{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.049766718506998445, "eval_steps": 500, "global_step": 40, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.0, "learning_rate": 5.000000000000001e-07, "logits/chosen": -2.0463106632232666, "logits/rejected": -1.8486888408660889, "logps/chosen": -311.57489013671875, "logps/rejected": -415.468505859375, "loss": 0.671, "rewards/accuracies": 0.75, "rewards/chosen": -0.021808436140418053, "rewards/margins": 0.04626159369945526, "rewards/rejected": -0.06807002425193787, "step": 1 }, { "epoch": 0.0, "learning_rate": 1.0000000000000002e-06, "logits/chosen": -2.060105562210083, "logits/rejected": -2.106396198272705, "logps/chosen": -220.380615234375, "logps/rejected": -404.501220703125, "loss": 0.6745, "rewards/accuracies": 0.6875, "rewards/chosen": -0.02021012455224991, "rewards/margins": 0.03883519023656845, "rewards/rejected": -0.05904531478881836, "step": 2 }, { "epoch": 0.0, "learning_rate": 1.5e-06, "logits/chosen": -2.239086151123047, "logits/rejected": -2.0575525760650635, "logps/chosen": -163.8677978515625, "logps/rejected": -446.79180908203125, "loss": 0.67, "rewards/accuracies": 0.6875, "rewards/chosen": -0.024322081357240677, "rewards/margins": 0.04956565052270889, "rewards/rejected": -0.07388772815465927, "step": 3 }, { "epoch": 0.0, "learning_rate": 2.0000000000000003e-06, "logits/chosen": -2.1058502197265625, "logits/rejected": -2.344322681427002, "logps/chosen": -318.9349670410156, "logps/rejected": -378.9231262207031, "loss": 0.6576, "rewards/accuracies": 0.9375, "rewards/chosen": -0.027724243700504303, "rewards/margins": 0.07309143990278244, "rewards/rejected": -0.10081568360328674, "step": 4 }, { "epoch": 0.01, "learning_rate": 2.5e-06, "logits/chosen": -2.287339687347412, "logits/rejected": -2.3221607208251953, "logps/chosen": -221.33163452148438, "logps/rejected": -388.6357421875, "loss": 0.6498, "rewards/accuracies": 0.8125, "rewards/chosen": -0.03215093910694122, "rewards/margins": 0.09096165001392365, "rewards/rejected": -0.12311258912086487, "step": 5 }, { "epoch": 0.01, "learning_rate": 3e-06, "logits/chosen": -1.98830246925354, "logits/rejected": -1.9178718328475952, "logps/chosen": -244.0532684326172, "logps/rejected": -373.9111328125, "loss": 0.6395, "rewards/accuracies": 0.8125, "rewards/chosen": -0.0789945125579834, "rewards/margins": 0.11430712044239044, "rewards/rejected": -0.19330161809921265, "step": 6 }, { "epoch": 0.01, "learning_rate": 3.5000000000000004e-06, "logits/chosen": -2.0385985374450684, "logits/rejected": -1.9476547241210938, "logps/chosen": -347.14227294921875, "logps/rejected": -426.24603271484375, "loss": 0.6148, "rewards/accuracies": 0.9375, "rewards/chosen": -0.0816906988620758, "rewards/margins": 0.1658005714416504, "rewards/rejected": -0.2474912703037262, "step": 7 }, { "epoch": 0.01, "learning_rate": 4.000000000000001e-06, "logits/chosen": -1.695872187614441, "logits/rejected": -1.7980302572250366, "logps/chosen": -287.9168395996094, "logps/rejected": -355.14129638671875, "loss": 0.5785, "rewards/accuracies": 1.0, "rewards/chosen": -0.07900090515613556, "rewards/margins": 0.25094836950302124, "rewards/rejected": -0.329949289560318, "step": 8 }, { "epoch": 0.01, "learning_rate": 4.5e-06, "logits/chosen": -1.9222339391708374, "logits/rejected": -1.8972643613815308, "logps/chosen": -211.97979736328125, "logps/rejected": -405.848388671875, "loss": 0.5229, "rewards/accuracies": 1.0, "rewards/chosen": -0.09827227890491486, "rewards/margins": 0.3891267478466034, "rewards/rejected": -0.48739901185035706, "step": 9 }, { "epoch": 0.01, "learning_rate": 5e-06, "logits/chosen": -1.9703621864318848, "logits/rejected": -1.7560880184173584, "logps/chosen": -277.2589416503906, "logps/rejected": -445.6544494628906, "loss": 0.4853, "rewards/accuracies": 1.0, "rewards/chosen": -0.1426083892583847, "rewards/margins": 0.4861423373222351, "rewards/rejected": -0.628750741481781, "step": 10 }, { "epoch": 0.01, "learning_rate": 5.500000000000001e-06, "logits/chosen": -2.0542986392974854, "logits/rejected": -2.110658884048462, "logps/chosen": -283.9775390625, "logps/rejected": -394.361572265625, "loss": 0.4476, "rewards/accuracies": 1.0, "rewards/chosen": -0.14311671257019043, "rewards/margins": 0.6046640872955322, "rewards/rejected": -0.7477807998657227, "step": 11 }, { "epoch": 0.01, "learning_rate": 6e-06, "logits/chosen": -2.0650949478149414, "logits/rejected": -2.0736098289489746, "logps/chosen": -235.07675170898438, "logps/rejected": -448.17919921875, "loss": 0.3701, "rewards/accuracies": 1.0, "rewards/chosen": -0.23901399970054626, "rewards/margins": 0.9068580269813538, "rewards/rejected": -1.1458721160888672, "step": 12 }, { "epoch": 0.02, "learning_rate": 6.5000000000000004e-06, "logits/chosen": -2.011099100112915, "logits/rejected": -2.106595754623413, "logps/chosen": -289.19970703125, "logps/rejected": -417.64208984375, "loss": 0.3675, "rewards/accuracies": 1.0, "rewards/chosen": -0.26357388496398926, "rewards/margins": 0.8955046534538269, "rewards/rejected": -1.159078598022461, "step": 13 }, { "epoch": 0.02, "learning_rate": 7.000000000000001e-06, "logits/chosen": -2.1066994667053223, "logits/rejected": -1.7827305793762207, "logps/chosen": -169.27044677734375, "logps/rejected": -446.7305908203125, "loss": 0.254, "rewards/accuracies": 1.0, "rewards/chosen": -0.2715146541595459, "rewards/margins": 1.3259609937667847, "rewards/rejected": -1.5974756479263306, "step": 14 }, { "epoch": 0.02, "learning_rate": 7.5e-06, "logits/chosen": -1.9394950866699219, "logits/rejected": -1.9831421375274658, "logps/chosen": -194.98016357421875, "logps/rejected": -470.9692687988281, "loss": 0.2298, "rewards/accuracies": 1.0, "rewards/chosen": -0.4095136225223541, "rewards/margins": 1.628616452217102, "rewards/rejected": -2.038130044937134, "step": 15 }, { "epoch": 0.02, "learning_rate": 8.000000000000001e-06, "logits/chosen": -1.7606948614120483, "logits/rejected": -2.0152878761291504, "logps/chosen": -253.0347900390625, "logps/rejected": -428.3993835449219, "loss": 0.2528, "rewards/accuracies": 1.0, "rewards/chosen": -0.581798255443573, "rewards/margins": 1.3869069814682007, "rewards/rejected": -1.968705177307129, "step": 16 }, { "epoch": 0.02, "learning_rate": 8.500000000000002e-06, "logits/chosen": -2.0609982013702393, "logits/rejected": -1.9835376739501953, "logps/chosen": -265.3518371582031, "logps/rejected": -501.8542785644531, "loss": 0.1997, "rewards/accuracies": 1.0, "rewards/chosen": -0.681609570980072, "rewards/margins": 2.0102298259735107, "rewards/rejected": -2.6918392181396484, "step": 17 }, { "epoch": 0.02, "learning_rate": 9e-06, "logits/chosen": -2.117746114730835, "logits/rejected": -2.218081474304199, "logps/chosen": -202.6907501220703, "logps/rejected": -413.13604736328125, "loss": 0.1582, "rewards/accuracies": 1.0, "rewards/chosen": -0.6984068751335144, "rewards/margins": 1.8970444202423096, "rewards/rejected": -2.5954513549804688, "step": 18 }, { "epoch": 0.02, "learning_rate": 9.5e-06, "logits/chosen": -2.385519504547119, "logits/rejected": -1.8222811222076416, "logps/chosen": -175.93809509277344, "logps/rejected": -469.17657470703125, "loss": 0.0951, "rewards/accuracies": 1.0, "rewards/chosen": -0.7677481174468994, "rewards/margins": 2.7113454341888428, "rewards/rejected": -3.479093551635742, "step": 19 }, { "epoch": 0.02, "learning_rate": 1e-05, "logits/chosen": -2.187441349029541, "logits/rejected": -2.1218032836914062, "logps/chosen": -301.7166442871094, "logps/rejected": -490.854736328125, "loss": 0.0944, "rewards/accuracies": 1.0, "rewards/chosen": -1.2999629974365234, "rewards/margins": 3.2447361946105957, "rewards/rejected": -4.544699192047119, "step": 20 }, { "epoch": 0.03, "learning_rate": 1.05e-05, "logits/chosen": -1.967098355293274, "logits/rejected": -1.9263262748718262, "logps/chosen": -176.0479278564453, "logps/rejected": -472.8221435546875, "loss": 0.0459, "rewards/accuracies": 1.0, "rewards/chosen": -0.8402897119522095, "rewards/margins": 4.1183624267578125, "rewards/rejected": -4.958651542663574, "step": 21 }, { "epoch": 0.03, "learning_rate": 1.1000000000000001e-05, "logits/chosen": -2.188459873199463, "logits/rejected": -2.156071424484253, "logps/chosen": -243.41683959960938, "logps/rejected": -465.08685302734375, "loss": 0.1389, "rewards/accuracies": 0.9375, "rewards/chosen": -1.5002142190933228, "rewards/margins": 4.108778953552246, "rewards/rejected": -5.608993053436279, "step": 22 }, { "epoch": 0.03, "learning_rate": 1.1500000000000002e-05, "logits/chosen": -1.7361280918121338, "logits/rejected": -1.8734406232833862, "logps/chosen": -274.4007568359375, "logps/rejected": -465.7174987792969, "loss": 0.074, "rewards/accuracies": 1.0, "rewards/chosen": -1.6840157508850098, "rewards/margins": 4.5255866050720215, "rewards/rejected": -6.209602355957031, "step": 23 }, { "epoch": 0.03, "learning_rate": 1.2e-05, "logits/chosen": -1.8966342210769653, "logits/rejected": -1.772972583770752, "logps/chosen": -275.3910217285156, "logps/rejected": -462.07720947265625, "loss": 0.0381, "rewards/accuracies": 1.0, "rewards/chosen": -2.017077922821045, "rewards/margins": 4.740576267242432, "rewards/rejected": -6.757654666900635, "step": 24 }, { "epoch": 0.03, "learning_rate": 1.25e-05, "logits/chosen": -2.2573952674865723, "logits/rejected": -1.9883058071136475, "logps/chosen": -172.16043090820312, "logps/rejected": -391.4390563964844, "loss": 0.0666, "rewards/accuracies": 1.0, "rewards/chosen": -1.9856681823730469, "rewards/margins": 4.858271598815918, "rewards/rejected": -6.843939781188965, "step": 25 }, { "epoch": 0.03, "learning_rate": 1.3000000000000001e-05, "logits/chosen": -2.0475640296936035, "logits/rejected": -1.8690011501312256, "logps/chosen": -184.4323272705078, "logps/rejected": -468.86175537109375, "loss": 0.0805, "rewards/accuracies": 0.9375, "rewards/chosen": -2.08463716506958, "rewards/margins": 6.207521438598633, "rewards/rejected": -8.292158126831055, "step": 26 }, { "epoch": 0.03, "learning_rate": 1.3500000000000001e-05, "logits/chosen": -1.9634509086608887, "logits/rejected": -2.260413646697998, "logps/chosen": -398.9476013183594, "logps/rejected": -540.3334350585938, "loss": 0.029, "rewards/accuracies": 1.0, "rewards/chosen": -3.3595480918884277, "rewards/margins": 6.726611614227295, "rewards/rejected": -10.086159706115723, "step": 27 }, { "epoch": 0.03, "learning_rate": 1.4000000000000001e-05, "logits/chosen": -1.9712620973587036, "logits/rejected": -1.827807068824768, "logps/chosen": -221.26559448242188, "logps/rejected": -483.7889709472656, "loss": 0.0063, "rewards/accuracies": 1.0, "rewards/chosen": -2.3635034561157227, "rewards/margins": 7.733745098114014, "rewards/rejected": -10.097248077392578, "step": 28 }, { "epoch": 0.04, "learning_rate": 1.45e-05, "logits/chosen": -1.542866587638855, "logits/rejected": -1.7318823337554932, "logps/chosen": -267.33880615234375, "logps/rejected": -558.8424072265625, "loss": 0.0746, "rewards/accuracies": 0.9375, "rewards/chosen": -3.391514301300049, "rewards/margins": 9.136863708496094, "rewards/rejected": -12.528377532958984, "step": 29 }, { "epoch": 0.04, "learning_rate": 1.5e-05, "logits/chosen": -2.037372589111328, "logits/rejected": -1.8312575817108154, "logps/chosen": -288.1588134765625, "logps/rejected": -670.7269287109375, "loss": 0.0007, "rewards/accuracies": 1.0, "rewards/chosen": -3.795253276824951, "rewards/margins": 11.858222961425781, "rewards/rejected": -15.653478622436523, "step": 30 }, { "epoch": 0.04, "learning_rate": 1.55e-05, "logits/chosen": -1.5374680757522583, "logits/rejected": -1.8392678499221802, "logps/chosen": -281.0010070800781, "logps/rejected": -553.43017578125, "loss": 0.0149, "rewards/accuracies": 1.0, "rewards/chosen": -3.836461067199707, "rewards/margins": 9.380411148071289, "rewards/rejected": -13.216873168945312, "step": 31 }, { "epoch": 0.04, "learning_rate": 1.6000000000000003e-05, "logits/chosen": -2.033914566040039, "logits/rejected": -1.9804089069366455, "logps/chosen": -261.31103515625, "logps/rejected": -545.380126953125, "loss": 0.0123, "rewards/accuracies": 1.0, "rewards/chosen": -4.321026802062988, "rewards/margins": 9.96163272857666, "rewards/rejected": -14.282659530639648, "step": 32 }, { "epoch": 0.04, "learning_rate": 1.65e-05, "logits/chosen": -1.652747631072998, "logits/rejected": -1.6527116298675537, "logps/chosen": -250.82638549804688, "logps/rejected": -536.0584106445312, "loss": 0.0017, "rewards/accuracies": 1.0, "rewards/chosen": -4.393105506896973, "rewards/margins": 11.006621360778809, "rewards/rejected": -15.399725914001465, "step": 33 }, { "epoch": 0.04, "learning_rate": 1.7000000000000003e-05, "logits/chosen": -1.5870184898376465, "logits/rejected": -1.9295532703399658, "logps/chosen": -276.73248291015625, "logps/rejected": -588.6087646484375, "loss": 0.2671, "rewards/accuracies": 0.875, "rewards/chosen": -5.017908573150635, "rewards/margins": 12.34096908569336, "rewards/rejected": -17.35887908935547, "step": 34 }, { "epoch": 0.04, "learning_rate": 1.75e-05, "logits/chosen": -1.9227735996246338, "logits/rejected": -1.954828143119812, "logps/chosen": -266.1369323730469, "logps/rejected": -550.742919921875, "loss": 0.0035, "rewards/accuracies": 1.0, "rewards/chosen": -4.8841400146484375, "rewards/margins": 13.669473648071289, "rewards/rejected": -18.553613662719727, "step": 35 }, { "epoch": 0.04, "learning_rate": 1.8e-05, "logits/chosen": -1.6637210845947266, "logits/rejected": -1.3936012983322144, "logps/chosen": -256.7511901855469, "logps/rejected": -591.1787109375, "loss": 0.0022, "rewards/accuracies": 1.0, "rewards/chosen": -6.027996063232422, "rewards/margins": 13.286176681518555, "rewards/rejected": -19.314172744750977, "step": 36 }, { "epoch": 0.05, "learning_rate": 1.85e-05, "logits/chosen": -1.6170068979263306, "logits/rejected": -1.6388572454452515, "logps/chosen": -257.0255126953125, "logps/rejected": -605.7335205078125, "loss": 0.0008, "rewards/accuracies": 1.0, "rewards/chosen": -5.929409027099609, "rewards/margins": 14.035564422607422, "rewards/rejected": -19.9649715423584, "step": 37 }, { "epoch": 0.05, "learning_rate": 1.9e-05, "logits/chosen": -1.967097520828247, "logits/rejected": -1.778485894203186, "logps/chosen": -239.90533447265625, "logps/rejected": -570.31787109375, "loss": 0.0839, "rewards/accuracies": 0.9375, "rewards/chosen": -6.8724775314331055, "rewards/margins": 13.892675399780273, "rewards/rejected": -20.765153884887695, "step": 38 }, { "epoch": 0.05, "learning_rate": 1.9500000000000003e-05, "logits/chosen": -1.6693722009658813, "logits/rejected": -1.573314905166626, "logps/chosen": -349.86480712890625, "logps/rejected": -702.0185546875, "loss": 0.0008, "rewards/accuracies": 1.0, "rewards/chosen": -7.373364448547363, "rewards/margins": 15.567668914794922, "rewards/rejected": -22.941036224365234, "step": 39 }, { "epoch": 0.05, "learning_rate": 2e-05, "logits/chosen": -1.556606650352478, "logits/rejected": -1.394392967224121, "logps/chosen": -213.81959533691406, "logps/rejected": -637.4287109375, "loss": 0.041, "rewards/accuracies": 1.0, "rewards/chosen": -4.418168067932129, "rewards/margins": 17.626686096191406, "rewards/rejected": -22.04485321044922, "step": 40 } ], "logging_steps": 1, "max_steps": 200, "num_train_epochs": 1, "save_steps": 40, "total_flos": 0.0, "trial_name": null, "trial_params": null }