| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 625, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 7.936507936507937e-08, | |
| "logits/chosen": 0.0711287260055542, | |
| "logits/rejected": 0.20400863885879517, | |
| "logps/chosen": -313.75396728515625, | |
| "logps/rejected": -420.73980712890625, | |
| "loss": 0.2285, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 7.936507936507937e-07, | |
| "logits/chosen": 0.2657856047153473, | |
| "logits/rejected": 0.23175562918186188, | |
| "logps/chosen": -354.52484130859375, | |
| "logps/rejected": -365.6746520996094, | |
| "loss": 0.203, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -0.0005699184257537127, | |
| "rewards/margins": 2.8342270525172353e-05, | |
| "rewards/rejected": -0.0005982607253827155, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.5873015873015873e-06, | |
| "logits/chosen": 0.16931554675102234, | |
| "logits/rejected": 0.265936940908432, | |
| "logps/chosen": -376.02166748046875, | |
| "logps/rejected": -396.56524658203125, | |
| "loss": 0.2032, | |
| "rewards/accuracies": 0.3062500059604645, | |
| "rewards/chosen": -0.0009135094587691128, | |
| "rewards/margins": -1.457883081457112e-05, | |
| "rewards/rejected": -0.000898930593393743, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.380952380952381e-06, | |
| "logits/chosen": 0.1378197968006134, | |
| "logits/rejected": 0.2801808714866638, | |
| "logps/chosen": -394.52301025390625, | |
| "logps/rejected": -406.01971435546875, | |
| "loss": 0.2045, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -0.0023785685189068317, | |
| "rewards/margins": -4.536235792329535e-05, | |
| "rewards/rejected": -0.0023332058917731047, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.1746031746031746e-06, | |
| "logits/chosen": 0.22031190991401672, | |
| "logits/rejected": 0.2555253505706787, | |
| "logps/chosen": -413.9378356933594, | |
| "logps/rejected": -384.2980041503906, | |
| "loss": 0.2078, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.0029828939586877823, | |
| "rewards/margins": 0.0004020760825369507, | |
| "rewards/rejected": -0.0033849701285362244, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 3.968253968253968e-06, | |
| "logits/chosen": 0.17930440604686737, | |
| "logits/rejected": 0.27743226289749146, | |
| "logps/chosen": -437.57220458984375, | |
| "logps/rejected": -428.25726318359375, | |
| "loss": 0.2109, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -0.0068945749662816525, | |
| "rewards/margins": 0.0003428836935199797, | |
| "rewards/rejected": -0.0072374590672552586, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.761904761904762e-06, | |
| "logits/chosen": 0.1960189789533615, | |
| "logits/rejected": 0.20567241311073303, | |
| "logps/chosen": -412.0870666503906, | |
| "logps/rejected": -429.5210876464844, | |
| "loss": 0.2106, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.01030620839446783, | |
| "rewards/margins": 0.0008126860484480858, | |
| "rewards/rejected": -0.011118894442915916, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 4.998086282661188e-06, | |
| "logits/chosen": 0.18384099006652832, | |
| "logits/rejected": 0.16322512924671173, | |
| "logps/chosen": -480.0482482910156, | |
| "logps/rejected": -449.12872314453125, | |
| "loss": 0.2125, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.012364747002720833, | |
| "rewards/margins": 0.0012391259660944343, | |
| "rewards/rejected": -0.013603871688246727, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.988720025682995e-06, | |
| "logits/chosen": 0.17162439227104187, | |
| "logits/rejected": 0.3024311661720276, | |
| "logps/chosen": -425.10821533203125, | |
| "logps/rejected": -405.69903564453125, | |
| "loss": 0.203, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.002422564197331667, | |
| "rewards/margins": 0.0019961665384471416, | |
| "rewards/rejected": -0.004418730735778809, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 4.9715789537359126e-06, | |
| "logits/chosen": 0.14976395666599274, | |
| "logits/rejected": 0.293326199054718, | |
| "logps/chosen": -450.09283447265625, | |
| "logps/rejected": -463.3335876464844, | |
| "loss": 0.1928, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.004968429449945688, | |
| "rewards/margins": 0.004259931854903698, | |
| "rewards/rejected": 0.0007084974786266685, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.946716615897932e-06, | |
| "logits/chosen": 0.14262747764587402, | |
| "logits/rejected": 0.14231689274311066, | |
| "logps/chosen": -409.68719482421875, | |
| "logps/rejected": -385.9997253417969, | |
| "loss": 0.2044, | |
| "rewards/accuracies": 0.42500001192092896, | |
| "rewards/chosen": 0.0011129004415124655, | |
| "rewards/margins": 0.0027904310263693333, | |
| "rewards/rejected": -0.001677530468441546, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.9142106826480114e-06, | |
| "logits/chosen": 0.09172149002552032, | |
| "logits/rejected": 0.18723782896995544, | |
| "logps/chosen": -376.23382568359375, | |
| "logps/rejected": -382.14312744140625, | |
| "loss": 0.2117, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": 0.0011799463536590338, | |
| "rewards/margins": 0.006289638578891754, | |
| "rewards/rejected": -0.005109691992402077, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.874162703221823e-06, | |
| "logits/chosen": 0.21056421101093292, | |
| "logits/rejected": 0.10638086497783661, | |
| "logps/chosen": -455.8184509277344, | |
| "logps/rejected": -461.58575439453125, | |
| "loss": 0.2024, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": 0.0018554453272372484, | |
| "rewards/margins": 0.007339824922382832, | |
| "rewards/rejected": -0.005484379827976227, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.826697788369752e-06, | |
| "logits/chosen": 0.0843835398554802, | |
| "logits/rejected": 0.10346312820911407, | |
| "logps/chosen": -420.7416076660156, | |
| "logps/rejected": -426.8565979003906, | |
| "loss": 0.1938, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": 0.0005354422028176486, | |
| "rewards/margins": 0.012502538040280342, | |
| "rewards/rejected": -0.011967095546424389, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.7719642195082224e-06, | |
| "logits/chosen": 0.03641371801495552, | |
| "logits/rejected": 0.10479255020618439, | |
| "logps/chosen": -401.6028137207031, | |
| "logps/rejected": -464.2025451660156, | |
| "loss": 0.1958, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.024544300511479378, | |
| "rewards/margins": 0.018662814050912857, | |
| "rewards/rejected": -0.043207116425037384, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.710132985485355e-06, | |
| "logits/chosen": 0.11580105125904083, | |
| "logits/rejected": 0.021813513711094856, | |
| "logps/chosen": -430.5687561035156, | |
| "logps/rejected": -515.4051513671875, | |
| "loss": 0.1976, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -0.04684633016586304, | |
| "rewards/margins": 0.043844569474458694, | |
| "rewards/rejected": -0.09069089591503143, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.641397248408122e-06, | |
| "logits/chosen": -0.0415755994617939, | |
| "logits/rejected": 0.07411660254001617, | |
| "logps/chosen": -467.6756286621094, | |
| "logps/rejected": -472.20001220703125, | |
| "loss": 0.1972, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.0470997616648674, | |
| "rewards/margins": 0.02540501020848751, | |
| "rewards/rejected": -0.07250477373600006, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.5659717401997655e-06, | |
| "logits/chosen": -0.03600798547267914, | |
| "logits/rejected": 0.027341466397047043, | |
| "logps/chosen": -504.47894287109375, | |
| "logps/rejected": -516.9270629882812, | |
| "loss": 0.1948, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": -0.05515031889081001, | |
| "rewards/margins": 0.03458093851804733, | |
| "rewards/rejected": -0.08973126113414764, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.4840920917726425e-06, | |
| "logits/chosen": -0.039528343826532364, | |
| "logits/rejected": -0.019377198070287704, | |
| "logps/chosen": -484.7760314941406, | |
| "logps/rejected": -515.7113037109375, | |
| "loss": 0.1955, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -0.05891401320695877, | |
| "rewards/margins": 0.03348463773727417, | |
| "rewards/rejected": -0.09239865839481354, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.396014096912182e-06, | |
| "logits/chosen": 0.041537586599588394, | |
| "logits/rejected": -0.029888898134231567, | |
| "logps/chosen": -474.3414001464844, | |
| "logps/rejected": -515.9990234375, | |
| "loss": 0.1919, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.07046758383512497, | |
| "rewards/margins": 0.03831402584910393, | |
| "rewards/rejected": -0.1087816134095192, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.302012913171584e-06, | |
| "logits/chosen": -0.017491130158305168, | |
| "logits/rejected": -0.04960538074374199, | |
| "logps/chosen": -484.7154846191406, | |
| "logps/rejected": -524.5064086914062, | |
| "loss": 0.1935, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -0.07989688217639923, | |
| "rewards/margins": 0.04493844509124756, | |
| "rewards/rejected": -0.12483533471822739, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.202382202273702e-06, | |
| "logits/chosen": -0.13748802244663239, | |
| "logits/rejected": -0.001641835318878293, | |
| "logps/chosen": -507.37506103515625, | |
| "logps/rejected": -535.3721923828125, | |
| "loss": 0.1804, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.07145627588033676, | |
| "rewards/margins": 0.051401056349277496, | |
| "rewards/rejected": -0.12285731732845306, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.097433212705492e-06, | |
| "logits/chosen": -0.021799543872475624, | |
| "logits/rejected": -0.074518583714962, | |
| "logps/chosen": -477.99981689453125, | |
| "logps/rejected": -518.2962646484375, | |
| "loss": 0.2019, | |
| "rewards/accuracies": 0.4437499940395355, | |
| "rewards/chosen": -0.09213604032993317, | |
| "rewards/margins": 0.04488346725702286, | |
| "rewards/rejected": -0.13701951503753662, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 3.987493807371033e-06, | |
| "logits/chosen": -0.16067767143249512, | |
| "logits/rejected": 0.01047535240650177, | |
| "logps/chosen": -489.7818298339844, | |
| "logps/rejected": -525.8330688476562, | |
| "loss": 0.1852, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": -0.0610770583152771, | |
| "rewards/margins": 0.04805826395750046, | |
| "rewards/rejected": -0.10913531482219696, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.872907439340758e-06, | |
| "logits/chosen": -0.035437412559986115, | |
| "logits/rejected": -0.07582642138004303, | |
| "logps/chosen": -440.7459411621094, | |
| "logps/rejected": -492.434326171875, | |
| "loss": 0.177, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": -0.04167747497558594, | |
| "rewards/margins": 0.04895929619669914, | |
| "rewards/rejected": -0.09063677489757538, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.75403207889666e-06, | |
| "logits/chosen": -0.09768973290920258, | |
| "logits/rejected": -0.0913015678524971, | |
| "logps/chosen": -445.75750732421875, | |
| "logps/rejected": -473.06671142578125, | |
| "loss": 0.1765, | |
| "rewards/accuracies": 0.45625001192092896, | |
| "rewards/chosen": -0.04659276083111763, | |
| "rewards/margins": 0.03554708510637283, | |
| "rewards/rejected": -0.08213984221220016, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.631239095225417e-06, | |
| "logits/chosen": -0.002584155648946762, | |
| "logits/rejected": -0.0503731295466423, | |
| "logps/chosen": -443.8072204589844, | |
| "logps/rejected": -522.3037109375, | |
| "loss": 0.1903, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.057579755783081055, | |
| "rewards/margins": 0.042577676475048065, | |
| "rewards/rejected": -0.10015741735696793, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.5049120962530608e-06, | |
| "logits/chosen": -0.07919616252183914, | |
| "logits/rejected": -0.051559675484895706, | |
| "logps/chosen": -442.90020751953125, | |
| "logps/rejected": -506.71026611328125, | |
| "loss": 0.1801, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": -0.04894575849175453, | |
| "rewards/margins": 0.060835689306259155, | |
| "rewards/rejected": -0.1097814291715622, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.3754457302455464e-06, | |
| "logits/chosen": -0.05536097288131714, | |
| "logits/rejected": -0.05594850331544876, | |
| "logps/chosen": -491.963623046875, | |
| "logps/rejected": -476.1817932128906, | |
| "loss": 0.1987, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.05905281379818916, | |
| "rewards/margins": 0.03382394462823868, | |
| "rewards/rejected": -0.09287675470113754, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.2432444529190714e-06, | |
| "logits/chosen": -0.06521332263946533, | |
| "logits/rejected": 0.04030764847993851, | |
| "logps/chosen": -503.37115478515625, | |
| "logps/rejected": -539.033935546875, | |
| "loss": 0.1758, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": -0.06568136066198349, | |
| "rewards/margins": 0.04754118248820305, | |
| "rewards/rejected": -0.11322255432605743, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.1087212639117057e-06, | |
| "logits/chosen": -0.007789143826812506, | |
| "logits/rejected": -0.08723556995391846, | |
| "logps/chosen": -448.15185546875, | |
| "logps/rejected": -472.66046142578125, | |
| "loss": 0.195, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -0.04991251975297928, | |
| "rewards/margins": 0.03361000493168831, | |
| "rewards/rejected": -0.08352252840995789, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.9722964165636263e-06, | |
| "logits/chosen": -0.10513603687286377, | |
| "logits/rejected": -0.02427390217781067, | |
| "logps/chosen": -456.4302673339844, | |
| "logps/rejected": -480.01904296875, | |
| "loss": 0.1777, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": -0.04909784346818924, | |
| "rewards/margins": 0.032266296446323395, | |
| "rewards/rejected": -0.08136413991451263, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 2.8343961050366275e-06, | |
| "logits/chosen": -0.06911730021238327, | |
| "logits/rejected": -0.02331097424030304, | |
| "logps/chosen": -412.3163146972656, | |
| "logps/rejected": -419.2935485839844, | |
| "loss": 0.1807, | |
| "rewards/accuracies": 0.45625001192092896, | |
| "rewards/chosen": -0.031493835151195526, | |
| "rewards/margins": 0.03578699007630348, | |
| "rewards/rejected": -0.06728082150220871, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 2.695451132874385e-06, | |
| "logits/chosen": 0.030731942504644394, | |
| "logits/rejected": -0.028069671243429184, | |
| "logps/chosen": -477.44512939453125, | |
| "logps/rejected": -501.154541015625, | |
| "loss": 0.1925, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -0.045379411429166794, | |
| "rewards/margins": 0.04541900008916855, | |
| "rewards/rejected": -0.09079841524362564, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.5558955671628964e-06, | |
| "logits/chosen": -0.0826629176735878, | |
| "logits/rejected": -0.04008691757917404, | |
| "logps/chosen": -429.1664123535156, | |
| "logps/rejected": -467.46173095703125, | |
| "loss": 0.1838, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.05404701828956604, | |
| "rewards/margins": 0.04224228113889694, | |
| "rewards/rejected": -0.09628931432962418, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.4161653824955654e-06, | |
| "logits/chosen": -0.056598931550979614, | |
| "logits/rejected": -0.10974061489105225, | |
| "logps/chosen": -513.0817260742188, | |
| "logps/rejected": -546.1875, | |
| "loss": 0.1864, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.06718280166387558, | |
| "rewards/margins": 0.05270420387387276, | |
| "rewards/rejected": -0.11988700926303864, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.2766970989791697e-06, | |
| "logits/chosen": -0.04662217199802399, | |
| "logits/rejected": -0.05574822425842285, | |
| "logps/chosen": -480.88848876953125, | |
| "logps/rejected": -530.1246337890625, | |
| "loss": 0.1915, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.06772033870220184, | |
| "rewards/margins": 0.03862085938453674, | |
| "rewards/rejected": -0.10634119808673859, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.1379264185356545e-06, | |
| "logits/chosen": -0.020046677440404892, | |
| "logits/rejected": 0.09056108444929123, | |
| "logps/chosen": -486.93115234375, | |
| "logps/rejected": -528.6192626953125, | |
| "loss": 0.1814, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -0.054186802357435226, | |
| "rewards/margins": 0.05737306550145149, | |
| "rewards/rejected": -0.11155986785888672, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 2.000286863759934e-06, | |
| "logits/chosen": -0.09826477617025375, | |
| "logits/rejected": -0.07946722954511642, | |
| "logps/chosen": -470.00811767578125, | |
| "logps/rejected": -511.309814453125, | |
| "loss": 0.1828, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.06328146904706955, | |
| "rewards/margins": 0.05404907464981079, | |
| "rewards/rejected": -0.11733055114746094, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.8642084235859764e-06, | |
| "logits/chosen": -0.15358105301856995, | |
| "logits/rejected": -0.07723913341760635, | |
| "logps/chosen": -438.9100646972656, | |
| "logps/rejected": -479.63092041015625, | |
| "loss": 0.1972, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": -0.05276118963956833, | |
| "rewards/margins": 0.04282676801085472, | |
| "rewards/rejected": -0.09558796137571335, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.7301162099921013e-06, | |
| "logits/chosen": -0.06951303780078888, | |
| "logits/rejected": -0.16843798756599426, | |
| "logps/chosen": -422.1139221191406, | |
| "logps/rejected": -459.0763244628906, | |
| "loss": 0.1995, | |
| "rewards/accuracies": 0.45625001192092896, | |
| "rewards/chosen": -0.05717768147587776, | |
| "rewards/margins": 0.040271807461977005, | |
| "rewards/rejected": -0.09744948893785477, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1.5984291299420117e-06, | |
| "logits/chosen": -0.1256350576877594, | |
| "logits/rejected": -0.08233270049095154, | |
| "logps/chosen": -404.26263427734375, | |
| "logps/rejected": -455.4290466308594, | |
| "loss": 0.1892, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.04233751818537712, | |
| "rewards/margins": 0.05102572590112686, | |
| "rewards/rejected": -0.09336324036121368, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.4695585767104092e-06, | |
| "logits/chosen": -0.0601225309073925, | |
| "logits/rejected": -0.12005972862243652, | |
| "logps/chosen": -482.4237365722656, | |
| "logps/rejected": -531.4822998046875, | |
| "loss": 0.1715, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.045725829899311066, | |
| "rewards/margins": 0.04561304301023483, | |
| "rewards/rejected": -0.0913388729095459, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3439071446815452e-06, | |
| "logits/chosen": -0.15923841297626495, | |
| "logits/rejected": 0.011970462277531624, | |
| "logps/chosen": -441.70648193359375, | |
| "logps/rejected": -458.97052001953125, | |
| "loss": 0.186, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -0.03852088749408722, | |
| "rewards/margins": 0.038460638374090195, | |
| "rewards/rejected": -0.07698152959346771, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.2218673716356919e-06, | |
| "logits/chosen": -0.05337870121002197, | |
| "logits/rejected": -0.16203683614730835, | |
| "logps/chosen": -411.7693786621094, | |
| "logps/rejected": -472.7538146972656, | |
| "loss": 0.1868, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -0.04310641810297966, | |
| "rewards/margins": 0.053300343453884125, | |
| "rewards/rejected": -0.09640677273273468, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1.103820512452661e-06, | |
| "logits/chosen": -0.16054467856884003, | |
| "logits/rejected": -0.10418369621038437, | |
| "logps/chosen": -471.15740966796875, | |
| "logps/rejected": -499.28997802734375, | |
| "loss": 0.1866, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.044102419167757034, | |
| "rewards/margins": 0.04977239668369293, | |
| "rewards/rejected": -0.09387481212615967, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 9.901353480633468e-07, | |
| "logits/chosen": -0.01945580169558525, | |
| "logits/rejected": -0.08020860701799393, | |
| "logps/chosen": -478.95135498046875, | |
| "logps/rejected": -510.05645751953125, | |
| "loss": 0.1876, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -0.05099078267812729, | |
| "rewards/margins": 0.04500190168619156, | |
| "rewards/rejected": -0.09599269926548004, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.811670333701544e-07, | |
| "logits/chosen": -0.09503830224275589, | |
| "logits/rejected": -0.12620458006858826, | |
| "logps/chosen": -442.7535095214844, | |
| "logps/rejected": -478.792236328125, | |
| "loss": 0.1797, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": -0.039722852408885956, | |
| "rewards/margins": 0.04783398285508156, | |
| "rewards/rejected": -0.08755683898925781, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.772559877354341e-07, | |
| "logits/chosen": -0.1489885151386261, | |
| "logits/rejected": -0.052212201058864594, | |
| "logps/chosen": -408.1684265136719, | |
| "logps/rejected": -442.21466064453125, | |
| "loss": 0.1972, | |
| "rewards/accuracies": 0.4312500059604645, | |
| "rewards/chosen": -0.05573519319295883, | |
| "rewards/margins": 0.033387087285518646, | |
| "rewards/rejected": -0.08912228047847748, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 6.787268315040604e-07, | |
| "logits/chosen": -0.05187498405575752, | |
| "logits/rejected": -0.04689168184995651, | |
| "logps/chosen": -473.5567932128906, | |
| "logps/rejected": -553.9025268554688, | |
| "loss": 0.1722, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -0.047236062586307526, | |
| "rewards/margins": 0.06369349360466003, | |
| "rewards/rejected": -0.11092956364154816, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 5.858873718824829e-07, | |
| "logits/chosen": -0.1269344836473465, | |
| "logits/rejected": -0.12827740609645844, | |
| "logps/chosen": -436.95001220703125, | |
| "logps/rejected": -530.323486328125, | |
| "loss": 0.1828, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.03250400722026825, | |
| "rewards/margins": 0.06746237725019455, | |
| "rewards/rejected": -0.0999663770198822, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.990276413423817e-07, | |
| "logits/chosen": -0.05357852578163147, | |
| "logits/rejected": -0.0927656888961792, | |
| "logps/chosen": -499.13262939453125, | |
| "logps/rejected": -548.8838500976562, | |
| "loss": 0.1697, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.04802858829498291, | |
| "rewards/margins": 0.05423368886113167, | |
| "rewards/rejected": -0.10226227343082428, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.184189915529796e-07, | |
| "logits/chosen": -0.11798721551895142, | |
| "logits/rejected": -0.09332814067602158, | |
| "logps/chosen": -465.55816650390625, | |
| "logps/rejected": -533.0569458007812, | |
| "loss": 0.1967, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.042215488851070404, | |
| "rewards/margins": 0.05858848616480827, | |
| "rewards/rejected": -0.10080397129058838, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 3.4431324567258176e-07, | |
| "logits/chosen": -0.1948605328798294, | |
| "logits/rejected": -0.08446381241083145, | |
| "logps/chosen": -440.0462951660156, | |
| "logps/rejected": -479.0397033691406, | |
| "loss": 0.1885, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": -0.045319609344005585, | |
| "rewards/margins": 0.04320630431175232, | |
| "rewards/rejected": -0.08852590620517731, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 2.769419116476052e-07, | |
| "logits/chosen": -0.2041836529970169, | |
| "logits/rejected": -0.06282895803451538, | |
| "logps/chosen": -415.84576416015625, | |
| "logps/rejected": -452.8704528808594, | |
| "loss": 0.1829, | |
| "rewards/accuracies": 0.4437499940395355, | |
| "rewards/chosen": -0.050526998937129974, | |
| "rewards/margins": 0.03672551363706589, | |
| "rewards/rejected": -0.08725249767303467, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.1651545897676512e-07, | |
| "logits/chosen": -0.056377578526735306, | |
| "logits/rejected": -0.0602828674018383, | |
| "logps/chosen": -482.2828674316406, | |
| "logps/rejected": -510.7273864746094, | |
| "loss": 0.194, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.04934944957494736, | |
| "rewards/margins": 0.04233521968126297, | |
| "rewards/rejected": -0.09168466180562973, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.6322266119983222e-07, | |
| "logits/chosen": -0.08606644719839096, | |
| "logits/rejected": -0.05609310790896416, | |
| "logps/chosen": -435.7982482910156, | |
| "logps/rejected": -460.5846252441406, | |
| "loss": 0.185, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.037545233964920044, | |
| "rewards/margins": 0.0525357648730278, | |
| "rewards/rejected": -0.09008099883794785, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.1723000616502167e-07, | |
| "logits/chosen": -0.21670150756835938, | |
| "logits/rejected": -0.034546859562397, | |
| "logps/chosen": -498.1796875, | |
| "logps/rejected": -531.9613037109375, | |
| "loss": 0.1779, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.05661952495574951, | |
| "rewards/margins": 0.05806880071759224, | |
| "rewards/rejected": -0.11468833684921265, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 7.868117591737585e-08, | |
| "logits/chosen": -0.15769873559474945, | |
| "logits/rejected": -0.11322434991598129, | |
| "logps/chosen": -501.90802001953125, | |
| "logps/rejected": -510.9298400878906, | |
| "loss": 0.1865, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -0.0535421147942543, | |
| "rewards/margins": 0.04890746995806694, | |
| "rewards/rejected": -0.10244959592819214, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 4.769659783295383e-08, | |
| "logits/chosen": -0.1612723022699356, | |
| "logits/rejected": -0.09152115881443024, | |
| "logps/chosen": -478.5709533691406, | |
| "logps/rejected": -501.6288146972656, | |
| "loss": 0.1926, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.04939908906817436, | |
| "rewards/margins": 0.04336509853601456, | |
| "rewards/rejected": -0.09276418387889862, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.4373068401120358e-08, | |
| "logits/chosen": -0.16256621479988098, | |
| "logits/rejected": 0.01677759736776352, | |
| "logps/chosen": -480.29571533203125, | |
| "logps/rejected": -480.0000915527344, | |
| "loss": 0.1777, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": -0.059523385018110275, | |
| "rewards/margins": 0.03656711056828499, | |
| "rewards/rejected": -0.09609050303697586, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 8.78345083022425e-09, | |
| "logits/chosen": -0.2046932876110077, | |
| "logits/rejected": -0.010766489431262016, | |
| "logps/chosen": -467.3860778808594, | |
| "logps/rejected": -513.6978149414062, | |
| "loss": 0.195, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.0569658987224102, | |
| "rewards/margins": 0.04467035457491875, | |
| "rewards/rejected": -0.10163625329732895, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 9.764474213677654e-10, | |
| "logits/chosen": -0.15736037492752075, | |
| "logits/rejected": -0.16347061097621918, | |
| "logps/chosen": -454.09942626953125, | |
| "logps/rejected": -512.8070678710938, | |
| "loss": 0.1987, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": -0.06590863317251205, | |
| "rewards/margins": 0.044044043868780136, | |
| "rewards/rejected": -0.10995267331600189, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 625, | |
| "total_flos": 0.0, | |
| "train_loss": 0.1911162176847458, | |
| "train_runtime": 4612.7069, | |
| "train_samples_per_second": 4.336, | |
| "train_steps_per_second": 0.135 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 625, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |