| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9983361064891847, | |
| "eval_steps": 500, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 1151.6129333090275, | |
| "learning_rate": 3.333333333333333e-10, | |
| "logits/chosen": -4.106247425079346, | |
| "logits/rejected": -4.200438499450684, | |
| "logps/chosen": -382.81439208984375, | |
| "logps/rejected": -357.65960693359375, | |
| "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": 1175.0279383615512, | |
| "learning_rate": 3.3333333333333334e-09, | |
| "logits/chosen": -4.217203617095947, | |
| "logits/rejected": -4.32081413269043, | |
| "logps/chosen": -334.6714172363281, | |
| "logps/rejected": -313.4322509765625, | |
| "loss": 0.7317, | |
| "rewards/accuracies": 0.3923611044883728, | |
| "rewards/chosen": -0.044815655797719955, | |
| "rewards/margins": -0.03612741455435753, | |
| "rewards/rejected": -0.008688241243362427, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 1253.9538230101048, | |
| "learning_rate": 4.983095894354858e-09, | |
| "logits/chosen": -4.266427040100098, | |
| "logits/rejected": -4.4187798500061035, | |
| "logps/chosen": -313.9143371582031, | |
| "logps/rejected": -288.782470703125, | |
| "loss": 0.732, | |
| "rewards/accuracies": 0.4781250059604645, | |
| "rewards/chosen": -0.000932177877984941, | |
| "rewards/margins": -0.012739461846649647, | |
| "rewards/rejected": 0.011807283386588097, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 1147.308654200848, | |
| "learning_rate": 4.849231551964771e-09, | |
| "logits/chosen": -4.301184177398682, | |
| "logits/rejected": -4.36545991897583, | |
| "logps/chosen": -308.3140563964844, | |
| "logps/rejected": -285.6829833984375, | |
| "loss": 0.7352, | |
| "rewards/accuracies": 0.46562498807907104, | |
| "rewards/chosen": -0.018803134560585022, | |
| "rewards/margins": 0.008298242464661598, | |
| "rewards/rejected": -0.02710137702524662, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 1152.01480268075, | |
| "learning_rate": 4.588719528532341e-09, | |
| "logits/chosen": -4.186487674713135, | |
| "logits/rejected": -4.270204544067383, | |
| "logps/chosen": -332.467041015625, | |
| "logps/rejected": -307.530517578125, | |
| "loss": 0.722, | |
| "rewards/accuracies": 0.5531250238418579, | |
| "rewards/chosen": 0.012754167430102825, | |
| "rewards/margins": 0.04999501258134842, | |
| "rewards/rejected": -0.03724084421992302, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 1161.340877574017, | |
| "learning_rate": 4.215604094671834e-09, | |
| "logits/chosen": -4.197475910186768, | |
| "logits/rejected": -4.366654396057129, | |
| "logps/chosen": -333.69940185546875, | |
| "logps/rejected": -309.05511474609375, | |
| "loss": 0.7263, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.022939234972000122, | |
| "rewards/margins": -0.01653190515935421, | |
| "rewards/rejected": 0.03947114199399948, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 1135.5326964714952, | |
| "learning_rate": 3.7500000000000005e-09, | |
| "logits/chosen": -4.186155796051025, | |
| "logits/rejected": -4.269167900085449, | |
| "logps/chosen": -323.9843444824219, | |
| "logps/rejected": -308.364990234375, | |
| "loss": 0.7133, | |
| "rewards/accuracies": 0.484375, | |
| "rewards/chosen": 0.009113344363868237, | |
| "rewards/margins": -0.005971288774162531, | |
| "rewards/rejected": 0.015084633603692055, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 1170.598014061488, | |
| "learning_rate": 3.2170080817777257e-09, | |
| "logits/chosen": -4.14601469039917, | |
| "logits/rejected": -4.301178932189941, | |
| "logps/chosen": -344.8753967285156, | |
| "logps/rejected": -314.6096496582031, | |
| "loss": 0.7182, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": 0.08767497539520264, | |
| "rewards/margins": 0.08200599253177643, | |
| "rewards/rejected": 0.005668987520039082, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 1088.822875661876, | |
| "learning_rate": 2.6453620722761897e-09, | |
| "logits/chosen": -4.220850944519043, | |
| "logits/rejected": -4.3867106437683105, | |
| "logps/chosen": -334.7141418457031, | |
| "logps/rejected": -305.6958923339844, | |
| "loss": 0.7175, | |
| "rewards/accuracies": 0.515625, | |
| "rewards/chosen": 0.06255482137203217, | |
| "rewards/margins": 0.04721928387880325, | |
| "rewards/rejected": 0.015335534699261189, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 1120.1472438016, | |
| "learning_rate": 2.0658795558326744e-09, | |
| "logits/chosen": -4.226521968841553, | |
| "logits/rejected": -4.347161293029785, | |
| "logps/chosen": -327.5687561035156, | |
| "logps/rejected": -308.19744873046875, | |
| "loss": 0.7128, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": 0.03935312479734421, | |
| "rewards/margins": 0.012340927496552467, | |
| "rewards/rejected": 0.02701219543814659, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 1147.49194388717, | |
| "learning_rate": 1.5098005849021078e-09, | |
| "logits/chosen": -4.208071708679199, | |
| "logits/rejected": -4.417771339416504, | |
| "logps/chosen": -320.9470520019531, | |
| "logps/rejected": -290.018310546875, | |
| "loss": 0.712, | |
| "rewards/accuracies": 0.5218750238418579, | |
| "rewards/chosen": 0.060337893664836884, | |
| "rewards/margins": 0.03545590117573738, | |
| "rewards/rejected": 0.0248819962143898, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 1131.801124941311, | |
| "learning_rate": 1.0071035207430352e-09, | |
| "logits/chosen": -4.222798824310303, | |
| "logits/rejected": -4.275721549987793, | |
| "logps/chosen": -315.00030517578125, | |
| "logps/rejected": -303.7385559082031, | |
| "loss": 0.7022, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 0.08098876476287842, | |
| "rewards/margins": 0.037886131554841995, | |
| "rewards/rejected": 0.04310264065861702, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 1201.8232728734756, | |
| "learning_rate": 5.848888922025553e-10, | |
| "logits/chosen": -4.147335052490234, | |
| "logits/rejected": -4.3030242919921875, | |
| "logps/chosen": -339.2781677246094, | |
| "logps/rejected": -313.5749206542969, | |
| "loss": 0.7129, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 0.08869560062885284, | |
| "rewards/margins": 0.048744406551122665, | |
| "rewards/rejected": 0.03995119035243988, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 1182.4206440855232, | |
| "learning_rate": 2.659183991914696e-10, | |
| "logits/chosen": -4.173482418060303, | |
| "logits/rejected": -4.367284297943115, | |
| "logps/chosen": -324.4493713378906, | |
| "logps/rejected": -305.2012634277344, | |
| "loss": 0.7, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.08546491712331772, | |
| "rewards/margins": 0.01908993348479271, | |
| "rewards/rejected": 0.06637498736381531, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 1158.9214495930962, | |
| "learning_rate": 6.738782355044049e-11, | |
| "logits/chosen": -4.290203094482422, | |
| "logits/rejected": -4.35637092590332, | |
| "logps/chosen": -308.0101318359375, | |
| "logps/rejected": -297.37701416015625, | |
| "loss": 0.7082, | |
| "rewards/accuracies": 0.5406249761581421, | |
| "rewards/chosen": 0.12010886520147324, | |
| "rewards/margins": 0.05074785277247429, | |
| "rewards/rejected": 0.06936100870370865, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 1208.2615961890754, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -4.257304668426514, | |
| "logits/rejected": -4.3310017585754395, | |
| "logps/chosen": -312.7613830566406, | |
| "logps/rejected": -300.3546447753906, | |
| "loss": 0.7025, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.10821112245321274, | |
| "rewards/margins": 0.047078292816877365, | |
| "rewards/rejected": 0.06113281846046448, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 150, | |
| "total_flos": 0.0, | |
| "train_loss": 0.7162073644002279, | |
| "train_runtime": 4512.4874, | |
| "train_samples_per_second": 8.52, | |
| "train_steps_per_second": 0.033 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 150, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |