { |
| "epoch": 0.7248506220001959, |
| "grad_norm": 145.34722900390625, |
| "learning_rate": 8.789127532963639e-07, |
| "logits/chosen": -1.1216288805007935, |
| "logits/rejected": -0.8633120656013489, |
| "logps/chosen": -379.6500244140625, |
| "logps/rejected": -340.3668212890625, |
| "loss": 0.4937, |
| "rewards/accuracies": 0.762499988079071, |
| "rewards/chosen": -0.894076406955719, |
| "rewards/margins": 1.335474967956543, |
| "rewards/rejected": -2.2295515537261963, |
| "step": 1850 |
| }, |
| { |
| "epoch": 0.7444411793515525, |
| "grad_norm": 85.14502716064453, |
| "learning_rate": 8.720060207942547e-07, |
| "logits/chosen": -1.1888411045074463, |
| "logits/rejected": -0.8253584504127502, |
| "logps/chosen": -358.1293029785156, |
| "logps/rejected": -322.2793273925781, |
| "loss": 0.453, |
| "rewards/accuracies": 0.7774999737739563, |
| "rewards/chosen": -0.7874532341957092, |
| "rewards/margins": 1.4413965940475464, |
| "rewards/rejected": -2.2288498878479004, |
| "step": 1900 |
| }, |
| { |
| "epoch": 0.7640317367029092, |
| "grad_norm": 59.02031707763672, |
| "learning_rate": 8.649365050563955e-07, |
| "logits/chosen": -1.0339776277542114, |
| "logits/rejected": -0.7359542846679688, |
| "logps/chosen": -379.0245666503906, |
| "logps/rejected": -335.61248779296875, |
| "loss": 0.383, |
| "rewards/accuracies": 0.824999988079071, |
| "rewards/chosen": -0.9282081127166748, |
| "rewards/margins": 1.724035382270813, |
| "rewards/rejected": -2.6522433757781982, |
| "step": 1950 |
| }, |
| { |
| "epoch": 0.7836222940542659, |
| "grad_norm": 100.15818786621094, |
| "learning_rate": 8.577072995774679e-07, |
| "logits/chosen": -1.0461410284042358, |
| "logits/rejected": -0.7923344373703003, |
| "logps/chosen": -358.7364196777344, |
| "logps/rejected": -319.2044372558594, |
| "loss": 0.5073, |
| "rewards/accuracies": 0.7450000047683716, |
| "rewards/chosen": -1.07187819480896, |
| "rewards/margins": 1.472280740737915, |
| "rewards/rejected": -2.544158697128296, |
| "step": 2000 |
| }, |
| { |
| "epoch": 0.7836222940542659, |
| "eval_logits/chosen": -1.0298668146133423, |
| "eval_logits/rejected": -0.857089638710022, |
| "eval_logps/chosen": -361.0681457519531, |
| "eval_logps/rejected": -309.28570556640625, |
| "eval_loss": 0.4695436656475067, |
| "eval_rewards/accuracies": 0.7756036520004272, |
| "eval_rewards/chosen": -0.8754721283912659, |
| "eval_rewards/margins": 1.5098369121551514, |
| "eval_rewards/rejected": -2.3853089809417725, |
| "eval_runtime": 181.4637, |
| "eval_samples_per_second": 13.237, |
| "eval_steps_per_second": 6.618, |
| "step": 2000 |
| }, |
| { |
| "epoch": 0.8032128514056225, |
| "grad_norm": 109.68525695800781, |
| "learning_rate": 8.503215677295522e-07, |
| "logits/chosen": -1.1481313705444336, |
| "logits/rejected": -0.8855399489402771, |
| "logps/chosen": -377.3407897949219, |
| "logps/rejected": -321.94903564453125, |
| "loss": 0.4, |
| "rewards/accuracies": 0.8075000047683716, |
| "rewards/chosen": -0.8187921047210693, |
| "rewards/margins": 1.6883336305618286, |
| "rewards/rejected": -2.5071256160736084, |
| "step": 2050 |
| }, |
| { |
| "epoch": 0.8228034087569791, |
| "grad_norm": 113.09908294677734, |
| "learning_rate": 8.427825413778904e-07, |
| "logits/chosen": -1.182418704032898, |
| "logits/rejected": -1.0226842164993286, |
| "logps/chosen": -370.88800048828125, |
| "logps/rejected": -316.3885803222656, |
| "loss": 0.4302, |
| "rewards/accuracies": 0.7925000190734863, |
| "rewards/chosen": -0.8482898473739624, |
| "rewards/margins": 1.790774941444397, |
| "rewards/rejected": -2.6390650272369385, |
| "step": 2100 |
| }, |
| { |
| "epoch": 0.8423939661083358, |
| "grad_norm": 115.06123352050781, |
| "learning_rate": 8.350935194666808e-07, |
| "logits/chosen": -1.19704008102417, |
| "logits/rejected": -0.9841734170913696, |
| "logps/chosen": -344.68988037109375, |
| "logps/rejected": -300.1245422363281, |
| "loss": 0.4607, |
| "rewards/accuracies": 0.762499988079071, |
| "rewards/chosen": -0.8847514390945435, |
| "rewards/margins": 1.6138572692871094, |
| "rewards/rejected": -2.4986085891723633, |
| "step": 2150 |
| }, |
| { |
| "epoch": 0.8619845234596925, |
| "grad_norm": 91.62993621826172, |
| "learning_rate": 8.272578665755176e-07, |
| "logits/chosen": -1.1561052799224854, |
| "logits/rejected": -0.892173171043396, |
| "logps/chosen": -357.5077819824219, |
| "logps/rejected": -313.1963806152344, |
| "loss": 0.4934, |
| "rewards/accuracies": 0.7574999928474426, |
| "rewards/chosen": -0.9803217053413391, |
| "rewards/margins": 1.5475577116012573, |
| "rewards/rejected": -2.527879238128662, |
| "step": 2200 |
| }, |
| { |
| "epoch": 0.8619845234596925, |
| "eval_logits/chosen": -1.0830334424972534, |
| "eval_logits/rejected": -0.9147012233734131, |
| "eval_logps/chosen": -361.2430114746094, |
| "eval_logps/rejected": -310.0792541503906, |
| "eval_loss": 0.4646490514278412, |
| "eval_rewards/accuracies": 0.7768526077270508, |
| "eval_rewards/chosen": -0.8929603099822998, |
| "eval_rewards/margins": 1.5717030763626099, |
| "eval_rewards/rejected": -2.464663505554199, |
| "eval_runtime": 183.2831, |
| "eval_samples_per_second": 13.105, |
| "eval_steps_per_second": 6.553, |
| "step": 2200 |
| }, |
| { |
| "epoch": 0.8815750808110491, |
| "grad_norm": 72.61209869384766, |
| "learning_rate": 8.192790114471106e-07, |
| "logits/chosen": -1.1451025009155273, |
| "logits/rejected": -0.9403939247131348, |
| "logps/chosen": -370.800048828125, |
| "logps/rejected": -303.31817626953125, |
| "loss": 0.4352, |
| "rewards/accuracies": 0.7875000238418579, |
| "rewards/chosen": -0.7773234844207764, |
| "rewards/margins": 1.5500972270965576, |
| "rewards/rejected": -2.327420711517334, |
| "step": 2250 |
| }, |
| { |
| "epoch": 0.9011656381624057, |
| "grad_norm": 70.87027740478516, |
| "learning_rate": 8.111604454869285e-07, |
| "logits/chosen": -1.17734956741333, |
| "logits/rejected": -0.9346526861190796, |
| "logps/chosen": -345.1195068359375, |
| "logps/rejected": -306.4920654296875, |
| "loss": 0.4703, |
| "rewards/accuracies": 0.7674999833106995, |
| "rewards/chosen": -0.9427626729011536, |
| "rewards/margins": 1.593432903289795, |
| "rewards/rejected": -2.536195755004883, |
| "step": 2300 |
| }, |
| { |
| "epoch": 0.9207561955137624, |
| "grad_norm": 90.32552337646484, |
| "learning_rate": 8.029057212354218e-07, |
| "logits/chosen": -1.0256972312927246, |
| "logits/rejected": -0.9072439074516296, |
| "logps/chosen": -364.8022155761719, |
| "logps/rejected": -313.5185241699219, |
| "loss": 0.4846, |
| "rewards/accuracies": 0.7699999809265137, |
| "rewards/chosen": -1.1208609342575073, |
| "rewards/margins": 1.5102592706680298, |
| "rewards/rejected": -2.631119966506958, |
| "step": 2350 |
| }, |
| { |
| "epoch": 0.940346752865119, |
| "grad_norm": 143.2118377685547, |
| "learning_rate": 7.945184508134936e-07, |
| "logits/chosen": -1.1666345596313477, |
| "logits/rejected": -0.9767065644264221, |
| "logps/chosen": -391.8117980957031, |
| "logps/rejected": -319.5590515136719, |
| "loss": 0.4403, |
| "rewards/accuracies": 0.7875000238418579, |
| "rewards/chosen": -0.9298499822616577, |
| "rewards/margins": 1.5455856323242188, |
| "rewards/rejected": -2.475435495376587, |
| "step": 2400 |
| }, |
| { |
| "epoch": 0.940346752865119, |
| "eval_logits/chosen": -1.10908043384552, |
| "eval_logits/rejected": -0.9499452710151672, |
| "eval_logps/chosen": -362.099609375, |
| "eval_logps/rejected": -311.1372375488281, |
| "eval_loss": 0.46884483098983765, |
| "eval_rewards/accuracies": 0.7722731232643127, |
| "eval_rewards/chosen": -0.9786169528961182, |
| "eval_rewards/margins": 1.5918443202972412, |
| "eval_rewards/rejected": -2.5704612731933594, |
| "eval_runtime": 181.3567, |
| "eval_samples_per_second": 13.245, |
| "eval_steps_per_second": 6.622, |
| "step": 2400 |
| }, |
| { |
| "epoch": 0.9599373102164757, |
| "grad_norm": 192.61532592773438, |
| "learning_rate": 7.860023043419004e-07, |
| "logits/chosen": -1.1775944232940674, |
| "logits/rejected": -0.9322341084480286, |
| "logps/chosen": -363.9286193847656, |
| "logps/rejected": -324.30621337890625, |
| "loss": 0.4438, |
| "rewards/accuracies": 0.8174999952316284, |
| "rewards/chosen": -1.010249376296997, |
| "rewards/margins": 1.8364521265029907, |
| "rewards/rejected": -2.8467013835906982, |
| "step": 2450 |
| }, |
| { |
| "epoch": 0.9795278675678323, |
| "grad_norm": 154.35006713867188, |
| "learning_rate": 7.773610083352717e-07, |
| "logits/chosen": -0.9925041794776917, |
| "logits/rejected": -0.813052773475647, |
| "logps/chosen": -363.6401062011719, |
| "logps/rejected": -323.7989501953125, |
| "loss": 0.5359, |
| "rewards/accuracies": 0.7225000262260437, |
| "rewards/chosen": -1.0799657106399536, |
| "rewards/margins": 1.3734463453292847, |
| "rewards/rejected": -2.453411817550659, |
| "step": 2500 |
| }, |
| { |
| "epoch": 0.9991184249191889, |
| "grad_norm": 196.2805633544922, |
| "learning_rate": 7.685983440714535e-07, |
| "logits/chosen": -1.093634009361267, |
| "logits/rejected": -0.8999694585800171, |
| "logps/chosen": -393.096435546875, |
| "logps/rejected": -340.03387451171875, |
| "loss": 0.4938, |
| "rewards/accuracies": 0.7699999809265137, |
| "rewards/chosen": -1.0175117254257202, |
| "rewards/margins": 1.5324511528015137, |
| "rewards/rejected": -2.5499629974365234, |
| "step": 2550 |
| }, |
| { |
| "epoch": 1.0184151239102752, |
| "grad_norm": 34.11044692993164, |
| "learning_rate": 7.597181459368884e-07, |
| "logits/chosen": -1.1935698986053467, |
| "logits/rejected": -0.9796786308288574, |
| "logps/chosen": -363.0191955566406, |
| "logps/rejected": -347.12249755859375, |
| "loss": 0.129, |
| "rewards/accuracies": 0.9670050740242004, |
| "rewards/chosen": -0.2811751663684845, |
| "rewards/margins": 3.3759641647338867, |
| "rewards/rejected": -3.657139301300049, |
| "step": 2600 |
| }, |
| { |
| "epoch": 1.0184151239102752, |
| "eval_logits/chosen": -1.2456655502319336, |
| "eval_logits/rejected": -1.0991578102111816, |
| "eval_logps/chosen": -363.8660888671875, |
| "eval_logps/rejected": -313.871826171875, |
| "eval_loss": 0.47346433997154236, |
| "eval_rewards/accuracies": 0.7718567848205566, |
| "eval_rewards/chosen": -1.1552621126174927, |
| "eval_rewards/margins": 1.68865966796875, |
| "eval_rewards/rejected": -2.8439218997955322, |
| "eval_runtime": 181.5327, |
| "eval_samples_per_second": 13.232, |
| "eval_steps_per_second": 6.616, |
| "step": 2600 |
| }, |
| { |
| "epoch": 1.0380056812616318, |
| "grad_norm": 75.25180053710938, |
| "learning_rate": 7.50724299748756e-07, |
| "logits/chosen": -1.2610849142074585, |
| "logits/rejected": -1.150437355041504, |
| "logps/chosen": -375.75, |
| "logps/rejected": -315.3388366699219, |
| "loss": 0.146, |
| "rewards/accuracies": 0.9674999713897705, |
| "rewards/chosen": -0.38413771986961365, |
| "rewards/margins": 3.210354804992676, |
| "rewards/rejected": -3.5944931507110596, |
| "step": 2650 |
| }, |
| { |
| "epoch": 1.0575962386129885, |
| "grad_norm": 19.142627716064453, |
| "learning_rate": 7.416207410546075e-07, |
| "logits/chosen": -1.4331450462341309, |
| "logits/rejected": -1.1033505201339722, |
| "logps/chosen": -341.708251953125, |
| "logps/rejected": -322.0888977050781, |
| "loss": 0.1627, |
| "rewards/accuracies": 0.9449999928474426, |
| "rewards/chosen": -0.5315707921981812, |
| "rewards/margins": 3.344498872756958, |
| "rewards/rejected": -3.876070022583008, |
| "step": 2700 |
| }, |
| { |
| "epoch": 1.0771867959643453, |
| "grad_norm": 18.423038482666016, |
| "learning_rate": 7.324114534102414e-07, |
| "logits/chosen": -1.2948585748672485, |
| "logits/rejected": -1.1851221323013306, |
| "logps/chosen": -384.01495361328125, |
| "logps/rejected": -340.6744384765625, |
| "loss": 0.1212, |
| "rewards/accuracies": 0.9549999833106995, |
| "rewards/chosen": -0.458404541015625, |
| "rewards/margins": 3.760544776916504, |
| "rewards/rejected": -4.218949317932129, |
| "step": 2750 |
| }, |
| { |
| "epoch": 1.096777353315702, |
| "grad_norm": 15.992474555969238, |
| "learning_rate": 7.231004666365688e-07, |
| "logits/chosen": -1.5737820863723755, |
| "logits/rejected": -1.3861126899719238, |
| "logps/chosen": -370.8525695800781, |
| "logps/rejected": -342.98675537109375, |
| "loss": 0.1362, |
| "rewards/accuracies": 0.949999988079071, |
| "rewards/chosen": -0.7288404703140259, |
| "rewards/margins": 3.653024911880493, |
| "rewards/rejected": -4.381865978240967, |
| "step": 2800 |
| }, |
| { |
| "epoch": 1.096777353315702, |
| "eval_logits/chosen": -1.4936131238937378, |
| "eval_logits/rejected": -1.3645743131637573, |
| "eval_logps/chosen": -369.6283874511719, |
| "eval_logps/rejected": -322.2799072265625, |
| "eval_loss": 0.5014060735702515, |
| "eval_rewards/accuracies": 0.7772689461708069, |
| "eval_rewards/chosen": -1.7314954996109009, |
| "eval_rewards/margins": 1.9532350301742554, |
| "eval_rewards/rejected": -3.6847305297851562, |
| "eval_runtime": 183.7133, |
| "eval_samples_per_second": 13.075, |
| "eval_steps_per_second": 6.537, |
| "step": 2800 |
| }, |
| { |
| "epoch": 1.1163679106670585, |
| "grad_norm": 5.409101963043213, |
| "learning_rate": 7.136918550562358e-07, |
| "logits/chosen": -1.522940993309021, |
| "logits/rejected": -1.4030097723007202, |
| "logps/chosen": -366.8924560546875, |
| "logps/rejected": -333.6314697265625, |
| "loss": 0.1492, |
| "rewards/accuracies": 0.9449999928474426, |
| "rewards/chosen": -0.8802686929702759, |
| "rewards/margins": 3.616269826889038, |
| "rewards/rejected": -4.496538162231445, |
| "step": 2850 |
| }, |
| { |
| "epoch": 1.1359584680184152, |
| "grad_norm": 54.6688232421875, |
| "learning_rate": 7.041897357107727e-07, |
| "logits/chosen": -1.6211928129196167, |
| "logits/rejected": -1.4868953227996826, |
| "logps/chosen": -359.833740234375, |
| "logps/rejected": -310.93939208984375, |
| "loss": 0.1698, |
| "rewards/accuracies": 0.9424999952316284, |
| "rewards/chosen": -0.7522315979003906, |
| "rewards/margins": 3.544074058532715, |
| "rewards/rejected": -4.296305179595947, |
| "step": 2900 |
| }, |
| { |
| "epoch": 1.1555490253697718, |
| "grad_norm": 56.798030853271484, |
| "learning_rate": 6.945982665590479e-07, |
| "logits/chosen": -1.4974216222763062, |
| "logits/rejected": -1.2985191345214844, |
| "logps/chosen": -377.2979736328125, |
| "logps/rejected": -348.5942077636719, |
| "loss": 0.146, |
| "rewards/accuracies": 0.9399999976158142, |
| "rewards/chosen": -0.8809213042259216, |
| "rewards/margins": 3.612612247467041, |
| "rewards/rejected": -4.493533134460449, |
| "step": 2950 |
| }, |
| { |
| "epoch": 1.1751395827211284, |
| "grad_norm": 22.727725982666016, |
| "learning_rate": 6.849216446578215e-07, |
| "logits/chosen": -1.590391993522644, |
| "logits/rejected": -1.433569073677063, |
| "logps/chosen": -387.9193115234375, |
| "logps/rejected": -361.5350646972656, |
| "loss": 0.1149, |
| "rewards/accuracies": 0.9725000262260437, |
| "rewards/chosen": -1.0071202516555786, |
| "rewards/margins": 3.8817479610443115, |
| "rewards/rejected": -4.88886833190918, |
| "step": 3000 |
| }, |
| { |
| "epoch": 1.1751395827211284, |
| "eval_logits/chosen": -1.6452481746673584, |
| "eval_logits/rejected": -1.5406179428100586, |
| "eval_logps/chosen": -371.4642333984375, |
| "eval_logps/rejected": -325.0046691894531, |
| "eval_loss": 0.5132338404655457, |
| "eval_rewards/accuracies": 0.770191490650177, |
| "eval_rewards/chosen": -1.9150793552398682, |
| "eval_rewards/margins": 2.042127847671509, |
| "eval_rewards/rejected": -3.957206964492798, |
| "eval_runtime": 182.0948, |
| "eval_samples_per_second": 13.191, |
| "eval_steps_per_second": 6.595, |
| "step": 3000 |
| }, |
| { |
| "epoch": 1.194730140072485, |
| "grad_norm": 25.459936141967773, |
| "learning_rate": 6.751641043251853e-07, |
| "logits/chosen": -1.5482993125915527, |
| "logits/rejected": -1.4024869203567505, |
| "logps/chosen": -392.9327697753906, |
| "logps/rejected": -338.0341796875, |
| "loss": 0.1499, |
| "rewards/accuracies": 0.949999988079071, |
| "rewards/chosen": -0.9234389662742615, |
| "rewards/margins": 3.838209629058838, |
| "rewards/rejected": -4.761648178100586, |
| "step": 3050 |
| }, |
| { |
| "epoch": 1.2143206974238416, |
| "grad_norm": 71.29440307617188, |
| "learning_rate": 6.653299152877016e-07, |
| "logits/chosen": -1.7307244539260864, |
| "logits/rejected": -1.550933837890625, |
| "logps/chosen": -357.46661376953125, |
| "logps/rejected": -342.1903076171875, |
| "loss": 0.1345, |
| "rewards/accuracies": 0.9549999833106995, |
| "rewards/chosen": -1.0262528657913208, |
| "rewards/margins": 3.8201754093170166, |
| "rewards/rejected": -4.846428871154785, |
| "step": 3100 |
| }, |
| { |
| "epoch": 1.2339112547751983, |
| "grad_norm": 25.778587341308594, |
| "learning_rate": 6.554233808120463e-07, |
| "logits/chosen": -1.6679961681365967, |
| "logits/rejected": -1.5830594301223755, |
| "logps/chosen": -386.9595642089844, |
| "logps/rejected": -344.4521789550781, |
| "loss": 0.1438, |
| "rewards/accuracies": 0.9649999737739563, |
| "rewards/chosen": -1.1483721733093262, |
| "rewards/margins": 3.911073684692383, |
| "rewards/rejected": -5.059445858001709, |
| "step": 3150 |
| }, |
| { |
| "epoch": 1.2535018121265549, |
| "grad_norm": 24.407367706298828, |
| "learning_rate": 6.454488358219756e-07, |
| "logits/chosen": -1.6095058917999268, |
| "logits/rejected": -1.5268667936325073, |
| "logps/chosen": -364.7217102050781, |
| "logps/rejected": -321.9256591796875, |
| "loss": 0.1198, |
| "rewards/accuracies": 0.9599999785423279, |
| "rewards/chosen": -0.9407815337181091, |
| "rewards/margins": 3.9222023487091064, |
| "rewards/rejected": -4.862983703613281, |
| "step": 3200 |
| }, |
| { |
| "epoch": 1.2535018121265549, |
| "eval_logits/chosen": -1.6486918926239014, |
| "eval_logits/rejected": -1.56648588180542, |
| "eval_logps/chosen": -373.4631652832031, |
| "eval_logps/rejected": -328.3328552246094, |
| "eval_loss": 0.5195496678352356, |
| "eval_rewards/accuracies": 0.7751873731613159, |
| "eval_rewards/chosen": -2.114971399307251, |
| "eval_rewards/margins": 2.175050735473633, |
| "eval_rewards/rejected": -4.290021896362305, |
| "eval_runtime": 182.0558, |
| "eval_samples_per_second": 13.194, |
| "eval_steps_per_second": 6.597, |
| "step": 3200 |
| }, |
| { |
| "epoch": 1.2730923694779117, |
| "grad_norm": 61.75260925292969, |
| "learning_rate": 6.354106450014404e-07, |
| "logits/chosen": -1.6261045932769775, |
| "logits/rejected": -1.468245267868042, |
| "logps/chosen": -394.1415100097656, |
| "logps/rejected": -353.34942626953125, |
| "loss": 0.1666, |
| "rewards/accuracies": 0.9549999833106995, |
| "rewards/chosen": -1.1136356592178345, |
| "rewards/margins": 4.16200065612793, |
| "rewards/rejected": -5.275636196136475, |
| "step": 3250 |
| }, |
| { |
| "epoch": 1.2926829268292683, |
| "grad_norm": 19.361291885375977, |
| "learning_rate": 6.253132008846786e-07, |
| "logits/chosen": -1.6353635787963867, |
| "logits/rejected": -1.5276310443878174, |
| "logps/chosen": -366.8433532714844, |
| "logps/rejected": -341.7208251953125, |
| "loss": 0.1418, |
| "rewards/accuracies": 0.9549999833106995, |
| "rewards/chosen": -0.9791030883789062, |
| "rewards/margins": 3.893622636795044, |
| "rewards/rejected": -4.872725963592529, |
| "step": 3300 |
| }, |
| { |
| "epoch": 1.312273484180625, |
| "grad_norm": 11.505748748779297, |
| "learning_rate": 6.15160921934118e-07, |
| "logits/chosen": -1.6867657899856567, |
| "logits/rejected": -1.5593910217285156, |
| "logps/chosen": -365.0918884277344, |
| "logps/rejected": -318.346923828125, |
| "loss": 0.1254, |
| "rewards/accuracies": 0.9549999833106995, |
| "rewards/chosen": -0.9864305257797241, |
| "rewards/margins": 3.980782985687256, |
| "rewards/rejected": -4.9672136306762695, |
| "step": 3350 |
| }, |
| { |
| "epoch": 1.3318640415319816, |
| "grad_norm": 48.97773742675781, |
| "learning_rate": 6.049582506069384e-07, |
| "logits/chosen": -1.657313585281372, |
| "logits/rejected": -1.4994758367538452, |
| "logps/chosen": -359.9202880859375, |
| "logps/rejected": -347.42828369140625, |
| "loss": 0.1331, |
| "rewards/accuracies": 0.949999988079071, |
| "rewards/chosen": -1.2405058145523071, |
| "rewards/margins": 3.8140506744384766, |
| "rewards/rejected": -5.054556369781494, |
| "step": 3400 |
| }, |
| { |
| "epoch": 1.3318640415319816, |
| "eval_logits/chosen": -1.716418981552124, |
| "eval_logits/rejected": -1.6371155977249146, |
| "eval_logps/chosen": -372.63067626953125, |
| "eval_logps/rejected": -327.14337158203125, |
| "eval_loss": 0.5230005979537964, |
| "eval_rewards/accuracies": 0.7793505191802979, |
| "eval_rewards/chosen": -2.0317211151123047, |
| "eval_rewards/margins": 2.1393544673919678, |
| "eval_rewards/rejected": -4.17107629776001, |
| "eval_runtime": 182.2547, |
| "eval_samples_per_second": 13.179, |
| "eval_steps_per_second": 6.59, |
| "step": 3400 |
| }, |
| { |
| "epoch": 1.3514545988833382, |
| "grad_norm": 107.73078155517578, |
| "learning_rate": 5.947096514111293e-07, |
| "logits/chosen": -1.6291078329086304, |
| "logits/rejected": -1.5198391675949097, |
| "logps/chosen": -383.50958251953125, |
| "logps/rejected": -363.100341796875, |
| "loss": 0.1277, |
| "rewards/accuracies": 0.949999988079071, |
| "rewards/chosen": -1.0857621431350708, |
| "rewards/margins": 4.09164571762085, |
| "rewards/rejected": -5.177408218383789, |
| "step": 3450 |
| }, |
| { |
| "epoch": 1.3710451562346948, |
| "grad_norm": 15.163590431213379, |
| "learning_rate": 5.844196089519004e-07, |
| "logits/chosen": -1.6963342428207397, |
| "logits/rejected": -1.5234904289245605, |
| "logps/chosen": -367.5895690917969, |
| "logps/rejected": -334.65203857421875, |
| "loss": 0.1338, |
| "rewards/accuracies": 0.9524999856948853, |
| "rewards/chosen": -1.2845149040222168, |
| "rewards/margins": 4.080843448638916, |
| "rewards/rejected": -5.365358829498291, |
| "step": 3500 |
| }, |
| { |
| "epoch": 1.3906357135860516, |
| "grad_norm": 183.16290283203125, |
| "learning_rate": 5.740926259692986e-07, |
| "logits/chosen": -1.6925588846206665, |
| "logits/rejected": -1.4511497020721436, |
| "logps/chosen": -342.0213317871094, |
| "logps/rejected": -328.423583984375, |
| "loss": 0.1684, |
| "rewards/accuracies": 0.9150000214576721, |
| "rewards/chosen": -1.2314543724060059, |
| "rewards/margins": 4.002154350280762, |
| "rewards/rejected": -5.233608245849609, |
| "step": 3550 |
| }, |
| { |
| "epoch": 1.4102262709374083, |
| "grad_norm": 130.83932495117188, |
| "learning_rate": 5.637332213678889e-07, |
| "logits/chosen": -1.747999668121338, |
| "logits/rejected": -1.552089810371399, |
| "logps/chosen": -357.76312255859375, |
| "logps/rejected": -321.6129150390625, |
| "loss": 0.1506, |
| "rewards/accuracies": 0.9574999809265137, |
| "rewards/chosen": -1.2750509977340698, |
| "rewards/margins": 4.070872783660889, |
| "rewards/rejected": -5.345923900604248, |
| "step": 3600 |
| }, |
| { |
| "epoch": 1.4102262709374083, |
| "eval_logits/chosen": -1.6983522176742554, |
| "eval_logits/rejected": -1.6196881532669067, |
| "eval_logps/chosen": -374.8834228515625, |
| "eval_logps/rejected": -329.9042053222656, |
| "eval_loss": 0.5274552702903748, |
| "eval_rewards/accuracies": 0.7756036520004272, |
| "eval_rewards/chosen": -2.256995916366577, |
| "eval_rewards/margins": 2.1901612281799316, |
| "eval_rewards/rejected": -4.4471564292907715, |
| "eval_runtime": 182.0012, |
| "eval_samples_per_second": 13.198, |
| "eval_steps_per_second": 6.599, |
| "step": 3600 |
| }, |
| { |
| "epoch": 1.4298168282887649, |
| "grad_norm": 62.9024658203125, |
| "learning_rate": 5.53345928239361e-07, |
| "logits/chosen": -1.6344012022018433, |
| "logits/rejected": -1.3985601663589478, |
| "logps/chosen": -371.9190979003906, |
| "logps/rejected": -362.4126892089844, |
| "loss": 0.1324, |
| "rewards/accuracies": 0.9549999833106995, |
| "rewards/chosen": -1.3249839544296265, |
| "rewards/margins": 4.2258501052856445, |
| "rewards/rejected": -5.550833702087402, |
| "step": 3650 |
| }, |
| { |
| "epoch": 1.4494073856401215, |
| "grad_norm": 63.766395568847656, |
| "learning_rate": 5.429352918789295e-07, |
| "logits/chosen": -1.6781551837921143, |
| "logits/rejected": -1.6595940589904785, |
| "logps/chosen": -359.553955078125, |
| "logps/rejected": -333.27191162109375, |
| "loss": 0.1255, |
| "rewards/accuracies": 0.9574999809265137, |
| "rewards/chosen": -1.3423984050750732, |
| "rewards/margins": 4.154109477996826, |
| "rewards/rejected": -5.496507167816162, |
| "step": 3700 |
| }, |
| { |
| "epoch": 1.4689979429914781, |
| "grad_norm": 24.34297752380371, |
| "learning_rate": 5.325058677963933e-07, |
| "logits/chosen": -1.7628166675567627, |
| "logits/rejected": -1.5162057876586914, |
| "logps/chosen": -379.5002746582031, |
| "logps/rejected": -361.4959411621094, |
| "loss": 0.1278, |
| "rewards/accuracies": 0.9424999952316284, |
| "rewards/chosen": -1.2505874633789062, |
| "rewards/margins": 4.234759330749512, |
| "rewards/rejected": -5.485346794128418, |
| "step": 3750 |
| }, |
| { |
| "epoch": 1.4885885003428347, |
| "grad_norm": 89.56745910644531, |
| "learning_rate": 5.220622197227254e-07, |
| "logits/chosen": -1.6744155883789062, |
| "logits/rejected": -1.6061152219772339, |
| "logps/chosen": -395.9300842285156, |
| "logps/rejected": -356.6335144042969, |
| "loss": 0.1219, |
| "rewards/accuracies": 0.9474999904632568, |
| "rewards/chosen": -1.2993007898330688, |
| "rewards/margins": 4.109463214874268, |
| "rewards/rejected": -5.408763885498047, |
| "step": 3800 |
| }, |
| { |
| "epoch": 1.4885885003428347, |
| "eval_logits/chosen": -1.755277156829834, |
| "eval_logits/rejected": -1.6812604665756226, |
| "eval_logps/chosen": -375.90478515625, |
| "eval_logps/rejected": -331.6509704589844, |
| "eval_loss": 0.5298057794570923, |
| "eval_rewards/accuracies": 0.7793505191802979, |
| "eval_rewards/chosen": -2.3591370582580566, |
| "eval_rewards/margins": 2.262698173522949, |
| "eval_rewards/rejected": -4.621835708618164, |
| "eval_runtime": 181.8002, |
| "eval_samples_per_second": 13.212, |
| "eval_steps_per_second": 6.606, |
| "step": 3800 |
| }, |
| { |
| "epoch": 1.5081790576941914, |
| "grad_norm": 57.11545944213867, |
| "learning_rate": 5.116089176130647e-07, |
| "logits/chosen": -1.8247219324111938, |
| "logits/rejected": -1.707238793373108, |
| "logps/chosen": -365.5677795410156, |
| "logps/rejected": -348.8277587890625, |
| "loss": 0.1248, |
| "rewards/accuracies": 0.9649999737739563, |
| "rewards/chosen": -1.4670463800430298, |
| "rewards/margins": 4.245115756988525, |
| "rewards/rejected": -5.712162017822266, |
| "step": 3850 |
| }, |
| { |
| "epoch": 1.527769615045548, |
| "grad_norm": 134.8478546142578, |
| "learning_rate": 5.011505356469849e-07, |
| "logits/chosen": -1.800379991531372, |
| "logits/rejected": -1.6222047805786133, |
| "logps/chosen": -396.044677734375, |
| "logps/rejected": -343.588623046875, |
| "loss": 0.1579, |
| "rewards/accuracies": 0.9350000023841858, |
| "rewards/chosen": -1.327430248260498, |
| "rewards/margins": 4.10135555267334, |
| "rewards/rejected": -5.428785800933838, |
| "step": 3900 |
| }, |
| { |
| "epoch": 1.5473601723969046, |
| "grad_norm": 47.98903274536133, |
| "learning_rate": 4.906916502269153e-07, |
| "logits/chosen": -1.8289942741394043, |
| "logits/rejected": -1.7046104669570923, |
| "logps/chosen": -360.898681640625, |
| "logps/rejected": -347.4443359375, |
| "loss": 0.1689, |
| "rewards/accuracies": 0.9424999952316284, |
| "rewards/chosen": -1.565004825592041, |
| "rewards/margins": 4.049954414367676, |
| "rewards/rejected": -5.6149582862854, |
| "step": 3950 |
| }, |
| { |
| "epoch": 1.5669507297482612, |
| "grad_norm": 36.070865631103516, |
| "learning_rate": 4.802368379755871e-07, |
| "logits/chosen": -1.901904582977295, |
| "logits/rejected": -1.7009533643722534, |
| "logps/chosen": -390.9181213378906, |
| "logps/rejected": -367.340087890625, |
| "loss": 0.1173, |
| "rewards/accuracies": 0.9649999737739563, |
| "rewards/chosen": -1.3107359409332275, |
| "rewards/margins": 4.3210906982421875, |
| "rewards/rejected": -5.631826877593994, |
| "step": 4000 |
| }, |
| { |
| "epoch": 1.5669507297482612, |
| "eval_logits/chosen": -1.83998441696167, |
| "eval_logits/rejected": -1.773179292678833, |
| "eval_logps/chosen": -375.8030700683594, |
| "eval_logps/rejected": -331.4768371582031, |
| "eval_loss": 0.5410642027854919, |
| "eval_rewards/accuracies": 0.779766857624054, |
| "eval_rewards/chosen": -2.348963737487793, |
| "eval_rewards/margins": 2.255459785461426, |
| "eval_rewards/rejected": -4.604423522949219, |
| "eval_runtime": 182.1217, |
| "eval_samples_per_second": 13.189, |
| "eval_steps_per_second": 6.594, |
| "step": 4000 |
| }, |
| { |
| "epoch": 1.5865412870996178, |
| "grad_norm": 79.75633239746094, |
| "learning_rate": 4.697906737333856e-07, |
| "logits/chosen": -1.890414834022522, |
| "logits/rejected": -1.7079460620880127, |
| "logps/chosen": -363.3149108886719, |
| "logps/rejected": -336.81964111328125, |
| "loss": 0.1136, |
| "rewards/accuracies": 0.9599999785423279, |
| "rewards/chosen": -1.327339768409729, |
| "rewards/margins": 4.425736904144287, |
| "rewards/rejected": -5.753076553344727, |
| "step": 4050 |
| }, |
| { |
| "epoch": 1.6061318444509747, |
| "grad_norm": 14.279524803161621, |
| "learning_rate": 4.593577285564805e-07, |
| "logits/chosen": -1.9139132499694824, |
| "logits/rejected": -1.7257100343704224, |
| "logps/chosen": -373.6604309082031, |
| "logps/rejected": -360.3436279296875, |
| "loss": 0.1217, |
| "rewards/accuracies": 0.9524999856948853, |
| "rewards/chosen": -1.4224177598953247, |
| "rewards/margins": 4.337459564208984, |
| "rewards/rejected": -5.759876728057861, |
| "step": 4100 |
| }, |
| { |
| "epoch": 1.6257224018023313, |
| "grad_norm": 18.292810440063477, |
| "learning_rate": 4.489425677166128e-07, |
| "logits/chosen": -1.8857932090759277, |
| "logits/rejected": -1.838065505027771, |
| "logps/chosen": -374.2807922363281, |
| "logps/rejected": -350.9076232910156, |
| "loss": 0.1116, |
| "rewards/accuracies": 0.9599999785423279, |
| "rewards/chosen": -1.2832770347595215, |
| "rewards/margins": 4.261753082275391, |
| "rewards/rejected": -5.545030117034912, |
| "step": 4150 |
| }, |
| { |
| "epoch": 1.645312959153688, |
| "grad_norm": 91.74571990966797, |
| "learning_rate": 4.385497487034136e-07, |
| "logits/chosen": -1.8253960609436035, |
| "logits/rejected": -1.80033540725708, |
| "logps/chosen": -360.9422607421875, |
| "logps/rejected": -348.8204040527344, |
| "loss": 0.1668, |
| "rewards/accuracies": 0.9449999928474426, |
| "rewards/chosen": -1.5318783521652222, |
| "rewards/margins": 4.232841968536377, |
| "rewards/rejected": -5.764720439910889, |
| "step": 4200 |
| }, |
| { |
| "epoch": 1.645312959153688, |
| "eval_logits/chosen": -1.825873851776123, |
| "eval_logits/rejected": -1.7672475576400757, |
| "eval_logps/chosen": -375.93621826171875, |
| "eval_logps/rejected": -332.0399475097656, |
| "eval_loss": 0.5490909814834595, |
| "eval_rewards/accuracies": 0.7781015634536743, |
| "eval_rewards/chosen": -2.362279176712036, |
| "eval_rewards/margins": 2.2984535694122314, |
| "eval_rewards/rejected": -4.660732746124268, |
| "eval_runtime": 181.3571, |
| "eval_samples_per_second": 13.245, |
| "eval_steps_per_second": 6.622, |
| "step": 4200 |
| }, |
| { |
| "epoch": 1.6649035165050445, |
| "grad_norm": 182.9907989501953, |
| "learning_rate": 4.281838192301266e-07, |
| "logits/chosen": -1.8667634725570679, |
| "logits/rejected": -1.7775866985321045, |
| "logps/chosen": -350.6466064453125, |
| "logps/rejected": -349.0397644042969, |
| "loss": 0.133, |
| "rewards/accuracies": 0.9424999952316284, |
| "rewards/chosen": -1.5763401985168457, |
| "rewards/margins": 4.427700996398926, |
| "rewards/rejected": -6.004040718078613, |
| "step": 4250 |
| }, |
| { |
| "epoch": 1.6844940738564014, |
| "grad_norm": 90.31336975097656, |
| "learning_rate": 4.1784931524360996e-07, |
| "logits/chosen": -1.8399044275283813, |
| "logits/rejected": -1.7566088438034058, |
| "logps/chosen": -357.8070983886719, |
| "logps/rejected": -337.547119140625, |
| "loss": 0.1338, |
| "rewards/accuracies": 0.9524999856948853, |
| "rewards/chosen": -1.5925673246383667, |
| "rewards/margins": 4.302795886993408, |
| "rewards/rejected": -5.895363807678223, |
| "step": 4300 |
| }, |
| { |
| "epoch": 1.704084631207758, |
| "grad_norm": 60.01850128173828, |
| "learning_rate": 4.075507589394862e-07, |
| "logits/chosen": -1.8707722425460815, |
| "logits/rejected": -1.7477329969406128, |
| "logps/chosen": -367.755126953125, |
| "logps/rejected": -334.7363586425781, |
| "loss": 0.1432, |
| "rewards/accuracies": 0.9399999976158142, |
| "rewards/chosen": -1.4867920875549316, |
| "rewards/margins": 4.018145561218262, |
| "rewards/rejected": -5.504937171936035, |
| "step": 4350 |
| }, |
| { |
| "epoch": 1.7236751885591146, |
| "grad_norm": 7.138427257537842, |
| "learning_rate": 3.972926567833095e-07, |
| "logits/chosen": -1.7986953258514404, |
| "logits/rejected": -1.716946005821228, |
| "logps/chosen": -377.9021911621094, |
| "logps/rejected": -367.26483154296875, |
| "loss": 0.1184, |
| "rewards/accuracies": 0.949999988079071, |
| "rewards/chosen": -1.5053943395614624, |
| "rewards/margins": 4.413090705871582, |
| "rewards/rejected": -5.918485164642334, |
| "step": 4400 |
| }, |
| { |
| "epoch": 1.7236751885591146, |
| "eval_logits/chosen": -1.7733187675476074, |
| "eval_logits/rejected": -1.7134273052215576, |
| "eval_logps/chosen": -377.1297607421875, |
| "eval_logps/rejected": -333.46954345703125, |
| "eval_loss": 0.5378764271736145, |
| "eval_rewards/accuracies": 0.7814321517944336, |
| "eval_rewards/chosen": -2.4816348552703857, |
| "eval_rewards/margins": 2.3220555782318115, |
| "eval_rewards/rejected": -4.803690433502197, |
| "eval_runtime": 182.8904, |
| "eval_samples_per_second": 13.134, |
| "eval_steps_per_second": 6.567, |
| "step": 4400 |
| }, |
| { |
| "epoch": 1.7432657459104712, |
| "grad_norm": 18.62919807434082, |
| "learning_rate": 3.8707949753861716e-07, |
| "logits/chosen": -1.8632616996765137, |
| "logits/rejected": -1.7376039028167725, |
| "logps/chosen": -342.5689697265625, |
| "logps/rejected": -335.2397766113281, |
| "loss": 0.1411, |
| "rewards/accuracies": 0.9574999809265137, |
| "rewards/chosen": -1.3847324848175049, |
| "rewards/margins": 4.106419563293457, |
| "rewards/rejected": -5.491152286529541, |
| "step": 4450 |
| }, |
| { |
| "epoch": 1.7628563032618279, |
| "grad_norm": 23.290203094482422, |
| "learning_rate": 3.76915750302725e-07, |
| "logits/chosen": -1.7581045627593994, |
| "logits/rejected": -1.6029043197631836, |
| "logps/chosen": -373.6942138671875, |
| "logps/rejected": -338.0625, |
| "loss": 0.1141, |
| "rewards/accuracies": 0.9649999737739563, |
| "rewards/chosen": -1.4231642484664917, |
| "rewards/margins": 4.145116806030273, |
| "rewards/rejected": -5.5682806968688965, |
| "step": 4500 |
| }, |
| { |
| "epoch": 1.7824468606131845, |
| "grad_norm": 113.75978088378906, |
| "learning_rate": 3.668058625511305e-07, |
| "logits/chosen": -1.7640550136566162, |
| "logits/rejected": -1.6161187887191772, |
| "logps/chosen": -369.67816162109375, |
| "logps/rejected": -363.20465087890625, |
| "loss": 0.1117, |
| "rewards/accuracies": 0.9599999785423279, |
| "rewards/chosen": -1.3803114891052246, |
| "rewards/margins": 4.385107040405273, |
| "rewards/rejected": -5.765418529510498, |
| "step": 4550 |
| }, |
| { |
| "epoch": 1.802037417964541, |
| "grad_norm": 7.093522548675537, |
| "learning_rate": 3.567542581913762e-07, |
| "logits/chosen": -1.8656272888183594, |
| "logits/rejected": -1.8022925853729248, |
| "logps/chosen": -395.6029968261719, |
| "logps/rejected": -359.8951110839844, |
| "loss": 0.1466, |
| "rewards/accuracies": 0.9524999856948853, |
| "rewards/chosen": -1.3073208332061768, |
| "rewards/margins": 4.359652519226074, |
| "rewards/rejected": -5.666974067687988, |
| "step": 4600 |
| }, |
| { |
| "epoch": 1.802037417964541, |
| "eval_logits/chosen": -1.739977240562439, |
| "eval_logits/rejected": -1.677179217338562, |
| "eval_logps/chosen": -376.30279541015625, |
| "eval_logps/rejected": -332.7083740234375, |
| "eval_loss": 0.5302212834358215, |
| "eval_rewards/accuracies": 0.7843464016914368, |
| "eval_rewards/chosen": -2.3989357948303223, |
| "eval_rewards/margins": 2.3286385536193848, |
| "eval_rewards/rejected": -4.727574348449707, |
| "eval_runtime": 181.7513, |
| "eval_samples_per_second": 13.216, |
| "eval_steps_per_second": 6.608, |
| "step": 4600 |
| }, |
| { |
| "epoch": 1.8216279753158977, |
| "grad_norm": 168.2427978515625, |
| "learning_rate": 3.467653356272264e-07, |
| "logits/chosen": -1.7600982189178467, |
| "logits/rejected": -1.75065016746521, |
| "logps/chosen": -368.1263732910156, |
| "logps/rejected": -329.2120056152344, |
| "loss": 0.1393, |
| "rewards/accuracies": 0.9549999833106995, |
| "rewards/chosen": -1.325424313545227, |
| "rewards/margins": 4.229908466339111, |
| "rewards/rejected": -5.555332660675049, |
| "step": 4650 |
| }, |
| { |
| "epoch": 1.8412185326672543, |
| "grad_norm": 56.983543395996094, |
| "learning_rate": 3.368434658340035e-07, |
| "logits/chosen": -1.6969270706176758, |
| "logits/rejected": -1.5315319299697876, |
| "logps/chosen": -393.845947265625, |
| "logps/rejected": -366.9394836425781, |
| "loss": 0.1201, |
| "rewards/accuracies": 0.9524999856948853, |
| "rewards/chosen": -1.3427032232284546, |
| "rewards/margins": 4.412410259246826, |
| "rewards/rejected": -5.755113124847412, |
| "step": 4700 |
| }, |
| { |
| "epoch": 1.860809090018611, |
| "grad_norm": 46.993534088134766, |
| "learning_rate": 3.269929904459265e-07, |
| "logits/chosen": -1.6990869045257568, |
| "logits/rejected": -1.6050549745559692, |
| "logps/chosen": -379.8760070800781, |
| "logps/rejected": -358.8374938964844, |
| "loss": 0.1347, |
| "rewards/accuracies": 0.9524999856948853, |
| "rewards/chosen": -1.5243346691131592, |
| "rewards/margins": 4.38024377822876, |
| "rewards/rejected": -5.904578685760498, |
| "step": 4750 |
| }, |
| { |
| "epoch": 1.8803996473699676, |
| "grad_norm": 22.50926399230957, |
| "learning_rate": 3.1721821985628946e-07, |
| "logits/chosen": -1.7507199048995972, |
| "logits/rejected": -1.587468147277832, |
| "logps/chosen": -342.8749084472656, |
| "logps/rejected": -330.3369140625, |
| "loss": 0.154, |
| "rewards/accuracies": 0.9574999809265137, |
| "rewards/chosen": -1.4566929340362549, |
| "rewards/margins": 4.056729793548584, |
| "rewards/rejected": -5.513422012329102, |
| "step": 4800 |
| }, |
| { |
| "epoch": 1.8803996473699676, |
| "eval_logits/chosen": -1.8217155933380127, |
| "eval_logits/rejected": -1.761517882347107, |
| "eval_logps/chosen": -376.3056335449219, |
| "eval_logps/rejected": -332.8399658203125, |
| "eval_loss": 0.528506338596344, |
| "eval_rewards/accuracies": 0.7843464016914368, |
| "eval_rewards/chosen": -2.3992199897766113, |
| "eval_rewards/margins": 2.3415098190307617, |
| "eval_rewards/rejected": -4.740729808807373, |
| "eval_runtime": 182.1926, |
| "eval_samples_per_second": 13.184, |
| "eval_steps_per_second": 6.592, |
| "step": 4800 |
| }, |
| { |
| "epoch": 1.8999902047213242, |
| "grad_norm": 41.3251953125, |
| "learning_rate": 3.075234313313095e-07, |
| "logits/chosen": -1.822758674621582, |
| "logits/rejected": -1.6095324754714966, |
| "logps/chosen": -368.2705078125, |
| "logps/rejected": -363.7851867675781, |
| "loss": 0.1114, |
| "rewards/accuracies": 0.9599999785423279, |
| "rewards/chosen": -1.1875016689300537, |
| "rewards/margins": 4.392082691192627, |
| "rewards/rejected": -5.579584121704102, |
| "step": 4850 |
| }, |
| { |
| "epoch": 1.9195807620726808, |
| "grad_norm": 33.448936462402344, |
| "learning_rate": 2.9791286713847106e-07, |
| "logits/chosen": -1.874271035194397, |
| "logits/rejected": -1.7505204677581787, |
| "logps/chosen": -380.5478820800781, |
| "logps/rejected": -355.3659362792969, |
| "loss": 0.1139, |
| "rewards/accuracies": 0.9549999833106995, |
| "rewards/chosen": -1.4798643589019775, |
| "rewards/margins": 4.445069789886475, |
| "rewards/rejected": -5.924932956695557, |
| "step": 4900 |
| }, |
| { |
| "epoch": 1.9391713194240376, |
| "grad_norm": 53.873172760009766, |
| "learning_rate": 2.883907326901849e-07, |
| "logits/chosen": -1.9296669960021973, |
| "logits/rejected": -1.7788227796554565, |
| "logps/chosen": -355.0895690917969, |
| "logps/rejected": -336.3532409667969, |
| "loss": 0.136, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": -1.5472720861434937, |
| "rewards/margins": 4.275333881378174, |
| "rewards/rejected": -5.822606086730957, |
| "step": 4950 |
| }, |
| { |
| "epoch": 1.9587618767753943, |
| "grad_norm": 66.68231964111328, |
| "learning_rate": 2.7896119470357394e-07, |
| "logits/chosen": -1.8380649089813232, |
| "logits/rejected": -1.626741647720337, |
| "logps/chosen": -367.21307373046875, |
| "logps/rejected": -350.3079833984375, |
| "loss": 0.1328, |
| "rewards/accuracies": 0.9574999809265137, |
| "rewards/chosen": -1.383242130279541, |
| "rewards/margins": 4.56887674331665, |
| "rewards/rejected": -5.95211935043335, |
| "step": 5000 |
| }, |
| { |
| "epoch": 1.9587618767753943, |
| "eval_logits/chosen": -1.7905840873718262, |
| "eval_logits/rejected": -1.730137586593628, |
| "eval_logps/chosen": -377.50054931640625, |
| "eval_logps/rejected": -334.88006591796875, |
| "eval_loss": 0.5351826548576355, |
| "eval_rewards/accuracies": 0.7847626805305481, |
| "eval_rewards/chosen": -2.5187087059020996, |
| "eval_rewards/margins": 2.4260356426239014, |
| "eval_rewards/rejected": -4.94474458694458, |
| "eval_runtime": 182.3765, |
| "eval_samples_per_second": 13.171, |
| "eval_steps_per_second": 6.585, |
| "step": 5000 |
| }, |
| { |
| "epoch": 1.9783524341267509, |
| "grad_norm": 9.479029655456543, |
| "learning_rate": 2.696283793771921e-07, |
| "logits/chosen": -1.8894351720809937, |
| "logits/rejected": -1.6450564861297607, |
| "logps/chosen": -372.2572326660156, |
| "logps/rejected": -358.4980773925781, |
| "loss": 0.1521, |
| "rewards/accuracies": 0.9325000047683716, |
| "rewards/chosen": -1.5351108312606812, |
| "rewards/margins": 4.4576897621154785, |
| "rewards/rejected": -5.992800235748291, |
| "step": 5050 |
| }, |
| { |
| "epoch": 1.9979429914781075, |
| "grad_norm": 84.90283203125, |
| "learning_rate": 2.603963705854731e-07, |
| "logits/chosen": -1.905731201171875, |
| "logits/rejected": -1.6645513772964478, |
| "logps/chosen": -361.50701904296875, |
| "logps/rejected": -353.17059326171875, |
| "loss": 0.1461, |
| "rewards/accuracies": 0.9424999952316284, |
| "rewards/chosen": -1.4440213441848755, |
| "rewards/margins": 4.272027492523193, |
| "rewards/rejected": -5.716048717498779, |
| "step": 5100 |
| } |
| ], |
| "logging_steps": 50, |
| "max_steps": 7659, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "save_steps": 300, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 0.0, |
| "train_batch_size": 2, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
