| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 100, | |
| "global_step": 478, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": -2.847970962524414, | |
| "logits/rejected": -2.79160213470459, | |
| "logps/chosen": -284.9612731933594, | |
| "logps/rejected": -276.45928955078125, | |
| "loss": 0.0866, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -2.754711151123047, | |
| "logits/rejected": -2.7527668476104736, | |
| "logps/chosen": -249.9619140625, | |
| "logps/rejected": -223.05694580078125, | |
| "loss": 0.0866, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": -0.00014172698138281703, | |
| "rewards/margins": -5.154276004759595e-05, | |
| "rewards/rejected": -9.018417040351778e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -2.7448391914367676, | |
| "logits/rejected": -2.7453367710113525, | |
| "logps/chosen": -257.4018249511719, | |
| "logps/rejected": -247.55197143554688, | |
| "loss": 0.0866, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.0005366196855902672, | |
| "rewards/margins": 0.0017497055232524872, | |
| "rewards/rejected": -0.0012130856048315763, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.8008646965026855, | |
| "logits/rejected": -2.753528356552124, | |
| "logps/chosen": -300.37725830078125, | |
| "logps/rejected": -261.90411376953125, | |
| "loss": 0.086, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.0019977777265012264, | |
| "rewards/margins": 0.010120971128344536, | |
| "rewards/rejected": -0.008123192936182022, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.7638278007507324, | |
| "logits/rejected": -2.7516701221466064, | |
| "logps/chosen": -256.6244201660156, | |
| "logps/rejected": -274.86651611328125, | |
| "loss": 0.0851, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.00190549751278013, | |
| "rewards/margins": 0.025926370173692703, | |
| "rewards/rejected": -0.027831867337226868, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.999733114418725e-07, | |
| "logits/chosen": -2.767770767211914, | |
| "logits/rejected": -2.7401928901672363, | |
| "logps/chosen": -284.50543212890625, | |
| "logps/rejected": -256.618408203125, | |
| "loss": 0.0834, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.02425984852015972, | |
| "rewards/margins": 0.06489187479019165, | |
| "rewards/rejected": -0.08915172517299652, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.990398100856366e-07, | |
| "logits/chosen": -2.7361197471618652, | |
| "logits/rejected": -2.724640130996704, | |
| "logps/chosen": -282.0168151855469, | |
| "logps/rejected": -256.8324279785156, | |
| "loss": 0.0805, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.06132340431213379, | |
| "rewards/margins": 0.10652206093072891, | |
| "rewards/rejected": -0.1678454577922821, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.967775735898179e-07, | |
| "logits/chosen": -2.7822699546813965, | |
| "logits/rejected": -2.739877223968506, | |
| "logps/chosen": -291.2421875, | |
| "logps/rejected": -274.023193359375, | |
| "loss": 0.0781, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.24106796085834503, | |
| "rewards/margins": 0.17975760996341705, | |
| "rewards/rejected": -0.4208255410194397, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.931986719649298e-07, | |
| "logits/chosen": -2.7829222679138184, | |
| "logits/rejected": -2.7552578449249268, | |
| "logps/chosen": -290.30157470703125, | |
| "logps/rejected": -332.9333801269531, | |
| "loss": 0.0744, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.29948586225509644, | |
| "rewards/margins": 0.3068729043006897, | |
| "rewards/rejected": -0.6063587665557861, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.883222001996351e-07, | |
| "logits/chosen": -2.8093197345733643, | |
| "logits/rejected": -2.7847156524658203, | |
| "logps/chosen": -308.68927001953125, | |
| "logps/rejected": -327.748046875, | |
| "loss": 0.0733, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.3332245349884033, | |
| "rewards/margins": 0.4322693943977356, | |
| "rewards/rejected": -0.7654938697814941, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.821741763807186e-07, | |
| "logits/chosen": -2.776167392730713, | |
| "logits/rejected": -2.743318557739258, | |
| "logps/chosen": -356.5120544433594, | |
| "logps/rejected": -373.7995910644531, | |
| "loss": 0.0717, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.6754709482192993, | |
| "rewards/margins": 0.440376341342926, | |
| "rewards/rejected": -1.1158472299575806, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": -2.728982448577881, | |
| "eval_logits/rejected": -2.712980031967163, | |
| "eval_logps/chosen": -316.240478515625, | |
| "eval_logps/rejected": -367.9753112792969, | |
| "eval_loss": 0.073273666203022, | |
| "eval_rewards/accuracies": 0.7265625, | |
| "eval_rewards/chosen": -0.5920084714889526, | |
| "eval_rewards/margins": 0.5142118334770203, | |
| "eval_rewards/rejected": -1.1062203645706177, | |
| "eval_runtime": 53.5143, | |
| "eval_samples_per_second": 37.373, | |
| "eval_steps_per_second": 0.598, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.747874028753375e-07, | |
| "logits/chosen": -2.6736016273498535, | |
| "logits/rejected": -2.6673741340637207, | |
| "logps/chosen": -319.99285888671875, | |
| "logps/rejected": -394.5204772949219, | |
| "loss": 0.0675, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.557295560836792, | |
| "rewards/margins": 0.660796582698822, | |
| "rewards/rejected": -1.2180922031402588, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.662012913161997e-07, | |
| "logits/chosen": -2.6141879558563232, | |
| "logits/rejected": -2.5749917030334473, | |
| "logps/chosen": -359.4350280761719, | |
| "logps/rejected": -392.105712890625, | |
| "loss": 0.0709, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.9189525842666626, | |
| "rewards/margins": 0.5283113718032837, | |
| "rewards/rejected": -1.4472639560699463, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.5646165232345103e-07, | |
| "logits/chosen": -2.5792384147644043, | |
| "logits/rejected": -2.5439562797546387, | |
| "logps/chosen": -372.8755798339844, | |
| "logps/rejected": -363.8757629394531, | |
| "loss": 0.0701, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.8074409365653992, | |
| "rewards/margins": 0.42888420820236206, | |
| "rewards/rejected": -1.2363251447677612, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.456204510851956e-07, | |
| "logits/chosen": -2.4518723487854004, | |
| "logits/rejected": -2.429705858230591, | |
| "logps/chosen": -335.4825744628906, | |
| "logps/rejected": -358.73992919921875, | |
| "loss": 0.0687, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.8712352514266968, | |
| "rewards/margins": 0.5185465812683105, | |
| "rewards/rejected": -1.3897819519042969, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.337355301007335e-07, | |
| "logits/chosen": -2.3888745307922363, | |
| "logits/rejected": -2.3675789833068848, | |
| "logps/chosen": -345.4061584472656, | |
| "logps/rejected": -378.5643005371094, | |
| "loss": 0.0674, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.8017293810844421, | |
| "rewards/margins": 0.5842651128768921, | |
| "rewards/rejected": -1.385994553565979, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.2087030056579986e-07, | |
| "logits/chosen": -2.339261531829834, | |
| "logits/rejected": -2.3114781379699707, | |
| "logps/chosen": -338.3687744140625, | |
| "logps/rejected": -375.3154296875, | |
| "loss": 0.0684, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.7738892436027527, | |
| "rewards/margins": 0.6639295816421509, | |
| "rewards/rejected": -1.4378188848495483, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.070934040463998e-07, | |
| "logits/chosen": -2.2538902759552, | |
| "logits/rejected": -2.2150468826293945, | |
| "logps/chosen": -380.0686950683594, | |
| "logps/rejected": -427.46514892578125, | |
| "loss": 0.0679, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.1141223907470703, | |
| "rewards/margins": 0.6542497873306274, | |
| "rewards/rejected": -1.7683721780776978, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.9247834624635404e-07, | |
| "logits/chosen": -2.331003427505493, | |
| "logits/rejected": -2.2696776390075684, | |
| "logps/chosen": -374.4627380371094, | |
| "logps/rejected": -422.1844177246094, | |
| "loss": 0.066, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.980134129524231, | |
| "rewards/margins": 0.7622623443603516, | |
| "rewards/rejected": -1.742396593093872, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.7710310482256523e-07, | |
| "logits/chosen": -2.2278285026550293, | |
| "logits/rejected": -2.187769651412964, | |
| "logps/chosen": -367.53619384765625, | |
| "logps/rejected": -409.8011169433594, | |
| "loss": 0.0653, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.1032383441925049, | |
| "rewards/margins": 0.6265621781349182, | |
| "rewards/rejected": -1.7298005819320679, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.610497133404795e-07, | |
| "logits/chosen": -2.2965497970581055, | |
| "logits/rejected": -2.2077629566192627, | |
| "logps/chosen": -437.07275390625, | |
| "logps/rejected": -473.34228515625, | |
| "loss": 0.0672, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.024962067604065, | |
| "rewards/margins": 0.9094398617744446, | |
| "rewards/rejected": -1.9344017505645752, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": -2.23770809173584, | |
| "eval_logits/rejected": -2.204430103302002, | |
| "eval_logps/chosen": -350.147216796875, | |
| "eval_logps/rejected": -429.3445129394531, | |
| "eval_loss": 0.06620719283819199, | |
| "eval_rewards/accuracies": 0.7421875, | |
| "eval_rewards/chosen": -0.9310759902000427, | |
| "eval_rewards/margins": 0.7888363599777222, | |
| "eval_rewards/rejected": -1.7199124097824097, | |
| "eval_runtime": 53.5039, | |
| "eval_samples_per_second": 37.38, | |
| "eval_steps_per_second": 0.598, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.4440382358952115e-07, | |
| "logits/chosen": -2.2368922233581543, | |
| "logits/rejected": -2.164578437805176, | |
| "logps/chosen": -367.9010009765625, | |
| "logps/rejected": -403.06689453125, | |
| "loss": 0.0688, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.0069431066513062, | |
| "rewards/margins": 0.6194397211074829, | |
| "rewards/rejected": -1.6263830661773682, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.272542485937368e-07, | |
| "logits/chosen": -2.223553419113159, | |
| "logits/rejected": -2.1863625049591064, | |
| "logps/chosen": -383.9325256347656, | |
| "logps/rejected": -408.54156494140625, | |
| "loss": 0.066, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.893720805644989, | |
| "rewards/margins": 0.7813506126403809, | |
| "rewards/rejected": -1.6750714778900146, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.096924887558854e-07, | |
| "logits/chosen": -2.1857941150665283, | |
| "logits/rejected": -2.1605677604675293, | |
| "logps/chosen": -385.0683288574219, | |
| "logps/rejected": -465.76214599609375, | |
| "loss": 0.065, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.0325820446014404, | |
| "rewards/margins": 0.7163792848587036, | |
| "rewards/rejected": -1.7489612102508545, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.9181224366319943e-07, | |
| "logits/chosen": -2.1450023651123047, | |
| "logits/rejected": -2.1016509532928467, | |
| "logps/chosen": -375.78448486328125, | |
| "logps/rejected": -456.2076110839844, | |
| "loss": 0.0663, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.2208166122436523, | |
| "rewards/margins": 0.755986213684082, | |
| "rewards/rejected": -1.9768028259277344, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.7370891215954565e-07, | |
| "logits/chosen": -2.084066390991211, | |
| "logits/rejected": -2.000739574432373, | |
| "logps/chosen": -408.6034240722656, | |
| "logps/rejected": -460.944091796875, | |
| "loss": 0.0656, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.496781587600708, | |
| "rewards/margins": 0.7584124207496643, | |
| "rewards/rejected": -2.2551939487457275, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.55479083351317e-07, | |
| "logits/chosen": -2.1052582263946533, | |
| "logits/rejected": -2.0781798362731934, | |
| "logps/chosen": -385.717529296875, | |
| "logps/rejected": -427.49053955078125, | |
| "loss": 0.0675, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -1.171087622642517, | |
| "rewards/margins": 0.6191247701644897, | |
| "rewards/rejected": -1.7902123928070068, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.3722002126275822e-07, | |
| "logits/chosen": -2.060842514038086, | |
| "logits/rejected": -2.032809257507324, | |
| "logps/chosen": -411.20904541015625, | |
| "logps/rejected": -466.87945556640625, | |
| "loss": 0.0644, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.3635554313659668, | |
| "rewards/margins": 0.7496285438537598, | |
| "rewards/rejected": -2.1131839752197266, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.19029145890313e-07, | |
| "logits/chosen": -2.0691192150115967, | |
| "logits/rejected": -2.014458417892456, | |
| "logps/chosen": -368.3045349121094, | |
| "logps/rejected": -434.04840087890625, | |
| "loss": 0.0659, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.2689889669418335, | |
| "rewards/margins": 0.7694240808486938, | |
| "rewards/rejected": -2.0384132862091064, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 2.0100351342479216e-07, | |
| "logits/chosen": -2.0620741844177246, | |
| "logits/rejected": -2.0015225410461426, | |
| "logps/chosen": -380.07489013671875, | |
| "logps/rejected": -437.5985412597656, | |
| "loss": 0.0641, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.3346366882324219, | |
| "rewards/margins": 0.6989216208457947, | |
| "rewards/rejected": -2.0335583686828613, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.8323929841460178e-07, | |
| "logits/chosen": -2.0831615924835205, | |
| "logits/rejected": -2.0331509113311768, | |
| "logps/chosen": -385.72332763671875, | |
| "logps/rejected": -458.3067932128906, | |
| "loss": 0.0648, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.294547200202942, | |
| "rewards/margins": 0.7150087356567383, | |
| "rewards/rejected": -2.0095558166503906, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": -2.1098170280456543, | |
| "eval_logits/rejected": -2.072664976119995, | |
| "eval_logps/chosen": -382.6705322265625, | |
| "eval_logps/rejected": -471.1217041015625, | |
| "eval_loss": 0.06434247642755508, | |
| "eval_rewards/accuracies": 0.7734375, | |
| "eval_rewards/chosen": -1.256308674812317, | |
| "eval_rewards/margins": 0.8813759088516235, | |
| "eval_rewards/rejected": -2.1376843452453613, | |
| "eval_runtime": 53.4979, | |
| "eval_samples_per_second": 37.385, | |
| "eval_steps_per_second": 0.598, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.6583128063291573e-07, | |
| "logits/chosen": -2.038897752761841, | |
| "logits/rejected": -1.9522100687026978, | |
| "logps/chosen": -395.3385009765625, | |
| "logps/rejected": -463.807861328125, | |
| "loss": 0.0619, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -1.2215960025787354, | |
| "rewards/margins": 0.9447793960571289, | |
| "rewards/rejected": -2.1663756370544434, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.488723393865766e-07, | |
| "logits/chosen": -2.0476739406585693, | |
| "logits/rejected": -2.003208875656128, | |
| "logps/chosen": -440.6886291503906, | |
| "logps/rejected": -456.0061950683594, | |
| "loss": 0.0649, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.443062663078308, | |
| "rewards/margins": 0.7629345655441284, | |
| "rewards/rejected": -2.2059974670410156, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3245295796480788e-07, | |
| "logits/chosen": -2.0648202896118164, | |
| "logits/rejected": -2.0035393238067627, | |
| "logps/chosen": -399.42724609375, | |
| "logps/rejected": -452.73468017578125, | |
| "loss": 0.0643, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.472140908241272, | |
| "rewards/margins": 0.7188968062400818, | |
| "rewards/rejected": -2.191037654876709, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.1666074087171627e-07, | |
| "logits/chosen": -2.010235548019409, | |
| "logits/rejected": -1.9665842056274414, | |
| "logps/chosen": -401.20391845703125, | |
| "logps/rejected": -479.63140869140625, | |
| "loss": 0.0655, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.4996058940887451, | |
| "rewards/margins": 0.8774793744087219, | |
| "rewards/rejected": -2.3770852088928223, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0157994641835734e-07, | |
| "logits/chosen": -1.9906257390975952, | |
| "logits/rejected": -1.9489715099334717, | |
| "logps/chosen": -377.80291748046875, | |
| "logps/rejected": -473.23565673828125, | |
| "loss": 0.0621, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.5185272693634033, | |
| "rewards/margins": 0.806858241558075, | |
| "rewards/rejected": -2.325385570526123, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.729103716819111e-08, | |
| "logits/chosen": -2.0000805854797363, | |
| "logits/rejected": -1.9243812561035156, | |
| "logps/chosen": -406.80120849609375, | |
| "logps/rejected": -481.9264221191406, | |
| "loss": 0.0596, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.3953304290771484, | |
| "rewards/margins": 1.0305477380752563, | |
| "rewards/rejected": -2.4258780479431152, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.387025063449081e-08, | |
| "logits/chosen": -1.9636814594268799, | |
| "logits/rejected": -1.941389799118042, | |
| "logps/chosen": -379.83538818359375, | |
| "logps/rejected": -478.5880432128906, | |
| "loss": 0.0611, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.2458791732788086, | |
| "rewards/margins": 0.9973210096359253, | |
| "rewards/rejected": -2.2432003021240234, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 6.138919252022435e-08, | |
| "logits/chosen": -2.004906415939331, | |
| "logits/rejected": -1.9661741256713867, | |
| "logps/chosen": -411.92626953125, | |
| "logps/rejected": -458.4734802246094, | |
| "loss": 0.0625, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.46096932888031, | |
| "rewards/margins": 0.755499005317688, | |
| "rewards/rejected": -2.216468334197998, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.991445467064689e-08, | |
| "logits/chosen": -1.9821668863296509, | |
| "logits/rejected": -1.9519681930541992, | |
| "logps/chosen": -385.4888000488281, | |
| "logps/rejected": -463.5560607910156, | |
| "loss": 0.0612, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -1.282893419265747, | |
| "rewards/margins": 0.933008074760437, | |
| "rewards/rejected": -2.2159013748168945, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.9507259776993954e-08, | |
| "logits/chosen": -2.0385870933532715, | |
| "logits/rejected": -2.0015995502471924, | |
| "logps/chosen": -436.51800537109375, | |
| "logps/rejected": -499.8250427246094, | |
| "loss": 0.0636, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.3635799884796143, | |
| "rewards/margins": 0.9055550694465637, | |
| "rewards/rejected": -2.269134998321533, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": -2.047151565551758, | |
| "eval_logits/rejected": -2.0084266662597656, | |
| "eval_logps/chosen": -391.10418701171875, | |
| "eval_logps/rejected": -488.8221740722656, | |
| "eval_loss": 0.06317802518606186, | |
| "eval_rewards/accuracies": 0.7734375, | |
| "eval_rewards/chosen": -1.3406453132629395, | |
| "eval_rewards/margins": 0.9740438461303711, | |
| "eval_rewards/rejected": -2.3146889209747314, | |
| "eval_runtime": 53.4405, | |
| "eval_samples_per_second": 37.425, | |
| "eval_steps_per_second": 0.599, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.022313472693447e-08, | |
| "logits/chosen": -2.0089824199676514, | |
| "logits/rejected": -1.990161657333374, | |
| "logps/chosen": -397.4513854980469, | |
| "logps/rejected": -437.2802734375, | |
| "loss": 0.064, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.4276202917099, | |
| "rewards/margins": 0.6205279231071472, | |
| "rewards/rejected": -2.0481481552124023, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.2111614344599684e-08, | |
| "logits/chosen": -2.084846019744873, | |
| "logits/rejected": -2.0091195106506348, | |
| "logps/chosen": -399.70892333984375, | |
| "logps/rejected": -477.39215087890625, | |
| "loss": 0.061, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.2201029062271118, | |
| "rewards/margins": 1.0571900606155396, | |
| "rewards/rejected": -2.2772929668426514, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.521597710086439e-08, | |
| "logits/chosen": -2.076000928878784, | |
| "logits/rejected": -2.0238547325134277, | |
| "logps/chosen": -406.68243408203125, | |
| "logps/rejected": -459.352783203125, | |
| "loss": 0.0611, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.3172928094863892, | |
| "rewards/margins": 0.819710910320282, | |
| "rewards/rejected": -2.1370038986206055, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 9.57301420397924e-09, | |
| "logits/chosen": -1.9988025426864624, | |
| "logits/rejected": -1.979654312133789, | |
| "logps/chosen": -415.54779052734375, | |
| "logps/rejected": -451.9610900878906, | |
| "loss": 0.0628, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.372821569442749, | |
| "rewards/margins": 0.7440255880355835, | |
| "rewards/rejected": -2.116847276687622, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 5.212833302556258e-09, | |
| "logits/chosen": -2.028916120529175, | |
| "logits/rejected": -1.9799381494522095, | |
| "logps/chosen": -409.3075866699219, | |
| "logps/rejected": -491.86444091796875, | |
| "loss": 0.0605, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.326899766921997, | |
| "rewards/margins": 0.9430659413337708, | |
| "rewards/rejected": -2.269965648651123, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.158697848236607e-09, | |
| "logits/chosen": -1.9905271530151367, | |
| "logits/rejected": -1.9591131210327148, | |
| "logps/chosen": -391.362060546875, | |
| "logps/rejected": -461.91204833984375, | |
| "loss": 0.0604, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.2606008052825928, | |
| "rewards/margins": 0.8890584111213684, | |
| "rewards/rejected": -2.1496591567993164, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 4.269029751107489e-10, | |
| "logits/chosen": -2.0331690311431885, | |
| "logits/rejected": -1.9691358804702759, | |
| "logps/chosen": -417.90203857421875, | |
| "logps/rejected": -465.673828125, | |
| "loss": 0.0625, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.3203939199447632, | |
| "rewards/margins": 0.9791151285171509, | |
| "rewards/rejected": -2.299509048461914, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 478, | |
| "total_flos": 0.0, | |
| "train_loss": 0.06801126423490596, | |
| "train_runtime": 3957.1126, | |
| "train_samples_per_second": 15.449, | |
| "train_steps_per_second": 0.121 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 478, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |