| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984301412872841, | |
| "eval_steps": 100, | |
| "global_step": 477, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 6.25e-07, | |
| "logits/chosen": 0.2953376770019531, | |
| "logits/rejected": 0.3199407160282135, | |
| "logps/chosen": -295.2257080078125, | |
| "logps/rejected": -290.8670959472656, | |
| "loss": 0.1859, | |
| "rewards/accuracies": 0.4000000059604645, | |
| "rewards/chosen": 0.00015557199367322028, | |
| "rewards/margins": 0.0003326966252643615, | |
| "rewards/rejected": -0.00017712473345454782, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 1.25e-06, | |
| "logits/chosen": 0.26482802629470825, | |
| "logits/rejected": 0.33679550886154175, | |
| "logps/chosen": -318.49652099609375, | |
| "logps/rejected": -291.0345458984375, | |
| "loss": 0.1869, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.002573266625404358, | |
| "rewards/margins": 0.011565858498215675, | |
| "rewards/rejected": -0.014139125123620033, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 1.875e-06, | |
| "logits/chosen": 0.3993307650089264, | |
| "logits/rejected": 0.45834311842918396, | |
| "logps/chosen": -301.6546936035156, | |
| "logps/rejected": -295.427734375, | |
| "loss": 0.1703, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.05365907400846481, | |
| "rewards/margins": 0.07345211505889893, | |
| "rewards/rejected": -0.12711118161678314, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 2.5e-06, | |
| "logits/chosen": 0.35423707962036133, | |
| "logits/rejected": 0.4464724063873291, | |
| "logps/chosen": -333.11279296875, | |
| "logps/rejected": -334.0306091308594, | |
| "loss": 0.1389, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.2748720049858093, | |
| "rewards/margins": 0.14294137060642242, | |
| "rewards/rejected": -0.41781339049339294, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 2.999839121261416e-06, | |
| "logits/chosen": 0.35561248660087585, | |
| "logits/rejected": 0.3796083331108093, | |
| "logps/chosen": -374.3013916015625, | |
| "logps/rejected": -392.155517578125, | |
| "loss": 0.0917, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.6061999201774597, | |
| "rewards/margins": 0.2705405652523041, | |
| "rewards/rejected": -0.8767404556274414, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 2.994211988057582e-06, | |
| "logits/chosen": 0.17622683942317963, | |
| "logits/rejected": 0.313021719455719, | |
| "logps/chosen": -346.9971618652344, | |
| "logps/rejected": -349.66387939453125, | |
| "loss": 0.0903, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.5485752820968628, | |
| "rewards/margins": 0.28624793887138367, | |
| "rewards/rejected": -0.8348232507705688, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 2.9805753939568693e-06, | |
| "logits/chosen": 0.1983145773410797, | |
| "logits/rejected": 0.23167672753334045, | |
| "logps/chosen": -356.7235412597656, | |
| "logps/rejected": -386.5905456542969, | |
| "loss": 0.0815, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.5771406888961792, | |
| "rewards/margins": 0.28289300203323364, | |
| "rewards/rejected": -0.8600338101387024, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 2.959002435526626e-06, | |
| "logits/chosen": 0.35488566756248474, | |
| "logits/rejected": 0.4059143662452698, | |
| "logps/chosen": -371.07623291015625, | |
| "logps/rejected": -389.4164123535156, | |
| "loss": 0.0655, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.755893349647522, | |
| "rewards/margins": 0.5531964302062988, | |
| "rewards/rejected": -1.3090897798538208, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 2.929608750821129e-06, | |
| "logits/chosen": 0.2614609897136688, | |
| "logits/rejected": 0.4048990309238434, | |
| "logps/chosen": -387.1451110839844, | |
| "logps/rejected": -373.1450500488281, | |
| "loss": 0.0699, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.7152177095413208, | |
| "rewards/margins": 0.39531436562538147, | |
| "rewards/rejected": -1.1105320453643799, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 2.892551899524109e-06, | |
| "logits/chosen": 0.35849034786224365, | |
| "logits/rejected": 0.3978693187236786, | |
| "logps/chosen": -395.839599609375, | |
| "logps/rejected": -456.36016845703125, | |
| "loss": 0.0642, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -1.0831067562103271, | |
| "rewards/margins": 0.505523681640625, | |
| "rewards/rejected": -1.5886304378509521, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": 0.4300641715526581, | |
| "eval_logits/rejected": 0.45788079500198364, | |
| "eval_logps/chosen": -443.42120361328125, | |
| "eval_logps/rejected": -556.7063598632812, | |
| "eval_loss": 0.04867542162537575, | |
| "eval_rewards/accuracies": 0.71875, | |
| "eval_rewards/chosen": -1.3983652591705322, | |
| "eval_rewards/margins": 1.0524473190307617, | |
| "eval_rewards/rejected": -2.450812578201294, | |
| "eval_runtime": 73.3058, | |
| "eval_samples_per_second": 27.283, | |
| "eval_steps_per_second": 0.437, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 2.848030518377739e-06, | |
| "logits/chosen": 0.2927325367927551, | |
| "logits/rejected": 0.3830544054508209, | |
| "logps/chosen": -459.0276794433594, | |
| "logps/rejected": -510.6993103027344, | |
| "loss": 0.0405, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.4556727409362793, | |
| "rewards/margins": 0.8043819665908813, | |
| "rewards/rejected": -2.260054588317871, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 2.7962832564252724e-06, | |
| "logits/chosen": 0.15991483628749847, | |
| "logits/rejected": 0.2417908012866974, | |
| "logps/chosen": -477.8702087402344, | |
| "logps/rejected": -537.2095947265625, | |
| "loss": 0.0489, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.243250846862793, | |
| "rewards/margins": 0.6618646383285522, | |
| "rewards/rejected": -1.9051154851913452, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 2.7375874957747644e-06, | |
| "logits/chosen": 0.06199001148343086, | |
| "logits/rejected": 0.19189774990081787, | |
| "logps/chosen": -411.61602783203125, | |
| "logps/rejected": -418.01019287109375, | |
| "loss": 0.0665, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.8180139660835266, | |
| "rewards/margins": 0.39107373356819153, | |
| "rewards/rejected": -1.209087610244751, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 2.672257864741005e-06, | |
| "logits/chosen": 0.16753675043582916, | |
| "logits/rejected": 0.2094758003950119, | |
| "logps/chosen": -402.02484130859375, | |
| "logps/rejected": -429.792724609375, | |
| "loss": 0.0608, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.8829413652420044, | |
| "rewards/margins": 0.4748268127441406, | |
| "rewards/rejected": -1.3577684164047241, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 2.600644551335706e-06, | |
| "logits/chosen": 0.12230317294597626, | |
| "logits/rejected": 0.2780611515045166, | |
| "logps/chosen": -399.07012939453125, | |
| "logps/rejected": -447.2755432128906, | |
| "loss": 0.0409, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.1799192428588867, | |
| "rewards/margins": 0.7245117425918579, | |
| "rewards/rejected": -1.9044307470321655, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 2.5231314261461732e-06, | |
| "logits/chosen": 0.15223169326782227, | |
| "logits/rejected": 0.2353695183992386, | |
| "logps/chosen": -423.8455505371094, | |
| "logps/rejected": -493.4649353027344, | |
| "loss": 0.0439, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.3035107851028442, | |
| "rewards/margins": 0.7260502576828003, | |
| "rewards/rejected": -2.0295610427856445, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 2.440133984664454e-06, | |
| "logits/chosen": 0.07280103117227554, | |
| "logits/rejected": 0.15081152319908142, | |
| "logps/chosen": -472.6494140625, | |
| "logps/rejected": -495.68768310546875, | |
| "loss": 0.0448, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -1.1380369663238525, | |
| "rewards/margins": 0.558665931224823, | |
| "rewards/rejected": -1.6967029571533203, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 2.3520971200967337e-06, | |
| "logits/chosen": 0.1431509107351303, | |
| "logits/rejected": 0.12718148529529572, | |
| "logps/chosen": -421.31591796875, | |
| "logps/rejected": -481.0022888183594, | |
| "loss": 0.0398, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.277320146560669, | |
| "rewards/margins": 0.7172085642814636, | |
| "rewards/rejected": -1.9945287704467773, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 2.2594927385914546e-06, | |
| "logits/chosen": 0.1423339694738388, | |
| "logits/rejected": 0.20066264271736145, | |
| "logps/chosen": -385.1085205078125, | |
| "logps/rejected": -413.58660888671875, | |
| "loss": 0.044, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.1722856760025024, | |
| "rewards/margins": 0.5436893105506897, | |
| "rewards/rejected": -1.715975046157837, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 2.1628172296692954e-06, | |
| "logits/chosen": 0.08901094645261765, | |
| "logits/rejected": 0.13271097838878632, | |
| "logps/chosen": -385.1289367675781, | |
| "logps/rejected": -454.6835021972656, | |
| "loss": 0.047, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.0956504344940186, | |
| "rewards/margins": 0.6720448732376099, | |
| "rewards/rejected": -1.767695426940918, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": 0.11227821558713913, | |
| "eval_logits/rejected": 0.12742303311824799, | |
| "eval_logps/chosen": -415.04937744140625, | |
| "eval_logps/rejected": -496.232666015625, | |
| "eval_loss": 0.0460943840444088, | |
| "eval_rewards/accuracies": 0.7421875, | |
| "eval_rewards/chosen": -1.1146469116210938, | |
| "eval_rewards/margins": 0.7314289212226868, | |
| "eval_rewards/rejected": -1.8460756540298462, | |
| "eval_runtime": 73.4078, | |
| "eval_samples_per_second": 27.245, | |
| "eval_steps_per_second": 0.436, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 2.062588805414343e-06, | |
| "logits/chosen": -0.006222477648407221, | |
| "logits/rejected": 0.02312941662967205, | |
| "logps/chosen": -435.7500915527344, | |
| "logps/rejected": -491.56561279296875, | |
| "loss": 0.0379, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.1337292194366455, | |
| "rewards/margins": 0.7197057008743286, | |
| "rewards/rejected": -1.8534349203109741, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 1.9593447226892386e-06, | |
| "logits/chosen": -0.17459270358085632, | |
| "logits/rejected": -0.05563253164291382, | |
| "logps/chosen": -466.9278259277344, | |
| "logps/rejected": -490.50567626953125, | |
| "loss": 0.033, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.4599158763885498, | |
| "rewards/margins": 0.615450382232666, | |
| "rewards/rejected": -2.075366497039795, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 1.853638403264141e-06, | |
| "logits/chosen": -0.037995658814907074, | |
| "logits/rejected": -0.011381956748664379, | |
| "logps/chosen": -448.68939208984375, | |
| "logps/rejected": -532.2055053710938, | |
| "loss": 0.0365, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.6667238473892212, | |
| "rewards/margins": 0.8385307192802429, | |
| "rewards/rejected": -2.5052547454833984, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 1.7460364672965328e-06, | |
| "logits/chosen": -0.0721597820520401, | |
| "logits/rejected": -0.08736585080623627, | |
| "logps/chosen": -470.03045654296875, | |
| "logps/rejected": -542.5087280273438, | |
| "loss": 0.0484, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.167931318283081, | |
| "rewards/margins": 0.7865667939186096, | |
| "rewards/rejected": -1.954498052597046, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 1.637115696063402e-06, | |
| "logits/chosen": -0.07914544641971588, | |
| "logits/rejected": -0.05659206956624985, | |
| "logps/chosen": -430.75726318359375, | |
| "logps/rejected": -472.98284912109375, | |
| "loss": 0.0475, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.1496978998184204, | |
| "rewards/margins": 0.6665709018707275, | |
| "rewards/rejected": -1.8162685632705688, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 1.5274599402265162e-06, | |
| "logits/chosen": -0.23472614586353302, | |
| "logits/rejected": -0.11506778001785278, | |
| "logps/chosen": -493.5909118652344, | |
| "logps/rejected": -528.9447021484375, | |
| "loss": 0.0421, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.5179877281188965, | |
| "rewards/margins": 0.8330425024032593, | |
| "rewards/rejected": -2.3510303497314453, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 1.4176569902035088e-06, | |
| "logits/chosen": -0.17047128081321716, | |
| "logits/rejected": -0.09592470526695251, | |
| "logps/chosen": -456.50042724609375, | |
| "logps/rejected": -494.05059814453125, | |
| "loss": 0.0513, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.2792171239852905, | |
| "rewards/margins": 0.6969503164291382, | |
| "rewards/rejected": -1.9761674404144287, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 1.308295425420593e-06, | |
| "logits/chosen": -0.24125584959983826, | |
| "logits/rejected": -0.21810472011566162, | |
| "logps/chosen": -445.63916015625, | |
| "logps/rejected": -531.8533325195312, | |
| "loss": 0.0463, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.3838244676589966, | |
| "rewards/margins": 0.7514439225196838, | |
| "rewards/rejected": -2.135268211364746, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 1.1999614593359337e-06, | |
| "logits/chosen": -0.2246803492307663, | |
| "logits/rejected": -0.19805817306041718, | |
| "logps/chosen": -426.7688903808594, | |
| "logps/rejected": -472.63134765625, | |
| "loss": 0.0407, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.2982797622680664, | |
| "rewards/margins": 0.6589063405990601, | |
| "rewards/rejected": -1.9571861028671265, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.0932357971453745e-06, | |
| "logits/chosen": -0.28935036063194275, | |
| "logits/rejected": -0.16529107093811035, | |
| "logps/chosen": -460.61138916015625, | |
| "logps/rejected": -460.7236328125, | |
| "loss": 0.0401, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.4895880222320557, | |
| "rewards/margins": 0.5250975489616394, | |
| "rewards/rejected": -2.0146853923797607, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": -0.1928369402885437, | |
| "eval_logits/rejected": -0.17653387784957886, | |
| "eval_logps/chosen": -431.743896484375, | |
| "eval_logps/rejected": -528.1251831054688, | |
| "eval_loss": 0.04077731445431709, | |
| "eval_rewards/accuracies": 0.71484375, | |
| "eval_rewards/chosen": -1.2815920114517212, | |
| "eval_rewards/margins": 0.8834086656570435, | |
| "eval_rewards/rejected": -2.1650009155273438, | |
| "eval_runtime": 75.4051, | |
| "eval_samples_per_second": 26.523, | |
| "eval_steps_per_second": 0.424, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 9.886905230142433e-07, | |
| "logits/chosen": -0.19448669254779816, | |
| "logits/rejected": -0.2653646469116211, | |
| "logps/chosen": -470.20654296875, | |
| "logps/rejected": -532.0836181640625, | |
| "loss": 0.0436, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.4866724014282227, | |
| "rewards/margins": 0.6113816499710083, | |
| "rewards/rejected": -2.0980541706085205, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 8.868860335206678e-07, | |
| "logits/chosen": -0.274607390165329, | |
| "logits/rejected": -0.18271437287330627, | |
| "logps/chosen": -476.87591552734375, | |
| "logps/rejected": -476.43304443359375, | |
| "loss": 0.0457, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -1.2201156616210938, | |
| "rewards/margins": 0.4418094754219055, | |
| "rewards/rejected": -1.6619250774383545, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 7.883680337481599e-07, | |
| "logits/chosen": -0.17901699244976044, | |
| "logits/rejected": -0.19661930203437805, | |
| "logps/chosen": -414.8687438964844, | |
| "logps/rejected": -532.6778564453125, | |
| "loss": 0.0437, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.1104294061660767, | |
| "rewards/margins": 0.8232589960098267, | |
| "rewards/rejected": -1.9336884021759033, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 6.936646121293654e-07, | |
| "logits/chosen": -0.21059110760688782, | |
| "logits/rejected": -0.22660131752490997, | |
| "logps/chosen": -434.880859375, | |
| "logps/rejected": -541.180908203125, | |
| "loss": 0.0366, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -1.4544843435287476, | |
| "rewards/margins": 0.8150385618209839, | |
| "rewards/rejected": -2.2695229053497314, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 6.032834097207889e-07, | |
| "logits/chosen": -0.17869403958320618, | |
| "logits/rejected": -0.1880768984556198, | |
| "logps/chosen": -464.3429260253906, | |
| "logps/rejected": -565.3632202148438, | |
| "loss": 0.0355, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.5690315961837769, | |
| "rewards/margins": 0.8189191818237305, | |
| "rewards/rejected": -2.387950897216797, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 5.177088990820725e-07, | |
| "logits/chosen": -0.2201962023973465, | |
| "logits/rejected": -0.17389468848705292, | |
| "logps/chosen": -494.4137268066406, | |
| "logps/rejected": -554.2338256835938, | |
| "loss": 0.0348, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.467558741569519, | |
| "rewards/margins": 0.9708374738693237, | |
| "rewards/rejected": -2.4383959770202637, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 4.3739978734594494e-07, | |
| "logits/chosen": -0.2049982100725174, | |
| "logits/rejected": -0.13553449511528015, | |
| "logps/chosen": -510.56610107421875, | |
| "logps/rejected": -544.7944946289062, | |
| "loss": 0.0343, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.5760037899017334, | |
| "rewards/margins": 0.8068605661392212, | |
| "rewards/rejected": -2.382864236831665, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 3.627865573992087e-07, | |
| "logits/chosen": -0.23506097495555878, | |
| "logits/rejected": -0.1945139467716217, | |
| "logps/chosen": -473.996337890625, | |
| "logps/rejected": -546.2703857421875, | |
| "loss": 0.0411, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.4194319248199463, | |
| "rewards/margins": 0.9198926091194153, | |
| "rewards/rejected": -2.339324474334717, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 2.9426916035484166e-07, | |
| "logits/chosen": -0.18075546622276306, | |
| "logits/rejected": -0.12812264263629913, | |
| "logps/chosen": -451.8818359375, | |
| "logps/rejected": -511.1201171875, | |
| "loss": 0.0428, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.2848039865493774, | |
| "rewards/margins": 0.7686089277267456, | |
| "rewards/rejected": -2.053412675857544, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 2.322148716843081e-07, | |
| "logits/chosen": -0.15861022472381592, | |
| "logits/rejected": -0.1434582620859146, | |
| "logps/chosen": -434.2018127441406, | |
| "logps/rejected": -514.0694580078125, | |
| "loss": 0.0435, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -1.3711330890655518, | |
| "rewards/margins": 0.6456871032714844, | |
| "rewards/rejected": -2.016820192337036, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": -0.08009739220142365, | |
| "eval_logits/rejected": -0.07131167501211166, | |
| "eval_logps/chosen": -420.7032470703125, | |
| "eval_logps/rejected": -517.1834106445312, | |
| "eval_loss": 0.04270826652646065, | |
| "eval_rewards/accuracies": 0.7265625, | |
| "eval_rewards/chosen": -1.1711856126785278, | |
| "eval_rewards/margins": 0.8843976259231567, | |
| "eval_rewards/rejected": -2.0555832386016846, | |
| "eval_runtime": 75.7653, | |
| "eval_samples_per_second": 26.397, | |
| "eval_steps_per_second": 0.422, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 1.7695632250191002e-07, | |
| "logits/chosen": -0.16302308440208435, | |
| "logits/rejected": -0.1545868217945099, | |
| "logps/chosen": -431.54083251953125, | |
| "logps/rejected": -492.18975830078125, | |
| "loss": 0.0419, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.1856681108474731, | |
| "rewards/margins": 0.7893311381340027, | |
| "rewards/rejected": -1.9749990701675415, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 1.2878971655412515e-07, | |
| "logits/chosen": -0.15071699023246765, | |
| "logits/rejected": -0.1457306295633316, | |
| "logps/chosen": -424.42547607421875, | |
| "logps/rejected": -474.6815490722656, | |
| "loss": 0.0388, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.1596481800079346, | |
| "rewards/margins": 0.791610836982727, | |
| "rewards/rejected": -1.9512590169906616, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 8.797324247145411e-08, | |
| "logits/chosen": -0.16887474060058594, | |
| "logits/rejected": -0.11855659633874893, | |
| "logps/chosen": -482.5933532714844, | |
| "logps/rejected": -524.8614501953125, | |
| "loss": 0.0433, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.3214952945709229, | |
| "rewards/margins": 0.7315466403961182, | |
| "rewards/rejected": -2.053041934967041, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 5.472568979361853e-08, | |
| "logits/chosen": -0.1303589642047882, | |
| "logits/rejected": -0.1282893717288971, | |
| "logps/chosen": -428.30877685546875, | |
| "logps/rejected": -485.59185791015625, | |
| "loss": 0.0406, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.3397411108016968, | |
| "rewards/margins": 0.8013665080070496, | |
| "rewards/rejected": -2.1411080360412598, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 2.922527618666465e-08, | |
| "logits/chosen": -0.20207130908966064, | |
| "logits/rejected": -0.07939404994249344, | |
| "logps/chosen": -461.0880432128906, | |
| "logps/rejected": -495.314453125, | |
| "loss": 0.0414, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -1.184661626815796, | |
| "rewards/margins": 0.7377935647964478, | |
| "rewards/rejected": -1.922455072402954, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 1.1608692138469379e-08, | |
| "logits/chosen": -0.18077249825000763, | |
| "logits/rejected": -0.11532274633646011, | |
| "logps/chosen": -430.862060546875, | |
| "logps/rejected": -464.39385986328125, | |
| "loss": 0.0406, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.2867480516433716, | |
| "rewards/margins": 0.6147353053092957, | |
| "rewards/rejected": -1.9014835357666016, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 1.970368253390198e-09, | |
| "logits/chosen": -0.21146085858345032, | |
| "logits/rejected": -0.20344269275665283, | |
| "logps/chosen": -457.4112243652344, | |
| "logps/rejected": -528.634521484375, | |
| "loss": 0.039, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.1956093311309814, | |
| "rewards/margins": 0.7480987906455994, | |
| "rewards/rejected": -1.943708062171936, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 477, | |
| "total_flos": 0.0, | |
| "train_loss": 0.05794826206171288, | |
| "train_runtime": 4397.3222, | |
| "train_samples_per_second": 13.903, | |
| "train_steps_per_second": 0.108 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 477, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |