| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9962157048249763, | |
| "eval_steps": 500, | |
| "global_step": 162, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 44.790242854161896, | |
| "learning_rate": 2.941176470588235e-08, | |
| "logits/chosen": 0.4138435125350952, | |
| "logits/rejected": 0.3073309361934662, | |
| "logps/chosen": -238.74684143066406, | |
| "logps/rejected": -277.3367919921875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 32.66922851838542, | |
| "learning_rate": 2.941176470588235e-07, | |
| "logits/chosen": 0.19251321256160736, | |
| "logits/rejected": 0.15595921874046326, | |
| "logps/chosen": -266.190673828125, | |
| "logps/rejected": -288.5094299316406, | |
| "loss": 0.6912, | |
| "rewards/accuracies": 0.5384615659713745, | |
| "rewards/chosen": -0.04362406209111214, | |
| "rewards/margins": 0.010117193683981895, | |
| "rewards/rejected": -0.05374125763773918, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 34.2665441677138, | |
| "learning_rate": 4.99472085783721e-07, | |
| "logits/chosen": 0.4772653877735138, | |
| "logits/rejected": 0.34988027811050415, | |
| "logps/chosen": -288.5440368652344, | |
| "logps/rejected": -317.422607421875, | |
| "loss": 0.6966, | |
| "rewards/accuracies": 0.48846152424812317, | |
| "rewards/chosen": -0.29237106442451477, | |
| "rewards/margins": 0.041345465928316116, | |
| "rewards/rejected": -0.3337165117263794, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 51.989784907303765, | |
| "learning_rate": 4.901488388458247e-07, | |
| "logits/chosen": 0.08469453454017639, | |
| "logits/rejected": 0.02361711673438549, | |
| "logps/chosen": -257.55908203125, | |
| "logps/rejected": -290.6418151855469, | |
| "loss": 0.6904, | |
| "rewards/accuracies": 0.5153846144676208, | |
| "rewards/chosen": 0.0216812863945961, | |
| "rewards/margins": 0.06838896870613098, | |
| "rewards/rejected": -0.04670768231153488, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 48.07967092856285, | |
| "learning_rate": 4.695964991097616e-07, | |
| "logits/chosen": 0.5012978315353394, | |
| "logits/rejected": 0.3107348382472992, | |
| "logps/chosen": -293.4065856933594, | |
| "logps/rejected": -318.2868347167969, | |
| "loss": 0.6923, | |
| "rewards/accuracies": 0.5153846144676208, | |
| "rewards/chosen": -0.3082655370235443, | |
| "rewards/margins": 0.04468757286667824, | |
| "rewards/rejected": -0.35295310616493225, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 50.8642865419175, | |
| "learning_rate": 4.3877607113930516e-07, | |
| "logits/chosen": 0.5789575576782227, | |
| "logits/rejected": 0.747968852519989, | |
| "logps/chosen": -291.36279296875, | |
| "logps/rejected": -306.2597961425781, | |
| "loss": 0.696, | |
| "rewards/accuracies": 0.5269230604171753, | |
| "rewards/chosen": -0.29860785603523254, | |
| "rewards/margins": 0.016954706981778145, | |
| "rewards/rejected": -0.31556254625320435, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 42.68053454562829, | |
| "learning_rate": 3.991286838919086e-07, | |
| "logits/chosen": 0.5176121592521667, | |
| "logits/rejected": 0.48673737049102783, | |
| "logps/chosen": -285.60284423828125, | |
| "logps/rejected": -301.834228515625, | |
| "loss": 0.6896, | |
| "rewards/accuracies": 0.4961538314819336, | |
| "rewards/chosen": -0.20814552903175354, | |
| "rewards/margins": 0.02328580990433693, | |
| "rewards/rejected": -0.23143133521080017, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 37.59954622299739, | |
| "learning_rate": 3.52508205130354e-07, | |
| "logits/chosen": 0.47016510367393494, | |
| "logits/rejected": 0.6350060105323792, | |
| "logps/chosen": -298.3149719238281, | |
| "logps/rejected": -311.8184509277344, | |
| "loss": 0.6954, | |
| "rewards/accuracies": 0.5423076748847961, | |
| "rewards/chosen": -0.33209383487701416, | |
| "rewards/margins": 0.01967799849808216, | |
| "rewards/rejected": -0.35177183151245117, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 40.19774755409481, | |
| "learning_rate": 3.010945566265912e-07, | |
| "logits/chosen": 0.8041943311691284, | |
| "logits/rejected": 0.9286781549453735, | |
| "logps/chosen": -320.7852783203125, | |
| "logps/rejected": -339.48193359375, | |
| "loss": 0.6856, | |
| "rewards/accuracies": 0.557692289352417, | |
| "rewards/chosen": -0.5020374655723572, | |
| "rewards/margins": 0.013489325530827045, | |
| "rewards/rejected": -0.5155267715454102, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 40.3298911710128, | |
| "learning_rate": 2.4729178344249006e-07, | |
| "logits/chosen": 0.526244044303894, | |
| "logits/rejected": 0.5612362027168274, | |
| "logps/chosen": -289.9747009277344, | |
| "logps/rejected": -304.9601135253906, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.5230769515037537, | |
| "rewards/chosen": -0.2381971925497055, | |
| "rewards/margins": 0.04978089779615402, | |
| "rewards/rejected": -0.2879781126976013, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 35.40524465285595, | |
| "learning_rate": 1.9361564345465145e-07, | |
| "logits/chosen": 0.3361697196960449, | |
| "logits/rejected": 0.5479218363761902, | |
| "logps/chosen": -272.99285888671875, | |
| "logps/rejected": -300.7643737792969, | |
| "loss": 0.6878, | |
| "rewards/accuracies": 0.5153846144676208, | |
| "rewards/chosen": -0.1911364644765854, | |
| "rewards/margins": 0.03422596678137779, | |
| "rewards/rejected": -0.22536242008209229, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 38.98659700871813, | |
| "learning_rate": 1.4257597331216208e-07, | |
| "logits/chosen": 0.6074225902557373, | |
| "logits/rejected": 0.7285165786743164, | |
| "logps/chosen": -311.04827880859375, | |
| "logps/rejected": -335.56396484375, | |
| "loss": 0.685, | |
| "rewards/accuracies": 0.5423076748847961, | |
| "rewards/chosen": -0.4663804769515991, | |
| "rewards/margins": 0.05671105906367302, | |
| "rewards/rejected": -0.5230914950370789, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 50.7969521151244, | |
| "learning_rate": 9.655933126436563e-08, | |
| "logits/chosen": 0.5686596035957336, | |
| "logits/rejected": 0.6905936002731323, | |
| "logps/chosen": -276.1257629394531, | |
| "logps/rejected": -292.55120849609375, | |
| "loss": 0.7065, | |
| "rewards/accuracies": 0.5538461804389954, | |
| "rewards/chosen": -0.14772899448871613, | |
| "rewards/margins": 0.05048359930515289, | |
| "rewards/rejected": -0.19821257889270782, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 44.99508540990744, | |
| "learning_rate": 5.771740434959277e-08, | |
| "logits/chosen": 0.7891207337379456, | |
| "logits/rejected": 0.6963477730751038, | |
| "logps/chosen": -289.924072265625, | |
| "logps/rejected": -315.6805419921875, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.2571752965450287, | |
| "rewards/margins": 0.07165656983852386, | |
| "rewards/rejected": -0.32883188128471375, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 41.07192587696782, | |
| "learning_rate": 2.7866397900677185e-08, | |
| "logits/chosen": 0.728725016117096, | |
| "logits/rejected": 0.6798302531242371, | |
| "logps/chosen": -313.1536865234375, | |
| "logps/rejected": -325.9017639160156, | |
| "loss": 0.686, | |
| "rewards/accuracies": 0.5423076748847961, | |
| "rewards/chosen": -0.5450281500816345, | |
| "rewards/margins": 0.03127431869506836, | |
| "rewards/rejected": -0.5763024687767029, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 51.88495068918687, | |
| "learning_rate": 8.402111802159412e-09, | |
| "logits/chosen": 0.7722111344337463, | |
| "logits/rejected": 0.8273798227310181, | |
| "logps/chosen": -296.1044921875, | |
| "logps/rejected": -326.81695556640625, | |
| "loss": 0.6812, | |
| "rewards/accuracies": 0.607692301273346, | |
| "rewards/chosen": -0.4110731780529022, | |
| "rewards/margins": 0.07831522077322006, | |
| "rewards/rejected": -0.4893884062767029, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 43.53404843175348, | |
| "learning_rate": 2.3467443900582197e-10, | |
| "logits/chosen": 0.9644160866737366, | |
| "logits/rejected": 1.0424695014953613, | |
| "logps/chosen": -288.78082275390625, | |
| "logps/rejected": -310.8374328613281, | |
| "loss": 0.685, | |
| "rewards/accuracies": 0.5346153974533081, | |
| "rewards/chosen": -0.3611108064651489, | |
| "rewards/margins": 0.10043878108263016, | |
| "rewards/rejected": -0.4615496098995209, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 162, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6911558170377472, | |
| "train_runtime": 23474.6733, | |
| "train_samples_per_second": 0.9, | |
| "train_steps_per_second": 0.007 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 162, | |
| "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 | |
| } | |