| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 939, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.003194888178913738, | |
| "grad_norm": 1.7594991820231036, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -0.84765625, | |
| "logits/rejected": -0.89453125, | |
| "logps/chosen": -15.1875, | |
| "logps/rejected": -16.25, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03194888178913738, | |
| "grad_norm": 3.2874000198076727, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -0.890625, | |
| "logits/rejected": -0.875, | |
| "logps/chosen": -16.125, | |
| "logps/rejected": -16.5, | |
| "loss": 0.6917, | |
| "rewards/accuracies": 0.0972222238779068, | |
| "rewards/chosen": 0.0001735687255859375, | |
| "rewards/margins": 0.00130462646484375, | |
| "rewards/rejected": -0.001129150390625, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 6.20375008201732, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -0.88671875, | |
| "logits/rejected": -0.8984375, | |
| "logps/chosen": -16.25, | |
| "logps/rejected": -16.625, | |
| "loss": 0.692, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": -7.62939453125e-05, | |
| "rewards/margins": -0.00023746490478515625, | |
| "rewards/rejected": 0.0001583099365234375, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 4.664908763761001, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -0.859375, | |
| "logits/rejected": -0.875, | |
| "logps/chosen": -16.125, | |
| "logps/rejected": -16.25, | |
| "loss": 0.6913, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": 0.0020751953125, | |
| "rewards/margins": 0.00019550323486328125, | |
| "rewards/rejected": 0.0018768310546875, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 2.4116741194375373, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -0.87890625, | |
| "logits/rejected": -0.8671875, | |
| "logps/chosen": -16.375, | |
| "logps/rejected": -16.5, | |
| "loss": 0.6908, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": 0.00811767578125, | |
| "rewards/margins": 0.0003108978271484375, | |
| "rewards/rejected": 0.0078125, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 3.308574154629096, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -0.87109375, | |
| "logits/rejected": -0.84765625, | |
| "logps/chosen": -14.9375, | |
| "logps/rejected": -15.6875, | |
| "loss": 0.691, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.0166015625, | |
| "rewards/margins": 0.0029754638671875, | |
| "rewards/rejected": 0.01361083984375, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 1.6245800907371344, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -0.859375, | |
| "logits/rejected": -0.85546875, | |
| "logps/chosen": -16.125, | |
| "logps/rejected": -15.0, | |
| "loss": 0.6908, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": 0.0198974609375, | |
| "rewards/margins": -0.0006256103515625, | |
| "rewards/rejected": 0.0205078125, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 2.6711590089749127, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -0.8828125, | |
| "logits/rejected": -0.8828125, | |
| "logps/chosen": -16.5, | |
| "logps/rejected": -16.125, | |
| "loss": 0.6904, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.0224609375, | |
| "rewards/margins": 0.00579833984375, | |
| "rewards/rejected": 0.0167236328125, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 2.3762135272130664, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -0.8515625, | |
| "logits/rejected": -0.84375, | |
| "logps/chosen": -14.9375, | |
| "logps/rejected": -15.25, | |
| "loss": 0.6856, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": 0.057373046875, | |
| "rewards/margins": 0.007568359375, | |
| "rewards/rejected": 0.0498046875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 3.4884770405641, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -0.86328125, | |
| "logits/rejected": -0.8671875, | |
| "logps/chosen": -14.125, | |
| "logps/rejected": -14.0, | |
| "loss": 0.6812, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.142578125, | |
| "rewards/margins": 0.02001953125, | |
| "rewards/rejected": 0.12255859375, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 4.958656955255395, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -0.859375, | |
| "logits/rejected": -0.89453125, | |
| "logps/chosen": -13.5625, | |
| "logps/rejected": -14.5, | |
| "loss": 0.6847, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.1826171875, | |
| "rewards/margins": 0.0277099609375, | |
| "rewards/rejected": 0.1552734375, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 4.274608448473679, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -0.8515625, | |
| "logits/rejected": -0.8359375, | |
| "logps/chosen": -13.3125, | |
| "logps/rejected": -13.4375, | |
| "loss": 0.6855, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": 0.2392578125, | |
| "rewards/margins": 0.0169677734375, | |
| "rewards/rejected": 0.22265625, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 6.050646800243223, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -0.85546875, | |
| "logits/rejected": -0.84375, | |
| "logps/chosen": -12.3125, | |
| "logps/rejected": -12.6875, | |
| "loss": 0.6831, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.3046875, | |
| "rewards/margins": 0.0400390625, | |
| "rewards/rejected": 0.263671875, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 0.584658593263476, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -0.828125, | |
| "logits/rejected": -0.875, | |
| "logps/chosen": -15.125, | |
| "logps/rejected": -17.0, | |
| "loss": 0.6783, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": 0.19140625, | |
| "rewards/margins": 0.0830078125, | |
| "rewards/rejected": 0.10888671875, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 2.1563062155083426, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -0.8671875, | |
| "logits/rejected": -0.8515625, | |
| "logps/chosen": -14.125, | |
| "logps/rejected": -15.1875, | |
| "loss": 0.6758, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": 0.2275390625, | |
| "rewards/margins": 0.06640625, | |
| "rewards/rejected": 0.1611328125, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 2.43580270301018, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -0.83984375, | |
| "logits/rejected": -0.84375, | |
| "logps/chosen": -13.6875, | |
| "logps/rejected": -14.5, | |
| "loss": 0.6744, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.232421875, | |
| "rewards/margins": 0.041259765625, | |
| "rewards/rejected": 0.19140625, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 4.371745508926852, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -0.828125, | |
| "logits/rejected": -0.8046875, | |
| "logps/chosen": -14.75, | |
| "logps/rejected": -14.5625, | |
| "loss": 0.6823, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.18359375, | |
| "rewards/margins": 0.0028533935546875, | |
| "rewards/rejected": 0.1806640625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 5.276282548018104, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -0.796875, | |
| "logits/rejected": -0.79296875, | |
| "logps/chosen": -12.875, | |
| "logps/rejected": -13.1875, | |
| "loss": 0.6852, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.234375, | |
| "rewards/margins": 0.03857421875, | |
| "rewards/rejected": 0.1953125, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 3.1471054069237403, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -0.82421875, | |
| "logits/rejected": -0.80078125, | |
| "logps/chosen": -15.125, | |
| "logps/rejected": -15.5, | |
| "loss": 0.668, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.1435546875, | |
| "rewards/margins": 0.0712890625, | |
| "rewards/rejected": 0.072265625, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 5.526336807387032, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -0.77734375, | |
| "logits/rejected": -0.7890625, | |
| "logps/chosen": -14.375, | |
| "logps/rejected": -15.625, | |
| "loss": 0.6698, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.13671875, | |
| "rewards/margins": 0.060546875, | |
| "rewards/rejected": 0.076171875, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 6.08437673538068, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -0.80078125, | |
| "logits/rejected": -0.8046875, | |
| "logps/chosen": -13.625, | |
| "logps/rejected": -14.375, | |
| "loss": 0.6795, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.20703125, | |
| "rewards/margins": 0.053466796875, | |
| "rewards/rejected": 0.1533203125, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 1.6033107403006597, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -0.80078125, | |
| "logits/rejected": -0.7890625, | |
| "logps/chosen": -12.875, | |
| "logps/rejected": -14.0625, | |
| "loss": 0.6702, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.291015625, | |
| "rewards/margins": 0.09326171875, | |
| "rewards/rejected": 0.197265625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 1.7290270560712724, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -0.79296875, | |
| "logits/rejected": -0.78515625, | |
| "logps/chosen": -13.25, | |
| "logps/rejected": -13.5625, | |
| "loss": 0.6655, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.2236328125, | |
| "rewards/margins": 0.04052734375, | |
| "rewards/rejected": 0.1826171875, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 3.4471878821701267, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -0.7890625, | |
| "logits/rejected": -0.796875, | |
| "logps/chosen": -15.625, | |
| "logps/rejected": -16.5, | |
| "loss": 0.6706, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.138671875, | |
| "rewards/margins": 0.11083984375, | |
| "rewards/rejected": 0.02783203125, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 4.299918979944633, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -0.76171875, | |
| "logits/rejected": -0.765625, | |
| "logps/chosen": -14.875, | |
| "logps/rejected": -15.8125, | |
| "loss": 0.6774, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.0732421875, | |
| "rewards/margins": 0.08251953125, | |
| "rewards/rejected": -0.0093994140625, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 4.20769528530577, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -0.8046875, | |
| "logits/rejected": -0.7890625, | |
| "logps/chosen": -14.5625, | |
| "logps/rejected": -14.375, | |
| "loss": 0.6723, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.1806640625, | |
| "rewards/margins": 0.022705078125, | |
| "rewards/rejected": 0.158203125, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 4.95860147750649, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -0.7578125, | |
| "logits/rejected": -0.7421875, | |
| "logps/chosen": -15.9375, | |
| "logps/rejected": -17.0, | |
| "loss": 0.6511, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.111328125, | |
| "rewards/margins": 0.1474609375, | |
| "rewards/rejected": -0.0361328125, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 1.5354295895251373, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -0.76171875, | |
| "logits/rejected": -0.7890625, | |
| "logps/chosen": -12.8125, | |
| "logps/rejected": -15.0625, | |
| "loss": 0.6686, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.21875, | |
| "rewards/margins": 0.10888671875, | |
| "rewards/rejected": 0.1103515625, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 1.9206310698954268, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -0.75390625, | |
| "logits/rejected": -0.76953125, | |
| "logps/chosen": -12.75, | |
| "logps/rejected": -14.5, | |
| "loss": 0.6707, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.2314453125, | |
| "rewards/margins": 0.1171875, | |
| "rewards/rejected": 0.1142578125, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 2.1151877094244007, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -0.78515625, | |
| "logits/rejected": -0.8046875, | |
| "logps/chosen": -13.5, | |
| "logps/rejected": -15.4375, | |
| "loss": 0.6668, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.166015625, | |
| "rewards/margins": 0.0947265625, | |
| "rewards/rejected": 0.0712890625, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 1.721440878382303, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -0.796875, | |
| "logits/rejected": -0.8125, | |
| "logps/chosen": -14.5, | |
| "logps/rejected": -16.125, | |
| "loss": 0.6753, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.1171875, | |
| "rewards/margins": 0.1220703125, | |
| "rewards/rejected": -0.0047607421875, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 0.9947102905281956, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -0.76171875, | |
| "logits/rejected": -0.78125, | |
| "logps/chosen": -16.5, | |
| "logps/rejected": -16.625, | |
| "loss": 0.6691, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.1416015625, | |
| "rewards/margins": 0.10986328125, | |
| "rewards/rejected": 0.0322265625, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -0.765625, | |
| "eval_logits/rejected": -0.7734375, | |
| "eval_logps/chosen": -14.5625, | |
| "eval_logps/rejected": -15.6875, | |
| "eval_loss": 0.6690624952316284, | |
| "eval_rewards/accuracies": 0.0892857164144516, | |
| "eval_rewards/chosen": 0.0439453125, | |
| "eval_rewards/margins": 0.06787109375, | |
| "eval_rewards/rejected": -0.0240478515625, | |
| "eval_runtime": 12.3648, | |
| "eval_samples_per_second": 16.175, | |
| "eval_steps_per_second": 0.566, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 1.8828691509066227, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -0.75390625, | |
| "logits/rejected": -0.76171875, | |
| "logps/chosen": -16.125, | |
| "logps/rejected": -17.5, | |
| "loss": 0.6494, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.048828125, | |
| "rewards/margins": 0.1767578125, | |
| "rewards/rejected": -0.1279296875, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 2.5848517949930514, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -0.75, | |
| "logits/rejected": -0.7421875, | |
| "logps/chosen": -15.9375, | |
| "logps/rejected": -18.25, | |
| "loss": 0.6366, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.10107421875, | |
| "rewards/margins": 0.2421875, | |
| "rewards/rejected": -0.1416015625, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 3.781049848834596, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -0.77734375, | |
| "logits/rejected": -0.7578125, | |
| "logps/chosen": -14.3125, | |
| "logps/rejected": -16.0, | |
| "loss": 0.6541, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.11083984375, | |
| "rewards/margins": 0.1484375, | |
| "rewards/rejected": -0.038330078125, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 0.6719434060927373, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -0.8125, | |
| "logits/rejected": -0.77734375, | |
| "logps/chosen": -16.5, | |
| "logps/rejected": -16.625, | |
| "loss": 0.6624, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.10205078125, | |
| "rewards/margins": 0.2041015625, | |
| "rewards/rejected": -0.1025390625, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 0.5710729927303285, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -0.75390625, | |
| "logits/rejected": -0.76171875, | |
| "logps/chosen": -17.0, | |
| "logps/rejected": -18.875, | |
| "loss": 0.6608, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.033203125, | |
| "rewards/margins": 0.16015625, | |
| "rewards/rejected": -0.193359375, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 0.9018327997611894, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -0.75390625, | |
| "logits/rejected": -0.75390625, | |
| "logps/chosen": -16.375, | |
| "logps/rejected": -18.5, | |
| "loss": 0.6564, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -0.09130859375, | |
| "rewards/margins": 0.15234375, | |
| "rewards/rejected": -0.2431640625, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 0.20118163629327215, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -0.70703125, | |
| "logits/rejected": -0.71875, | |
| "logps/chosen": -19.25, | |
| "logps/rejected": -22.25, | |
| "loss": 0.6286, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -0.271484375, | |
| "rewards/margins": 0.294921875, | |
| "rewards/rejected": -0.56640625, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 2.184217252117973, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -0.7265625, | |
| "logits/rejected": -0.71484375, | |
| "logps/chosen": -17.625, | |
| "logps/rejected": -19.75, | |
| "loss": 0.6601, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.2060546875, | |
| "rewards/margins": 0.1962890625, | |
| "rewards/rejected": -0.40234375, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 5.41441250386917, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -0.67578125, | |
| "logits/rejected": -0.69140625, | |
| "logps/chosen": -17.75, | |
| "logps/rejected": -19.25, | |
| "loss": 0.6346, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -0.177734375, | |
| "rewards/margins": 0.1494140625, | |
| "rewards/rejected": -0.328125, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 1.1705115214695994, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -0.6796875, | |
| "logits/rejected": -0.67578125, | |
| "logps/chosen": -16.625, | |
| "logps/rejected": -16.625, | |
| "loss": 0.664, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.1357421875, | |
| "rewards/margins": 0.023193359375, | |
| "rewards/rejected": -0.158203125, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 1.5227177159924228, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -0.66015625, | |
| "logits/rejected": -0.66796875, | |
| "logps/chosen": -20.75, | |
| "logps/rejected": -23.125, | |
| "loss": 0.6474, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -0.53125, | |
| "rewards/margins": 0.150390625, | |
| "rewards/rejected": -0.6796875, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 1.703321680381799, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -0.61328125, | |
| "logits/rejected": -0.62109375, | |
| "logps/chosen": -20.0, | |
| "logps/rejected": -23.125, | |
| "loss": 0.6258, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -0.5, | |
| "rewards/margins": 0.2197265625, | |
| "rewards/rejected": -0.71875, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 1.1510493475143673, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -0.640625, | |
| "logits/rejected": -0.65625, | |
| "logps/chosen": -19.5, | |
| "logps/rejected": -22.125, | |
| "loss": 0.6509, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -0.431640625, | |
| "rewards/margins": 0.220703125, | |
| "rewards/rejected": -0.65234375, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 1.3183934075792523, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -0.64453125, | |
| "logits/rejected": -0.6484375, | |
| "logps/chosen": -20.375, | |
| "logps/rejected": -23.625, | |
| "loss": 0.6478, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -0.462890625, | |
| "rewards/margins": 0.1689453125, | |
| "rewards/rejected": -0.6328125, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 5.704668608100729, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -0.6796875, | |
| "logits/rejected": -0.6953125, | |
| "logps/chosen": -20.75, | |
| "logps/rejected": -24.125, | |
| "loss": 0.6427, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -0.43359375, | |
| "rewards/margins": 0.302734375, | |
| "rewards/rejected": -0.73828125, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 2.935017286186551, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -0.671875, | |
| "logits/rejected": -0.6953125, | |
| "logps/chosen": -19.5, | |
| "logps/rejected": -23.375, | |
| "loss": 0.647, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": -0.28125, | |
| "rewards/margins": 0.380859375, | |
| "rewards/rejected": -0.6640625, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 0.486060756917544, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -0.62890625, | |
| "logits/rejected": -0.63671875, | |
| "logps/chosen": -17.5, | |
| "logps/rejected": -18.75, | |
| "loss": 0.6485, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -0.24609375, | |
| "rewards/margins": 0.111328125, | |
| "rewards/rejected": -0.357421875, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 0.010927695001285508, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -0.64453125, | |
| "logits/rejected": -0.66015625, | |
| "logps/chosen": -20.25, | |
| "logps/rejected": -21.625, | |
| "loss": 0.6604, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.451171875, | |
| "rewards/margins": 0.14453125, | |
| "rewards/rejected": -0.59375, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 1.8999960547821024, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -0.62109375, | |
| "logits/rejected": -0.6328125, | |
| "logps/chosen": -18.625, | |
| "logps/rejected": -21.375, | |
| "loss": 0.6581, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -0.326171875, | |
| "rewards/margins": 0.2265625, | |
| "rewards/rejected": -0.55078125, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 2.135118341421029, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -0.62109375, | |
| "logits/rejected": -0.625, | |
| "logps/chosen": -19.625, | |
| "logps/rejected": -24.0, | |
| "loss": 0.6277, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -0.345703125, | |
| "rewards/margins": 0.359375, | |
| "rewards/rejected": -0.70703125, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 1.1880487203363566, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -0.625, | |
| "logits/rejected": -0.60546875, | |
| "logps/chosen": -18.25, | |
| "logps/rejected": -19.125, | |
| "loss": 0.6294, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.2353515625, | |
| "rewards/margins": 0.171875, | |
| "rewards/rejected": -0.40625, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 0.16870359343576236, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -0.63671875, | |
| "logits/rejected": -0.625, | |
| "logps/chosen": -19.5, | |
| "logps/rejected": -23.375, | |
| "loss": 0.6364, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -0.361328125, | |
| "rewards/margins": 0.294921875, | |
| "rewards/rejected": -0.65625, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 1.3708538666044734, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -0.625, | |
| "logits/rejected": -0.5859375, | |
| "logps/chosen": -19.875, | |
| "logps/rejected": -20.375, | |
| "loss": 0.6439, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -0.345703125, | |
| "rewards/margins": 0.1728515625, | |
| "rewards/rejected": -0.51953125, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 2.8331688198449148, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -0.640625, | |
| "logits/rejected": -0.640625, | |
| "logps/chosen": -20.0, | |
| "logps/rejected": -23.75, | |
| "loss": 0.6363, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -0.423828125, | |
| "rewards/margins": 0.298828125, | |
| "rewards/rejected": -0.72265625, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 0.8122610251667061, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -0.6171875, | |
| "logits/rejected": -0.625, | |
| "logps/chosen": -19.0, | |
| "logps/rejected": -20.625, | |
| "loss": 0.6424, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": -0.404296875, | |
| "rewards/margins": 0.11962890625, | |
| "rewards/rejected": -0.5234375, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 3.8931571635129187, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -0.61328125, | |
| "logits/rejected": -0.62109375, | |
| "logps/chosen": -20.125, | |
| "logps/rejected": -23.625, | |
| "loss": 0.6467, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -0.44921875, | |
| "rewards/margins": 0.306640625, | |
| "rewards/rejected": -0.75390625, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 0.8880645735078768, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -0.609375, | |
| "logits/rejected": -0.58203125, | |
| "logps/chosen": -21.375, | |
| "logps/rejected": -23.75, | |
| "loss": 0.6363, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -0.546875, | |
| "rewards/margins": 0.259765625, | |
| "rewards/rejected": -0.80859375, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 3.981951196386728, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -0.61328125, | |
| "logits/rejected": -0.6171875, | |
| "logps/chosen": -21.375, | |
| "logps/rejected": -23.75, | |
| "loss": 0.6466, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -0.53125, | |
| "rewards/margins": 0.26171875, | |
| "rewards/rejected": -0.79296875, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 2.0863403594125844, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -0.57421875, | |
| "logits/rejected": -0.50390625, | |
| "logps/chosen": -24.0, | |
| "logps/rejected": -25.625, | |
| "loss": 0.6328, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.83984375, | |
| "rewards/margins": 0.208984375, | |
| "rewards/rejected": -1.046875, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 9.496432924682129, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -0.61328125, | |
| "logits/rejected": -0.58984375, | |
| "logps/chosen": -24.25, | |
| "logps/rejected": -28.0, | |
| "loss": 0.6435, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -0.6796875, | |
| "rewards/margins": 0.384765625, | |
| "rewards/rejected": -1.0625, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 4.992354200554088, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -0.59375, | |
| "logits/rejected": -0.61328125, | |
| "logps/chosen": -19.375, | |
| "logps/rejected": -25.75, | |
| "loss": 0.6344, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -0.458984375, | |
| "rewards/margins": 0.51171875, | |
| "rewards/rejected": -0.96875, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -0.59375, | |
| "eval_logits/rejected": -0.6015625, | |
| "eval_logps/chosen": -19.875, | |
| "eval_logps/rejected": -21.875, | |
| "eval_loss": 0.6564062237739563, | |
| "eval_rewards/accuracies": 0.0892857164144516, | |
| "eval_rewards/chosen": -0.48046875, | |
| "eval_rewards/margins": 0.1611328125, | |
| "eval_rewards/rejected": -0.64453125, | |
| "eval_runtime": 12.2837, | |
| "eval_samples_per_second": 16.282, | |
| "eval_steps_per_second": 0.57, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 0.47756859491215975, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -0.5703125, | |
| "logits/rejected": -0.609375, | |
| "logps/chosen": -21.875, | |
| "logps/rejected": -27.25, | |
| "loss": 0.6334, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.62890625, | |
| "rewards/margins": 0.33203125, | |
| "rewards/rejected": -0.9609375, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 7.692990866640667, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -0.59765625, | |
| "logits/rejected": -0.6171875, | |
| "logps/chosen": -21.0, | |
| "logps/rejected": -23.75, | |
| "loss": 0.6428, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -0.466796875, | |
| "rewards/margins": 0.2578125, | |
| "rewards/rejected": -0.7265625, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 1.6845955257446075, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": -0.5390625, | |
| "logits/rejected": -0.58984375, | |
| "logps/chosen": -20.875, | |
| "logps/rejected": -24.0, | |
| "loss": 0.6515, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -0.578125, | |
| "rewards/margins": 0.1943359375, | |
| "rewards/rejected": -0.7734375, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 0.6884600902567845, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -0.498046875, | |
| "logits/rejected": -0.5, | |
| "logps/chosen": -23.25, | |
| "logps/rejected": -26.5, | |
| "loss": 0.6295, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -0.796875, | |
| "rewards/margins": 0.271484375, | |
| "rewards/rejected": -1.0703125, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 1.2231931860750773, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -0.54296875, | |
| "logits/rejected": -0.5390625, | |
| "logps/chosen": -24.875, | |
| "logps/rejected": -29.625, | |
| "loss": 0.6261, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -0.87890625, | |
| "rewards/margins": 0.41796875, | |
| "rewards/rejected": -1.296875, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 11.078142492071423, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -0.58984375, | |
| "logits/rejected": -0.546875, | |
| "logps/chosen": -27.125, | |
| "logps/rejected": -29.75, | |
| "loss": 0.6355, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.92578125, | |
| "rewards/margins": 0.380859375, | |
| "rewards/rejected": -1.3046875, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 0.7348700969108466, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -0.53515625, | |
| "logits/rejected": -0.51171875, | |
| "logps/chosen": -25.375, | |
| "logps/rejected": -29.125, | |
| "loss": 0.6226, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -0.97265625, | |
| "rewards/margins": 0.37109375, | |
| "rewards/rejected": -1.34375, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 1.187887622769852, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -0.5546875, | |
| "logits/rejected": -0.5390625, | |
| "logps/chosen": -28.75, | |
| "logps/rejected": -33.0, | |
| "loss": 0.6269, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -1.21875, | |
| "rewards/margins": 0.427734375, | |
| "rewards/rejected": -1.6484375, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 0.03453262127447758, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -0.484375, | |
| "logits/rejected": -0.5234375, | |
| "logps/chosen": -26.25, | |
| "logps/rejected": -33.25, | |
| "loss": 0.6364, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -1.1171875, | |
| "rewards/margins": 0.4921875, | |
| "rewards/rejected": -1.609375, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 1.3083004341496636, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -0.546875, | |
| "logits/rejected": -0.5078125, | |
| "logps/chosen": -25.75, | |
| "logps/rejected": -28.125, | |
| "loss": 0.6324, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -0.98046875, | |
| "rewards/margins": 0.26171875, | |
| "rewards/rejected": -1.2421875, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 1.8028119184923288, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -0.51171875, | |
| "logits/rejected": -0.4921875, | |
| "logps/chosen": -26.875, | |
| "logps/rejected": -31.25, | |
| "loss": 0.6229, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -1.0859375, | |
| "rewards/margins": 0.421875, | |
| "rewards/rejected": -1.5078125, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 0.124855255033815, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -0.462890625, | |
| "logits/rejected": -0.478515625, | |
| "logps/chosen": -26.125, | |
| "logps/rejected": -29.375, | |
| "loss": 0.6101, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -1.078125, | |
| "rewards/margins": 0.291015625, | |
| "rewards/rejected": -1.3671875, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 2.603483259161549, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": -0.5546875, | |
| "logits/rejected": -0.51953125, | |
| "logps/chosen": -29.0, | |
| "logps/rejected": -33.0, | |
| "loss": 0.6202, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -1.203125, | |
| "rewards/margins": 0.4765625, | |
| "rewards/rejected": -1.6796875, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 0.9178503377613045, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": -0.54296875, | |
| "logits/rejected": -0.5, | |
| "logps/chosen": -29.625, | |
| "logps/rejected": -35.25, | |
| "loss": 0.6058, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -1.2890625, | |
| "rewards/margins": 0.6015625, | |
| "rewards/rejected": -1.890625, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 8.273469175480448, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.455078125, | |
| "logits/rejected": -0.515625, | |
| "logps/chosen": -29.25, | |
| "logps/rejected": -38.0, | |
| "loss": 0.6037, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -1.34375, | |
| "rewards/margins": 0.77734375, | |
| "rewards/rejected": -2.125, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 2.8566426933872373, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -0.5078125, | |
| "logits/rejected": -0.486328125, | |
| "logps/chosen": -30.875, | |
| "logps/rejected": -35.75, | |
| "loss": 0.6208, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -1.4375, | |
| "rewards/margins": 0.46484375, | |
| "rewards/rejected": -1.90625, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 8.212277089600386, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -0.482421875, | |
| "logits/rejected": -0.43359375, | |
| "logps/chosen": -31.125, | |
| "logps/rejected": -35.75, | |
| "loss": 0.6351, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -1.546875, | |
| "rewards/margins": 0.46484375, | |
| "rewards/rejected": -2.015625, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 0.2370418122400776, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -0.4609375, | |
| "logits/rejected": -0.4765625, | |
| "logps/chosen": -30.5, | |
| "logps/rejected": -38.0, | |
| "loss": 0.6213, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -1.515625, | |
| "rewards/margins": 0.55859375, | |
| "rewards/rejected": -2.078125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 4.880574588659062, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -0.4140625, | |
| "logits/rejected": -0.416015625, | |
| "logps/chosen": -30.625, | |
| "logps/rejected": -37.25, | |
| "loss": 0.6321, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -1.515625, | |
| "rewards/margins": 0.65625, | |
| "rewards/rejected": -2.171875, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 2.095030195849023, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -0.50390625, | |
| "logits/rejected": -0.42578125, | |
| "logps/chosen": -33.75, | |
| "logps/rejected": -36.0, | |
| "loss": 0.6275, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -1.6171875, | |
| "rewards/margins": 0.416015625, | |
| "rewards/rejected": -2.03125, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 0.19334958288060325, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -0.458984375, | |
| "logits/rejected": -0.474609375, | |
| "logps/chosen": -30.125, | |
| "logps/rejected": -38.25, | |
| "loss": 0.6521, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": -1.4375, | |
| "rewards/margins": 0.6484375, | |
| "rewards/rejected": -2.09375, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 0.15216033874375934, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -0.4921875, | |
| "logits/rejected": -0.44140625, | |
| "logps/chosen": -32.25, | |
| "logps/rejected": -38.25, | |
| "loss": 0.62, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -1.46875, | |
| "rewards/margins": 0.81640625, | |
| "rewards/rejected": -2.28125, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 0.7437150470780799, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -0.421875, | |
| "logits/rejected": -0.404296875, | |
| "logps/chosen": -32.5, | |
| "logps/rejected": -36.75, | |
| "loss": 0.6256, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -1.671875, | |
| "rewards/margins": 0.423828125, | |
| "rewards/rejected": -2.09375, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 1.280179770444938, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -0.443359375, | |
| "logits/rejected": -0.41796875, | |
| "logps/chosen": -31.75, | |
| "logps/rejected": -37.0, | |
| "loss": 0.6072, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -1.59375, | |
| "rewards/margins": 0.50390625, | |
| "rewards/rejected": -2.09375, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 2.0151756760762396, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -0.458984375, | |
| "logits/rejected": -0.423828125, | |
| "logps/chosen": -32.25, | |
| "logps/rejected": -35.25, | |
| "loss": 0.6298, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -1.59375, | |
| "rewards/margins": 0.306640625, | |
| "rewards/rejected": -1.8984375, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 0.9264195339363805, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -0.40625, | |
| "logits/rejected": -0.431640625, | |
| "logps/chosen": -30.875, | |
| "logps/rejected": -35.5, | |
| "loss": 0.6411, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -1.5703125, | |
| "rewards/margins": 0.3984375, | |
| "rewards/rejected": -1.9609375, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 6.430978702389866, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -0.419921875, | |
| "logits/rejected": -0.41796875, | |
| "logps/chosen": -34.0, | |
| "logps/rejected": -41.5, | |
| "loss": 0.6154, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -1.7890625, | |
| "rewards/margins": 0.7109375, | |
| "rewards/rejected": -2.5, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 2.8804581682473938, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -0.37109375, | |
| "logits/rejected": -0.380859375, | |
| "logps/chosen": -31.375, | |
| "logps/rejected": -36.0, | |
| "loss": 0.6245, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -1.65625, | |
| "rewards/margins": 0.42578125, | |
| "rewards/rejected": -2.09375, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 0.1960808455329659, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -0.41015625, | |
| "logits/rejected": -0.390625, | |
| "logps/chosen": -30.5, | |
| "logps/rejected": -34.75, | |
| "loss": 0.6091, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -1.4921875, | |
| "rewards/margins": 0.408203125, | |
| "rewards/rejected": -1.8984375, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 0.4729918310991834, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -0.47265625, | |
| "logits/rejected": -0.453125, | |
| "logps/chosen": -33.25, | |
| "logps/rejected": -42.5, | |
| "loss": 0.625, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": -1.6796875, | |
| "rewards/margins": 0.8828125, | |
| "rewards/rejected": -2.5625, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 0.4814000798263685, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": -0.462890625, | |
| "logits/rejected": -0.40234375, | |
| "logps/chosen": -35.0, | |
| "logps/rejected": -38.0, | |
| "loss": 0.6426, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -1.8046875, | |
| "rewards/margins": 0.396484375, | |
| "rewards/rejected": -2.203125, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -0.408203125, | |
| "eval_logits/rejected": -0.41796875, | |
| "eval_logps/chosen": -32.25, | |
| "eval_logps/rejected": -34.25, | |
| "eval_loss": 0.6540625095367432, | |
| "eval_rewards/accuracies": 0.0892857164144516, | |
| "eval_rewards/chosen": -1.71875, | |
| "eval_rewards/margins": 0.169921875, | |
| "eval_rewards/rejected": -1.8828125, | |
| "eval_runtime": 15.0164, | |
| "eval_samples_per_second": 13.319, | |
| "eval_steps_per_second": 0.466, | |
| "step": 939 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 939, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "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 | |
| } | |