| { | |
| "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": 24.277064761737957, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -3.765625, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -111.5, | |
| "logps/rejected": -113.5, | |
| "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": 22.01509678618656, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -3.6875, | |
| "logits/rejected": -3.75, | |
| "logps/chosen": -119.5, | |
| "logps/rejected": -135.0, | |
| "loss": 0.6913, | |
| "rewards/accuracies": 0.1111111119389534, | |
| "rewards/chosen": 0.0048828125, | |
| "rewards/margins": 0.006256103515625, | |
| "rewards/rejected": -0.0013885498046875, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 26.23315038989938, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.796875, | |
| "logps/chosen": -140.0, | |
| "logps/rejected": -150.0, | |
| "loss": 0.6917, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": 0.0087890625, | |
| "rewards/margins": -0.000316619873046875, | |
| "rewards/rejected": 0.009033203125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 26.06572383011012, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.75, | |
| "logps/chosen": -134.0, | |
| "logps/rejected": -140.0, | |
| "loss": 0.691, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": 0.025634765625, | |
| "rewards/margins": 0.00750732421875, | |
| "rewards/rejected": 0.0181884765625, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 13.166840981977872, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -3.765625, | |
| "logits/rejected": -3.75, | |
| "logps/chosen": -143.0, | |
| "logps/rejected": -142.0, | |
| "loss": 0.692, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.0556640625, | |
| "rewards/margins": 0.00250244140625, | |
| "rewards/rejected": 0.05322265625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 30.306245976533457, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.75, | |
| "logps/chosen": -122.5, | |
| "logps/rejected": -137.0, | |
| "loss": 0.6876, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": 0.0966796875, | |
| "rewards/margins": 0.0106201171875, | |
| "rewards/rejected": 0.0859375, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 26.83559964765963, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -3.828125, | |
| "logits/rejected": -3.765625, | |
| "logps/chosen": -130.0, | |
| "logps/rejected": -131.0, | |
| "loss": 0.6847, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.1923828125, | |
| "rewards/margins": 0.0322265625, | |
| "rewards/rejected": 0.16015625, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 27.35473267545649, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -122.0, | |
| "logps/rejected": -129.0, | |
| "loss": 0.677, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.19140625, | |
| "rewards/margins": 0.03955078125, | |
| "rewards/rejected": 0.15234375, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 31.93177456499877, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -3.796875, | |
| "logits/rejected": -3.828125, | |
| "logps/chosen": -136.0, | |
| "logps/rejected": -150.0, | |
| "loss": 0.6783, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": 0.353515625, | |
| "rewards/margins": 0.050537109375, | |
| "rewards/rejected": 0.302734375, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 25.116095410119712, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -133.0, | |
| "logps/rejected": -143.0, | |
| "loss": 0.679, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.37890625, | |
| "rewards/margins": 0.06591796875, | |
| "rewards/rejected": 0.3125, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 32.30648339723812, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -3.8125, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -121.5, | |
| "logps/rejected": -123.0, | |
| "loss": 0.6674, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": 0.41015625, | |
| "rewards/margins": 0.057373046875, | |
| "rewards/rejected": 0.353515625, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 19.894965188757386, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -132.0, | |
| "logps/rejected": -140.0, | |
| "loss": 0.6687, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.322265625, | |
| "rewards/margins": 0.1318359375, | |
| "rewards/rejected": 0.19140625, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 22.763609842360637, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -3.796875, | |
| "logits/rejected": -3.828125, | |
| "logps/chosen": -122.0, | |
| "logps/rejected": -126.5, | |
| "loss": 0.6629, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.51171875, | |
| "rewards/margins": 0.0673828125, | |
| "rewards/rejected": 0.4453125, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 28.99821873200043, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -3.703125, | |
| "logits/rejected": -3.765625, | |
| "logps/chosen": -113.5, | |
| "logps/rejected": -138.0, | |
| "loss": 0.6736, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": 0.6015625, | |
| "rewards/margins": 0.1171875, | |
| "rewards/rejected": 0.486328125, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 113.14344588395304, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.796875, | |
| "logps/chosen": -128.0, | |
| "logps/rejected": -137.0, | |
| "loss": 0.6608, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.2470703125, | |
| "rewards/margins": 0.06787109375, | |
| "rewards/rejected": 0.1796875, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 26.586734580838513, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -3.796875, | |
| "logits/rejected": -3.828125, | |
| "logps/chosen": -129.0, | |
| "logps/rejected": -134.0, | |
| "loss": 0.6597, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": 0.1611328125, | |
| "rewards/margins": 0.1328125, | |
| "rewards/rejected": 0.02783203125, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 18.377318511019165, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -3.78125, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -131.0, | |
| "logps/rejected": -141.0, | |
| "loss": 0.6446, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.031005859375, | |
| "rewards/margins": 0.197265625, | |
| "rewards/rejected": -0.166015625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 14.396584252817917, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -3.8125, | |
| "logits/rejected": -3.828125, | |
| "logps/chosen": -111.0, | |
| "logps/rejected": -112.0, | |
| "loss": 0.636, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.259765625, | |
| "rewards/margins": 0.1708984375, | |
| "rewards/rejected": 0.08935546875, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 18.99460682217856, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -3.796875, | |
| "logits/rejected": -3.8125, | |
| "logps/chosen": -118.0, | |
| "logps/rejected": -124.0, | |
| "loss": 0.6389, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": -0.1005859375, | |
| "rewards/margins": 0.240234375, | |
| "rewards/rejected": -0.33984375, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 16.222491503696077, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -3.765625, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -123.0, | |
| "logps/rejected": -136.0, | |
| "loss": 0.642, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": -0.045166015625, | |
| "rewards/margins": 0.1435546875, | |
| "rewards/rejected": -0.1884765625, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 26.791655684959107, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -3.78125, | |
| "logits/rejected": -3.765625, | |
| "logps/chosen": -120.0, | |
| "logps/rejected": -127.5, | |
| "loss": 0.6475, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -0.09521484375, | |
| "rewards/margins": 0.15625, | |
| "rewards/rejected": -0.251953125, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 68.37542169612291, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -3.8125, | |
| "logits/rejected": -3.828125, | |
| "logps/chosen": -126.5, | |
| "logps/rejected": -128.0, | |
| "loss": 0.6515, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": -0.09033203125, | |
| "rewards/margins": 0.1728515625, | |
| "rewards/rejected": -0.263671875, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 17.049745693631767, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -3.96875, | |
| "logits/rejected": -4.0, | |
| "logps/chosen": -138.0, | |
| "logps/rejected": -150.0, | |
| "loss": 0.6227, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -0.7109375, | |
| "rewards/margins": 0.294921875, | |
| "rewards/rejected": -1.0078125, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 32.3667179695025, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -3.984375, | |
| "logits/rejected": -3.96875, | |
| "logps/chosen": -158.0, | |
| "logps/rejected": -148.0, | |
| "loss": 0.6975, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -0.546875, | |
| "rewards/margins": 0.1083984375, | |
| "rewards/rejected": -0.65625, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 19.236192318267502, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -3.8125, | |
| "logits/rejected": -3.828125, | |
| "logps/chosen": -123.5, | |
| "logps/rejected": -138.0, | |
| "loss": 0.6271, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -0.392578125, | |
| "rewards/margins": 0.244140625, | |
| "rewards/rejected": -0.63671875, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 31.43001379061512, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.8125, | |
| "logps/chosen": -119.5, | |
| "logps/rejected": -131.0, | |
| "loss": 0.6765, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": 0.1376953125, | |
| "rewards/margins": 0.330078125, | |
| "rewards/rejected": -0.193359375, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 41.23944776953403, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -3.765625, | |
| "logits/rejected": -3.796875, | |
| "logps/chosen": -134.0, | |
| "logps/rejected": -145.0, | |
| "loss": 0.6747, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": 0.1875, | |
| "rewards/margins": 0.177734375, | |
| "rewards/rejected": 0.0093994140625, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 29.299081181603164, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -3.828125, | |
| "logits/rejected": -3.828125, | |
| "logps/chosen": -117.0, | |
| "logps/rejected": -121.5, | |
| "loss": 0.6623, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": 0.1630859375, | |
| "rewards/margins": 0.16015625, | |
| "rewards/rejected": 0.003814697265625, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 26.452757757346387, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -3.734375, | |
| "logits/rejected": -3.765625, | |
| "logps/chosen": -122.0, | |
| "logps/rejected": -125.5, | |
| "loss": 0.6555, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": 0.07177734375, | |
| "rewards/margins": 0.08251953125, | |
| "rewards/rejected": -0.0101318359375, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 26.718085170404358, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -135.0, | |
| "logps/rejected": -145.0, | |
| "loss": 0.6607, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -0.16796875, | |
| "rewards/margins": 0.20703125, | |
| "rewards/rejected": -0.375, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 23.457732968099574, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -140.0, | |
| "logps/rejected": -146.0, | |
| "loss": 0.6655, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -0.423828125, | |
| "rewards/margins": 0.07275390625, | |
| "rewards/rejected": -0.49609375, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 18.948652581153286, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -3.8125, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -142.0, | |
| "logps/rejected": -146.0, | |
| "loss": 0.6675, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -0.205078125, | |
| "rewards/margins": 0.09619140625, | |
| "rewards/rejected": -0.30078125, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -3.75, | |
| "eval_logits/rejected": -3.75, | |
| "eval_logps/chosen": -110.0, | |
| "eval_logps/rejected": -116.5, | |
| "eval_loss": 0.6621875166893005, | |
| "eval_rewards/accuracies": 0.1785714328289032, | |
| "eval_rewards/chosen": -0.0888671875, | |
| "eval_rewards/margins": 0.1787109375, | |
| "eval_rewards/rejected": -0.267578125, | |
| "eval_runtime": 12.1475, | |
| "eval_samples_per_second": 16.464, | |
| "eval_steps_per_second": 0.576, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 12.780435839397535, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -144.0, | |
| "logps/rejected": -154.0, | |
| "loss": 0.5925, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.06494140625, | |
| "rewards/margins": 0.84765625, | |
| "rewards/rejected": -0.91015625, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 2.3499904446036433, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -129.0, | |
| "logps/rejected": -146.0, | |
| "loss": 0.54, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -0.50390625, | |
| "rewards/margins": 0.6640625, | |
| "rewards/rejected": -1.171875, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 5.297851157977379, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.984375, | |
| "logps/chosen": -133.0, | |
| "logps/rejected": -143.0, | |
| "loss": 0.5106, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -0.2890625, | |
| "rewards/margins": 0.8828125, | |
| "rewards/rejected": -1.171875, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 15.356893947406501, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -127.0, | |
| "logps/rejected": -158.0, | |
| "loss": 0.5021, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -0.41015625, | |
| "rewards/margins": 1.3515625, | |
| "rewards/rejected": -1.7578125, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 7.611884421622937, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.96875, | |
| "logps/chosen": -145.0, | |
| "logps/rejected": -178.0, | |
| "loss": 0.5444, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -1.25, | |
| "rewards/margins": 1.109375, | |
| "rewards/rejected": -2.359375, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 14.348371026442878, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -4.0, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -158.0, | |
| "logps/rejected": -170.0, | |
| "loss": 0.5288, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": -1.6640625, | |
| "rewards/margins": 1.0625, | |
| "rewards/rejected": -2.734375, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 2.947572631593703, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -129.0, | |
| "logps/rejected": -141.0, | |
| "loss": 0.5245, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": -1.1171875, | |
| "rewards/margins": 0.9453125, | |
| "rewards/rejected": -2.0625, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 8.284381008128047, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -3.796875, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -146.0, | |
| "logps/rejected": -165.0, | |
| "loss": 0.5533, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -1.265625, | |
| "rewards/margins": 1.0078125, | |
| "rewards/rejected": -2.265625, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 3.2439367204723837, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -160.0, | |
| "logps/rejected": -176.0, | |
| "loss": 0.5292, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -1.671875, | |
| "rewards/margins": 1.1328125, | |
| "rewards/rejected": -2.796875, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 23.291826626822854, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -149.0, | |
| "logps/rejected": -167.0, | |
| "loss": 0.4924, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -1.625, | |
| "rewards/margins": 1.2578125, | |
| "rewards/rejected": -2.875, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 11.526112822046265, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -3.828125, | |
| "logits/rejected": -3.96875, | |
| "logps/chosen": -148.0, | |
| "logps/rejected": -175.0, | |
| "loss": 0.4906, | |
| "rewards/accuracies": 0.38749998807907104, | |
| "rewards/chosen": -1.1328125, | |
| "rewards/margins": 1.8671875, | |
| "rewards/rejected": -3.0, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 7.079418300270812, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -3.765625, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -137.0, | |
| "logps/rejected": -163.0, | |
| "loss": 0.5138, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -1.625, | |
| "rewards/margins": 1.1328125, | |
| "rewards/rejected": -2.765625, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 23.51606757056333, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -3.921875, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -149.0, | |
| "logps/rejected": -164.0, | |
| "loss": 0.5057, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -1.890625, | |
| "rewards/margins": 1.140625, | |
| "rewards/rejected": -3.03125, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 11.56064802099853, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -3.8125, | |
| "logits/rejected": -3.8125, | |
| "logps/chosen": -144.0, | |
| "logps/rejected": -174.0, | |
| "loss": 0.5141, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -1.4609375, | |
| "rewards/margins": 1.6640625, | |
| "rewards/rejected": -3.125, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 11.236096530357507, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.75, | |
| "logps/chosen": -152.0, | |
| "logps/rejected": -165.0, | |
| "loss": 0.5067, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -1.9609375, | |
| "rewards/margins": 1.15625, | |
| "rewards/rejected": -3.125, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 3.370024966920903, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -3.921875, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -141.0, | |
| "logps/rejected": -163.0, | |
| "loss": 0.523, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -2.09375, | |
| "rewards/margins": 1.53125, | |
| "rewards/rejected": -3.625, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 9.978824743164745, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -3.828125, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -145.0, | |
| "logps/rejected": -163.0, | |
| "loss": 0.5252, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -2.5, | |
| "rewards/margins": 0.96484375, | |
| "rewards/rejected": -3.453125, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 14.326969919306727, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.984375, | |
| "logps/chosen": -152.0, | |
| "logps/rejected": -172.0, | |
| "loss": 0.5329, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -2.515625, | |
| "rewards/margins": 1.734375, | |
| "rewards/rejected": -4.25, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 6.967846297592076, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -4.03125, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -176.0, | |
| "logps/rejected": -205.0, | |
| "loss": 0.539, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -2.75, | |
| "rewards/margins": 1.53125, | |
| "rewards/rejected": -4.28125, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 10.478631685036758, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -3.96875, | |
| "logits/rejected": -3.953125, | |
| "logps/chosen": -166.0, | |
| "logps/rejected": -181.0, | |
| "loss": 0.5138, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -2.390625, | |
| "rewards/margins": 1.53125, | |
| "rewards/rejected": -3.921875, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 3.9485916094762277, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -3.8125, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -138.0, | |
| "logps/rejected": -165.0, | |
| "loss": 0.5187, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -1.78125, | |
| "rewards/margins": 1.7890625, | |
| "rewards/rejected": -3.578125, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 9.691106682722046, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.96875, | |
| "logps/chosen": -149.0, | |
| "logps/rejected": -163.0, | |
| "loss": 0.5132, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -1.890625, | |
| "rewards/margins": 1.0078125, | |
| "rewards/rejected": -2.890625, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 6.194603425697891, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -158.0, | |
| "logps/rejected": -181.0, | |
| "loss": 0.5271, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -2.328125, | |
| "rewards/margins": 1.2890625, | |
| "rewards/rejected": -3.609375, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 4.8547868881787535, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -158.0, | |
| "logps/rejected": -180.0, | |
| "loss": 0.5112, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -2.265625, | |
| "rewards/margins": 1.6015625, | |
| "rewards/rejected": -3.875, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 6.042788937636992, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -157.0, | |
| "logps/rejected": -203.0, | |
| "loss": 0.5342, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -2.8125, | |
| "rewards/margins": 1.65625, | |
| "rewards/rejected": -4.46875, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 17.608362227397397, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -140.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.5249, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -2.125, | |
| "rewards/margins": 1.609375, | |
| "rewards/rejected": -3.734375, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 24.398412082273794, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -3.78125, | |
| "logits/rejected": -3.734375, | |
| "logps/chosen": -151.0, | |
| "logps/rejected": -158.0, | |
| "loss": 0.52, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -2.953125, | |
| "rewards/margins": 1.171875, | |
| "rewards/rejected": -4.125, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 17.069584368649224, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -3.796875, | |
| "logits/rejected": -3.8125, | |
| "logps/chosen": -165.0, | |
| "logps/rejected": -181.0, | |
| "loss": 0.5235, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -3.28125, | |
| "rewards/margins": 1.5390625, | |
| "rewards/rejected": -4.8125, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 2.1180201542039883, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -3.765625, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -150.0, | |
| "logps/rejected": -176.0, | |
| "loss": 0.5272, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -3.15625, | |
| "rewards/margins": 1.6015625, | |
| "rewards/rejected": -4.75, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 3.4777849704272694, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -152.0, | |
| "logps/rejected": -160.0, | |
| "loss": 0.5547, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -2.6875, | |
| "rewards/margins": 1.203125, | |
| "rewards/rejected": -3.890625, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 27.898326722256666, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -3.5625, | |
| "logits/rejected": -3.5625, | |
| "logps/chosen": -160.0, | |
| "logps/rejected": -174.0, | |
| "loss": 0.5474, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -3.390625, | |
| "rewards/margins": 1.1015625, | |
| "rewards/rejected": -4.5, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -3.8125, | |
| "eval_logits/rejected": -3.828125, | |
| "eval_logps/chosen": -144.0, | |
| "eval_logps/rejected": -151.0, | |
| "eval_loss": 0.6862499713897705, | |
| "eval_rewards/accuracies": 0.1428571492433548, | |
| "eval_rewards/chosen": -3.4375, | |
| "eval_rewards/margins": 0.33203125, | |
| "eval_rewards/rejected": -3.78125, | |
| "eval_runtime": 12.0804, | |
| "eval_samples_per_second": 16.556, | |
| "eval_steps_per_second": 0.579, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 2.410305409475697, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -3.796875, | |
| "logits/rejected": -3.8125, | |
| "logps/chosen": -165.0, | |
| "logps/rejected": -183.0, | |
| "loss": 0.5177, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -3.171875, | |
| "rewards/margins": 1.609375, | |
| "rewards/rejected": -4.78125, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 0.8113508868000807, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.765625, | |
| "logps/chosen": -155.0, | |
| "logps/rejected": -177.0, | |
| "loss": 0.4748, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -2.515625, | |
| "rewards/margins": 2.234375, | |
| "rewards/rejected": -4.75, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 0.8022386849359633, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": -3.578125, | |
| "logits/rejected": -3.578125, | |
| "logps/chosen": -165.0, | |
| "logps/rejected": -182.0, | |
| "loss": 0.494, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -3.546875, | |
| "rewards/margins": 2.0625, | |
| "rewards/rejected": -5.59375, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 1.4127272526100374, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -3.5, | |
| "logits/rejected": -3.59375, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -198.0, | |
| "loss": 0.4836, | |
| "rewards/accuracies": 0.4000000059604645, | |
| "rewards/chosen": -3.0625, | |
| "rewards/margins": 3.265625, | |
| "rewards/rejected": -6.3125, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 1.6015839862908308, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -3.484375, | |
| "logits/rejected": -3.59375, | |
| "logps/chosen": -166.0, | |
| "logps/rejected": -197.0, | |
| "loss": 0.4914, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 2.328125, | |
| "rewards/rejected": -6.46875, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 1.3365523456961093, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -3.40625, | |
| "logits/rejected": -3.421875, | |
| "logps/chosen": -161.0, | |
| "logps/rejected": -193.0, | |
| "loss": 0.5066, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -3.578125, | |
| "rewards/margins": 2.546875, | |
| "rewards/rejected": -6.125, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 0.3732248620636056, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -3.390625, | |
| "logits/rejected": -3.4375, | |
| "logps/chosen": -165.0, | |
| "logps/rejected": -191.0, | |
| "loss": 0.508, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -3.984375, | |
| "rewards/margins": 2.1875, | |
| "rewards/rejected": -6.15625, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 1.3702802041665394, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -3.578125, | |
| "logits/rejected": -3.65625, | |
| "logps/chosen": -168.0, | |
| "logps/rejected": -199.0, | |
| "loss": 0.4658, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -3.375, | |
| "rewards/margins": 2.40625, | |
| "rewards/rejected": -5.78125, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 4.552571161813631, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -3.484375, | |
| "logits/rejected": -3.59375, | |
| "logps/chosen": -170.0, | |
| "logps/rejected": -208.0, | |
| "loss": 0.4617, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -4.0, | |
| "rewards/margins": 2.34375, | |
| "rewards/rejected": -6.34375, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 1.124329278277091, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -3.453125, | |
| "logits/rejected": -3.5, | |
| "logps/chosen": -173.0, | |
| "logps/rejected": -215.0, | |
| "loss": 0.5046, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -3.859375, | |
| "rewards/margins": 2.953125, | |
| "rewards/rejected": -6.8125, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 0.9258502818824418, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -3.34375, | |
| "logits/rejected": -3.515625, | |
| "logps/chosen": -163.0, | |
| "logps/rejected": -202.0, | |
| "loss": 0.4876, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -3.984375, | |
| "rewards/margins": 2.71875, | |
| "rewards/rejected": -6.6875, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 0.5020087699809023, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -3.546875, | |
| "logits/rejected": -3.515625, | |
| "logps/chosen": -176.0, | |
| "logps/rejected": -192.0, | |
| "loss": 0.5051, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -4.3125, | |
| "rewards/margins": 2.171875, | |
| "rewards/rejected": -6.5, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 1.2704607759015243, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": -3.515625, | |
| "logits/rejected": -3.46875, | |
| "logps/chosen": -177.0, | |
| "logps/rejected": -210.0, | |
| "loss": 0.4615, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -4.0625, | |
| "rewards/margins": 2.375, | |
| "rewards/rejected": -6.4375, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 0.2652885305653395, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": -3.546875, | |
| "logits/rejected": -3.53125, | |
| "logps/chosen": -166.0, | |
| "logps/rejected": -178.0, | |
| "loss": 0.5083, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -4.21875, | |
| "rewards/margins": 1.46875, | |
| "rewards/rejected": -5.65625, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 0.47184139297382116, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -3.40625, | |
| "logits/rejected": -3.546875, | |
| "logps/chosen": -147.0, | |
| "logps/rejected": -197.0, | |
| "loss": 0.4696, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -3.890625, | |
| "rewards/margins": 2.84375, | |
| "rewards/rejected": -6.71875, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 0.27130707751259775, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -3.5625, | |
| "logits/rejected": -3.609375, | |
| "logps/chosen": -164.0, | |
| "logps/rejected": -212.0, | |
| "loss": 0.4639, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": -3.78125, | |
| "rewards/margins": 3.703125, | |
| "rewards/rejected": -7.46875, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 0.8280931838884058, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -3.59375, | |
| "logits/rejected": -3.640625, | |
| "logps/chosen": -168.0, | |
| "logps/rejected": -200.0, | |
| "loss": 0.5019, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 2.203125, | |
| "rewards/rejected": -6.78125, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 0.4060588791740879, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -3.390625, | |
| "logits/rejected": -3.53125, | |
| "logps/chosen": -186.0, | |
| "logps/rejected": -220.0, | |
| "loss": 0.4955, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -4.40625, | |
| "rewards/margins": 2.421875, | |
| "rewards/rejected": -6.8125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 0.3435065796213597, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -3.5, | |
| "logits/rejected": -3.5625, | |
| "logps/chosen": -162.0, | |
| "logps/rejected": -189.0, | |
| "loss": 0.4954, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 2.4375, | |
| "rewards/rejected": -6.5625, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 0.5535707114884407, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -3.5, | |
| "logits/rejected": -3.65625, | |
| "logps/chosen": -182.0, | |
| "logps/rejected": -213.0, | |
| "loss": 0.5022, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -5.0, | |
| "rewards/margins": 2.1875, | |
| "rewards/rejected": -7.21875, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 1.08090734300108, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -3.53125, | |
| "logits/rejected": -3.59375, | |
| "logps/chosen": -178.0, | |
| "logps/rejected": -208.0, | |
| "loss": 0.5132, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 2.328125, | |
| "rewards/rejected": -6.59375, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 0.33498681500442445, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -3.453125, | |
| "logits/rejected": -3.484375, | |
| "logps/chosen": -166.0, | |
| "logps/rejected": -202.0, | |
| "loss": 0.4933, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -4.1875, | |
| "rewards/margins": 2.59375, | |
| "rewards/rejected": -6.78125, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 0.2206671231026912, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -3.53125, | |
| "logits/rejected": -3.609375, | |
| "logps/chosen": -176.0, | |
| "logps/rejected": -217.0, | |
| "loss": 0.4845, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -4.375, | |
| "rewards/margins": 2.96875, | |
| "rewards/rejected": -7.34375, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 0.45488082207971153, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -3.578125, | |
| "logits/rejected": -3.625, | |
| "logps/chosen": -177.0, | |
| "logps/rejected": -216.0, | |
| "loss": 0.4979, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 2.75, | |
| "rewards/rejected": -7.5, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 1.3853289418261023, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -3.625, | |
| "logits/rejected": -3.65625, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -221.0, | |
| "loss": 0.474, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -5.25, | |
| "rewards/margins": 2.28125, | |
| "rewards/rejected": -7.53125, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 0.42469154475289916, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -3.5625, | |
| "logits/rejected": -3.59375, | |
| "logps/chosen": -166.0, | |
| "logps/rejected": -185.0, | |
| "loss": 0.5236, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -4.875, | |
| "rewards/margins": 1.4921875, | |
| "rewards/rejected": -6.375, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 0.5486851422310616, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -3.53125, | |
| "logits/rejected": -3.5625, | |
| "logps/chosen": -183.0, | |
| "logps/rejected": -203.0, | |
| "loss": 0.487, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -4.59375, | |
| "rewards/margins": 2.453125, | |
| "rewards/rejected": -7.0625, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 0.6798328903999439, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -3.484375, | |
| "logits/rejected": -3.53125, | |
| "logps/chosen": -176.0, | |
| "logps/rejected": -220.0, | |
| "loss": 0.502, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 3.484375, | |
| "rewards/rejected": -8.0, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 0.48478821436430236, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -3.640625, | |
| "logits/rejected": -3.734375, | |
| "logps/chosen": -195.0, | |
| "logps/rejected": -226.0, | |
| "loss": 0.5222, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -5.28125, | |
| "rewards/margins": 2.09375, | |
| "rewards/rejected": -7.375, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 0.5314704772510245, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -3.515625, | |
| "logits/rejected": -3.515625, | |
| "logps/chosen": -170.0, | |
| "logps/rejected": -197.0, | |
| "loss": 0.4831, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -4.375, | |
| "rewards/margins": 2.765625, | |
| "rewards/rejected": -7.15625, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 0.8748401307399186, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": -3.4375, | |
| "logits/rejected": -3.421875, | |
| "logps/chosen": -170.0, | |
| "logps/rejected": -197.0, | |
| "loss": 0.4948, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -4.09375, | |
| "rewards/margins": 2.859375, | |
| "rewards/rejected": -6.96875, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -3.46875, | |
| "eval_logits/rejected": -3.5, | |
| "eval_logps/chosen": -163.0, | |
| "eval_logps/rejected": -174.0, | |
| "eval_loss": 0.7490624785423279, | |
| "eval_rewards/accuracies": 0.1964285671710968, | |
| "eval_rewards/chosen": -5.375, | |
| "eval_rewards/margins": 0.6640625, | |
| "eval_rewards/rejected": -6.03125, | |
| "eval_runtime": 15.1057, | |
| "eval_samples_per_second": 13.24, | |
| "eval_steps_per_second": 0.463, | |
| "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 | |
| } | |