| { | |
| "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": 6813.720901215927, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": 5002.53564453125, | |
| "logits/rejected": 4591.72021484375, | |
| "logps/chosen": -265.6396789550781, | |
| "logps/rejected": -206.22401428222656, | |
| "loss": 503.0168, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02093692750588851, | |
| "grad_norm": 6935.582411893378, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": 5977.00244140625, | |
| "logits/rejected": 5015.09765625, | |
| "logps/chosen": -292.2739562988281, | |
| "logps/rejected": -259.0675048828125, | |
| "loss": 549.3314, | |
| "rewards/accuracies": 0.4583333432674408, | |
| "rewards/chosen": 0.0028684597928076982, | |
| "rewards/margins": 0.0003068610094487667, | |
| "rewards/rejected": 0.0025615987833589315, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04187385501177702, | |
| "grad_norm": 3127.701327967131, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": 5864.72802734375, | |
| "logits/rejected": 4844.66650390625, | |
| "logps/chosen": -283.6185302734375, | |
| "logps/rejected": -243.01513671875, | |
| "loss": 523.6729, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.06974941492080688, | |
| "rewards/margins": 0.002566719427704811, | |
| "rewards/rejected": 0.06718270480632782, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06281078251766553, | |
| "grad_norm": 2074.6624019206756, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": 5557.00439453125, | |
| "logits/rejected": 4840.8193359375, | |
| "logps/chosen": -256.04144287109375, | |
| "logps/rejected": -223.24319458007812, | |
| "loss": 494.8969, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": 0.16196031868457794, | |
| "rewards/margins": 0.006370093673467636, | |
| "rewards/rejected": 0.1555902361869812, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08374771002355404, | |
| "grad_norm": 1576.9313876726794, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": 5559.41650390625, | |
| "logits/rejected": 4830.68994140625, | |
| "logps/chosen": -251.85604858398438, | |
| "logps/rejected": -236.69235229492188, | |
| "loss": 496.3616, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": 0.22101028263568878, | |
| "rewards/margins": 0.0033943026792258024, | |
| "rewards/rejected": 0.217616006731987, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10468463752944256, | |
| "grad_norm": 1640.8995064962066, | |
| "learning_rate": 4.999731868769026e-07, | |
| "logits/chosen": 5797.015625, | |
| "logits/rejected": 5035.59326171875, | |
| "logps/chosen": -253.31631469726562, | |
| "logps/rejected": -242.33651733398438, | |
| "loss": 492.3395, | |
| "rewards/accuracies": 0.5218750238418579, | |
| "rewards/chosen": 0.2530141770839691, | |
| "rewards/margins": 0.0011868119472637773, | |
| "rewards/rejected": 0.2518273890018463, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12562156503533106, | |
| "grad_norm": 1480.537194221153, | |
| "learning_rate": 4.990353313429303e-07, | |
| "logits/chosen": 5540.62158203125, | |
| "logits/rejected": 4727.97607421875, | |
| "logps/chosen": -248.41848754882812, | |
| "logps/rejected": -221.98336791992188, | |
| "loss": 466.3484, | |
| "rewards/accuracies": 0.5218750238418579, | |
| "rewards/chosen": 0.30004000663757324, | |
| "rewards/margins": 0.005084425210952759, | |
| "rewards/rejected": 0.29495561122894287, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14655849254121958, | |
| "grad_norm": 1439.2312712606115, | |
| "learning_rate": 4.967625656594781e-07, | |
| "logits/chosen": 5331.10791015625, | |
| "logits/rejected": 4925.798828125, | |
| "logps/chosen": -247.7162628173828, | |
| "logps/rejected": -237.12353515625, | |
| "loss": 465.379, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": 0.30591654777526855, | |
| "rewards/margins": 0.009119048714637756, | |
| "rewards/rejected": 0.2967974543571472, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16749542004710807, | |
| "grad_norm": 1420.7895768869655, | |
| "learning_rate": 4.93167072587771e-07, | |
| "logits/chosen": 5414.23828125, | |
| "logits/rejected": 4648.93115234375, | |
| "logps/chosen": -242.97555541992188, | |
| "logps/rejected": -222.2198486328125, | |
| "loss": 464.9691, | |
| "rewards/accuracies": 0.5531250238418579, | |
| "rewards/chosen": 0.33111369609832764, | |
| "rewards/margins": 0.022587263956665993, | |
| "rewards/rejected": 0.3085264265537262, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.1884323475529966, | |
| "grad_norm": 1456.2100554164706, | |
| "learning_rate": 4.882681251368548e-07, | |
| "logits/chosen": 5535.8984375, | |
| "logits/rejected": 4687.62255859375, | |
| "logps/chosen": -242.0615234375, | |
| "logps/rejected": -220.8953399658203, | |
| "loss": 465.3252, | |
| "rewards/accuracies": 0.5406249761581421, | |
| "rewards/chosen": 0.35646066069602966, | |
| "rewards/margins": 0.015677783638238907, | |
| "rewards/rejected": 0.34078291058540344, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "grad_norm": 1536.760185739677, | |
| "learning_rate": 4.820919832540181e-07, | |
| "logits/chosen": 5591.662109375, | |
| "logits/rejected": 5172.16259765625, | |
| "logps/chosen": -243.97738647460938, | |
| "logps/rejected": -236.50595092773438, | |
| "loss": 460.8633, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.3589131832122803, | |
| "rewards/margins": -0.021181438118219376, | |
| "rewards/rejected": 0.38009461760520935, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23030620256477363, | |
| "grad_norm": 1408.5395404483565, | |
| "learning_rate": 4.7467175306295647e-07, | |
| "logits/chosen": 5752.2939453125, | |
| "logits/rejected": 4664.9677734375, | |
| "logps/chosen": -236.39962768554688, | |
| "logps/rejected": -224.271484375, | |
| "loss": 463.9909, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.382229745388031, | |
| "rewards/margins": 0.03622853755950928, | |
| "rewards/rejected": 0.34600120782852173, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.2512431300706621, | |
| "grad_norm": 1430.8052441244079, | |
| "learning_rate": 4.6604720940421207e-07, | |
| "logits/chosen": 5345.5830078125, | |
| "logits/rejected": 4974.67724609375, | |
| "logps/chosen": -230.8021240234375, | |
| "logps/rejected": -229.9935302734375, | |
| "loss": 472.1473, | |
| "rewards/accuracies": 0.503125011920929, | |
| "rewards/chosen": 0.36376887559890747, | |
| "rewards/margins": -0.005053712520748377, | |
| "rewards/rejected": 0.36882254481315613, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2721800575765506, | |
| "grad_norm": 1613.4321490419964, | |
| "learning_rate": 4.5626458262912735e-07, | |
| "logits/chosen": 5406.8876953125, | |
| "logits/rejected": 4819.064453125, | |
| "logps/chosen": -233.12362670898438, | |
| "logps/rejected": -232.2109375, | |
| "loss": 465.5601, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.35393673181533813, | |
| "rewards/margins": -0.014040246605873108, | |
| "rewards/rejected": 0.36797699332237244, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29311698508243916, | |
| "grad_norm": 1397.7266878069634, | |
| "learning_rate": 4.453763107901675e-07, | |
| "logits/chosen": 5204.806640625, | |
| "logits/rejected": 4634.70947265625, | |
| "logps/chosen": -235.88473510742188, | |
| "logps/rejected": -216.02548217773438, | |
| "loss": 475.0084, | |
| "rewards/accuracies": 0.546875, | |
| "rewards/chosen": 0.36020082235336304, | |
| "rewards/margins": 0.009127211757004261, | |
| "rewards/rejected": 0.35107359290122986, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31405391258832765, | |
| "grad_norm": 1771.9334786366776, | |
| "learning_rate": 4.3344075855595097e-07, | |
| "logits/chosen": 5751.51171875, | |
| "logits/rejected": 4749.1025390625, | |
| "logps/chosen": -253.08169555664062, | |
| "logps/rejected": -215.91110229492188, | |
| "loss": 453.9507, | |
| "rewards/accuracies": 0.5531250238418579, | |
| "rewards/chosen": 0.3603329658508301, | |
| "rewards/margins": 0.004536592401564121, | |
| "rewards/rejected": 0.35579636693000793, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33499084009421615, | |
| "grad_norm": 1436.0408028073434, | |
| "learning_rate": 4.2052190435769554e-07, | |
| "logits/chosen": 5464.77099609375, | |
| "logits/rejected": 4553.0888671875, | |
| "logps/chosen": -241.3671112060547, | |
| "logps/rejected": -218.6112518310547, | |
| "loss": 470.9686, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.4322701394557953, | |
| "rewards/margins": 0.054599881172180176, | |
| "rewards/rejected": 0.3776702284812927, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.3559277676001047, | |
| "grad_norm": 1452.849884308315, | |
| "learning_rate": 4.0668899744407567e-07, | |
| "logits/chosen": 5712.3564453125, | |
| "logits/rejected": 4878.48193359375, | |
| "logps/chosen": -248.1728515625, | |
| "logps/rejected": -206.45004272460938, | |
| "loss": 457.2006, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.34142884612083435, | |
| "rewards/margins": -0.0005211926763877273, | |
| "rewards/rejected": 0.34195005893707275, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.3768646951059932, | |
| "grad_norm": 1345.993393031555, | |
| "learning_rate": 3.920161866827889e-07, | |
| "logits/chosen": 5387.048828125, | |
| "logits/rejected": 4639.03271484375, | |
| "logps/chosen": -250.1925811767578, | |
| "logps/rejected": -211.57275390625, | |
| "loss": 463.2538, | |
| "rewards/accuracies": 0.5218750238418579, | |
| "rewards/chosen": 0.37102702260017395, | |
| "rewards/margins": -0.013053232803940773, | |
| "rewards/rejected": 0.3840802311897278, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.39780162261188173, | |
| "grad_norm": 1364.64342210683, | |
| "learning_rate": 3.765821230985757e-07, | |
| "logits/chosen": 5516.2822265625, | |
| "logits/rejected": 4661.30517578125, | |
| "logps/chosen": -231.8093719482422, | |
| "logps/rejected": -209.94967651367188, | |
| "loss": 460.1928, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.37942126393318176, | |
| "rewards/margins": 8.549987978767604e-05, | |
| "rewards/rejected": 0.3793357312679291, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "grad_norm": 1501.9460217734675, | |
| "learning_rate": 3.604695382782159e-07, | |
| "logits/chosen": 5149.3212890625, | |
| "logits/rejected": 4487.83251953125, | |
| "logps/chosen": -217.900634765625, | |
| "logps/rejected": -206.22817993164062, | |
| "loss": 462.1984, | |
| "rewards/accuracies": 0.5406249761581421, | |
| "rewards/chosen": 0.3645021319389343, | |
| "rewards/margins": -0.0018228956032544374, | |
| "rewards/rejected": 0.3663250207901001, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4396754776236587, | |
| "grad_norm": 1429.3016977538932, | |
| "learning_rate": 3.4376480090239047e-07, | |
| "logits/chosen": 5780.9111328125, | |
| "logits/rejected": 4931.05810546875, | |
| "logps/chosen": -248.72396850585938, | |
| "logps/rejected": -224.09872436523438, | |
| "loss": 466.6646, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": 0.38043394684791565, | |
| "rewards/margins": 0.004385782405734062, | |
| "rewards/rejected": 0.37604817748069763, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46061240512954726, | |
| "grad_norm": 1443.7077268756489, | |
| "learning_rate": 3.265574537815398e-07, | |
| "logits/chosen": 5190.0830078125, | |
| "logits/rejected": 4833.0302734375, | |
| "logps/chosen": -234.4996795654297, | |
| "logps/rejected": -221.47488403320312, | |
| "loss": 474.1076, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.36616024374961853, | |
| "rewards/margins": -0.022525813430547714, | |
| "rewards/rejected": 0.38868606090545654, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48154933263543576, | |
| "grad_norm": 1524.9948933221324, | |
| "learning_rate": 3.0893973387735683e-07, | |
| "logits/chosen": 5389.8935546875, | |
| "logits/rejected": 4572.7626953125, | |
| "logps/chosen": -228.4185028076172, | |
| "logps/rejected": -209.15811157226562, | |
| "loss": 446.1821, | |
| "rewards/accuracies": 0.5531250238418579, | |
| "rewards/chosen": 0.39850276708602905, | |
| "rewards/margins": 0.035766832530498505, | |
| "rewards/rejected": 0.36273595690727234, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5024862601413242, | |
| "grad_norm": 1425.7617156287813, | |
| "learning_rate": 2.910060778827554e-07, | |
| "logits/chosen": 5539.8349609375, | |
| "logits/rejected": 4716.81201171875, | |
| "logps/chosen": -227.1032257080078, | |
| "logps/rejected": -210.68612670898438, | |
| "loss": 466.6096, | |
| "rewards/accuracies": 0.534375011920929, | |
| "rewards/chosen": 0.38473668694496155, | |
| "rewards/margins": -0.006211507134139538, | |
| "rewards/rejected": 0.3909481465816498, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5234231876472127, | |
| "grad_norm": 1407.421320970437, | |
| "learning_rate": 2.7285261601056697e-07, | |
| "logits/chosen": 5670.39453125, | |
| "logits/rejected": 4789.2001953125, | |
| "logps/chosen": -244.33895874023438, | |
| "logps/rejected": -223.86453247070312, | |
| "loss": 459.7925, | |
| "rewards/accuracies": 0.5531250238418579, | |
| "rewards/chosen": 0.4160972535610199, | |
| "rewards/margins": 0.03399023413658142, | |
| "rewards/rejected": 0.3821069598197937, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5443601151531012, | |
| "grad_norm": 1368.2023715670355, | |
| "learning_rate": 2.5457665670441937e-07, | |
| "logits/chosen": 5489.0439453125, | |
| "logits/rejected": 4975.576171875, | |
| "logps/chosen": -245.1668243408203, | |
| "logps/rejected": -221.49661254882812, | |
| "loss": 458.9599, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.4090227484703064, | |
| "rewards/margins": 0.013166209682822227, | |
| "rewards/rejected": 0.3958565592765808, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5652970426589898, | |
| "grad_norm": 1496.3498583187165, | |
| "learning_rate": 2.3627616503391812e-07, | |
| "logits/chosen": 5568.6240234375, | |
| "logits/rejected": 5110.2138671875, | |
| "logps/chosen": -233.6811065673828, | |
| "logps/rejected": -220.8893585205078, | |
| "loss": 447.9319, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.4093276560306549, | |
| "rewards/margins": 0.014256368391215801, | |
| "rewards/rejected": 0.3950712978839874, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5862339701648783, | |
| "grad_norm": 1386.5146227495984, | |
| "learning_rate": 2.1804923757009882e-07, | |
| "logits/chosen": 5578.41455078125, | |
| "logits/rejected": 4647.2431640625, | |
| "logps/chosen": -241.32284545898438, | |
| "logps/rejected": -207.7650146484375, | |
| "loss": 448.7999, | |
| "rewards/accuracies": 0.534375011920929, | |
| "rewards/chosen": 0.3940281867980957, | |
| "rewards/margins": 0.003837400348857045, | |
| "rewards/rejected": 0.3901907801628113, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6071708976707668, | |
| "grad_norm": 1565.300342790805, | |
| "learning_rate": 1.9999357655598891e-07, | |
| "logits/chosen": 5835.33984375, | |
| "logits/rejected": 5479.173828125, | |
| "logps/chosen": -249.04843139648438, | |
| "logps/rejected": -245.552734375, | |
| "loss": 468.733, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.4238561689853668, | |
| "rewards/margins": -0.09983213245868683, | |
| "rewards/rejected": 0.5236883163452148, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "grad_norm": 1477.060718071622, | |
| "learning_rate": 1.8220596619089573e-07, | |
| "logits/chosen": 5350.3173828125, | |
| "logits/rejected": 5039.15966796875, | |
| "logps/chosen": -233.1553192138672, | |
| "logps/rejected": -227.7581024169922, | |
| "loss": 463.6427, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.3819466233253479, | |
| "rewards/margins": -0.04966121166944504, | |
| "rewards/rejected": 0.43160781264305115, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6490447526825438, | |
| "grad_norm": 1467.0428814727877, | |
| "learning_rate": 1.647817538357072e-07, | |
| "logits/chosen": 5974.5576171875, | |
| "logits/rejected": 5248.5712890625, | |
| "logps/chosen": -246.29739379882812, | |
| "logps/rejected": -230.0133819580078, | |
| "loss": 451.9249, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.4453219473361969, | |
| "rewards/margins": 0.049888670444488525, | |
| "rewards/rejected": 0.395433247089386, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6699816801884323, | |
| "grad_norm": 1539.7938960746517, | |
| "learning_rate": 1.478143389201113e-07, | |
| "logits/chosen": 5526.9794921875, | |
| "logits/rejected": 4800.4169921875, | |
| "logps/chosen": -232.0591583251953, | |
| "logps/rejected": -218.22909545898438, | |
| "loss": 457.9831, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.3980428874492645, | |
| "rewards/margins": -0.017790189012885094, | |
| "rewards/rejected": 0.41583308577537537, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6909186076943209, | |
| "grad_norm": 1416.4774770653923, | |
| "learning_rate": 1.3139467229135998e-07, | |
| "logits/chosen": 5347.95751953125, | |
| "logits/rejected": 4977.25830078125, | |
| "logps/chosen": -237.83120727539062, | |
| "logps/rejected": -223.06503295898438, | |
| "loss": 456.2857, | |
| "rewards/accuracies": 0.515625, | |
| "rewards/chosen": 0.3895305097103119, | |
| "rewards/margins": -0.023113062605261803, | |
| "rewards/rejected": 0.41264358162879944, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7118555352002094, | |
| "grad_norm": 1354.3420498241383, | |
| "learning_rate": 1.1561076868822755e-07, | |
| "logits/chosen": 5184.0341796875, | |
| "logits/rejected": 4512.36376953125, | |
| "logps/chosen": -215.9802703857422, | |
| "logps/rejected": -202.09982299804688, | |
| "loss": 441.2616, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.4158664345741272, | |
| "rewards/margins": 0.02483862265944481, | |
| "rewards/rejected": 0.3910277783870697, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7327924627060979, | |
| "grad_norm": 1399.0816486347255, | |
| "learning_rate": 1.0054723495346482e-07, | |
| "logits/chosen": 5854.7001953125, | |
| "logits/rejected": 4766.1748046875, | |
| "logps/chosen": -257.1776428222656, | |
| "logps/rejected": -228.104736328125, | |
| "loss": 471.4627, | |
| "rewards/accuracies": 0.534375011920929, | |
| "rewards/chosen": 0.4373684823513031, | |
| "rewards/margins": 0.03503155708312988, | |
| "rewards/rejected": 0.402336984872818, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7537293902119864, | |
| "grad_norm": 1352.127266112888, | |
| "learning_rate": 8.628481651367875e-08, | |
| "logits/chosen": 5555.2060546875, | |
| "logits/rejected": 4852.7666015625, | |
| "logps/chosen": -235.2597198486328, | |
| "logps/rejected": -216.91415405273438, | |
| "loss": 443.8227, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.42463842034339905, | |
| "rewards/margins": 0.01125773973762989, | |
| "rewards/rejected": 0.4133806824684143, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7746663177178749, | |
| "grad_norm": 1283.6956906144683, | |
| "learning_rate": 7.289996455765748e-08, | |
| "logits/chosen": 5428.78271484375, | |
| "logits/rejected": 4644.25390625, | |
| "logps/chosen": -251.05245971679688, | |
| "logps/rejected": -221.3125, | |
| "loss": 453.7341, | |
| "rewards/accuracies": 0.578125, | |
| "rewards/chosen": 0.3774058520793915, | |
| "rewards/margins": 0.02088645100593567, | |
| "rewards/rejected": 0.3565194010734558, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7956032452237635, | |
| "grad_norm": 1389.3936014370584, | |
| "learning_rate": 6.046442623320145e-08, | |
| "logits/chosen": 5676.46826171875, | |
| "logits/rejected": 4961.1416015625, | |
| "logps/chosen": -245.52609252929688, | |
| "logps/rejected": -210.39529418945312, | |
| "loss": 472.2963, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.39175355434417725, | |
| "rewards/margins": -0.012869933620095253, | |
| "rewards/rejected": 0.40462350845336914, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.816540172729652, | |
| "grad_norm": 1904.7699931128577, | |
| "learning_rate": 4.904486005914027e-08, | |
| "logits/chosen": 5950.8603515625, | |
| "logits/rejected": 5257.07080078125, | |
| "logps/chosen": -246.93521118164062, | |
| "logps/rejected": -232.27798461914062, | |
| "loss": 459.8493, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.44814223051071167, | |
| "rewards/margins": 0.023155368864536285, | |
| "rewards/rejected": 0.42498689889907837, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "grad_norm": 1314.6946141188957, | |
| "learning_rate": 3.8702478614051345e-08, | |
| "logits/chosen": 5522.93212890625, | |
| "logits/rejected": 4914.50146484375, | |
| "logps/chosen": -246.7040557861328, | |
| "logps/rejected": -212.8903045654297, | |
| "loss": 463.632, | |
| "rewards/accuracies": 0.528124988079071, | |
| "rewards/chosen": 0.4420008063316345, | |
| "rewards/margins": 0.01232604868710041, | |
| "rewards/rejected": 0.42967477440834045, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8584140277414289, | |
| "grad_norm": 1386.6765646261215, | |
| "learning_rate": 2.9492720416985e-08, | |
| "logits/chosen": 5542.53125, | |
| "logits/rejected": 4905.96435546875, | |
| "logps/chosen": -233.2512664794922, | |
| "logps/rejected": -216.3258819580078, | |
| "loss": 459.2343, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.4861605167388916, | |
| "rewards/margins": -0.16168272495269775, | |
| "rewards/rejected": 0.6478432416915894, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8793509552473174, | |
| "grad_norm": 1421.0973915303414, | |
| "learning_rate": 2.1464952759020856e-08, | |
| "logits/chosen": 5875.4296875, | |
| "logits/rejected": 5138.45947265625, | |
| "logps/chosen": -249.4132843017578, | |
| "logps/rejected": -230.8553009033203, | |
| "loss": 457.3266, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": 0.42565909028053284, | |
| "rewards/margins": 0.009328785352408886, | |
| "rewards/rejected": 0.4163302779197693, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9002878827532059, | |
| "grad_norm": 2530.194729647411, | |
| "learning_rate": 1.4662207078575684e-08, | |
| "logits/chosen": 5782.8046875, | |
| "logits/rejected": 4674.46826171875, | |
| "logps/chosen": -244.35745239257812, | |
| "logps/rejected": -218.1392364501953, | |
| "loss": 449.309, | |
| "rewards/accuracies": 0.578125, | |
| "rewards/chosen": 0.4324522614479065, | |
| "rewards/margins": 0.03925667330622673, | |
| "rewards/rejected": 0.39319556951522827, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9212248102590945, | |
| "grad_norm": 1331.0653218757734, | |
| "learning_rate": 9.12094829893642e-09, | |
| "logits/chosen": 5607.314453125, | |
| "logits/rejected": 5351.4794921875, | |
| "logps/chosen": -240.36508178710938, | |
| "logps/rejected": -235.7848358154297, | |
| "loss": 447.1441, | |
| "rewards/accuracies": 0.47187501192092896, | |
| "rewards/chosen": 0.3817844092845917, | |
| "rewards/margins": -0.020826738327741623, | |
| "rewards/rejected": 0.4026111960411072, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.942161737764983, | |
| "grad_norm": 1400.3246977795131, | |
| "learning_rate": 4.8708793644441086e-09, | |
| "logits/chosen": 5760.21923828125, | |
| "logits/rejected": 4969.435546875, | |
| "logps/chosen": -248.47109985351562, | |
| "logps/rejected": -217.5637969970703, | |
| "loss": 450.0296, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.41864126920700073, | |
| "rewards/margins": -0.020225150510668755, | |
| "rewards/rejected": 0.43886643648147583, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9630986652708715, | |
| "grad_norm": 1459.4316477205139, | |
| "learning_rate": 1.9347820230782295e-09, | |
| "logits/chosen": 5948.6728515625, | |
| "logits/rejected": 5038.36181640625, | |
| "logps/chosen": -268.30438232421875, | |
| "logps/rejected": -239.8975067138672, | |
| "loss": 486.2789, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.4469468593597412, | |
| "rewards/margins": -0.011204786598682404, | |
| "rewards/rejected": 0.4581516683101654, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98403559277676, | |
| "grad_norm": 1500.642382585496, | |
| "learning_rate": 3.2839470889836627e-10, | |
| "logits/chosen": 5545.70947265625, | |
| "logits/rejected": 4929.12255859375, | |
| "logps/chosen": -237.06051635742188, | |
| "logps/rejected": -213.99795532226562, | |
| "loss": 457.8427, | |
| "rewards/accuracies": 0.528124988079071, | |
| "rewards/chosen": 0.4377865195274353, | |
| "rewards/margins": -0.0002623729524202645, | |
| "rewards/rejected": 0.4380488991737366, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.998691442030882, | |
| "step": 477, | |
| "total_flos": 0.0, | |
| "train_loss": 465.78390522123135, | |
| "train_runtime": 16522.4635, | |
| "train_samples_per_second": 3.7, | |
| "train_steps_per_second": 0.029 | |
| } | |
| ], | |
| "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 | |
| } | |