| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.998691442030882, | |
| "eval_steps": 500, | |
| "global_step": 477, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.002093692750588851, | |
| "grad_norm": 1.7795599560645496, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": 4963.591796875, | |
| "logits/rejected": 4602.0634765625, | |
| "logps/chosen": -222.9984130859375, | |
| "logps/rejected": -158.74850463867188, | |
| "loss": 0.5, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02093692750588851, | |
| "grad_norm": 1.664277641531783, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": 6011.06005859375, | |
| "logits/rejected": 5064.162109375, | |
| "logps/chosen": -240.0664825439453, | |
| "logps/rejected": -208.95372009277344, | |
| "loss": 0.5, | |
| "rewards/accuracies": 0.5034722089767456, | |
| "rewards/chosen": 0.00021603053028229624, | |
| "rewards/margins": 0.0005849565495736897, | |
| "rewards/rejected": -0.0003689260338433087, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04187385501177702, | |
| "grad_norm": 1.6870127509120665, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": 5908.4755859375, | |
| "logits/rejected": 4882.568359375, | |
| "logps/chosen": -231.4028778076172, | |
| "logps/rejected": -196.9695281982422, | |
| "loss": 0.4999, | |
| "rewards/accuracies": 0.5531250238418579, | |
| "rewards/chosen": 0.0007881122874096036, | |
| "rewards/margins": 0.0005757558392360806, | |
| "rewards/rejected": 0.00021235644817352295, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06281078251766553, | |
| "grad_norm": 1.6176895769159747, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": 5731.59716796875, | |
| "logits/rejected": 4970.974609375, | |
| "logps/chosen": -205.80563354492188, | |
| "logps/rejected": -182.66543579101562, | |
| "loss": 0.4996, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.0034130061976611614, | |
| "rewards/margins": 0.003777171252295375, | |
| "rewards/rejected": -0.0003641647635959089, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08374771002355404, | |
| "grad_norm": 1.7199138258926487, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": 5881.6220703125, | |
| "logits/rejected": 5133.04638671875, | |
| "logps/chosen": -199.26327514648438, | |
| "logps/rejected": -189.4043426513672, | |
| "loss": 0.4984, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.0059957681223750114, | |
| "rewards/margins": 0.01181001029908657, | |
| "rewards/rejected": -0.005814241245388985, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10468463752944256, | |
| "grad_norm": 2.335343455285067, | |
| "learning_rate": 4.999731868769026e-07, | |
| "logits/chosen": 6244.0546875, | |
| "logits/rejected": 5419.8291015625, | |
| "logps/chosen": -200.3682403564453, | |
| "logps/rejected": -195.88189697265625, | |
| "loss": 0.4965, | |
| "rewards/accuracies": 0.640625, | |
| "rewards/chosen": -0.022861136123538017, | |
| "rewards/margins": 0.02738131210207939, | |
| "rewards/rejected": -0.05024244636297226, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12562156503533106, | |
| "grad_norm": 1.985885842678977, | |
| "learning_rate": 4.990353313429303e-07, | |
| "logits/chosen": 6085.1591796875, | |
| "logits/rejected": 5170.4423828125, | |
| "logps/chosen": -202.0069122314453, | |
| "logps/rejected": -189.2211456298828, | |
| "loss": 0.494, | |
| "rewards/accuracies": 0.640625, | |
| "rewards/chosen": -0.06540889292955399, | |
| "rewards/margins": 0.07115975767374039, | |
| "rewards/rejected": -0.13656863570213318, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14655849254121958, | |
| "grad_norm": 2.02720438835768, | |
| "learning_rate": 4.967625656594781e-07, | |
| "logits/chosen": 5949.5009765625, | |
| "logits/rejected": 5459.0986328125, | |
| "logps/chosen": -203.56887817382812, | |
| "logps/rejected": -203.3651885986328, | |
| "loss": 0.493, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.06014765053987503, | |
| "rewards/margins": 0.07473154366016388, | |
| "rewards/rejected": -0.1348791867494583, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16749542004710807, | |
| "grad_norm": 2.2298529628194035, | |
| "learning_rate": 4.93167072587771e-07, | |
| "logits/chosen": 6108.1171875, | |
| "logits/rejected": 5237.1611328125, | |
| "logps/chosen": -200.6221160888672, | |
| "logps/rejected": -197.19618225097656, | |
| "loss": 0.4914, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.08121322095394135, | |
| "rewards/margins": 0.09713159501552582, | |
| "rewards/rejected": -0.17834481596946716, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.1884323475529966, | |
| "grad_norm": 2.2234876584054275, | |
| "learning_rate": 4.882681251368548e-07, | |
| "logits/chosen": 6233.083984375, | |
| "logits/rejected": 5252.95068359375, | |
| "logps/chosen": -201.39984130859375, | |
| "logps/rejected": -200.0003204345703, | |
| "loss": 0.4901, | |
| "rewards/accuracies": 0.6656249761581421, | |
| "rewards/chosen": -0.08355093002319336, | |
| "rewards/margins": 0.12262304872274399, | |
| "rewards/rejected": -0.20617397129535675, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "grad_norm": 2.1862542481259792, | |
| "learning_rate": 4.820919832540181e-07, | |
| "logits/chosen": 6330.16259765625, | |
| "logits/rejected": 5832.1982421875, | |
| "logps/chosen": -203.98583984375, | |
| "logps/rejected": -211.58010864257812, | |
| "loss": 0.491, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.07759369909763336, | |
| "rewards/margins": 0.12538166344165802, | |
| "rewards/rejected": -0.20297536253929138, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23030620256477363, | |
| "grad_norm": 2.7054972610472565, | |
| "learning_rate": 4.7467175306295647e-07, | |
| "logits/chosen": 6421.41015625, | |
| "logits/rejected": 5193.19287109375, | |
| "logps/chosen": -198.57032775878906, | |
| "logps/rejected": -209.23104858398438, | |
| "loss": 0.4869, | |
| "rewards/accuracies": 0.721875011920929, | |
| "rewards/chosen": -0.0885535329580307, | |
| "rewards/margins": 0.1617024838924408, | |
| "rewards/rejected": -0.2502560019493103, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.2512431300706621, | |
| "grad_norm": 2.490243429862705, | |
| "learning_rate": 4.6604720940421207e-07, | |
| "logits/chosen": 6070.7314453125, | |
| "logits/rejected": 5633.78369140625, | |
| "logps/chosen": -196.9094696044922, | |
| "logps/rejected": -213.2709197998047, | |
| "loss": 0.4892, | |
| "rewards/accuracies": 0.6781250238418579, | |
| "rewards/chosen": -0.10950146615505219, | |
| "rewards/margins": 0.1391303390264511, | |
| "rewards/rejected": -0.2486318051815033, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2721800575765506, | |
| "grad_norm": 2.7610246742258715, | |
| "learning_rate": 4.5626458262912735e-07, | |
| "logits/chosen": 6097.5146484375, | |
| "logits/rejected": 5424.9013671875, | |
| "logps/chosen": -197.1998748779297, | |
| "logps/rejected": -216.65188598632812, | |
| "loss": 0.4879, | |
| "rewards/accuracies": 0.715624988079071, | |
| "rewards/chosen": -0.08192776143550873, | |
| "rewards/margins": 0.15857011079788208, | |
| "rewards/rejected": -0.2404978722333908, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29311698508243916, | |
| "grad_norm": 2.5211223485513323, | |
| "learning_rate": 4.453763107901675e-07, | |
| "logits/chosen": 5834.34814453125, | |
| "logits/rejected": 5199.22119140625, | |
| "logps/chosen": -204.2546844482422, | |
| "logps/rejected": -208.1846923828125, | |
| "loss": 0.4883, | |
| "rewards/accuracies": 0.684374988079071, | |
| "rewards/chosen": -0.10303564369678497, | |
| "rewards/margins": 0.19196416437625885, | |
| "rewards/rejected": -0.29499977827072144, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31405391258832765, | |
| "grad_norm": 2.8015563362051052, | |
| "learning_rate": 4.3344075855595097e-07, | |
| "logits/chosen": 6411.4541015625, | |
| "logits/rejected": 5283.06689453125, | |
| "logps/chosen": -216.62496948242188, | |
| "logps/rejected": -208.1221466064453, | |
| "loss": 0.4855, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.07201126962900162, | |
| "rewards/margins": 0.21424655616283417, | |
| "rewards/rejected": -0.286257803440094, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33499084009421615, | |
| "grad_norm": 3.380927510289668, | |
| "learning_rate": 4.2052190435769554e-07, | |
| "logits/chosen": 6130.6318359375, | |
| "logits/rejected": 5110.82080078125, | |
| "logps/chosen": -209.1565704345703, | |
| "logps/rejected": -219.35220336914062, | |
| "loss": 0.4824, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.10825005918741226, | |
| "rewards/margins": 0.2637324929237366, | |
| "rewards/rejected": -0.37198254466056824, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.3559277676001047, | |
| "grad_norm": 3.2112846087091484, | |
| "learning_rate": 4.0668899744407567e-07, | |
| "logits/chosen": 6476.86767578125, | |
| "logits/rejected": 5530.86767578125, | |
| "logps/chosen": -220.4525146484375, | |
| "logps/rejected": -208.67733764648438, | |
| "loss": 0.4833, | |
| "rewards/accuracies": 0.703125, | |
| "rewards/chosen": -0.12977644801139832, | |
| "rewards/margins": 0.23194870352745056, | |
| "rewards/rejected": -0.3617251515388489, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.3768646951059932, | |
| "grad_norm": 2.5422470233501713, | |
| "learning_rate": 3.920161866827889e-07, | |
| "logits/chosen": 6021.7919921875, | |
| "logits/rejected": 5178.1259765625, | |
| "logps/chosen": -220.57601928710938, | |
| "logps/rejected": -207.26339721679688, | |
| "loss": 0.4847, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.09060440957546234, | |
| "rewards/margins": 0.20774026215076447, | |
| "rewards/rejected": -0.2983446717262268, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.39780162261188173, | |
| "grad_norm": 5.303607389601098, | |
| "learning_rate": 3.765821230985757e-07, | |
| "logits/chosen": 6147.5517578125, | |
| "logits/rejected": 5167.7958984375, | |
| "logps/chosen": -208.33944702148438, | |
| "logps/rejected": -221.4107208251953, | |
| "loss": 0.4836, | |
| "rewards/accuracies": 0.7406250238418579, | |
| "rewards/chosen": -0.1318391114473343, | |
| "rewards/margins": 0.3124375343322754, | |
| "rewards/rejected": -0.4442766308784485, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "grad_norm": 2.6432475137123794, | |
| "learning_rate": 3.604695382782159e-07, | |
| "logits/chosen": 5748.025390625, | |
| "logits/rejected": 4992.97021484375, | |
| "logps/chosen": -193.7339630126953, | |
| "logps/rejected": -207.1187286376953, | |
| "loss": 0.4868, | |
| "rewards/accuracies": 0.703125, | |
| "rewards/chosen": -0.08920423686504364, | |
| "rewards/margins": 0.2186780869960785, | |
| "rewards/rejected": -0.30788230895996094, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4396754776236587, | |
| "grad_norm": 3.3944179691537864, | |
| "learning_rate": 3.4376480090239047e-07, | |
| "logits/chosen": 6434.7412109375, | |
| "logits/rejected": 5470.30224609375, | |
| "logps/chosen": -219.40087890625, | |
| "logps/rejected": -221.01614379882812, | |
| "loss": 0.4847, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -0.07487013936042786, | |
| "rewards/margins": 0.21102896332740784, | |
| "rewards/rejected": -0.2858991026878357, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46061240512954726, | |
| "grad_norm": 2.9009193632951207, | |
| "learning_rate": 3.265574537815398e-07, | |
| "logits/chosen": 5810.1513671875, | |
| "logits/rejected": 5406.72021484375, | |
| "logps/chosen": -212.43923950195312, | |
| "logps/rejected": -227.1441192626953, | |
| "loss": 0.4829, | |
| "rewards/accuracies": 0.753125011920929, | |
| "rewards/chosen": -0.105693019926548, | |
| "rewards/margins": 0.26153022050857544, | |
| "rewards/rejected": -0.36722320318222046, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48154933263543576, | |
| "grad_norm": 2.6189951178492614, | |
| "learning_rate": 3.0893973387735683e-07, | |
| "logits/chosen": 6018.71484375, | |
| "logits/rejected": 5094.3134765625, | |
| "logps/chosen": -207.9360809326172, | |
| "logps/rejected": -231.08480834960938, | |
| "loss": 0.4805, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.13205979764461517, | |
| "rewards/margins": 0.36629724502563477, | |
| "rewards/rejected": -0.49835705757141113, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5024862601413242, | |
| "grad_norm": 2.9120071931474514, | |
| "learning_rate": 2.910060778827554e-07, | |
| "logits/chosen": 6167.8330078125, | |
| "logits/rejected": 5222.9423828125, | |
| "logps/chosen": -206.72628784179688, | |
| "logps/rejected": -222.31649780273438, | |
| "loss": 0.4816, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.10643552243709564, | |
| "rewards/margins": 0.27724453806877136, | |
| "rewards/rejected": -0.3836800754070282, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5234231876472127, | |
| "grad_norm": 3.640526921626944, | |
| "learning_rate": 2.7285261601056697e-07, | |
| "logits/chosen": 6262.5078125, | |
| "logits/rejected": 5280.09619140625, | |
| "logps/chosen": -224.32699584960938, | |
| "logps/rejected": -236.7519989013672, | |
| "loss": 0.4832, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.11429344117641449, | |
| "rewards/margins": 0.28984737396240234, | |
| "rewards/rejected": -0.4041408598423004, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5443601151531012, | |
| "grad_norm": 3.480502498738401, | |
| "learning_rate": 2.5457665670441937e-07, | |
| "logits/chosen": 6063.48681640625, | |
| "logits/rejected": 5488.103515625, | |
| "logps/chosen": -227.7340850830078, | |
| "logps/rejected": -235.0940399169922, | |
| "loss": 0.4823, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -0.11375202983617783, | |
| "rewards/margins": 0.27239111065864563, | |
| "rewards/rejected": -0.38614311814308167, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5652970426589898, | |
| "grad_norm": 3.100193815819693, | |
| "learning_rate": 2.3627616503391812e-07, | |
| "logits/chosen": 6130.5126953125, | |
| "logits/rejected": 5626.0517578125, | |
| "logps/chosen": -217.3065643310547, | |
| "logps/rejected": -236.91329956054688, | |
| "loss": 0.4834, | |
| "rewards/accuracies": 0.753125011920929, | |
| "rewards/chosen": -0.1151232123374939, | |
| "rewards/margins": 0.2929432988166809, | |
| "rewards/rejected": -0.40806645154953003, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5862339701648783, | |
| "grad_norm": 3.6139410491493034, | |
| "learning_rate": 2.1804923757009882e-07, | |
| "logits/chosen": 6155.16357421875, | |
| "logits/rejected": 5117.0361328125, | |
| "logps/chosen": -224.9839630126953, | |
| "logps/rejected": -229.80514526367188, | |
| "loss": 0.4807, | |
| "rewards/accuracies": 0.7718750238418579, | |
| "rewards/chosen": -0.10252990573644638, | |
| "rewards/margins": 0.33688706159591675, | |
| "rewards/rejected": -0.4394169747829437, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6071708976707668, | |
| "grad_norm": 3.0356982001656423, | |
| "learning_rate": 1.9999357655598891e-07, | |
| "logits/chosen": 6409.4736328125, | |
| "logits/rejected": 6006.4580078125, | |
| "logps/chosen": -239.07656860351562, | |
| "logps/rejected": -265.92523193359375, | |
| "loss": 0.4842, | |
| "rewards/accuracies": 0.6968749761581421, | |
| "rewards/chosen": -0.15593001246452332, | |
| "rewards/margins": 0.2926904857158661, | |
| "rewards/rejected": -0.4486205577850342, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "grad_norm": 4.271112938322883, | |
| "learning_rate": 1.8220596619089573e-07, | |
| "logits/chosen": 5861.82177734375, | |
| "logits/rejected": 5508.8740234375, | |
| "logps/chosen": -222.1393585205078, | |
| "logps/rejected": -245.2710418701172, | |
| "loss": 0.4845, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.11530385166406631, | |
| "rewards/margins": 0.2702717185020447, | |
| "rewards/rejected": -0.3855755925178528, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6490447526825438, | |
| "grad_norm": 3.0832085833192004, | |
| "learning_rate": 1.647817538357072e-07, | |
| "logits/chosen": 6551.92529296875, | |
| "logits/rejected": 5745.2666015625, | |
| "logps/chosen": -234.02554321289062, | |
| "logps/rejected": -243.51119995117188, | |
| "loss": 0.4815, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.10852447897195816, | |
| "rewards/margins": 0.22780513763427734, | |
| "rewards/rejected": -0.3363296091556549, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6699816801884323, | |
| "grad_norm": 5.627304230331988, | |
| "learning_rate": 1.478143389201113e-07, | |
| "logits/chosen": 6057.1337890625, | |
| "logits/rejected": 5255.4736328125, | |
| "logps/chosen": -222.4619903564453, | |
| "logps/rejected": -239.05459594726562, | |
| "loss": 0.4817, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -0.09626330435276031, | |
| "rewards/margins": 0.28026360273361206, | |
| "rewards/rejected": -0.37652695178985596, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6909186076943209, | |
| "grad_norm": 4.070544491993295, | |
| "learning_rate": 1.3139467229135998e-07, | |
| "logits/chosen": 5855.13671875, | |
| "logits/rejected": 5428.80224609375, | |
| "logps/chosen": -231.0357208251953, | |
| "logps/rejected": -246.7843780517578, | |
| "loss": 0.484, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.11221712827682495, | |
| "rewards/margins": 0.30028775334358215, | |
| "rewards/rejected": -0.4125048518180847, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7118555352002094, | |
| "grad_norm": 4.9587556176352905, | |
| "learning_rate": 1.1561076868822755e-07, | |
| "logits/chosen": 5686.6806640625, | |
| "logits/rejected": 4937.55615234375, | |
| "logps/chosen": -216.03945922851562, | |
| "logps/rejected": -238.09805297851562, | |
| "loss": 0.4819, | |
| "rewards/accuracies": 0.7593749761581421, | |
| "rewards/chosen": -0.16062280535697937, | |
| "rewards/margins": 0.34666866064071655, | |
| "rewards/rejected": -0.5072914958000183, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7327924627060979, | |
| "grad_norm": 3.8596857041552393, | |
| "learning_rate": 1.0054723495346482e-07, | |
| "logits/chosen": 6366.03955078125, | |
| "logits/rejected": 5157.458984375, | |
| "logps/chosen": -255.4525604248047, | |
| "logps/rejected": -263.61639404296875, | |
| "loss": 0.4836, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.14537517726421356, | |
| "rewards/margins": 0.3442801833152771, | |
| "rewards/rejected": -0.48965534567832947, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7537293902119864, | |
| "grad_norm": 4.404236860707489, | |
| "learning_rate": 8.628481651367875e-08, | |
| "logits/chosen": 6031.80615234375, | |
| "logits/rejected": 5262.02734375, | |
| "logps/chosen": -231.36318969726562, | |
| "logps/rejected": -245.8846893310547, | |
| "loss": 0.4812, | |
| "rewards/accuracies": 0.765625, | |
| "rewards/chosen": -0.10767906904220581, | |
| "rewards/margins": 0.31250181794166565, | |
| "rewards/rejected": -0.42018088698387146, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7746663177178749, | |
| "grad_norm": 4.286082049703798, | |
| "learning_rate": 7.289996455765748e-08, | |
| "logits/chosen": 5882.796875, | |
| "logits/rejected": 5024.01171875, | |
| "logps/chosen": -250.63809204101562, | |
| "logps/rejected": -253.8474884033203, | |
| "loss": 0.4799, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.12153621017932892, | |
| "rewards/margins": 0.3065151572227478, | |
| "rewards/rejected": -0.4280513823032379, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7956032452237635, | |
| "grad_norm": 4.2771383904693545, | |
| "learning_rate": 6.046442623320145e-08, | |
| "logits/chosen": 6166.33447265625, | |
| "logits/rejected": 5370.65185546875, | |
| "logps/chosen": -244.5850067138672, | |
| "logps/rejected": -238.40518188476562, | |
| "loss": 0.4816, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.10372950881719589, | |
| "rewards/margins": 0.2737034261226654, | |
| "rewards/rejected": -0.3774329721927643, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.816540172729652, | |
| "grad_norm": 4.7598934741028565, | |
| "learning_rate": 4.904486005914027e-08, | |
| "logits/chosen": 6437.08251953125, | |
| "logits/rejected": 5671.439453125, | |
| "logps/chosen": -249.28125, | |
| "logps/rejected": -271.4849548339844, | |
| "loss": 0.4818, | |
| "rewards/accuracies": 0.753125011920929, | |
| "rewards/chosen": -0.1305597722530365, | |
| "rewards/margins": 0.3592670261859894, | |
| "rewards/rejected": -0.4898267686367035, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "grad_norm": 3.0633087764980194, | |
| "learning_rate": 3.8702478614051345e-08, | |
| "logits/chosen": 5975.5859375, | |
| "logits/rejected": 5307.0068359375, | |
| "logps/chosen": -249.6516876220703, | |
| "logps/rejected": -247.6414031982422, | |
| "loss": 0.4821, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.12782196700572968, | |
| "rewards/margins": 0.30399927496910095, | |
| "rewards/rejected": -0.4318212568759918, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8584140277414289, | |
| "grad_norm": 5.414232808843427, | |
| "learning_rate": 2.9492720416985e-08, | |
| "logits/chosen": 5987.8955078125, | |
| "logits/rejected": 5288.86669921875, | |
| "logps/chosen": -239.4297637939453, | |
| "logps/rejected": -252.67214965820312, | |
| "loss": 0.4809, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.14551712572574615, | |
| "rewards/margins": 0.2954447865486145, | |
| "rewards/rejected": -0.44096192717552185, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8793509552473174, | |
| "grad_norm": 3.7651155714131734, | |
| "learning_rate": 2.1464952759020856e-08, | |
| "logits/chosen": 6339.23046875, | |
| "logits/rejected": 5529.13037109375, | |
| "logps/chosen": -254.3148193359375, | |
| "logps/rejected": -267.2603759765625, | |
| "loss": 0.481, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": -0.12421433627605438, | |
| "rewards/margins": 0.3070390820503235, | |
| "rewards/rejected": -0.43125343322753906, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9002878827532059, | |
| "grad_norm": 3.0456608776131104, | |
| "learning_rate": 1.4662207078575684e-08, | |
| "logits/chosen": 6244.36669921875, | |
| "logits/rejected": 5038.8232421875, | |
| "logps/chosen": -249.1537628173828, | |
| "logps/rejected": -258.28509521484375, | |
| "loss": 0.4806, | |
| "rewards/accuracies": 0.765625, | |
| "rewards/chosen": -0.11752188205718994, | |
| "rewards/margins": 0.3456159830093384, | |
| "rewards/rejected": -0.46313780546188354, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9212248102590945, | |
| "grad_norm": 3.7152553962599533, | |
| "learning_rate": 9.12094829893642e-09, | |
| "logits/chosen": 6062.4091796875, | |
| "logits/rejected": 5768.86572265625, | |
| "logps/chosen": -247.6633758544922, | |
| "logps/rejected": -270.07781982421875, | |
| "loss": 0.4831, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.125531867146492, | |
| "rewards/margins": 0.2730921804904938, | |
| "rewards/rejected": -0.39862409234046936, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.942161737764983, | |
| "grad_norm": 3.385435845957556, | |
| "learning_rate": 4.8708793644441086e-09, | |
| "logits/chosen": 6208.9853515625, | |
| "logits/rejected": 5350.4443359375, | |
| "logps/chosen": -253.6076202392578, | |
| "logps/rejected": -252.9407196044922, | |
| "loss": 0.4804, | |
| "rewards/accuracies": 0.753125011920929, | |
| "rewards/chosen": -0.10648094117641449, | |
| "rewards/margins": 0.2990611791610718, | |
| "rewards/rejected": -0.40554213523864746, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9630986652708715, | |
| "grad_norm": 3.849103407950397, | |
| "learning_rate": 1.9347820230782295e-09, | |
| "logits/chosen": 6395.10302734375, | |
| "logits/rejected": 5409.55078125, | |
| "logps/chosen": -276.77349853515625, | |
| "logps/rejected": -277.45819091796875, | |
| "loss": 0.4832, | |
| "rewards/accuracies": 0.7093750238418579, | |
| "rewards/chosen": -0.13324801623821259, | |
| "rewards/margins": 0.28933390974998474, | |
| "rewards/rejected": -0.42258191108703613, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98403559277676, | |
| "grad_norm": 3.425749585050412, | |
| "learning_rate": 3.2839470889836627e-10, | |
| "logits/chosen": 5962.32763671875, | |
| "logits/rejected": 5300.6298828125, | |
| "logps/chosen": -245.10165405273438, | |
| "logps/rejected": -248.36233520507812, | |
| "loss": 0.481, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.1263255625963211, | |
| "rewards/margins": 0.26447486877441406, | |
| "rewards/rejected": -0.390800416469574, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.998691442030882, | |
| "step": 477, | |
| "total_flos": 0.0, | |
| "train_loss": 0.4856958399278813, | |
| "train_runtime": 14922.9586, | |
| "train_samples_per_second": 4.097, | |
| "train_steps_per_second": 0.032 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 477, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |