| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0265486725663717, | |
| "eval_steps": 500, | |
| "global_step": 29, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.035398230088495575, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -3.0738143920898438, | |
| "logits/rejected": -3.0764384269714355, | |
| "logps/chosen": -48.15470886230469, | |
| "logps/rejected": -73.88392639160156, | |
| "loss": 2.7726, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.07079646017699115, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -3.094489812850952, | |
| "logits/rejected": -3.090228319168091, | |
| "logps/chosen": -51.06940841674805, | |
| "logps/rejected": -76.33265686035156, | |
| "loss": 2.7726, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.10619469026548672, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -3.0312275886535645, | |
| "logits/rejected": -3.030165910720825, | |
| "logps/chosen": -47.49669647216797, | |
| "logps/rejected": -76.91014099121094, | |
| "loss": 2.7726, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.1415929203539823, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -3.0501389503479004, | |
| "logits/rejected": -3.054299831390381, | |
| "logps/chosen": -52.20546340942383, | |
| "logps/rejected": -76.84318542480469, | |
| "loss": 2.7726, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.17699115044247787, | |
| "grad_norm": 39.514007568359375, | |
| "learning_rate": 2e-05, | |
| "logits/chosen": -3.059661865234375, | |
| "logits/rejected": -3.0544543266296387, | |
| "logps/chosen": -54.500221252441406, | |
| "logps/rejected": -80.73619079589844, | |
| "loss": 2.7726, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.21238938053097345, | |
| "grad_norm": 35.310333251953125, | |
| "learning_rate": 4e-05, | |
| "logits/chosen": -3.0174481868743896, | |
| "logits/rejected": -3.0225086212158203, | |
| "logps/chosen": -45.962432861328125, | |
| "logps/rejected": -69.7489013671875, | |
| "loss": 2.7726, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.24778761061946902, | |
| "grad_norm": 33.619834899902344, | |
| "learning_rate": 6e-05, | |
| "logits/chosen": -3.0760531425476074, | |
| "logits/rejected": -3.075507402420044, | |
| "logps/chosen": -52.68150329589844, | |
| "logps/rejected": -82.66320037841797, | |
| "loss": 2.3724, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": 0.033591579645872116, | |
| "rewards/margins": 0.2135639488697052, | |
| "rewards/rejected": -0.17997236549854279, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.2831858407079646, | |
| "grad_norm": 26.519262313842773, | |
| "learning_rate": 8e-05, | |
| "logits/chosen": -3.141781806945801, | |
| "logits/rejected": -3.111323595046997, | |
| "logps/chosen": -49.25138854980469, | |
| "logps/rejected": -83.14276885986328, | |
| "loss": 1.6547, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.08978669345378876, | |
| "rewards/margins": 0.6971186399459839, | |
| "rewards/rejected": -0.6073319315910339, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.3185840707964602, | |
| "grad_norm": 17.522594451904297, | |
| "learning_rate": 0.0001, | |
| "logits/chosen": -3.1494789123535156, | |
| "logits/rejected": -3.107832431793213, | |
| "logps/chosen": -47.360626220703125, | |
| "logps/rejected": -87.75389099121094, | |
| "loss": 0.9735, | |
| "rewards/accuracies": 0.953125, | |
| "rewards/chosen": 0.19129924476146698, | |
| "rewards/margins": 1.4335689544677734, | |
| "rewards/rejected": -1.24226975440979, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.35398230088495575, | |
| "grad_norm": 7.8468241691589355, | |
| "learning_rate": 0.00012, | |
| "logits/chosen": -3.105159044265747, | |
| "logits/rejected": -3.0585837364196777, | |
| "logps/chosen": -42.66419982910156, | |
| "logps/rejected": -92.25341796875, | |
| "loss": 0.5496, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 0.2652404308319092, | |
| "rewards/margins": 2.590460777282715, | |
| "rewards/rejected": -2.3252203464508057, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3893805309734513, | |
| "grad_norm": 2.123518228530884, | |
| "learning_rate": 0.00014, | |
| "logits/chosen": -3.239974021911621, | |
| "logits/rejected": -3.114036798477173, | |
| "logps/chosen": -46.743621826171875, | |
| "logps/rejected": -125.77220153808594, | |
| "loss": 0.0432, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.3305644392967224, | |
| "rewards/margins": 5.1967926025390625, | |
| "rewards/rejected": -4.866228103637695, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.4247787610619469, | |
| "grad_norm": 0.18487615883350372, | |
| "learning_rate": 0.00016, | |
| "logits/chosen": -3.2138984203338623, | |
| "logits/rejected": -2.992119312286377, | |
| "logps/chosen": -54.21443557739258, | |
| "logps/rejected": -160.70346069335938, | |
| "loss": 0.132, | |
| "rewards/accuracies": 0.953125, | |
| "rewards/chosen": 0.07433129847049713, | |
| "rewards/margins": 8.329842567443848, | |
| "rewards/rejected": -8.255511283874512, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.46017699115044247, | |
| "grad_norm": 0.3735005855560303, | |
| "learning_rate": 0.00018, | |
| "logits/chosen": -3.314460277557373, | |
| "logits/rejected": -2.955761194229126, | |
| "logps/chosen": -52.52698516845703, | |
| "logps/rejected": -211.6153564453125, | |
| "loss": 0.0467, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -0.11871516704559326, | |
| "rewards/margins": 13.215730667114258, | |
| "rewards/rejected": -13.33444595336914, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.49557522123893805, | |
| "grad_norm": 0.000926459557376802, | |
| "learning_rate": 0.0002, | |
| "logits/chosen": -3.378767490386963, | |
| "logits/rejected": -2.9119057655334473, | |
| "logps/chosen": -59.550537109375, | |
| "logps/rejected": -289.6356201171875, | |
| "loss": 0.0433, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -0.4433783292770386, | |
| "rewards/margins": 20.33984375, | |
| "rewards/rejected": -20.78322410583496, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.5309734513274337, | |
| "grad_norm": 0.018438709899783134, | |
| "learning_rate": 0.00019863613034027224, | |
| "logits/chosen": -3.4022810459136963, | |
| "logits/rejected": -2.9307518005371094, | |
| "logps/chosen": -75.22628784179688, | |
| "logps/rejected": -371.2684326171875, | |
| "loss": 0.0867, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -2.2902748584747314, | |
| "rewards/margins": 27.06856918334961, | |
| "rewards/rejected": -29.358842849731445, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.5663716814159292, | |
| "grad_norm": 0.00022576068295165896, | |
| "learning_rate": 0.00019458172417006347, | |
| "logits/chosen": -3.4614078998565674, | |
| "logits/rejected": -3.028236150741577, | |
| "logps/chosen": -111.99066162109375, | |
| "logps/rejected": -499.0533142089844, | |
| "loss": 0.0433, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -6.10006856918335, | |
| "rewards/margins": 36.00136947631836, | |
| "rewards/rejected": -42.101436614990234, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.6017699115044248, | |
| "grad_norm": 0.0, | |
| "learning_rate": 0.0001879473751206489, | |
| "logits/chosen": -3.4879064559936523, | |
| "logits/rejected": -3.1346540451049805, | |
| "logps/chosen": -181.54132080078125, | |
| "logps/rejected": -627.1567993164062, | |
| "loss": 0.0433, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -12.846463203430176, | |
| "rewards/margins": 42.00312805175781, | |
| "rewards/rejected": -54.84959411621094, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.6371681415929203, | |
| "grad_norm": 0.0, | |
| "learning_rate": 0.00017891405093963938, | |
| "logits/chosen": -3.5802273750305176, | |
| "logits/rejected": -3.2139182090759277, | |
| "logps/chosen": -298.01898193359375, | |
| "logps/rejected": -730.3427734375, | |
| "loss": 0.0433, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -24.727779388427734, | |
| "rewards/margins": 40.633914947509766, | |
| "rewards/rejected": -65.3616943359375, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.672566371681416, | |
| "grad_norm": 0.0006348241004161537, | |
| "learning_rate": 0.00016772815716257412, | |
| "logits/chosen": -3.6362297534942627, | |
| "logits/rejected": -3.3277578353881836, | |
| "logps/chosen": -441.85595703125, | |
| "logps/rejected": -857.678466796875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -39.0582275390625, | |
| "rewards/margins": 38.73958969116211, | |
| "rewards/rejected": -77.79782104492188, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.7079646017699115, | |
| "grad_norm": 5.561949728871696e-05, | |
| "learning_rate": 0.00015469481581224272, | |
| "logits/chosen": -3.7110390663146973, | |
| "logits/rejected": -3.481760025024414, | |
| "logps/chosen": -620.394287109375, | |
| "logps/rejected": -953.712646484375, | |
| "loss": 0.0866, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -56.605865478515625, | |
| "rewards/margins": 30.84559440612793, | |
| "rewards/rejected": -87.45145416259766, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.7433628318584071, | |
| "grad_norm": 0.006302627269178629, | |
| "learning_rate": 0.00014016954246529696, | |
| "logits/chosen": -3.705810070037842, | |
| "logits/rejected": -3.480112314224243, | |
| "logps/chosen": -626.589111328125, | |
| "logps/rejected": -926.1605224609375, | |
| "loss": 0.2166, | |
| "rewards/accuracies": 0.921875, | |
| "rewards/chosen": -57.85133743286133, | |
| "rewards/margins": 27.39826202392578, | |
| "rewards/rejected": -85.24958801269531, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.7787610619469026, | |
| "grad_norm": 0.19819259643554688, | |
| "learning_rate": 0.00012454854871407994, | |
| "logits/chosen": -3.771535873413086, | |
| "logits/rejected": -3.57301664352417, | |
| "logps/chosen": -856.1668701171875, | |
| "logps/rejected": -1113.4576416015625, | |
| "loss": 0.0434, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -80.38790130615234, | |
| "rewards/margins": 23.13334846496582, | |
| "rewards/rejected": -103.52125549316406, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.8141592920353983, | |
| "grad_norm": 0.00016238712123595178, | |
| "learning_rate": 0.00010825793454723325, | |
| "logits/chosen": -3.9207518100738525, | |
| "logits/rejected": -3.6793296337127686, | |
| "logps/chosen": -834.5657958984375, | |
| "logps/rejected": -1170.9774169921875, | |
| "loss": 0.0433, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -78.40864562988281, | |
| "rewards/margins": 31.1939697265625, | |
| "rewards/rejected": -109.60260772705078, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.8495575221238938, | |
| "grad_norm": 0.0, | |
| "learning_rate": 9.174206545276677e-05, | |
| "logits/chosen": -4.076857089996338, | |
| "logits/rejected": -3.751131534576416, | |
| "logps/chosen": -736.5623168945312, | |
| "logps/rejected": -1179.674072265625, | |
| "loss": 0.13, | |
| "rewards/accuracies": 0.953125, | |
| "rewards/chosen": -69.1473159790039, | |
| "rewards/margins": 41.747955322265625, | |
| "rewards/rejected": -110.89527893066406, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.8849557522123894, | |
| "grad_norm": 0.0, | |
| "learning_rate": 7.54514512859201e-05, | |
| "logits/chosen": -4.139317035675049, | |
| "logits/rejected": -3.7883429527282715, | |
| "logps/chosen": -723.9154663085938, | |
| "logps/rejected": -1208.33154296875, | |
| "loss": 0.13, | |
| "rewards/accuracies": 0.953125, | |
| "rewards/chosen": -67.75996398925781, | |
| "rewards/margins": 45.80049133300781, | |
| "rewards/rejected": -113.56045532226562, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.9203539823008849, | |
| "grad_norm": 0.0, | |
| "learning_rate": 5.983045753470308e-05, | |
| "logits/chosen": -4.219160079956055, | |
| "logits/rejected": -3.8395681381225586, | |
| "logps/chosen": -831.74560546875, | |
| "logps/rejected": -1357.1932373046875, | |
| "loss": 0.0433, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -78.22254943847656, | |
| "rewards/margins": 49.974510192871094, | |
| "rewards/rejected": -128.19705200195312, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.9557522123893806, | |
| "grad_norm": 1.5598791378579335e-06, | |
| "learning_rate": 4.530518418775733e-05, | |
| "logits/chosen": -4.2556562423706055, | |
| "logits/rejected": -3.8611326217651367, | |
| "logps/chosen": -830.6559448242188, | |
| "logps/rejected": -1364.437744140625, | |
| "loss": 0.0433, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -78.08842468261719, | |
| "rewards/margins": 50.610897064208984, | |
| "rewards/rejected": -128.69931030273438, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.9911504424778761, | |
| "grad_norm": 0.0023740511387586594, | |
| "learning_rate": 3.227184283742591e-05, | |
| "logits/chosen": -4.302487373352051, | |
| "logits/rejected": -3.8732571601867676, | |
| "logps/chosen": -825.9217529296875, | |
| "logps/rejected": -1350.9755859375, | |
| "loss": 0.0433, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -77.16596984863281, | |
| "rewards/margins": 49.86052703857422, | |
| "rewards/rejected": -127.02650451660156, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 1.0265486725663717, | |
| "grad_norm": 0.0, | |
| "learning_rate": 2.1085949060360654e-05, | |
| "logits/chosen": -4.38674259185791, | |
| "logits/rejected": -3.9807748794555664, | |
| "logps/chosen": -908.9441528320312, | |
| "logps/rejected": -1436.7275390625, | |
| "loss": 0.13, | |
| "rewards/accuracies": 0.953125, | |
| "rewards/chosen": -85.50328063964844, | |
| "rewards/margins": 50.298500061035156, | |
| "rewards/rejected": -135.8017578125, | |
| "step": 29 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 29, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "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": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |