| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.8183306055646481, | |
| "eval_steps": 500, | |
| "global_step": 500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.016366612111292964, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.6304347826086957e-06, | |
| "logits/chosen": -1.2515310049057007, | |
| "logits/rejected": -0.8573959469795227, | |
| "logps/chosen": -455.1468811035156, | |
| "logps/rejected": -387.0008239746094, | |
| "loss": 0.6923, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.0002114105736836791, | |
| "rewards/margins": 0.00180077797267586, | |
| "rewards/rejected": -0.0015893673989921808, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03273322422258593, | |
| "grad_norm": 10.154256820678711, | |
| "learning_rate": 3.2608695652173914e-06, | |
| "logits/chosen": -1.145971417427063, | |
| "logits/rejected": -1.082183837890625, | |
| "logps/chosen": -446.249755859375, | |
| "logps/rejected": -426.87237548828125, | |
| "loss": 0.6942, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": -0.0005732345744036138, | |
| "rewards/margins": -0.002082281280308962, | |
| "rewards/rejected": 0.0015090465312823653, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.049099836333878884, | |
| "grad_norm": 8.393162727355957, | |
| "learning_rate": 5.4347826086956525e-06, | |
| "logits/chosen": -0.9614235162734985, | |
| "logits/rejected": -0.8527517318725586, | |
| "logps/chosen": -460.367431640625, | |
| "logps/rejected": -433.61285400390625, | |
| "loss": 0.6934, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": 0.0062791816890239716, | |
| "rewards/margins": -0.00045429245801642537, | |
| "rewards/rejected": 0.006733474787324667, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.06546644844517185, | |
| "grad_norm": 9.398829460144043, | |
| "learning_rate": 8.15217391304348e-06, | |
| "logits/chosen": -1.171631097793579, | |
| "logits/rejected": -1.0616101026535034, | |
| "logps/chosen": -421.583984375, | |
| "logps/rejected": -424.787109375, | |
| "loss": 0.6902, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.020799484103918076, | |
| "rewards/margins": 0.006221642717719078, | |
| "rewards/rejected": 0.014577841386198997, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.08183306055646482, | |
| "grad_norm": 9.43005657196045, | |
| "learning_rate": 1.0869565217391305e-05, | |
| "logits/chosen": -1.2450587749481201, | |
| "logits/rejected": -1.1417832374572754, | |
| "logps/chosen": -476.83538818359375, | |
| "logps/rejected": -444.6556091308594, | |
| "loss": 0.6883, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.042164258658885956, | |
| "rewards/margins": 0.010725794360041618, | |
| "rewards/rejected": 0.03143846243619919, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.09819967266775777, | |
| "grad_norm": 8.825221061706543, | |
| "learning_rate": 1.331521739130435e-05, | |
| "logits/chosen": -1.3223823308944702, | |
| "logits/rejected": NaN, | |
| "logps/chosen": -383.4811706542969, | |
| "logps/rejected": -454.43780517578125, | |
| "loss": 0.6965, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.03668512776494026, | |
| "rewards/margins": -0.003995636478066444, | |
| "rewards/rejected": 0.04068076238036156, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.11456628477905073, | |
| "grad_norm": 9.5093994140625, | |
| "learning_rate": 1.6032608695652173e-05, | |
| "logits/chosen": -1.2538468837738037, | |
| "logits/rejected": -1.2771422863006592, | |
| "logps/chosen": -468.79400634765625, | |
| "logps/rejected": -402.565673828125, | |
| "loss": 0.6864, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.04952714219689369, | |
| "rewards/margins": 0.01824185810983181, | |
| "rewards/rejected": 0.03128528222441673, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.1309328968903437, | |
| "grad_norm": 7.945922374725342, | |
| "learning_rate": 1.8750000000000002e-05, | |
| "logits/chosen": -1.386731743812561, | |
| "logits/rejected": -0.8966825604438782, | |
| "logps/chosen": -446.54736328125, | |
| "logps/rejected": -430.45025634765625, | |
| "loss": 0.6719, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.06767664849758148, | |
| "rewards/margins": 0.048766594380140305, | |
| "rewards/rejected": 0.01891004480421543, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.14729950900163666, | |
| "grad_norm": 8.496461868286133, | |
| "learning_rate": 2.1467391304347828e-05, | |
| "logits/chosen": -1.3730494976043701, | |
| "logits/rejected": -1.1826181411743164, | |
| "logps/chosen": -413.87835693359375, | |
| "logps/rejected": -385.1138916015625, | |
| "loss": 0.69, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.07122664153575897, | |
| "rewards/margins": 0.014941570349037647, | |
| "rewards/rejected": 0.0562850646674633, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.16366612111292964, | |
| "grad_norm": 10.744088172912598, | |
| "learning_rate": 2.4184782608695653e-05, | |
| "logits/chosen": -1.2254825830459595, | |
| "logits/rejected": -1.1136434078216553, | |
| "logps/chosen": -495.34295654296875, | |
| "logps/rejected": -435.0003967285156, | |
| "loss": 0.6618, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.1682208627462387, | |
| "rewards/margins": 0.08198796212673187, | |
| "rewards/rejected": 0.08623288571834564, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.18003273322422259, | |
| "grad_norm": 11.016289710998535, | |
| "learning_rate": 2.6902173913043476e-05, | |
| "logits/chosen": -1.3138881921768188, | |
| "logits/rejected": -1.0289314985275269, | |
| "logps/chosen": -476.26153564453125, | |
| "logps/rejected": -426.837646484375, | |
| "loss": 0.6355, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.21744687855243683, | |
| "rewards/margins": 0.15108919143676758, | |
| "rewards/rejected": 0.06635768711566925, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.19639934533551553, | |
| "grad_norm": 11.619248390197754, | |
| "learning_rate": 2.9619565217391305e-05, | |
| "logits/chosen": -1.2751747369766235, | |
| "logits/rejected": -1.1858090162277222, | |
| "logps/chosen": -444.4823303222656, | |
| "logps/rejected": -404.58123779296875, | |
| "loss": 0.6716, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.2203199863433838, | |
| "rewards/margins": 0.06618332862854004, | |
| "rewards/rejected": 0.15413665771484375, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2127659574468085, | |
| "grad_norm": 8.900009155273438, | |
| "learning_rate": 3.233695652173913e-05, | |
| "logits/chosen": -1.2365785837173462, | |
| "logits/rejected": -1.2031922340393066, | |
| "logps/chosen": -477.8089294433594, | |
| "logps/rejected": -456.80120849609375, | |
| "loss": 0.6794, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.20705166459083557, | |
| "rewards/margins": 0.060810189694166183, | |
| "rewards/rejected": 0.1462414562702179, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.22913256955810146, | |
| "grad_norm": 8.928816795349121, | |
| "learning_rate": 3.505434782608696e-05, | |
| "logits/chosen": -1.2421447038650513, | |
| "logits/rejected": -1.136252522468567, | |
| "logps/chosen": -547.4909057617188, | |
| "logps/rejected": -460.38275146484375, | |
| "loss": 0.6047, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.21343636512756348, | |
| "rewards/margins": 0.24689535796642303, | |
| "rewards/rejected": -0.033459004014730453, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.24549918166939444, | |
| "grad_norm": 11.977150917053223, | |
| "learning_rate": 3.777173913043478e-05, | |
| "logits/chosen": -1.338332176208496, | |
| "logits/rejected": -1.1015567779541016, | |
| "logps/chosen": -432.2693786621094, | |
| "logps/rejected": -423.0946350097656, | |
| "loss": 0.6946, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": 0.15119442343711853, | |
| "rewards/margins": 0.07870660722255707, | |
| "rewards/rejected": 0.07248781621456146, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.2618657937806874, | |
| "grad_norm": 13.040871620178223, | |
| "learning_rate": 4.0489130434782615e-05, | |
| "logits/chosen": -1.0111101865768433, | |
| "logits/rejected": NaN, | |
| "logps/chosen": -472.6917419433594, | |
| "logps/rejected": -426.6333923339844, | |
| "loss": 0.6598, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.24765905737876892, | |
| "rewards/margins": 0.13647916913032532, | |
| "rewards/rejected": 0.11117993295192719, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.27823240589198034, | |
| "grad_norm": 8.028325080871582, | |
| "learning_rate": 4.3206521739130434e-05, | |
| "logits/chosen": -0.9152072668075562, | |
| "logits/rejected": -0.8340185284614563, | |
| "logps/chosen": -438.38555908203125, | |
| "logps/rejected": -392.93426513671875, | |
| "loss": 0.6272, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.3285876512527466, | |
| "rewards/margins": 0.20840699970722198, | |
| "rewards/rejected": 0.1201806515455246, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.2945990180032733, | |
| "grad_norm": 9.436712265014648, | |
| "learning_rate": 4.592391304347826e-05, | |
| "logits/chosen": -1.0703661441802979, | |
| "logits/rejected": -1.0127884149551392, | |
| "logps/chosen": -449.83905029296875, | |
| "logps/rejected": -437.92822265625, | |
| "loss": 0.6684, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": 0.34186336398124695, | |
| "rewards/margins": 0.14258861541748047, | |
| "rewards/rejected": 0.19927474856376648, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3109656301145663, | |
| "grad_norm": 8.914007186889648, | |
| "learning_rate": 4.8641304347826086e-05, | |
| "logits/chosen": -0.8762575387954712, | |
| "logits/rejected": -0.5631248354911804, | |
| "logps/chosen": -473.806396484375, | |
| "logps/rejected": -443.6448669433594, | |
| "loss": 0.6243, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.3772050440311432, | |
| "rewards/margins": 0.24287982285022736, | |
| "rewards/rejected": 0.13432522118091583, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.32733224222585927, | |
| "grad_norm": 14.76502513885498, | |
| "learning_rate": 4.98483929654336e-05, | |
| "logits/chosen": NaN, | |
| "logits/rejected": -0.9821416139602661, | |
| "logps/chosen": -447.65386962890625, | |
| "logps/rejected": -400.1768493652344, | |
| "loss": 0.6805, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.2871859073638916, | |
| "rewards/margins": 0.17125040292739868, | |
| "rewards/rejected": 0.11593550443649292, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.3436988543371522, | |
| "grad_norm": 11.80213451385498, | |
| "learning_rate": 4.954517889630079e-05, | |
| "logits/chosen": -1.2535641193389893, | |
| "logits/rejected": -1.0489479303359985, | |
| "logps/chosen": -457.010986328125, | |
| "logps/rejected": -432.8558044433594, | |
| "loss": 0.6769, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.24471895396709442, | |
| "rewards/margins": 0.16205154359340668, | |
| "rewards/rejected": 0.08266742527484894, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.36006546644844517, | |
| "grad_norm": 7.534430980682373, | |
| "learning_rate": 4.924196482716799e-05, | |
| "logits/chosen": -1.5050650835037231, | |
| "logits/rejected": -1.5133517980575562, | |
| "logps/chosen": -445.3924865722656, | |
| "logps/rejected": -420.99114990234375, | |
| "loss": 0.7003, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.18017885088920593, | |
| "rewards/margins": 0.1387586146593094, | |
| "rewards/rejected": 0.04142025485634804, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.37643207855973815, | |
| "grad_norm": 7.207632541656494, | |
| "learning_rate": 4.893875075803517e-05, | |
| "logits/chosen": -1.3496973514556885, | |
| "logits/rejected": -1.2030962705612183, | |
| "logps/chosen": -436.5039978027344, | |
| "logps/rejected": -396.91534423828125, | |
| "loss": 0.6192, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.19479091465473175, | |
| "rewards/margins": 0.353401243686676, | |
| "rewards/rejected": -0.15861034393310547, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.39279869067103107, | |
| "grad_norm": 9.134820938110352, | |
| "learning_rate": 4.8665858095815645e-05, | |
| "logits/chosen": -1.2978178262710571, | |
| "logits/rejected": -1.1864020824432373, | |
| "logps/chosen": -486.7569885253906, | |
| "logps/rejected": -457.94586181640625, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": 0.29108190536499023, | |
| "rewards/margins": 0.15347807109355927, | |
| "rewards/rejected": 0.13760384917259216, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.40916530278232405, | |
| "grad_norm": 10.825005531311035, | |
| "learning_rate": 4.836264402668284e-05, | |
| "logits/chosen": -1.3271738290786743, | |
| "logits/rejected": -1.030890703201294, | |
| "logps/chosen": -421.89349365234375, | |
| "logps/rejected": -400.1674499511719, | |
| "loss": 0.6536, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.42532652616500854, | |
| "rewards/margins": 0.34351587295532227, | |
| "rewards/rejected": 0.08181065320968628, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.425531914893617, | |
| "grad_norm": 9.450189590454102, | |
| "learning_rate": 4.8059429957550034e-05, | |
| "logits/chosen": -1.0212231874465942, | |
| "logits/rejected": -1.0058345794677734, | |
| "logps/chosen": -422.1316833496094, | |
| "logps/rejected": -390.0425109863281, | |
| "loss": 0.6384, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.490867555141449, | |
| "rewards/margins": 0.35965341329574585, | |
| "rewards/rejected": 0.13121409714221954, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.44189852700491, | |
| "grad_norm": 8.231124877929688, | |
| "learning_rate": 4.7756215888417225e-05, | |
| "logits/chosen": -0.9735445976257324, | |
| "logits/rejected": -0.937743067741394, | |
| "logps/chosen": -552.4168701171875, | |
| "logps/rejected": -501.0535583496094, | |
| "loss": 0.6694, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.4536367952823639, | |
| "rewards/margins": 0.19916944205760956, | |
| "rewards/rejected": 0.2544674277305603, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.4582651391162029, | |
| "grad_norm": 9.207930564880371, | |
| "learning_rate": 4.7453001819284416e-05, | |
| "logits/chosen": -0.9774928092956543, | |
| "logits/rejected": -0.8485246896743774, | |
| "logps/chosen": -444.54827880859375, | |
| "logps/rejected": -419.5541076660156, | |
| "loss": 0.6607, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.3690536618232727, | |
| "rewards/margins": 0.22144463658332825, | |
| "rewards/rejected": 0.14760902523994446, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.4746317512274959, | |
| "grad_norm": 6.876308441162109, | |
| "learning_rate": 4.714978775015161e-05, | |
| "logits/chosen": -1.2095867395401, | |
| "logits/rejected": -0.8773800730705261, | |
| "logps/chosen": -437.869384765625, | |
| "logps/rejected": -419.9366149902344, | |
| "loss": 0.5867, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.5094749331474304, | |
| "rewards/margins": 0.5250319242477417, | |
| "rewards/rejected": -0.015556998550891876, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.4909983633387889, | |
| "grad_norm": 12.918950080871582, | |
| "learning_rate": 4.68465736810188e-05, | |
| "logits/chosen": -0.9076827764511108, | |
| "logits/rejected": -0.7181829214096069, | |
| "logps/chosen": -469.4864196777344, | |
| "logps/rejected": -420.38885498046875, | |
| "loss": 0.6833, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.5444287657737732, | |
| "rewards/margins": 0.2743019461631775, | |
| "rewards/rejected": 0.2701267600059509, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5073649754500819, | |
| "grad_norm": 4.9246110916137695, | |
| "learning_rate": 4.6543359611885996e-05, | |
| "logits/chosen": -1.214247226715088, | |
| "logits/rejected": -0.9910456538200378, | |
| "logps/chosen": -385.4250183105469, | |
| "logps/rejected": -390.26043701171875, | |
| "loss": 0.6419, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.4199371337890625, | |
| "rewards/margins": 0.34462255239486694, | |
| "rewards/rejected": 0.07531457394361496, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.5237315875613748, | |
| "grad_norm": 8.829249382019043, | |
| "learning_rate": 4.624014554275319e-05, | |
| "logits/chosen": -1.0396336317062378, | |
| "logits/rejected": -0.9439867734909058, | |
| "logps/chosen": -452.1678161621094, | |
| "logps/rejected": -379.70745849609375, | |
| "loss": 0.6396, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.472791850566864, | |
| "rewards/margins": 0.23509594798088074, | |
| "rewards/rejected": 0.23769588768482208, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.5400981996726678, | |
| "grad_norm": 13.249287605285645, | |
| "learning_rate": 4.593693147362038e-05, | |
| "logits/chosen": -0.927478015422821, | |
| "logits/rejected": -0.7090943455696106, | |
| "logps/chosen": -488.009765625, | |
| "logps/rejected": -473.47833251953125, | |
| "loss": 0.6833, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.4911763668060303, | |
| "rewards/margins": 0.25351986289024353, | |
| "rewards/rejected": 0.23765654861927032, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.5564648117839607, | |
| "grad_norm": 9.580172538757324, | |
| "learning_rate": 4.566403881140085e-05, | |
| "logits/chosen": -0.9790302515029907, | |
| "logits/rejected": -0.8282285928726196, | |
| "logps/chosen": -481.6971740722656, | |
| "logps/rejected": -441.3619689941406, | |
| "loss": 0.6922, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.4195849299430847, | |
| "rewards/margins": 0.210377499461174, | |
| "rewards/rejected": 0.20920750498771667, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.5728314238952537, | |
| "grad_norm": 6.872241973876953, | |
| "learning_rate": 4.536082474226804e-05, | |
| "logits/chosen": -0.9942235946655273, | |
| "logits/rejected": -0.824662983417511, | |
| "logps/chosen": -437.9869079589844, | |
| "logps/rejected": -417.3473205566406, | |
| "loss": 0.6071, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.49881643056869507, | |
| "rewards/margins": 0.41778793931007385, | |
| "rewards/rejected": 0.08102855831384659, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.5891980360065466, | |
| "grad_norm": 7.748345375061035, | |
| "learning_rate": 4.5057610673135234e-05, | |
| "logits/chosen": -1.0704625844955444, | |
| "logits/rejected": -0.7879603505134583, | |
| "logps/chosen": -428.53021240234375, | |
| "logps/rejected": -435.9026794433594, | |
| "loss": 0.6084, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.4484317898750305, | |
| "rewards/margins": 0.4125775396823883, | |
| "rewards/rejected": 0.035854239016771317, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.6055646481178396, | |
| "grad_norm": 7.428864479064941, | |
| "learning_rate": 4.475439660400243e-05, | |
| "logits/chosen": -1.098026990890503, | |
| "logits/rejected": -0.9232645034790039, | |
| "logps/chosen": -441.7071228027344, | |
| "logps/rejected": -402.95770263671875, | |
| "loss": 0.6901, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.37514036893844604, | |
| "rewards/margins": 0.2525361478328705, | |
| "rewards/rejected": 0.12260414659976959, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.6219312602291326, | |
| "grad_norm": 9.591557502746582, | |
| "learning_rate": 4.445118253486962e-05, | |
| "logits/chosen": -1.2409312725067139, | |
| "logits/rejected": -0.8442489504814148, | |
| "logps/chosen": -466.13311767578125, | |
| "logps/rejected": -420.9102478027344, | |
| "loss": 0.7684, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": 0.2286566197872162, | |
| "rewards/margins": 0.11457403004169464, | |
| "rewards/rejected": 0.11408257484436035, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.6382978723404256, | |
| "grad_norm": 7.025293350219727, | |
| "learning_rate": 4.414796846573681e-05, | |
| "logits/chosen": -1.1844158172607422, | |
| "logits/rejected": -1.1069843769073486, | |
| "logps/chosen": -428.5450134277344, | |
| "logps/rejected": -371.85614013671875, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.2637328803539276, | |
| "rewards/margins": 0.10889965295791626, | |
| "rewards/rejected": 0.15483322739601135, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.6546644844517185, | |
| "grad_norm": 19.2535343170166, | |
| "learning_rate": 4.3844754396604006e-05, | |
| "logits/chosen": -1.1141570806503296, | |
| "logits/rejected": -0.8798454999923706, | |
| "logps/chosen": -493.08197021484375, | |
| "logps/rejected": -465.0943298339844, | |
| "loss": 0.6545, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.4174181818962097, | |
| "rewards/margins": 0.25259822607040405, | |
| "rewards/rejected": 0.16481998562812805, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.6710310965630114, | |
| "grad_norm": 8.808697700500488, | |
| "learning_rate": 4.35415403274712e-05, | |
| "logits/chosen": -0.9878439903259277, | |
| "logits/rejected": -0.7299278974533081, | |
| "logps/chosen": -523.9139404296875, | |
| "logps/rejected": -482.0597229003906, | |
| "loss": 0.552, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.4945390820503235, | |
| "rewards/margins": 0.6175243854522705, | |
| "rewards/rejected": -0.1229853481054306, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.6873977086743044, | |
| "grad_norm": 9.99456787109375, | |
| "learning_rate": 4.323832625833839e-05, | |
| "logits/chosen": -1.0902565717697144, | |
| "logits/rejected": -0.9221500158309937, | |
| "logps/chosen": -443.0076599121094, | |
| "logps/rejected": -375.8013610839844, | |
| "loss": 0.5684, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.46177297830581665, | |
| "rewards/margins": 0.509793221950531, | |
| "rewards/rejected": -0.048020295798778534, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.7037643207855974, | |
| "grad_norm": 5.767345428466797, | |
| "learning_rate": 4.293511218920558e-05, | |
| "logits/chosen": -1.0450439453125, | |
| "logits/rejected": -0.7902520298957825, | |
| "logps/chosen": -458.2330017089844, | |
| "logps/rejected": -382.6787109375, | |
| "loss": 0.6126, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.5793684124946594, | |
| "rewards/margins": 0.35584205389022827, | |
| "rewards/rejected": 0.22352632880210876, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.7201309328968903, | |
| "grad_norm": 11.51710033416748, | |
| "learning_rate": 4.263189812007277e-05, | |
| "logits/chosen": -0.8710979223251343, | |
| "logits/rejected": -0.783151388168335, | |
| "logps/chosen": -451.32122802734375, | |
| "logps/rejected": -416.0005798339844, | |
| "loss": 0.6751, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.5284687280654907, | |
| "rewards/margins": 0.2338237315416336, | |
| "rewards/rejected": 0.2946449816226959, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.7364975450081833, | |
| "grad_norm": 8.408599853515625, | |
| "learning_rate": 4.232868405093997e-05, | |
| "logits/chosen": -0.9611997604370117, | |
| "logits/rejected": -0.8005222082138062, | |
| "logps/chosen": -390.06817626953125, | |
| "logps/rejected": -376.57183837890625, | |
| "loss": 0.723, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.35626164078712463, | |
| "rewards/margins": 0.07882154732942581, | |
| "rewards/rejected": 0.27744007110595703, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.7528641571194763, | |
| "grad_norm": 7.085683822631836, | |
| "learning_rate": 4.202546998180716e-05, | |
| "logits/chosen": -1.181065559387207, | |
| "logits/rejected": -0.9188621640205383, | |
| "logps/chosen": -446.1695861816406, | |
| "logps/rejected": -461.0537109375, | |
| "loss": 0.6376, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.6099992990493774, | |
| "rewards/margins": 0.40477481484413147, | |
| "rewards/rejected": 0.20522448420524597, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 7.778029918670654, | |
| "learning_rate": 4.172225591267435e-05, | |
| "logits/chosen": -1.1612136363983154, | |
| "logits/rejected": -0.7909379601478577, | |
| "logps/chosen": -415.356201171875, | |
| "logps/rejected": -405.0163879394531, | |
| "loss": 0.639, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.5554541349411011, | |
| "rewards/margins": 0.49387890100479126, | |
| "rewards/rejected": 0.06157524511218071, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.7855973813420621, | |
| "grad_norm": 9.144447326660156, | |
| "learning_rate": 4.141904184354154e-05, | |
| "logits/chosen": -1.1227318048477173, | |
| "logits/rejected": -0.8980533480644226, | |
| "logps/chosen": -470.1067810058594, | |
| "logps/rejected": -385.7353210449219, | |
| "loss": 0.6301, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.5011923909187317, | |
| "rewards/margins": 0.36913150548934937, | |
| "rewards/rejected": 0.1320609152317047, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.8019639934533551, | |
| "grad_norm": 9.528294563293457, | |
| "learning_rate": 4.111582777440873e-05, | |
| "logits/chosen": -1.4198505878448486, | |
| "logits/rejected": -1.1528406143188477, | |
| "logps/chosen": -430.28668212890625, | |
| "logps/rejected": -385.2845764160156, | |
| "loss": 0.634, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.6369161605834961, | |
| "rewards/margins": 0.4485607147216797, | |
| "rewards/rejected": 0.18835540115833282, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.8183306055646481, | |
| "grad_norm": 6.964690208435059, | |
| "learning_rate": 4.0812613705275924e-05, | |
| "logits/chosen": -1.1329933404922485, | |
| "logits/rejected": -1.066542625427246, | |
| "logps/chosen": -423.4967346191406, | |
| "logps/rejected": -376.92803955078125, | |
| "loss": 0.5904, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.6795145273208618, | |
| "rewards/margins": 0.3617789149284363, | |
| "rewards/rejected": 0.3177356719970703, | |
| "step": 500 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1833, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "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": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |