| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9945, | |
| "eval_steps": 500, | |
| "global_step": 153, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 28.229060496303497, | |
| "learning_rate": 3.125e-08, | |
| "logits/chosen": 0.5326807498931885, | |
| "logits/rejected": 0.5883637070655823, | |
| "logps/chosen": -185.19822692871094, | |
| "logps/rejected": -194.60989379882812, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 32.45486691880899, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": 0.33694812655448914, | |
| "logits/rejected": 0.18525859713554382, | |
| "logps/chosen": -228.18931579589844, | |
| "logps/rejected": -250.44186401367188, | |
| "loss": 0.6926, | |
| "rewards/accuracies": 0.44017094373703003, | |
| "rewards/chosen": -0.009515076875686646, | |
| "rewards/margins": 0.0016028096433728933, | |
| "rewards/rejected": -0.011117885820567608, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 32.74664763693393, | |
| "learning_rate": 4.989490450759331e-07, | |
| "logits/chosen": 0.37088415026664734, | |
| "logits/rejected": 0.4128836989402771, | |
| "logps/chosen": -254.14837646484375, | |
| "logps/rejected": -276.72271728515625, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.5384615659713745, | |
| "rewards/chosen": -0.197755366563797, | |
| "rewards/margins": 0.020086202770471573, | |
| "rewards/rejected": -0.21784158051013947, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 38.40726627675541, | |
| "learning_rate": 4.872270441827174e-07, | |
| "logits/chosen": 0.354877769947052, | |
| "logits/rejected": 0.33263731002807617, | |
| "logps/chosen": -253.26522827148438, | |
| "logps/rejected": -263.25042724609375, | |
| "loss": 0.6936, | |
| "rewards/accuracies": 0.5192307829856873, | |
| "rewards/chosen": -0.3028598725795746, | |
| "rewards/margins": 0.01474391482770443, | |
| "rewards/rejected": -0.3176037669181824, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 38.797592985526, | |
| "learning_rate": 4.6308512113530063e-07, | |
| "logits/chosen": 0.1634213924407959, | |
| "logits/rejected": 0.2352660596370697, | |
| "logps/chosen": -244.1608123779297, | |
| "logps/rejected": -260.97369384765625, | |
| "loss": 0.7013, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.0662732720375061, | |
| "rewards/margins": 0.021908778697252274, | |
| "rewards/rejected": -0.08818206936120987, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 35.745848630813434, | |
| "learning_rate": 4.277872161641681e-07, | |
| "logits/chosen": 0.44109058380126953, | |
| "logits/rejected": 0.4526838958263397, | |
| "logps/chosen": -233.4575653076172, | |
| "logps/rejected": -257.31494140625, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.5423076748847961, | |
| "rewards/chosen": -0.13867294788360596, | |
| "rewards/margins": 0.03718903288245201, | |
| "rewards/rejected": -0.17586196959018707, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 38.266936092602805, | |
| "learning_rate": 3.8318133624280046e-07, | |
| "logits/chosen": 0.3567802309989929, | |
| "logits/rejected": 0.4483684301376343, | |
| "logps/chosen": -273.6899719238281, | |
| "logps/rejected": -289.3863220214844, | |
| "loss": 0.6819, | |
| "rewards/accuracies": 0.5769230723381042, | |
| "rewards/chosen": -0.3489604592323303, | |
| "rewards/margins": 0.07096390426158905, | |
| "rewards/rejected": -0.4199243485927582, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 36.58606527761828, | |
| "learning_rate": 3.316028034595861e-07, | |
| "logits/chosen": 0.49843645095825195, | |
| "logits/rejected": 0.5629610419273376, | |
| "logps/chosen": -280.0177917480469, | |
| "logps/rejected": -301.7088317871094, | |
| "loss": 0.6845, | |
| "rewards/accuracies": 0.5538461804389954, | |
| "rewards/chosen": -0.45351117849349976, | |
| "rewards/margins": 0.03525887802243233, | |
| "rewards/rejected": -0.4887700378894806, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 37.11287003011496, | |
| "learning_rate": 2.7575199021178855e-07, | |
| "logits/chosen": 0.6807990670204163, | |
| "logits/rejected": 0.7718464136123657, | |
| "logps/chosen": -280.74658203125, | |
| "logps/rejected": -311.41009521484375, | |
| "loss": 0.676, | |
| "rewards/accuracies": 0.5923076868057251, | |
| "rewards/chosen": -0.41281428933143616, | |
| "rewards/margins": 0.10995330661535263, | |
| "rewards/rejected": -0.5227676033973694, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 50.701799655354534, | |
| "learning_rate": 2.1855294234408068e-07, | |
| "logits/chosen": 0.6627506613731384, | |
| "logits/rejected": 0.6323168277740479, | |
| "logps/chosen": -278.023681640625, | |
| "logps/rejected": -306.6815185546875, | |
| "loss": 0.6771, | |
| "rewards/accuracies": 0.557692289352417, | |
| "rewards/chosen": -0.4557516276836395, | |
| "rewards/margins": 0.07288946956396103, | |
| "rewards/rejected": -0.5286410450935364, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 29.961533632761956, | |
| "learning_rate": 1.6300029195778453e-07, | |
| "logits/chosen": 0.5369245409965515, | |
| "logits/rejected": 0.5473312735557556, | |
| "logps/chosen": -275.7201843261719, | |
| "logps/rejected": -282.8805236816406, | |
| "loss": 0.7042, | |
| "rewards/accuracies": 0.5115384459495544, | |
| "rewards/chosen": -0.3841624855995178, | |
| "rewards/margins": -0.015378502197563648, | |
| "rewards/rejected": -0.36878401041030884, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 31.11285576879039, | |
| "learning_rate": 1.1200247470632392e-07, | |
| "logits/chosen": 0.31445741653442383, | |
| "logits/rejected": 0.33911341428756714, | |
| "logps/chosen": -258.0011901855469, | |
| "logps/rejected": -270.0567932128906, | |
| "loss": 0.6857, | |
| "rewards/accuracies": 0.5423076748847961, | |
| "rewards/chosen": -0.2574421763420105, | |
| "rewards/margins": 0.022587427869439125, | |
| "rewards/rejected": -0.2800295948982239, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 31.866161116501196, | |
| "learning_rate": 6.822945986946385e-08, | |
| "logits/chosen": 0.6134840250015259, | |
| "logits/rejected": 0.691197395324707, | |
| "logps/chosen": -284.92510986328125, | |
| "logps/rejected": -306.45050048828125, | |
| "loss": 0.6837, | |
| "rewards/accuracies": 0.5423076748847961, | |
| "rewards/chosen": -0.5774205327033997, | |
| "rewards/margins": 0.037515509873628616, | |
| "rewards/rejected": -0.6149360537528992, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 41.02700179238128, | |
| "learning_rate": 3.397296523427806e-08, | |
| "logits/chosen": 1.0320525169372559, | |
| "logits/rejected": 0.8667150735855103, | |
| "logps/chosen": -293.99853515625, | |
| "logps/rejected": -317.0647888183594, | |
| "loss": 0.688, | |
| "rewards/accuracies": 0.5961538553237915, | |
| "rewards/chosen": -0.5775225162506104, | |
| "rewards/margins": 0.07906623929738998, | |
| "rewards/rejected": -0.6565887928009033, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 33.934196556449756, | |
| "learning_rate": 1.1026475173977978e-08, | |
| "logits/chosen": 0.4998157322406769, | |
| "logits/rejected": 0.41973716020584106, | |
| "logps/chosen": -282.9462585449219, | |
| "logps/rejected": -299.1942443847656, | |
| "loss": 0.6794, | |
| "rewards/accuracies": 0.557692289352417, | |
| "rewards/chosen": -0.4659727215766907, | |
| "rewards/margins": 0.056435175240039825, | |
| "rewards/rejected": -0.5224078297615051, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 38.33373569305899, | |
| "learning_rate": 5.913435276374834e-10, | |
| "logits/chosen": 0.42322245240211487, | |
| "logits/rejected": 0.43874579668045044, | |
| "logps/chosen": -272.926513671875, | |
| "logps/rejected": -309.4190368652344, | |
| "loss": 0.6797, | |
| "rewards/accuracies": 0.5807692408561707, | |
| "rewards/chosen": -0.5053122043609619, | |
| "rewards/margins": 0.1106579527258873, | |
| "rewards/rejected": -0.6159701943397522, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "step": 153, | |
| "total_flos": 0.0, | |
| "train_loss": 0.687320882978003, | |
| "train_runtime": 21824.845, | |
| "train_samples_per_second": 0.916, | |
| "train_steps_per_second": 0.007 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 153, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 200, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |