| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984, | |
| "eval_steps": 100, | |
| "global_step": 312, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0032, | |
| "grad_norm": 1009.1310262338153, | |
| "learning_rate": 1.5625e-08, | |
| "logits/chosen": -0.48170238733291626, | |
| "logits/rejected": -0.5971822738647461, | |
| "logps/chosen": -29.012826919555664, | |
| "logps/rejected": -37.434471130371094, | |
| "loss": 49.0548, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.032, | |
| "grad_norm": 969.0027881269435, | |
| "learning_rate": 1.5624999999999999e-07, | |
| "logits/chosen": -0.43042537569999695, | |
| "logits/rejected": -0.4711245000362396, | |
| "logps/chosen": -32.84762191772461, | |
| "logps/rejected": -34.32879638671875, | |
| "loss": 38.7199, | |
| "rewards/accuracies": 0.2777777910232544, | |
| "rewards/chosen": 5.492389027494937e-05, | |
| "rewards/margins": -5.384198811952956e-05, | |
| "rewards/rejected": 0.00010876586748054251, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.064, | |
| "grad_norm": 993.3615142178909, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -0.5954495072364807, | |
| "logits/rejected": -0.618130087852478, | |
| "logps/chosen": -48.733314514160156, | |
| "logps/rejected": -51.568695068359375, | |
| "loss": 41.7023, | |
| "rewards/accuracies": 0.42500001192092896, | |
| "rewards/chosen": 0.0010887670796364546, | |
| "rewards/margins": -3.9788112189853564e-05, | |
| "rewards/rejected": 0.0011285552754998207, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.096, | |
| "grad_norm": 964.1608085815409, | |
| "learning_rate": 4.6874999999999996e-07, | |
| "logits/chosen": -0.6574007272720337, | |
| "logits/rejected": -0.7291825413703918, | |
| "logps/chosen": -39.434776306152344, | |
| "logps/rejected": -42.335716247558594, | |
| "loss": 43.7887, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.001887084566988051, | |
| "rewards/margins": 2.423059231659863e-05, | |
| "rewards/rejected": 0.0018628539983183146, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.128, | |
| "grad_norm": 1181.1281208132718, | |
| "learning_rate": 4.857142857142857e-07, | |
| "logits/chosen": -0.6512608528137207, | |
| "logits/rejected": -0.6266627311706543, | |
| "logps/chosen": -37.72194290161133, | |
| "logps/rejected": -40.52619171142578, | |
| "loss": 40.7141, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.0024465853348374367, | |
| "rewards/margins": 0.00038118622615002096, | |
| "rewards/rejected": 0.0020653989631682634, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 1123.3502304939157, | |
| "learning_rate": 4.6785714285714283e-07, | |
| "logits/chosen": -0.5247907042503357, | |
| "logits/rejected": -0.5476827621459961, | |
| "logps/chosen": -37.13096237182617, | |
| "logps/rejected": -36.993751525878906, | |
| "loss": 41.8391, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": 0.0022364568430930376, | |
| "rewards/margins": 8.963019354268909e-05, | |
| "rewards/rejected": 0.0021468265913426876, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.192, | |
| "grad_norm": 1085.2090697086192, | |
| "learning_rate": 4.5e-07, | |
| "logits/chosen": -0.43748584389686584, | |
| "logits/rejected": -0.4264857769012451, | |
| "logps/chosen": -42.964988708496094, | |
| "logps/rejected": -37.55246353149414, | |
| "loss": 41.2449, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.001739932456985116, | |
| "rewards/margins": -0.00011503996938699856, | |
| "rewards/rejected": 0.0018549725646153092, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.224, | |
| "grad_norm": 1015.4302150816914, | |
| "learning_rate": 4.3214285714285713e-07, | |
| "logits/chosen": -0.5420335531234741, | |
| "logits/rejected": -0.5390623211860657, | |
| "logps/chosen": -39.6472282409668, | |
| "logps/rejected": -40.896602630615234, | |
| "loss": 42.2903, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": 0.0017996892565861344, | |
| "rewards/margins": -0.0006139087490737438, | |
| "rewards/rejected": 0.0024135981220752, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.256, | |
| "grad_norm": 1112.9628164769215, | |
| "learning_rate": 4.142857142857143e-07, | |
| "logits/chosen": -0.5414769053459167, | |
| "logits/rejected": -0.6392091512680054, | |
| "logps/chosen": -41.40280532836914, | |
| "logps/rejected": -38.21042251586914, | |
| "loss": 37.9115, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.002596390200778842, | |
| "rewards/margins": 0.000514088838826865, | |
| "rewards/rejected": 0.002082301303744316, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.288, | |
| "grad_norm": 1132.9412773578845, | |
| "learning_rate": 3.9642857142857137e-07, | |
| "logits/chosen": -0.5120124816894531, | |
| "logits/rejected": -0.46202293038368225, | |
| "logps/chosen": -30.624813079833984, | |
| "logps/rejected": -29.584774017333984, | |
| "loss": 41.9608, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.0020510884933173656, | |
| "rewards/margins": -0.00010951957665383816, | |
| "rewards/rejected": 0.00216060783714056, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 1065.2844763715325, | |
| "learning_rate": 3.785714285714285e-07, | |
| "logits/chosen": -0.6081634759902954, | |
| "logits/rejected": -0.5888960957527161, | |
| "logps/chosen": -34.334285736083984, | |
| "logps/rejected": -33.35486602783203, | |
| "loss": 39.9711, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.0022316346876323223, | |
| "rewards/margins": 0.0007666076417081058, | |
| "rewards/rejected": 0.0014650272205471992, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.352, | |
| "grad_norm": 1072.4046311311965, | |
| "learning_rate": 3.607142857142857e-07, | |
| "logits/chosen": -0.4411347806453705, | |
| "logits/rejected": -0.4419333040714264, | |
| "logps/chosen": -24.29958152770996, | |
| "logps/rejected": -25.08634376525879, | |
| "loss": 38.3559, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.00189125572796911, | |
| "rewards/margins": 0.00026331303524784744, | |
| "rewards/rejected": 0.001627942780032754, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.384, | |
| "grad_norm": 1108.2228357679985, | |
| "learning_rate": 3.4285714285714286e-07, | |
| "logits/chosen": -0.5807032585144043, | |
| "logits/rejected": -0.6294857859611511, | |
| "logps/chosen": -39.8912353515625, | |
| "logps/rejected": -42.09502029418945, | |
| "loss": 38.3583, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.0031242019031196833, | |
| "rewards/margins": 0.00099054048769176, | |
| "rewards/rejected": 0.002133661415427923, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.416, | |
| "grad_norm": 1149.96614138426, | |
| "learning_rate": 3.25e-07, | |
| "logits/chosen": -0.5135449767112732, | |
| "logits/rejected": -0.4764982759952545, | |
| "logps/chosen": -39.335472106933594, | |
| "logps/rejected": -40.464088439941406, | |
| "loss": 41.8146, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.0020922988187521696, | |
| "rewards/margins": 0.00032900949008762836, | |
| "rewards/rejected": 0.0017632890958338976, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.448, | |
| "grad_norm": 1137.1577747838844, | |
| "learning_rate": 3.0714285714285716e-07, | |
| "logits/chosen": -0.6154284477233887, | |
| "logits/rejected": -0.5684093236923218, | |
| "logps/chosen": -43.35015106201172, | |
| "logps/rejected": -45.817138671875, | |
| "loss": 43.4174, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.0029308353550732136, | |
| "rewards/margins": 0.00020970618061255664, | |
| "rewards/rejected": 0.0027211292181164026, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 1045.1524947853006, | |
| "learning_rate": 2.892857142857143e-07, | |
| "logits/chosen": -0.49414142966270447, | |
| "logits/rejected": -0.5363041758537292, | |
| "logps/chosen": -25.930377960205078, | |
| "logps/rejected": -29.026256561279297, | |
| "loss": 40.5417, | |
| "rewards/accuracies": 0.42500001192092896, | |
| "rewards/chosen": 0.0020943940617144108, | |
| "rewards/margins": -0.0005832756869494915, | |
| "rewards/rejected": 0.0026776697486639023, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.512, | |
| "grad_norm": 1137.7571627874984, | |
| "learning_rate": 2.714285714285714e-07, | |
| "logits/chosen": -0.45031625032424927, | |
| "logits/rejected": -0.4036481976509094, | |
| "logps/chosen": -35.83759307861328, | |
| "logps/rejected": -33.2865104675293, | |
| "loss": 41.7872, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.0025531931314617395, | |
| "rewards/margins": -0.00011138639092678204, | |
| "rewards/rejected": 0.0026645795442163944, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.544, | |
| "grad_norm": 1178.055387368166, | |
| "learning_rate": 2.5357142857142855e-07, | |
| "logits/chosen": -0.5122729539871216, | |
| "logits/rejected": -0.49646639823913574, | |
| "logps/chosen": -39.3989143371582, | |
| "logps/rejected": -42.247840881347656, | |
| "loss": 41.7091, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.002737321425229311, | |
| "rewards/margins": 0.000858407118357718, | |
| "rewards/rejected": 0.0018789141904562712, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.576, | |
| "grad_norm": 1222.0538303638675, | |
| "learning_rate": 2.357142857142857e-07, | |
| "logits/chosen": -0.5412944555282593, | |
| "logits/rejected": -0.5432535409927368, | |
| "logps/chosen": -41.7281494140625, | |
| "logps/rejected": -41.42877197265625, | |
| "loss": 45.593, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.0030562058091163635, | |
| "rewards/margins": 0.0002739687915891409, | |
| "rewards/rejected": 0.0027822370175272226, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.608, | |
| "grad_norm": 1059.5993351433744, | |
| "learning_rate": 2.1785714285714284e-07, | |
| "logits/chosen": -0.5831159353256226, | |
| "logits/rejected": -0.5751662254333496, | |
| "logps/chosen": -46.93434143066406, | |
| "logps/rejected": -43.748443603515625, | |
| "loss": 43.2402, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": 0.002087333472445607, | |
| "rewards/margins": 3.515209391480312e-05, | |
| "rewards/rejected": 0.0020521811675280333, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 1167.572795808609, | |
| "learning_rate": 2e-07, | |
| "logits/chosen": -0.48871344327926636, | |
| "logits/rejected": -0.5342381000518799, | |
| "logps/chosen": -31.179672241210938, | |
| "logps/rejected": -35.725318908691406, | |
| "loss": 40.539, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.0033684056252241135, | |
| "rewards/margins": 0.0005909694591537118, | |
| "rewards/rejected": 0.002777435816824436, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.672, | |
| "grad_norm": 1051.702249531706, | |
| "learning_rate": 1.8214285714285714e-07, | |
| "logits/chosen": -0.5266011357307434, | |
| "logits/rejected": -0.49189430475234985, | |
| "logps/chosen": -30.086624145507812, | |
| "logps/rejected": -30.298954010009766, | |
| "loss": 40.312, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.0023514502681791782, | |
| "rewards/margins": 0.00040291310870088637, | |
| "rewards/rejected": 0.0019485373049974442, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.704, | |
| "grad_norm": 1132.3748943763287, | |
| "learning_rate": 1.6428571428571429e-07, | |
| "logits/chosen": -0.6477980017662048, | |
| "logits/rejected": -0.5915540456771851, | |
| "logps/chosen": -52.536956787109375, | |
| "logps/rejected": -53.607810974121094, | |
| "loss": 42.3887, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.00340586481615901, | |
| "rewards/margins": 0.0009445315226912498, | |
| "rewards/rejected": 0.0024613335262984037, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.736, | |
| "grad_norm": 1128.7389141378083, | |
| "learning_rate": 1.4642857142857143e-07, | |
| "logits/chosen": -0.47689008712768555, | |
| "logits/rejected": -0.5440131425857544, | |
| "logps/chosen": -34.16132736206055, | |
| "logps/rejected": -37.00074005126953, | |
| "loss": 38.9005, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": 0.0027422527782619, | |
| "rewards/margins": -0.0002558851265348494, | |
| "rewards/rejected": 0.0029981378465890884, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.768, | |
| "grad_norm": 1083.3819182532363, | |
| "learning_rate": 1.2857142857142855e-07, | |
| "logits/chosen": -0.6017645597457886, | |
| "logits/rejected": -0.6154614686965942, | |
| "logps/chosen": -60.46274948120117, | |
| "logps/rejected": -53.796775817871094, | |
| "loss": 39.6015, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.003490505740046501, | |
| "rewards/margins": 0.0010509835556149483, | |
| "rewards/rejected": 0.002439522184431553, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 1158.4783936167528, | |
| "learning_rate": 1.107142857142857e-07, | |
| "logits/chosen": -0.5455101132392883, | |
| "logits/rejected": -0.5882077217102051, | |
| "logps/chosen": -51.72278594970703, | |
| "logps/rejected": -55.974029541015625, | |
| "loss": 40.6188, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.002927373396232724, | |
| "rewards/margins": 0.00027384894201532006, | |
| "rewards/rejected": 0.0026535247452557087, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.832, | |
| "grad_norm": 1119.1973257793477, | |
| "learning_rate": 9.285714285714286e-08, | |
| "logits/chosen": -0.5563905239105225, | |
| "logits/rejected": -0.5701907873153687, | |
| "logps/chosen": -36.81728744506836, | |
| "logps/rejected": -40.96973419189453, | |
| "loss": 36.8782, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.0024603630881756544, | |
| "rewards/margins": -0.000383538194000721, | |
| "rewards/rejected": 0.002843901515007019, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.864, | |
| "grad_norm": 1139.2702962270098, | |
| "learning_rate": 7.5e-08, | |
| "logits/chosen": -0.5481973886489868, | |
| "logits/rejected": -0.5287628173828125, | |
| "logps/chosen": -36.652503967285156, | |
| "logps/rejected": -36.31899642944336, | |
| "loss": 40.0096, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.002305429894477129, | |
| "rewards/margins": 0.0007444783695973456, | |
| "rewards/rejected": 0.0015609515830874443, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.896, | |
| "grad_norm": 1203.095436788835, | |
| "learning_rate": 5.714285714285714e-08, | |
| "logits/chosen": -0.7058078050613403, | |
| "logits/rejected": -0.6220946311950684, | |
| "logps/chosen": -41.582035064697266, | |
| "logps/rejected": -36.45862579345703, | |
| "loss": 42.6208, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.0033497163094580173, | |
| "rewards/margins": 0.0008428205619566143, | |
| "rewards/rejected": 0.0025068954564630985, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.928, | |
| "grad_norm": 1237.1888295476801, | |
| "learning_rate": 3.9285714285714285e-08, | |
| "logits/chosen": -0.6831991076469421, | |
| "logits/rejected": -0.6509202122688293, | |
| "logps/chosen": -54.39546585083008, | |
| "logps/rejected": -51.07060623168945, | |
| "loss": 39.5452, | |
| "rewards/accuracies": 0.42500001192092896, | |
| "rewards/chosen": 0.0032668504863977432, | |
| "rewards/margins": -0.00010241140989819542, | |
| "rewards/rejected": 0.0033692617435008287, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 1122.9875396625666, | |
| "learning_rate": 2.142857142857143e-08, | |
| "logits/chosen": -0.42987775802612305, | |
| "logits/rejected": -0.5015246868133545, | |
| "logps/chosen": -43.15519714355469, | |
| "logps/rejected": -45.89078140258789, | |
| "loss": 40.6293, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": 0.0028726663440465927, | |
| "rewards/margins": 0.00013752638187725097, | |
| "rewards/rejected": 0.002735140034928918, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.992, | |
| "grad_norm": 1133.8011206775125, | |
| "learning_rate": 3.571428571428571e-09, | |
| "logits/chosen": -0.6172088980674744, | |
| "logits/rejected": -0.6021891832351685, | |
| "logps/chosen": -42.21741485595703, | |
| "logps/rejected": -46.169761657714844, | |
| "loss": 39.9958, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.003974457737058401, | |
| "rewards/margins": 0.0007881852798163891, | |
| "rewards/rejected": 0.0031862719915807247, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.9984, | |
| "step": 312, | |
| "total_flos": 0.0, | |
| "train_loss": 40.90790059016301, | |
| "train_runtime": 2754.0963, | |
| "train_samples_per_second": 7.26, | |
| "train_steps_per_second": 0.113 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 312, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |