| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.2140127388535032, | |
| "eval_steps": 500, | |
| "global_step": 21, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01019108280254777, | |
| "grad_norm": 1.9706840515136719, | |
| "learning_rate": 8e-05, | |
| "log_odds_chosen": 0.043639570474624634, | |
| "log_odds_ratio": -0.6778613328933716, | |
| "logits/chosen": -0.5009905099868774, | |
| "logits/rejected": -0.5356172919273376, | |
| "logps/chosen": -2.954733371734619, | |
| "logps/rejected": -2.996232509613037, | |
| "loss": 3.6057, | |
| "nll_loss": 3.5378708839416504, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.2954733371734619, | |
| "rewards/margins": 0.00414988212287426, | |
| "rewards/rejected": -0.2996232211589813, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02038216560509554, | |
| "grad_norm": 1.9629229307174683, | |
| "learning_rate": 6e-05, | |
| "log_odds_chosen": 0.10935741662979126, | |
| "log_odds_ratio": -0.6454683542251587, | |
| "logits/chosen": -0.5241304636001587, | |
| "logits/rejected": -0.5285190939903259, | |
| "logps/chosen": -2.8890933990478516, | |
| "logps/rejected": -2.9919257164001465, | |
| "loss": 3.4851, | |
| "nll_loss": 3.420555830001831, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.28890931606292725, | |
| "rewards/margins": 0.010283238254487514, | |
| "rewards/rejected": -0.2991925776004791, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.030573248407643312, | |
| "grad_norm": 2.020477533340454, | |
| "learning_rate": 4e-05, | |
| "log_odds_chosen": 0.069330133497715, | |
| "log_odds_ratio": -0.6641735434532166, | |
| "logits/chosen": -0.4971535801887512, | |
| "logits/rejected": -0.5353123545646667, | |
| "logps/chosen": -2.731393575668335, | |
| "logps/rejected": -2.795962333679199, | |
| "loss": 3.3144, | |
| "nll_loss": 3.2479400634765625, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.2731393277645111, | |
| "rewards/margins": 0.006456875707954168, | |
| "rewards/rejected": -0.279596209526062, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.04076433121019108, | |
| "grad_norm": 2.3341853618621826, | |
| "learning_rate": 2e-05, | |
| "log_odds_chosen": 0.12229391932487488, | |
| "log_odds_ratio": -0.6381683945655823, | |
| "logits/chosen": -0.4815675914287567, | |
| "logits/rejected": -0.5111523866653442, | |
| "logps/chosen": -2.8219356536865234, | |
| "logps/rejected": -2.9366464614868164, | |
| "loss": 3.3456, | |
| "nll_loss": 3.281827211380005, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.28219354152679443, | |
| "rewards/margins": 0.011471111327409744, | |
| "rewards/rejected": -0.2936646640300751, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.050955414012738856, | |
| "grad_norm": 2.1892218589782715, | |
| "learning_rate": 0.0, | |
| "log_odds_chosen": 0.06512074917554855, | |
| "log_odds_ratio": -0.6671069860458374, | |
| "logits/chosen": -0.5229994654655457, | |
| "logits/rejected": -0.5436381697654724, | |
| "logps/chosen": -2.7552542686462402, | |
| "logps/rejected": -2.816582679748535, | |
| "loss": 3.2584, | |
| "nll_loss": 3.191640615463257, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.27552542090415955, | |
| "rewards/margins": 0.006132860202342272, | |
| "rewards/rejected": -0.2816582918167114, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.061146496815286625, | |
| "grad_norm": 1.755154013633728, | |
| "learning_rate": 9.79591836734694e-05, | |
| "log_odds_chosen": 0.09281320869922638, | |
| "log_odds_ratio": -0.6549628973007202, | |
| "logits/chosen": -0.5153884291648865, | |
| "logits/rejected": -0.5407605767250061, | |
| "logps/chosen": -2.8005495071411133, | |
| "logps/rejected": -2.8862051963806152, | |
| "loss": 3.2371, | |
| "nll_loss": 3.171649217605591, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.2800549566745758, | |
| "rewards/margins": 0.008565555326640606, | |
| "rewards/rejected": -0.2886205315589905, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.07133757961783439, | |
| "grad_norm": 1.8428237438201904, | |
| "learning_rate": 9.761904761904762e-05, | |
| "log_odds_chosen": 0.03281838819384575, | |
| "log_odds_ratio": -0.6809508800506592, | |
| "logits/chosen": -0.5399191975593567, | |
| "logits/rejected": -0.5274120569229126, | |
| "logps/chosen": -2.9202024936676025, | |
| "logps/rejected": -2.9510116577148438, | |
| "loss": 3.3829, | |
| "nll_loss": 3.314777374267578, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.2920202612876892, | |
| "rewards/margins": 0.003080902621150017, | |
| "rewards/rejected": -0.2951011657714844, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.08152866242038216, | |
| "grad_norm": 1.8741809129714966, | |
| "learning_rate": 9.727891156462585e-05, | |
| "log_odds_chosen": 0.09863043576478958, | |
| "log_odds_ratio": -0.6496734619140625, | |
| "logits/chosen": -0.5901607275009155, | |
| "logits/rejected": -0.5661267042160034, | |
| "logps/chosen": -2.7296760082244873, | |
| "logps/rejected": -2.8212223052978516, | |
| "loss": 3.1928, | |
| "nll_loss": 3.1278655529022217, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.2729676365852356, | |
| "rewards/margins": 0.009154656901955605, | |
| "rewards/rejected": -0.28212225437164307, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.09171974522292993, | |
| "grad_norm": 1.9783447980880737, | |
| "learning_rate": 9.693877551020408e-05, | |
| "log_odds_chosen": 0.05389028042554855, | |
| "log_odds_ratio": -0.6711492538452148, | |
| "logits/chosen": -0.5444136261940002, | |
| "logits/rejected": -0.564105212688446, | |
| "logps/chosen": -2.688671112060547, | |
| "logps/rejected": -2.7393460273742676, | |
| "loss": 3.1182, | |
| "nll_loss": 3.051093101501465, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.2688670754432678, | |
| "rewards/margins": 0.005067505408078432, | |
| "rewards/rejected": -0.27393460273742676, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.10191082802547771, | |
| "grad_norm": 3.7560782432556152, | |
| "learning_rate": 9.659863945578231e-05, | |
| "log_odds_chosen": 0.05389935150742531, | |
| "log_odds_ratio": -0.6712656617164612, | |
| "logits/chosen": -0.5685052871704102, | |
| "logits/rejected": -0.5835959911346436, | |
| "logps/chosen": -2.7232255935668945, | |
| "logps/rejected": -2.772873878479004, | |
| "loss": 3.1185, | |
| "nll_loss": 3.0513598918914795, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.27232253551483154, | |
| "rewards/margins": 0.004964826162904501, | |
| "rewards/rejected": -0.27728739380836487, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.11210191082802548, | |
| "grad_norm": 2.745220422744751, | |
| "learning_rate": 9.625850340136054e-05, | |
| "log_odds_chosen": 0.0562276728451252, | |
| "log_odds_ratio": -0.6704986095428467, | |
| "logits/chosen": -0.5921126008033752, | |
| "logits/rejected": -0.5893128514289856, | |
| "logps/chosen": -2.645206928253174, | |
| "logps/rejected": -2.696500539779663, | |
| "loss": 3.0034, | |
| "nll_loss": 2.936365842819214, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.26452070474624634, | |
| "rewards/margins": 0.005129385739564896, | |
| "rewards/rejected": -0.26965007185935974, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.12229299363057325, | |
| "grad_norm": 1.8458824157714844, | |
| "learning_rate": 9.591836734693878e-05, | |
| "log_odds_chosen": -0.026220712810754776, | |
| "log_odds_ratio": -0.7125887870788574, | |
| "logits/chosen": -0.5802856087684631, | |
| "logits/rejected": -0.6202648878097534, | |
| "logps/chosen": -2.6887969970703125, | |
| "logps/rejected": -2.6625289916992188, | |
| "loss": 3.0166, | |
| "nll_loss": 2.945356845855713, | |
| "rewards/accuracies": 0.40625, | |
| "rewards/chosen": -0.2688797116279602, | |
| "rewards/margins": -0.0026267748326063156, | |
| "rewards/rejected": -0.26625293493270874, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.13248407643312102, | |
| "grad_norm": 1.8597255945205688, | |
| "learning_rate": 9.557823129251701e-05, | |
| "log_odds_chosen": 0.0966324731707573, | |
| "log_odds_ratio": -0.6508103013038635, | |
| "logits/chosen": -0.5836719274520874, | |
| "logits/rejected": -0.6042653322219849, | |
| "logps/chosen": -2.5945496559143066, | |
| "logps/rejected": -2.6839816570281982, | |
| "loss": 2.891, | |
| "nll_loss": 2.8259291648864746, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.25945499539375305, | |
| "rewards/margins": 0.008943157270550728, | |
| "rewards/rejected": -0.26839813590049744, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.14267515923566879, | |
| "grad_norm": 1.9950586557388306, | |
| "learning_rate": 9.523809523809524e-05, | |
| "log_odds_chosen": 0.07333967834711075, | |
| "log_odds_ratio": -0.6641579270362854, | |
| "logits/chosen": -0.5741627812385559, | |
| "logits/rejected": -0.6016291379928589, | |
| "logps/chosen": -2.556469678878784, | |
| "logps/rejected": -2.6230132579803467, | |
| "loss": 2.8169, | |
| "nll_loss": 2.7504758834838867, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.2556469738483429, | |
| "rewards/margins": 0.006654377095401287, | |
| "rewards/rejected": -0.26230132579803467, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.15286624203821655, | |
| "grad_norm": 2.0696280002593994, | |
| "learning_rate": 9.489795918367348e-05, | |
| "log_odds_chosen": 0.12293906509876251, | |
| "log_odds_ratio": -0.6370925903320312, | |
| "logits/chosen": -0.5484431385993958, | |
| "logits/rejected": -0.5725057721138, | |
| "logps/chosen": -2.460176944732666, | |
| "logps/rejected": -2.5726397037506104, | |
| "loss": 2.7249, | |
| "nll_loss": 2.661222219467163, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.2460177093744278, | |
| "rewards/margins": 0.011246275156736374, | |
| "rewards/rejected": -0.25726398825645447, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.16305732484076432, | |
| "grad_norm": 3.213852882385254, | |
| "learning_rate": 9.455782312925171e-05, | |
| "log_odds_chosen": 0.0818968415260315, | |
| "log_odds_ratio": -0.6613866090774536, | |
| "logits/chosen": -0.5976958274841309, | |
| "logits/rejected": -0.6143887639045715, | |
| "logps/chosen": -2.413341522216797, | |
| "logps/rejected": -2.4865636825561523, | |
| "loss": 2.6428, | |
| "nll_loss": 2.5766918659210205, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.24133417010307312, | |
| "rewards/margins": 0.007322215940803289, | |
| "rewards/rejected": -0.24865636229515076, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.1732484076433121, | |
| "grad_norm": 3.263835906982422, | |
| "learning_rate": 9.421768707482994e-05, | |
| "log_odds_chosen": 0.12279009073972702, | |
| "log_odds_ratio": -0.6375764608383179, | |
| "logits/chosen": -0.5939255952835083, | |
| "logits/rejected": -0.5961549878120422, | |
| "logps/chosen": -2.3610243797302246, | |
| "logps/rejected": -2.47149395942688, | |
| "loss": 2.568, | |
| "nll_loss": 2.5042312145233154, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.23610244691371918, | |
| "rewards/margins": 0.011046944186091423, | |
| "rewards/rejected": -0.24714939296245575, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.18343949044585986, | |
| "grad_norm": 2.0416226387023926, | |
| "learning_rate": 9.387755102040817e-05, | |
| "log_odds_chosen": 0.1555405557155609, | |
| "log_odds_ratio": -0.6227366328239441, | |
| "logits/chosen": -0.586774468421936, | |
| "logits/rejected": -0.6310256123542786, | |
| "logps/chosen": -2.2520110607147217, | |
| "logps/rejected": -2.390634775161743, | |
| "loss": 2.4749, | |
| "nll_loss": 2.412618637084961, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.22520111501216888, | |
| "rewards/margins": 0.013862377032637596, | |
| "rewards/rejected": -0.23906347155570984, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.19363057324840766, | |
| "grad_norm": 1.8552919626235962, | |
| "learning_rate": 9.35374149659864e-05, | |
| "log_odds_chosen": 0.1357956975698471, | |
| "log_odds_ratio": -0.6316878199577332, | |
| "logits/chosen": -0.6498819589614868, | |
| "logits/rejected": -0.6668039560317993, | |
| "logps/chosen": -2.22139835357666, | |
| "logps/rejected": -2.3428542613983154, | |
| "loss": 2.3994, | |
| "nll_loss": 2.336277723312378, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.22213983535766602, | |
| "rewards/margins": 0.012145599350333214, | |
| "rewards/rejected": -0.23428542912006378, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.20382165605095542, | |
| "grad_norm": 1.9783782958984375, | |
| "learning_rate": 9.319727891156463e-05, | |
| "log_odds_chosen": 0.09560593217611313, | |
| "log_odds_ratio": -0.6527409553527832, | |
| "logits/chosen": -0.6574750542640686, | |
| "logits/rejected": -0.6695252656936646, | |
| "logps/chosen": -2.228286027908325, | |
| "logps/rejected": -2.3149209022521973, | |
| "loss": 2.4197, | |
| "nll_loss": 2.3543879985809326, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.22282862663269043, | |
| "rewards/margins": 0.008663482964038849, | |
| "rewards/rejected": -0.23149211704730988, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2140127388535032, | |
| "grad_norm": 2.270171880722046, | |
| "learning_rate": 9.285714285714286e-05, | |
| "log_odds_chosen": 0.13292132318019867, | |
| "log_odds_ratio": -0.6331073641777039, | |
| "logits/chosen": -0.6325170993804932, | |
| "logits/rejected": -0.6448156237602234, | |
| "logps/chosen": -2.0987019538879395, | |
| "logps/rejected": -2.2161264419555664, | |
| "loss": 2.2766, | |
| "nll_loss": 2.2132651805877686, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.20987018942832947, | |
| "rewards/margins": 0.01174245961010456, | |
| "rewards/rejected": -0.22161263227462769, | |
| "step": 21 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 294, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 1, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |