| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 50, | |
| "global_step": 352, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.028409090909090908, | |
| "grad_norm": 45.06143403082133, | |
| "learning_rate": 1.3888888888888888e-07, | |
| "logits/chosen": -2.859044075012207, | |
| "logits/rejected": -2.64284086227417, | |
| "logps/chosen": -390.5040588378906, | |
| "logps/rejected": -607.8244018554688, | |
| "loss": 0.6882, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.002557548927143216, | |
| "rewards/margins": 0.013725275173783302, | |
| "rewards/rejected": -0.01116772647947073, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.056818181818181816, | |
| "grad_norm": 19.067293358665005, | |
| "learning_rate": 2.7777777777777776e-07, | |
| "logits/chosen": -2.8423688411712646, | |
| "logits/rejected": -2.6946520805358887, | |
| "logps/chosen": -328.6228332519531, | |
| "logps/rejected": -774.904541015625, | |
| "loss": 0.5132, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": 0.08522307872772217, | |
| "rewards/margins": 0.6598724722862244, | |
| "rewards/rejected": -0.574649453163147, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08522727272727272, | |
| "grad_norm": 4.736974625657813, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.8571083545684814, | |
| "logits/rejected": -2.6747167110443115, | |
| "logps/chosen": -295.3691711425781, | |
| "logps/rejected": -1112.830078125, | |
| "loss": 0.1792, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.36912840604782104, | |
| "rewards/margins": 4.717885494232178, | |
| "rewards/rejected": -4.348757266998291, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.11363636363636363, | |
| "grad_norm": 1.4117830566891867, | |
| "learning_rate": 4.998023493068254e-07, | |
| "logits/chosen": -2.834780216217041, | |
| "logits/rejected": -2.644721508026123, | |
| "logps/chosen": -314.90582275390625, | |
| "logps/rejected": -2130.47900390625, | |
| "loss": 0.0415, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.31849661469459534, | |
| "rewards/margins": 14.544331550598145, | |
| "rewards/rejected": -14.225835800170898, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.14204545454545456, | |
| "grad_norm": 0.09320830768011595, | |
| "learning_rate": 4.975823666181255e-07, | |
| "logits/chosen": -2.814044713973999, | |
| "logits/rejected": -2.548596143722534, | |
| "logps/chosen": -437.7023010253906, | |
| "logps/rejected": -4283.9775390625, | |
| "loss": 0.0055, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7506707906723022, | |
| "rewards/margins": 35.5496940612793, | |
| "rewards/rejected": -36.30036926269531, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.14204545454545456, | |
| "eval_logits/chosen": -2.9176247119903564, | |
| "eval_logits/rejected": -2.528038263320923, | |
| "eval_logps/chosen": -455.486572265625, | |
| "eval_logps/rejected": -4884.189453125, | |
| "eval_loss": 0.0030528793577104807, | |
| "eval_rewards/accuracies": 0.9959677457809448, | |
| "eval_rewards/chosen": -0.8850486874580383, | |
| "eval_rewards/margins": 41.54896545410156, | |
| "eval_rewards/rejected": -42.43401336669922, | |
| "eval_runtime": 193.4773, | |
| "eval_samples_per_second": 20.183, | |
| "eval_steps_per_second": 0.32, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.17045454545454544, | |
| "grad_norm": 1.0625936242535339, | |
| "learning_rate": 4.929173350101024e-07, | |
| "logits/chosen": -2.990762710571289, | |
| "logits/rejected": -2.484398603439331, | |
| "logps/chosen": -463.26251220703125, | |
| "logps/rejected": -5167.1591796875, | |
| "loss": 0.0033, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -1.033675193786621, | |
| "rewards/margins": 44.20937728881836, | |
| "rewards/rejected": -45.2430534362793, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.19886363636363635, | |
| "grad_norm": 0.16389340246703632, | |
| "learning_rate": 4.858533249305336e-07, | |
| "logits/chosen": -2.814481735229492, | |
| "logits/rejected": -1.7779146432876587, | |
| "logps/chosen": -507.0504455566406, | |
| "logps/rejected": -5547.2275390625, | |
| "loss": 0.0031, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.2788976430892944, | |
| "rewards/margins": 48.066017150878906, | |
| "rewards/rejected": -49.34491729736328, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.22727272727272727, | |
| "grad_norm": 15.588141585226081, | |
| "learning_rate": 4.764600984163808e-07, | |
| "logits/chosen": -3.000598669052124, | |
| "logits/rejected": -1.848249077796936, | |
| "logps/chosen": -532.5420532226562, | |
| "logps/rejected": -6873.24853515625, | |
| "loss": 0.0039, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.2622119188308716, | |
| "rewards/margins": 60.732948303222656, | |
| "rewards/rejected": -61.99515914916992, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.2556818181818182, | |
| "grad_norm": 0.02323120317433541, | |
| "learning_rate": 4.6483042014491527e-07, | |
| "logits/chosen": -3.130722761154175, | |
| "logits/rejected": -2.385204792022705, | |
| "logps/chosen": -520.1354370117188, | |
| "logps/rejected": -6028.5537109375, | |
| "loss": 0.0138, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -1.5802205801010132, | |
| "rewards/margins": 52.33146286010742, | |
| "rewards/rejected": -53.91168212890625, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.2840909090909091, | |
| "grad_norm": 0.16442080948160046, | |
| "learning_rate": 4.510791413176912e-07, | |
| "logits/chosen": -2.8645615577697754, | |
| "logits/rejected": -1.0432738065719604, | |
| "logps/chosen": -532.5404052734375, | |
| "logps/rejected": -6479.7802734375, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.6307029724121094, | |
| "rewards/margins": 56.10929489135742, | |
| "rewards/rejected": -57.73999786376953, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2840909090909091, | |
| "eval_logits/chosen": -2.5817885398864746, | |
| "eval_logits/rejected": -0.1435869336128235, | |
| "eval_logps/chosen": -541.6495971679688, | |
| "eval_logps/rejected": -6361.2880859375, | |
| "eval_loss": 0.0003222826635465026, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -1.7466782331466675, | |
| "eval_rewards/margins": 55.458316802978516, | |
| "eval_rewards/rejected": -57.20499801635742, | |
| "eval_runtime": 190.7671, | |
| "eval_samples_per_second": 20.47, | |
| "eval_steps_per_second": 0.325, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3125, | |
| "grad_norm": 0.03669115199922144, | |
| "learning_rate": 4.353420654246546e-07, | |
| "logits/chosen": -2.299072742462158, | |
| "logits/rejected": 0.060857243835926056, | |
| "logps/chosen": -552.1079711914062, | |
| "logps/rejected": -6287.04736328125, | |
| "loss": 0.0024, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.766867995262146, | |
| "rewards/margins": 55.110923767089844, | |
| "rewards/rejected": -56.87778854370117, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.3409090909090909, | |
| "grad_norm": 0.03951441158135731, | |
| "learning_rate": 4.177746070897592e-07, | |
| "logits/chosen": -2.019587278366089, | |
| "logits/rejected": 0.7790449261665344, | |
| "logps/chosen": -561.4166870117188, | |
| "logps/rejected": -7041.1259765625, | |
| "loss": 0.0033, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.798998236656189, | |
| "rewards/margins": 62.23997116088867, | |
| "rewards/rejected": -64.03898620605469, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.3693181818181818, | |
| "grad_norm": 0.3141553784150294, | |
| "learning_rate": 3.9855025724292763e-07, | |
| "logits/chosen": -2.1000161170959473, | |
| "logits/rejected": 1.278324842453003, | |
| "logps/chosen": -563.4060668945312, | |
| "logps/rejected": -7105.3544921875, | |
| "loss": 0.0027, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -1.929544448852539, | |
| "rewards/margins": 62.7186164855957, | |
| "rewards/rejected": -64.64815521240234, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.3977272727272727, | |
| "grad_norm": 0.5821229491227392, | |
| "learning_rate": 3.7785886977585555e-07, | |
| "logits/chosen": -2.022021532058716, | |
| "logits/rejected": 1.353015422821045, | |
| "logps/chosen": -552.7689208984375, | |
| "logps/rejected": -7177.63427734375, | |
| "loss": 0.0017, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.867326021194458, | |
| "rewards/margins": 63.489112854003906, | |
| "rewards/rejected": -65.35643768310547, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.42613636363636365, | |
| "grad_norm": 0.7534850925649158, | |
| "learning_rate": 3.5590478660213206e-07, | |
| "logits/chosen": -1.8572208881378174, | |
| "logits/rejected": 1.6534769535064697, | |
| "logps/chosen": -621.11181640625, | |
| "logps/rejected": -7095.47802734375, | |
| "loss": 0.001, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -2.4304866790771484, | |
| "rewards/margins": 61.404510498046875, | |
| "rewards/rejected": -63.834999084472656, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.42613636363636365, | |
| "eval_logits/chosen": -1.9097061157226562, | |
| "eval_logits/rejected": 1.942201852798462, | |
| "eval_logps/chosen": -607.574462890625, | |
| "eval_logps/rejected": -7308.666015625, | |
| "eval_loss": 0.0001972884201677516, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -2.4059276580810547, | |
| "eval_rewards/margins": 64.27284240722656, | |
| "eval_rewards/rejected": -66.67876434326172, | |
| "eval_runtime": 193.1671, | |
| "eval_samples_per_second": 20.216, | |
| "eval_steps_per_second": 0.321, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.45454545454545453, | |
| "grad_norm": 0.5761003352865638, | |
| "learning_rate": 3.3290481963801696e-07, | |
| "logits/chosen": -2.1488444805145264, | |
| "logits/rejected": 1.8684966564178467, | |
| "logps/chosen": -605.7991943359375, | |
| "logps/rejected": -7330.78759765625, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.356612205505371, | |
| "rewards/margins": 64.78636169433594, | |
| "rewards/rejected": -67.14297485351562, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.48295454545454547, | |
| "grad_norm": 0.0136866819292242, | |
| "learning_rate": 3.0908610963322626e-07, | |
| "logits/chosen": -2.361186981201172, | |
| "logits/rejected": 1.371411919593811, | |
| "logps/chosen": -656.9031982421875, | |
| "logps/rejected": -7962.04443359375, | |
| "loss": 0.0059, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.6928484439849854, | |
| "rewards/margins": 69.64448547363281, | |
| "rewards/rejected": -72.33734130859375, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5113636363636364, | |
| "grad_norm": 0.00510201676789354, | |
| "learning_rate": 2.846838829972671e-07, | |
| "logits/chosen": -2.2552402019500732, | |
| "logits/rejected": 1.9691696166992188, | |
| "logps/chosen": -620.7752685546875, | |
| "logps/rejected": -7178.3388671875, | |
| "loss": 0.0021, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.5987088680267334, | |
| "rewards/margins": 63.054847717285156, | |
| "rewards/rejected": -65.65354919433594, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.5397727272727273, | |
| "grad_norm": 0.2102394829040594, | |
| "learning_rate": 2.5993912877423147e-07, | |
| "logits/chosen": -1.9847707748413086, | |
| "logits/rejected": 1.6557881832122803, | |
| "logps/chosen": -555.5148315429688, | |
| "logps/rejected": -7729.9189453125, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.0989906787872314, | |
| "rewards/margins": 68.28409576416016, | |
| "rewards/rejected": -70.38308715820312, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.5681818181818182, | |
| "grad_norm": 0.62932736058334, | |
| "learning_rate": 2.3509621870754504e-07, | |
| "logits/chosen": -1.2217730283737183, | |
| "logits/rejected": 2.607821226119995, | |
| "logps/chosen": -571.2921142578125, | |
| "logps/rejected": -6882.54833984375, | |
| "loss": 0.0018, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.0616230964660645, | |
| "rewards/margins": 60.45543670654297, | |
| "rewards/rejected": -62.517051696777344, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5681818181818182, | |
| "eval_logits/chosen": -1.0394700765609741, | |
| "eval_logits/rejected": 2.318869113922119, | |
| "eval_logps/chosen": -584.9549560546875, | |
| "eval_logps/rejected": -7401.896484375, | |
| "eval_loss": 0.0001572135806782171, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -2.179731845855713, | |
| "eval_rewards/margins": 65.43135070800781, | |
| "eval_rewards/rejected": -67.61107635498047, | |
| "eval_runtime": 192.1225, | |
| "eval_samples_per_second": 20.326, | |
| "eval_steps_per_second": 0.323, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5965909090909091, | |
| "grad_norm": 0.5427913238664086, | |
| "learning_rate": 2.1040049389819624e-07, | |
| "logits/chosen": -1.0268203020095825, | |
| "logits/rejected": 2.479326009750366, | |
| "logps/chosen": -601.0950927734375, | |
| "logps/rejected": -7280.90771484375, | |
| "loss": 0.0028, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.104518413543701, | |
| "rewards/margins": 64.38551330566406, | |
| "rewards/rejected": -66.49003601074219, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.625, | |
| "grad_norm": 2.0347197022348307, | |
| "learning_rate": 1.8609584188988133e-07, | |
| "logits/chosen": -0.5109957456588745, | |
| "logits/rejected": 2.219853401184082, | |
| "logps/chosen": -642.4463500976562, | |
| "logps/rejected": -6641.6826171875, | |
| "loss": 0.003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.7784571647644043, | |
| "rewards/margins": 57.14200973510742, | |
| "rewards/rejected": -59.92046356201172, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.6534090909090909, | |
| "grad_norm": 34.246565820645614, | |
| "learning_rate": 1.624222881090439e-07, | |
| "logits/chosen": -0.5791199207305908, | |
| "logits/rejected": 2.0261969566345215, | |
| "logps/chosen": -674.46044921875, | |
| "logps/rejected": -7846.1572265625, | |
| "loss": 0.0099, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -2.7979581356048584, | |
| "rewards/margins": 68.80779266357422, | |
| "rewards/rejected": -71.60575103759766, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.6818181818181818, | |
| "grad_norm": 0.31954869640074296, | |
| "learning_rate": 1.3961362544602212e-07, | |
| "logits/chosen": -0.41727203130722046, | |
| "logits/rejected": 2.4299378395080566, | |
| "logps/chosen": -630.17919921875, | |
| "logps/rejected": -6758.9619140625, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.541667938232422, | |
| "rewards/margins": 59.10721969604492, | |
| "rewards/rejected": -61.648887634277344, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7102272727272727, | |
| "grad_norm": 0.2930942520900322, | |
| "learning_rate": 1.1789510538684522e-07, | |
| "logits/chosen": -0.4175181984901428, | |
| "logits/rejected": 2.409147024154663, | |
| "logps/chosen": -591.2493896484375, | |
| "logps/rejected": -8521.580078125, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.392158031463623, | |
| "rewards/margins": 75.52781677246094, | |
| "rewards/rejected": -77.91997528076172, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.7102272727272727, | |
| "eval_logits/chosen": -0.3353612720966339, | |
| "eval_logits/rejected": 2.558471918106079, | |
| "eval_logps/chosen": -608.6731567382812, | |
| "eval_logps/rejected": -7398.6552734375, | |
| "eval_loss": 0.0001336697314400226, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -2.4169137477874756, | |
| "eval_rewards/margins": 65.16175079345703, | |
| "eval_rewards/rejected": -67.57865905761719, | |
| "eval_runtime": 191.4599, | |
| "eval_samples_per_second": 20.396, | |
| "eval_steps_per_second": 0.324, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.7386363636363636, | |
| "grad_norm": 0.18266288675107914, | |
| "learning_rate": 9.748121349736891e-08, | |
| "logits/chosen": -0.2991776764392853, | |
| "logits/rejected": 2.4152047634124756, | |
| "logps/chosen": -644.2991333007812, | |
| "logps/rejected": -7754.68359375, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.604710578918457, | |
| "rewards/margins": 68.16084289550781, | |
| "rewards/rejected": -70.76555633544922, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.7670454545454546, | |
| "grad_norm": 0.02843345452840696, | |
| "learning_rate": 7.857355122839673e-08, | |
| "logits/chosen": -0.5736527442932129, | |
| "logits/rejected": 2.2506680488586426, | |
| "logps/chosen": -621.7249145507812, | |
| "logps/rejected": -7290.41015625, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.408754348754883, | |
| "rewards/margins": 63.805747985839844, | |
| "rewards/rejected": -66.21451568603516, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.7954545454545454, | |
| "grad_norm": 0.06722410834807425, | |
| "learning_rate": 6.135884496044244e-08, | |
| "logits/chosen": -0.5607927441596985, | |
| "logits/rejected": 2.5057239532470703, | |
| "logps/chosen": -604.9693603515625, | |
| "logps/rejected": -7248.1767578125, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.257964611053467, | |
| "rewards/margins": 63.7749137878418, | |
| "rewards/rejected": -66.03287506103516, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.8238636363636364, | |
| "grad_norm": 0.007877917419381247, | |
| "learning_rate": 4.600710195020982e-08, | |
| "logits/chosen": -0.44481128454208374, | |
| "logits/rejected": 2.559147357940674, | |
| "logps/chosen": -628.8250732421875, | |
| "logps/rejected": -7324.6533203125, | |
| "loss": 0.0027, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.3968186378479004, | |
| "rewards/margins": 64.43711853027344, | |
| "rewards/rejected": -66.83393859863281, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.8522727272727273, | |
| "grad_norm": 0.5905211140332576, | |
| "learning_rate": 3.2669931390104374e-08, | |
| "logits/chosen": -0.4180712103843689, | |
| "logits/rejected": 2.4567205905914307, | |
| "logps/chosen": -577.669921875, | |
| "logps/rejected": -7993.05859375, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.214153289794922, | |
| "rewards/margins": 71.5081787109375, | |
| "rewards/rejected": -73.72233581542969, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.8522727272727273, | |
| "eval_logits/chosen": -0.2549493908882141, | |
| "eval_logits/rejected": 2.575054407119751, | |
| "eval_logps/chosen": -608.2271728515625, | |
| "eval_logps/rejected": -7665.2216796875, | |
| "eval_loss": 7.371166429948062e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -2.412454843521118, | |
| "eval_rewards/margins": 67.83187866210938, | |
| "eval_rewards/rejected": -70.24433898925781, | |
| "eval_runtime": 192.5021, | |
| "eval_samples_per_second": 20.285, | |
| "eval_steps_per_second": 0.322, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.8806818181818182, | |
| "grad_norm": 0.04237285499239183, | |
| "learning_rate": 2.147904716149135e-08, | |
| "logits/chosen": -0.1846945583820343, | |
| "logits/rejected": 2.5058579444885254, | |
| "logps/chosen": -589.5569458007812, | |
| "logps/rejected": -7777.6474609375, | |
| "loss": 0.0018, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.2749037742614746, | |
| "rewards/margins": 68.49655151367188, | |
| "rewards/rejected": -70.77145385742188, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "grad_norm": 0.016783305893164634, | |
| "learning_rate": 1.254496706805433e-08, | |
| "logits/chosen": -0.39586615562438965, | |
| "logits/rejected": 2.4627411365509033, | |
| "logps/chosen": -622.9832763671875, | |
| "logps/rejected": -7786.82421875, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.403102159500122, | |
| "rewards/margins": 68.9453125, | |
| "rewards/rejected": -71.3484115600586, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.9375, | |
| "grad_norm": 0.149850718402827, | |
| "learning_rate": 5.955921395237318e-09, | |
| "logits/chosen": -0.37074390053749084, | |
| "logits/rejected": 2.418095588684082, | |
| "logps/chosen": -577.0670776367188, | |
| "logps/rejected": -7757.859375, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.2179901599884033, | |
| "rewards/margins": 68.88307189941406, | |
| "rewards/rejected": -71.10105895996094, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.9659090909090909, | |
| "grad_norm": 0.24861872079475503, | |
| "learning_rate": 1.7769815745066474e-09, | |
| "logits/chosen": -0.6141588687896729, | |
| "logits/rejected": 2.2593562602996826, | |
| "logps/chosen": -593.7769775390625, | |
| "logps/rejected": -7168.8701171875, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.2479617595672607, | |
| "rewards/margins": 63.35308074951172, | |
| "rewards/rejected": -65.60104370117188, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.9943181818181818, | |
| "grad_norm": 0.890402381245667, | |
| "learning_rate": 4.9417557483610875e-11, | |
| "logits/chosen": -0.3333211839199066, | |
| "logits/rejected": 2.404022455215454, | |
| "logps/chosen": -621.5592041015625, | |
| "logps/rejected": -7657.5458984375, | |
| "loss": 0.0024, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -2.472914457321167, | |
| "rewards/margins": 67.62696075439453, | |
| "rewards/rejected": -70.0998764038086, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.9943181818181818, | |
| "eval_logits/chosen": -0.21884487569332123, | |
| "eval_logits/rejected": 2.5753109455108643, | |
| "eval_logps/chosen": -609.166015625, | |
| "eval_logps/rejected": -7672.18896484375, | |
| "eval_loss": 7.062770600896329e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -2.4218428134918213, | |
| "eval_rewards/margins": 67.89216613769531, | |
| "eval_rewards/rejected": -70.31401062011719, | |
| "eval_runtime": 191.9742, | |
| "eval_samples_per_second": 20.341, | |
| "eval_steps_per_second": 0.323, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 352, | |
| "total_flos": 0.0, | |
| "train_loss": 0.042729352997742935, | |
| "train_runtime": 10052.8601, | |
| "train_samples_per_second": 4.476, | |
| "train_steps_per_second": 0.035 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 352, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |