| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 62, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.016260162601626018, | |
| "grad_norm": 18.177886962890625, | |
| "learning_rate": 2e-05, | |
| "logits/chosen": -0.3472236394882202, | |
| "logits/rejected": -0.13716036081314087, | |
| "logps/chosen": -780.8181762695312, | |
| "logps/rejected": -909.20263671875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.032520325203252036, | |
| "grad_norm": 23.274246215820312, | |
| "learning_rate": 4e-05, | |
| "logits/chosen": -0.2127760350704193, | |
| "logits/rejected": -0.08323362469673157, | |
| "logps/chosen": -583.0169067382812, | |
| "logps/rejected": -715.5615234375, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.04878048780487805, | |
| "grad_norm": 20.149507522583008, | |
| "learning_rate": 6e-05, | |
| "logits/chosen": -0.18167662620544434, | |
| "logits/rejected": -0.04478086531162262, | |
| "logps/chosen": -941.0387573242188, | |
| "logps/rejected": -825.662841796875, | |
| "loss": 0.6976, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.025517277419567108, | |
| "rewards/margins": 0.022285467013716698, | |
| "rewards/rejected": 0.0032318076118826866, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.06504065040650407, | |
| "grad_norm": 16.67251205444336, | |
| "learning_rate": 8e-05, | |
| "logits/chosen": 0.6866837739944458, | |
| "logits/rejected": 0.971089243888855, | |
| "logps/chosen": -999.306640625, | |
| "logps/rejected": -386.5375671386719, | |
| "loss": 0.563, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.2688583433628082, | |
| "rewards/margins": 0.3312031030654907, | |
| "rewards/rejected": -0.062344741076231, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.08130081300813008, | |
| "grad_norm": 15.646084785461426, | |
| "learning_rate": 0.0001, | |
| "logits/chosen": 0.5107800364494324, | |
| "logits/rejected": 0.5942208766937256, | |
| "logps/chosen": -1051.1270751953125, | |
| "logps/rejected": -745.8003540039062, | |
| "loss": 0.647, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.3622299134731293, | |
| "rewards/margins": 0.34313660860061646, | |
| "rewards/rejected": 0.01909332349896431, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0975609756097561, | |
| "grad_norm": 38.70280456542969, | |
| "learning_rate": 0.00012, | |
| "logits/chosen": -0.31406939029693604, | |
| "logits/rejected": -0.24293695390224457, | |
| "logps/chosen": -845.9321899414062, | |
| "logps/rejected": -932.499755859375, | |
| "loss": 0.5175, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.5435073971748352, | |
| "rewards/margins": 0.47774890065193176, | |
| "rewards/rejected": 0.06575851887464523, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.11382113821138211, | |
| "grad_norm": 23.665071487426758, | |
| "learning_rate": 0.00014, | |
| "logits/chosen": -0.2646118402481079, | |
| "logits/rejected": -0.11520399153232574, | |
| "logps/chosen": -866.503173828125, | |
| "logps/rejected": -975.55126953125, | |
| "loss": 0.5487, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.6112838387489319, | |
| "rewards/margins": 0.4790405333042145, | |
| "rewards/rejected": 0.1322433352470398, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.13008130081300814, | |
| "grad_norm": 15.794047355651855, | |
| "learning_rate": 0.00016, | |
| "logits/chosen": -0.8256000876426697, | |
| "logits/rejected": -0.8912097811698914, | |
| "logps/chosen": -523.3858032226562, | |
| "logps/rejected": -1084.9468994140625, | |
| "loss": 0.4442, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.5804435610771179, | |
| "rewards/margins": 0.24081651866436005, | |
| "rewards/rejected": 0.33962705731391907, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.14634146341463414, | |
| "grad_norm": 13.538564682006836, | |
| "learning_rate": 0.00018, | |
| "logits/chosen": -0.11683523654937744, | |
| "logits/rejected": -0.0632472038269043, | |
| "logps/chosen": -652.114501953125, | |
| "logps/rejected": -551.6069946289062, | |
| "loss": 0.1564, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.6716469526290894, | |
| "rewards/margins": 2.151698350906372, | |
| "rewards/rejected": -0.4800514578819275, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.16260162601626016, | |
| "grad_norm": 3.9652626514434814, | |
| "learning_rate": 0.0002, | |
| "logits/chosen": 0.4062778949737549, | |
| "logits/rejected": 0.5438919067382812, | |
| "logps/chosen": -771.1934814453125, | |
| "logps/rejected": -616.55908203125, | |
| "loss": 0.0792, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 3.8721909523010254, | |
| "rewards/margins": 5.208758354187012, | |
| "rewards/rejected": -1.3365669250488281, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.17886178861788618, | |
| "grad_norm": 0.18261243402957916, | |
| "learning_rate": 0.0001999911398855782, | |
| "logits/chosen": -0.7774271965026855, | |
| "logits/rejected": -0.8629493117332458, | |
| "logps/chosen": -601.1015014648438, | |
| "logps/rejected": -1039.275146484375, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.0800025463104248, | |
| "rewards/margins": 6.853862762451172, | |
| "rewards/rejected": -5.773860454559326, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.1951219512195122, | |
| "grad_norm": 0.1421748697757721, | |
| "learning_rate": 0.00019996456111234527, | |
| "logits/chosen": 0.7899215817451477, | |
| "logits/rejected": 1.119359016418457, | |
| "logps/chosen": -1416.412353515625, | |
| "logps/rejected": -827.2066650390625, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 3.7505874633789062, | |
| "rewards/margins": 15.09115982055664, | |
| "rewards/rejected": -11.340574264526367, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.21138211382113822, | |
| "grad_norm": 3.4406840801239014, | |
| "learning_rate": 0.00019992026839012067, | |
| "logits/chosen": -0.8033453226089478, | |
| "logits/rejected": -0.877557098865509, | |
| "logps/chosen": -514.6026611328125, | |
| "logps/rejected": -1206.25537109375, | |
| "loss": 0.0102, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.7983558177947998, | |
| "rewards/margins": 23.49526596069336, | |
| "rewards/rejected": -21.696908950805664, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.22764227642276422, | |
| "grad_norm": 0.19398577511310577, | |
| "learning_rate": 0.0001998582695676762, | |
| "logits/chosen": 0.9254277944564819, | |
| "logits/rejected": 1.1634798049926758, | |
| "logps/chosen": -1028.993408203125, | |
| "logps/rejected": -955.4432983398438, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.5009795427322388, | |
| "rewards/margins": 17.867931365966797, | |
| "rewards/rejected": -18.368911743164062, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.24390243902439024, | |
| "grad_norm": 0.00010074722376884893, | |
| "learning_rate": 0.000199778575631345, | |
| "logits/chosen": 0.3904605507850647, | |
| "logits/rejected": 0.3719422519207001, | |
| "logps/chosen": -884.9620361328125, | |
| "logps/rejected": -1075.615966796875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.482113838195801, | |
| "rewards/margins": 21.95424461364746, | |
| "rewards/rejected": -24.436357498168945, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.2601626016260163, | |
| "grad_norm": 3.7136353057576343e-05, | |
| "learning_rate": 0.000199681200703075, | |
| "logits/chosen": 0.2578551769256592, | |
| "logits/rejected": 0.5335351824760437, | |
| "logps/chosen": -1073.548828125, | |
| "logps/rejected": -992.4033813476562, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.9434356689453125, | |
| "rewards/margins": 20.854663848876953, | |
| "rewards/rejected": -23.798099517822266, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.2764227642276423, | |
| "grad_norm": 8.596338147981442e-07, | |
| "learning_rate": 0.00019956616203792635, | |
| "logits/chosen": 0.5267460346221924, | |
| "logits/rejected": 0.4893237352371216, | |
| "logps/chosen": -987.3567504882812, | |
| "logps/rejected": -1127.171875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.0684036016464233, | |
| "rewards/margins": 32.558319091796875, | |
| "rewards/rejected": -33.62671661376953, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.2926829268292683, | |
| "grad_norm": 0.004051027819514275, | |
| "learning_rate": 0.00019943348002101371, | |
| "logits/chosen": 1.0484071969985962, | |
| "logits/rejected": 1.1081664562225342, | |
| "logps/chosen": -1105.1634521484375, | |
| "logps/rejected": -898.9759521484375, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.1622314453125, | |
| "rewards/margins": 23.434669494628906, | |
| "rewards/rejected": -26.596900939941406, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.3089430894308943, | |
| "grad_norm": 0.003306547412648797, | |
| "learning_rate": 0.00019928317816389417, | |
| "logits/chosen": 0.5566614866256714, | |
| "logits/rejected": 0.6963181495666504, | |
| "logps/chosen": -932.650390625, | |
| "logps/rejected": -1061.4989013671875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.36033821105957, | |
| "rewards/margins": 30.25779914855957, | |
| "rewards/rejected": -34.61813735961914, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.3252032520325203, | |
| "grad_norm": 1.3893560968369911e-08, | |
| "learning_rate": 0.00019911528310040074, | |
| "logits/chosen": 1.239579200744629, | |
| "logits/rejected": 1.046311855316162, | |
| "logps/chosen": -1079.0159912109375, | |
| "logps/rejected": -1033.2017822265625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.044548749923706, | |
| "rewards/margins": 41.88936233520508, | |
| "rewards/rejected": -40.844810485839844, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.34146341463414637, | |
| "grad_norm": 4.666223851756968e-09, | |
| "learning_rate": 0.00019892982458192288, | |
| "logits/chosen": 0.2726232409477234, | |
| "logits/rejected": 0.14665402472019196, | |
| "logps/chosen": -978.7222900390625, | |
| "logps/rejected": -1133.2047119140625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 11.054238319396973, | |
| "rewards/margins": 54.86410140991211, | |
| "rewards/rejected": -43.80986404418945, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.35772357723577236, | |
| "grad_norm": 4.876813477494579e-07, | |
| "learning_rate": 0.00019872683547213446, | |
| "logits/chosen": -0.16925190389156342, | |
| "logits/rejected": -0.19759103655815125, | |
| "logps/chosen": -965.187255859375, | |
| "logps/rejected": -1239.143798828125, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -14.977485656738281, | |
| "rewards/margins": 29.40732765197754, | |
| "rewards/rejected": -44.38481140136719, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.37398373983739835, | |
| "grad_norm": 37.638973236083984, | |
| "learning_rate": 0.00019850635174117033, | |
| "logits/chosen": 0.437714159488678, | |
| "logits/rejected": 0.4761970639228821, | |
| "logps/chosen": -1137.6966552734375, | |
| "logps/rejected": -1166.5640869140625, | |
| "loss": 0.4393, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -11.159793853759766, | |
| "rewards/margins": 32.14189529418945, | |
| "rewards/rejected": -43.301692962646484, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.3902439024390244, | |
| "grad_norm": 1.8173747229344173e-11, | |
| "learning_rate": 0.00019826841245925212, | |
| "logits/chosen": -0.7153763175010681, | |
| "logits/rejected": -0.6940470933914185, | |
| "logps/chosen": -938.263916015625, | |
| "logps/rejected": -1608.4205322265625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -24.817350387573242, | |
| "rewards/margins": 34.095001220703125, | |
| "rewards/rejected": -58.912349700927734, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.4065040650406504, | |
| "grad_norm": 83.79772186279297, | |
| "learning_rate": 0.0001980130597897651, | |
| "logits/chosen": 1.1592888832092285, | |
| "logits/rejected": 1.1738824844360352, | |
| "logps/chosen": -948.4622802734375, | |
| "logps/rejected": -865.396728515625, | |
| "loss": 0.3825, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.343675374984741, | |
| "rewards/margins": 26.49417495727539, | |
| "rewards/rejected": -29.837852478027344, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.42276422764227645, | |
| "grad_norm": 2.6143006834900007e-06, | |
| "learning_rate": 0.00019774033898178667, | |
| "logits/chosen": 0.5444796085357666, | |
| "logits/rejected": 0.47586876153945923, | |
| "logps/chosen": -932.6605834960938, | |
| "logps/rejected": -1091.639892578125, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.2753777503967285, | |
| "rewards/margins": 34.133514404296875, | |
| "rewards/rejected": -38.40888977050781, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.43902439024390244, | |
| "grad_norm": 0.0003061926399823278, | |
| "learning_rate": 0.00019745029836206813, | |
| "logits/chosen": -0.6794779896736145, | |
| "logits/rejected": -0.8602011203765869, | |
| "logps/chosen": -894.3270263671875, | |
| "logps/rejected": -1067.5921630859375, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -13.433198928833008, | |
| "rewards/margins": 17.333955764770508, | |
| "rewards/rejected": -30.767154693603516, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.45528455284552843, | |
| "grad_norm": 3.805017101399244e-08, | |
| "learning_rate": 0.00019714298932647098, | |
| "logits/chosen": 0.4980026185512543, | |
| "logits/rejected": 0.6999194025993347, | |
| "logps/chosen": -911.8473510742188, | |
| "logps/rejected": -1126.07421875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.5412168502807617, | |
| "rewards/margins": 29.520708084106445, | |
| "rewards/rejected": -30.06192398071289, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.4715447154471545, | |
| "grad_norm": 5.17633900187775e-08, | |
| "learning_rate": 0.00019681846633085967, | |
| "logits/chosen": -0.5973828434944153, | |
| "logits/rejected": -0.8376109600067139, | |
| "logps/chosen": -711.66259765625, | |
| "logps/rejected": -1186.1884765625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.467390537261963, | |
| "rewards/margins": 25.050704956054688, | |
| "rewards/rejected": -27.518096923828125, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.4878048780487805, | |
| "grad_norm": 0.00011633769463514909, | |
| "learning_rate": 0.0001964767868814516, | |
| "logits/chosen": 1.3797093629837036, | |
| "logits/rejected": 1.5397391319274902, | |
| "logps/chosen": -877.42333984375, | |
| "logps/rejected": -1003.4732666015625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 4.624107360839844, | |
| "rewards/margins": 29.784557342529297, | |
| "rewards/rejected": -25.160449981689453, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.5040650406504065, | |
| "grad_norm": 6.257723228486611e-09, | |
| "learning_rate": 0.00019611801152462715, | |
| "logits/chosen": 1.2731826305389404, | |
| "logits/rejected": 1.6379995346069336, | |
| "logps/chosen": -1053.573486328125, | |
| "logps/rejected": -1010.915283203125, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 11.018058776855469, | |
| "rewards/margins": 32.15219497680664, | |
| "rewards/rejected": -21.13413429260254, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.5203252032520326, | |
| "grad_norm": 0.00035472630406729877, | |
| "learning_rate": 0.00019574220383620055, | |
| "logits/chosen": 0.6649560928344727, | |
| "logits/rejected": 0.983564019203186, | |
| "logps/chosen": -872.1873168945312, | |
| "logps/rejected": -965.9480590820312, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 5.504961967468262, | |
| "rewards/margins": 23.669071197509766, | |
| "rewards/rejected": -18.164108276367188, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.5365853658536586, | |
| "grad_norm": 3.0934195820009336e-05, | |
| "learning_rate": 0.00019534943041015423, | |
| "logits/chosen": 0.49574941396713257, | |
| "logits/rejected": 0.5190873742103577, | |
| "logps/chosen": -708.9269409179688, | |
| "logps/rejected": -842.974365234375, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.209194660186768, | |
| "rewards/margins": 20.690357208251953, | |
| "rewards/rejected": -13.48116397857666, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.5528455284552846, | |
| "grad_norm": 0.0006856573163531721, | |
| "learning_rate": 0.00019493976084683813, | |
| "logits/chosen": 0.992796778678894, | |
| "logits/rejected": 1.1291236877441406, | |
| "logps/chosen": -673.6188354492188, | |
| "logps/rejected": -723.4482421875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 5.3715057373046875, | |
| "rewards/margins": 19.963485717773438, | |
| "rewards/rejected": -14.591980934143066, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.5691056910569106, | |
| "grad_norm": 5.983891969663091e-05, | |
| "learning_rate": 0.00019451326774063636, | |
| "logits/chosen": 0.7630600929260254, | |
| "logits/rejected": 0.910960853099823, | |
| "logps/chosen": -993.23828125, | |
| "logps/rejected": -1011.3184204101562, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.109509468078613, | |
| "rewards/margins": 24.603878021240234, | |
| "rewards/rejected": -17.494367599487305, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.5853658536585366, | |
| "grad_norm": 1.9749455532291904e-05, | |
| "learning_rate": 0.00019407002666710336, | |
| "logits/chosen": 1.8401339054107666, | |
| "logits/rejected": 1.9955703020095825, | |
| "logps/chosen": -1152.950927734375, | |
| "logps/rejected": -827.0269775390625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 15.768245697021484, | |
| "rewards/margins": 38.1776123046875, | |
| "rewards/rejected": -22.40936851501465, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.6016260162601627, | |
| "grad_norm": 0.0017285533249378204, | |
| "learning_rate": 0.00019361011616957164, | |
| "logits/chosen": 2.153351306915283, | |
| "logits/rejected": 2.235447883605957, | |
| "logps/chosen": -1090.1943359375, | |
| "logps/rejected": -682.7992553710938, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 11.726329803466797, | |
| "rewards/margins": 24.018630981445312, | |
| "rewards/rejected": -12.292303085327148, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.6178861788617886, | |
| "grad_norm": 0.00919501855969429, | |
| "learning_rate": 0.00019313361774523385, | |
| "logits/chosen": 0.47314736247062683, | |
| "logits/rejected": 0.557833731174469, | |
| "logps/chosen": -691.4217529296875, | |
| "logps/rejected": -673.1847534179688, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.087795257568359, | |
| "rewards/margins": 12.628225326538086, | |
| "rewards/rejected": -6.540430068969727, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.6341463414634146, | |
| "grad_norm": 0.002680833451449871, | |
| "learning_rate": 0.00019264061583070127, | |
| "logits/chosen": 0.20066705346107483, | |
| "logits/rejected": 0.2085224837064743, | |
| "logps/chosen": -693.7376098632812, | |
| "logps/rejected": -982.19091796875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.779763221740723, | |
| "rewards/margins": 22.904094696044922, | |
| "rewards/rejected": -15.124334335327148, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.6504065040650406, | |
| "grad_norm": 8.798202907200903e-05, | |
| "learning_rate": 0.00019213119778704128, | |
| "logits/chosen": 1.3898746967315674, | |
| "logits/rejected": 1.5520107746124268, | |
| "logps/chosen": -1247.770263671875, | |
| "logps/rejected": -916.4830322265625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 15.276836395263672, | |
| "rewards/margins": 34.69191360473633, | |
| "rewards/rejected": -19.415077209472656, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.6666666666666666, | |
| "grad_norm": 0.0009758697124198079, | |
| "learning_rate": 0.00019160545388429708, | |
| "logits/chosen": 2.345059633255005, | |
| "logits/rejected": 2.5746054649353027, | |
| "logps/chosen": -1102.5548095703125, | |
| "logps/rejected": -722.4332885742188, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 13.800348281860352, | |
| "rewards/margins": 32.747169494628906, | |
| "rewards/rejected": -18.946823120117188, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.6829268292682927, | |
| "grad_norm": 0.0016077810432761908, | |
| "learning_rate": 0.00019106347728549135, | |
| "logits/chosen": 0.9104095697402954, | |
| "logits/rejected": 0.9921329021453857, | |
| "logps/chosen": -753.8040771484375, | |
| "logps/rejected": -886.5813598632812, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 11.367500305175781, | |
| "rewards/margins": 27.856563568115234, | |
| "rewards/rejected": -16.489063262939453, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.6991869918699187, | |
| "grad_norm": 0.0004074655589647591, | |
| "learning_rate": 0.0001905053640301176, | |
| "logits/chosen": 0.5256392955780029, | |
| "logits/rejected": 0.4733426570892334, | |
| "logps/chosen": -715.4669189453125, | |
| "logps/rejected": -565.0441284179688, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.25009822845459, | |
| "rewards/margins": 21.391075134277344, | |
| "rewards/rejected": -15.14097785949707, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.7154471544715447, | |
| "grad_norm": 0.013145952485501766, | |
| "learning_rate": 0.00018993121301712193, | |
| "logits/chosen": 0.9358551502227783, | |
| "logits/rejected": 0.8306156992912292, | |
| "logps/chosen": -867.1063232421875, | |
| "logps/rejected": -973.7214965820312, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.3925018310546875, | |
| "rewards/margins": 21.35105323791504, | |
| "rewards/rejected": -13.958552360534668, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.7317073170731707, | |
| "grad_norm": 8.829876605886966e-05, | |
| "learning_rate": 0.00018934112598737777, | |
| "logits/chosen": 2.2844998836517334, | |
| "logits/rejected": 2.831254482269287, | |
| "logps/chosen": -1142.8726806640625, | |
| "logps/rejected": -776.1110229492188, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 17.17538833618164, | |
| "rewards/margins": 33.72625732421875, | |
| "rewards/rejected": -16.550867080688477, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.7479674796747967, | |
| "grad_norm": 0.02624354511499405, | |
| "learning_rate": 0.00018873520750565718, | |
| "logits/chosen": 0.1806122362613678, | |
| "logits/rejected": 0.31054702401161194, | |
| "logps/chosen": -692.7060546875, | |
| "logps/rejected": -1032.708740234375, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.434965133666992, | |
| "rewards/margins": 16.74932098388672, | |
| "rewards/rejected": -10.314356803894043, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.7642276422764228, | |
| "grad_norm": 4.268178963684477e-05, | |
| "learning_rate": 0.00018811356494210165, | |
| "logits/chosen": 1.1679103374481201, | |
| "logits/rejected": 1.0418663024902344, | |
| "logps/chosen": -720.220703125, | |
| "logps/rejected": -911.58837890625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.991888523101807, | |
| "rewards/margins": 21.064565658569336, | |
| "rewards/rejected": -13.072675704956055, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.7804878048780488, | |
| "grad_norm": 0.0009461237932555377, | |
| "learning_rate": 0.00018747630845319612, | |
| "logits/chosen": 0.13339552283287048, | |
| "logits/rejected": 0.3655449151992798, | |
| "logps/chosen": -420.11431884765625, | |
| "logps/rejected": -786.4783325195312, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 11.16606330871582, | |
| "rewards/margins": 30.41803741455078, | |
| "rewards/rejected": -19.251976013183594, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.7967479674796748, | |
| "grad_norm": 0.0033115639816969633, | |
| "learning_rate": 0.00018682355096224872, | |
| "logits/chosen": 0.4472777247428894, | |
| "logits/rejected": 0.3390260934829712, | |
| "logps/chosen": -536.7960205078125, | |
| "logps/rejected": -901.3749389648438, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 10.887458801269531, | |
| "rewards/margins": 27.701595306396484, | |
| "rewards/rejected": -16.814136505126953, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.8130081300813008, | |
| "grad_norm": 0.01153454091399908, | |
| "learning_rate": 0.0001861554081393806, | |
| "logits/chosen": 0.6489148139953613, | |
| "logits/rejected": 0.689254105091095, | |
| "logps/chosen": -738.5593872070312, | |
| "logps/rejected": -755.362060546875, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 10.205413818359375, | |
| "rewards/margins": 16.344358444213867, | |
| "rewards/rejected": -6.138944625854492, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.8292682926829268, | |
| "grad_norm": 0.001985176932066679, | |
| "learning_rate": 0.00018547199838102904, | |
| "logits/chosen": 0.144524484872818, | |
| "logits/rejected": 0.26266002655029297, | |
| "logps/chosen": -893.19482421875, | |
| "logps/rejected": -1031.27294921875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 9.087849617004395, | |
| "rewards/margins": 23.393884658813477, | |
| "rewards/rejected": -14.306035041809082, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.8455284552845529, | |
| "grad_norm": 0.00042794409091584384, | |
| "learning_rate": 0.0001847734427889671, | |
| "logits/chosen": 0.5121033191680908, | |
| "logits/rejected": 1.0676312446594238, | |
| "logps/chosen": -987.8340454101562, | |
| "logps/rejected": -830.7366943359375, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 11.409669876098633, | |
| "rewards/margins": 19.569660186767578, | |
| "rewards/rejected": -8.159988403320312, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.8617886178861789, | |
| "grad_norm": 0.0011688657104969025, | |
| "learning_rate": 0.00018405986514884434, | |
| "logits/chosen": 1.793473243713379, | |
| "logits/rejected": 1.9872632026672363, | |
| "logps/chosen": -926.424560546875, | |
| "logps/rejected": -618.4228515625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 11.011417388916016, | |
| "rewards/margins": 22.01776123046875, | |
| "rewards/rejected": -11.006343841552734, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.8780487804878049, | |
| "grad_norm": 0.005157554987818003, | |
| "learning_rate": 0.0001833313919082515, | |
| "logits/chosen": -0.02910199761390686, | |
| "logits/rejected": 0.14243453741073608, | |
| "logps/chosen": -725.36376953125, | |
| "logps/rejected": -997.5311279296875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 5.557222366333008, | |
| "rewards/margins": 15.359309196472168, | |
| "rewards/rejected": -9.802087783813477, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.8943089430894309, | |
| "grad_norm": 0.005044507794082165, | |
| "learning_rate": 0.00018258815215431396, | |
| "logits/chosen": 0.17898443341255188, | |
| "logits/rejected": 0.09989897906780243, | |
| "logps/chosen": -803.9798583984375, | |
| "logps/rejected": -925.3179321289062, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.798739433288574, | |
| "rewards/margins": 17.492319107055664, | |
| "rewards/rejected": -10.69357967376709, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.9105691056910569, | |
| "grad_norm": 0.0031374047975987196, | |
| "learning_rate": 0.0001818302775908169, | |
| "logits/chosen": 1.017639398574829, | |
| "logits/rejected": 1.2823631763458252, | |
| "logps/chosen": -824.6445922851562, | |
| "logps/rejected": -860.8942260742188, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.019498825073242, | |
| "rewards/margins": 16.16924285888672, | |
| "rewards/rejected": -10.149742126464844, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.926829268292683, | |
| "grad_norm": 0.00014241511235013604, | |
| "learning_rate": 0.0001810579025148674, | |
| "logits/chosen": 1.0959478616714478, | |
| "logits/rejected": 0.9008815288543701, | |
| "logps/chosen": -782.0526123046875, | |
| "logps/rejected": -916.8338623046875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.443077087402344, | |
| "rewards/margins": 24.263744354248047, | |
| "rewards/rejected": -15.820667266845703, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.943089430894309, | |
| "grad_norm": 5.913816494285129e-05, | |
| "learning_rate": 0.00018027116379309638, | |
| "logits/chosen": 0.2709883153438568, | |
| "logits/rejected": 0.29769933223724365, | |
| "logps/chosen": -735.5257568359375, | |
| "logps/rejected": -1044.0601806640625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.65300178527832, | |
| "rewards/margins": 18.755083084106445, | |
| "rewards/rejected": -10.102080345153809, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.959349593495935, | |
| "grad_norm": 0.01578771322965622, | |
| "learning_rate": 0.00017947020083740575, | |
| "logits/chosen": 1.5522100925445557, | |
| "logits/rejected": 1.7518442869186401, | |
| "logps/chosen": -1019.1099853515625, | |
| "logps/rejected": -624.6131591796875, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 10.32003402709961, | |
| "rewards/margins": 23.75770378112793, | |
| "rewards/rejected": -13.43766975402832, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.975609756097561, | |
| "grad_norm": 0.0010152229806408286, | |
| "learning_rate": 0.00017865515558026428, | |
| "logits/chosen": 0.8601479530334473, | |
| "logits/rejected": 0.819040060043335, | |
| "logps/chosen": -763.342041015625, | |
| "logps/rejected": -817.870849609375, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.2501859664917, | |
| "rewards/margins": 16.491539001464844, | |
| "rewards/rejected": -8.241353034973145, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.991869918699187, | |
| "grad_norm": 0.008696873672306538, | |
| "learning_rate": 0.0001778261724495566, | |
| "logits/chosen": 0.7409014701843262, | |
| "logits/rejected": 0.9245580434799194, | |
| "logps/chosen": -888.8350830078125, | |
| "logps/rejected": -796.002685546875, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 11.07230281829834, | |
| "rewards/margins": 22.53582000732422, | |
| "rewards/rejected": -11.463518142700195, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 2.3132517526391894e-05, | |
| "learning_rate": 0.00017698339834299061, | |
| "logits/chosen": 0.962340772151947, | |
| "logits/rejected": 1.369040608406067, | |
| "logps/chosen": -843.8861083984375, | |
| "logps/rejected": -833.0137329101562, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.60971736907959, | |
| "rewards/margins": 22.649456024169922, | |
| "rewards/rejected": -15.039739608764648, | |
| "step": 62 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 246, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 62, | |
| "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": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |