| { | |
| "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": 3.157389794520281, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -2.453125, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -25.25, | |
| "logps/rejected": -26.25, | |
| "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": 10.995392953199708, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -26.25, | |
| "logps/rejected": -27.5, | |
| "loss": 0.6913, | |
| "rewards/accuracies": 0.0972222238779068, | |
| "rewards/chosen": 0.0020904541015625, | |
| "rewards/margins": 0.0013885498046875, | |
| "rewards/rejected": 0.00069427490234375, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 14.756926960622804, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -26.375, | |
| "logps/rejected": -26.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": 0.010498046875, | |
| "rewards/margins": -0.0006256103515625, | |
| "rewards/rejected": 0.0111083984375, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 6.314461239279313, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -25.875, | |
| "logps/rejected": -26.75, | |
| "loss": 0.6903, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.0390625, | |
| "rewards/margins": 0.0037384033203125, | |
| "rewards/rejected": 0.035400390625, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 4.069457842491787, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -26.625, | |
| "logps/rejected": -25.125, | |
| "loss": 0.69, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": 0.0712890625, | |
| "rewards/margins": -0.00093841552734375, | |
| "rewards/rejected": 0.072265625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 8.686133525007243, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -23.375, | |
| "logps/rejected": -24.25, | |
| "loss": 0.6855, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": 0.203125, | |
| "rewards/margins": 0.02978515625, | |
| "rewards/rejected": 0.1728515625, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 1.9555966452506104, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.25, | |
| "logps/chosen": -23.5, | |
| "logps/rejected": -22.75, | |
| "loss": 0.6848, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": 0.275390625, | |
| "rewards/margins": 0.007232666015625, | |
| "rewards/rejected": 0.267578125, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 7.193018747273823, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -2.40625, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -24.375, | |
| "logps/rejected": -25.5, | |
| "loss": 0.6776, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.267578125, | |
| "rewards/margins": 0.0849609375, | |
| "rewards/rejected": 0.181640625, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 8.260204675450098, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -21.375, | |
| "logps/rejected": -22.25, | |
| "loss": 0.6711, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.423828125, | |
| "rewards/margins": 0.056884765625, | |
| "rewards/rejected": 0.3671875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 3.316086097526083, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -2.171875, | |
| "logits/rejected": -2.1875, | |
| "logps/chosen": -20.625, | |
| "logps/rejected": -22.0, | |
| "loss": 0.6649, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.4609375, | |
| "rewards/margins": 0.1298828125, | |
| "rewards/rejected": 0.330078125, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 4.11093829993774, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -2.1875, | |
| "logits/rejected": -2.203125, | |
| "logps/chosen": -23.875, | |
| "logps/rejected": -25.0, | |
| "loss": 0.6804, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.2001953125, | |
| "rewards/margins": 0.031982421875, | |
| "rewards/rejected": 0.16796875, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 15.094504412548241, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -2.125, | |
| "logits/rejected": -2.1875, | |
| "logps/chosen": -23.125, | |
| "logps/rejected": -25.125, | |
| "loss": 0.6786, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": 0.244140625, | |
| "rewards/margins": 0.1298828125, | |
| "rewards/rejected": 0.1142578125, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 14.996815100447646, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -2.09375, | |
| "logits/rejected": -2.109375, | |
| "logps/chosen": -23.875, | |
| "logps/rejected": -24.625, | |
| "loss": 0.6718, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.171875, | |
| "rewards/margins": 0.07568359375, | |
| "rewards/rejected": 0.09619140625, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 1.7038271869013584, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -25.375, | |
| "logps/rejected": -27.875, | |
| "loss": 0.6581, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": 0.2109375, | |
| "rewards/margins": 0.189453125, | |
| "rewards/rejected": 0.0211181640625, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 0.3354826155206458, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.171875, | |
| "logps/chosen": -25.0, | |
| "logps/rejected": -25.125, | |
| "loss": 0.6597, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.1728515625, | |
| "rewards/margins": 0.053955078125, | |
| "rewards/rejected": 0.11865234375, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 6.2912092048650194, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -2.203125, | |
| "logits/rejected": -2.171875, | |
| "logps/chosen": -26.625, | |
| "logps/rejected": -30.375, | |
| "loss": 0.6556, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": 0.0286865234375, | |
| "rewards/margins": 0.384765625, | |
| "rewards/rejected": -0.35546875, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 19.56501069511903, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -2.09375, | |
| "logits/rejected": -2.09375, | |
| "logps/chosen": -28.375, | |
| "logps/rejected": -30.25, | |
| "loss": 0.6509, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -0.12353515625, | |
| "rewards/margins": 0.197265625, | |
| "rewards/rejected": -0.3203125, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 0.2595057801833904, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -2.0, | |
| "logits/rejected": -1.9609375, | |
| "logps/chosen": -26.25, | |
| "logps/rejected": -27.0, | |
| "loss": 0.6754, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -0.0303955078125, | |
| "rewards/margins": 0.09619140625, | |
| "rewards/rejected": -0.126953125, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 24.845430445881203, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -2.171875, | |
| "logps/chosen": -27.125, | |
| "logps/rejected": -30.75, | |
| "loss": 0.6485, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": 0.01446533203125, | |
| "rewards/margins": 0.40625, | |
| "rewards/rejected": -0.392578125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 4.13928870532869, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -2.078125, | |
| "logits/rejected": -2.125, | |
| "logps/chosen": -27.75, | |
| "logps/rejected": -30.75, | |
| "loss": 0.6487, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -0.1650390625, | |
| "rewards/margins": 0.232421875, | |
| "rewards/rejected": -0.3984375, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 10.381558906601567, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -2.03125, | |
| "logits/rejected": -2.03125, | |
| "logps/chosen": -24.75, | |
| "logps/rejected": -26.25, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.1640625, | |
| "rewards/margins": 0.166015625, | |
| "rewards/rejected": -0.002227783203125, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 6.8578046959990395, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -1.90625, | |
| "logits/rejected": -2.046875, | |
| "logps/chosen": -23.5, | |
| "logps/rejected": -26.125, | |
| "loss": 0.6704, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.177734375, | |
| "rewards/margins": 0.1416015625, | |
| "rewards/rejected": 0.03564453125, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 8.373906679362355, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -1.921875, | |
| "logits/rejected": -1.9765625, | |
| "logps/chosen": -24.875, | |
| "logps/rejected": -27.25, | |
| "loss": 0.671, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": 0.08203125, | |
| "rewards/margins": 0.2109375, | |
| "rewards/rejected": -0.12890625, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 4.973122145274651, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -1.984375, | |
| "logits/rejected": -2.046875, | |
| "logps/chosen": -28.0, | |
| "logps/rejected": -32.5, | |
| "loss": 0.6406, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -0.181640625, | |
| "rewards/margins": 0.3671875, | |
| "rewards/rejected": -0.55078125, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 6.651592437544626, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -1.9375, | |
| "logits/rejected": -1.90625, | |
| "logps/chosen": -29.125, | |
| "logps/rejected": -30.125, | |
| "loss": 0.7045, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -0.298828125, | |
| "rewards/margins": 0.06982421875, | |
| "rewards/rejected": -0.37109375, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 24.012496220678372, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -1.8828125, | |
| "logits/rejected": -1.7890625, | |
| "logps/chosen": -29.0, | |
| "logps/rejected": -29.5, | |
| "loss": 0.6706, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -0.2392578125, | |
| "rewards/margins": 0.09521484375, | |
| "rewards/rejected": -0.333984375, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 5.6557617413477645, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -1.734375, | |
| "logits/rejected": -1.8046875, | |
| "logps/chosen": -29.75, | |
| "logps/rejected": -35.5, | |
| "loss": 0.6185, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -0.40625, | |
| "rewards/margins": 0.50390625, | |
| "rewards/rejected": -0.91015625, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 6.821519941669513, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -29.25, | |
| "logps/rejected": -31.0, | |
| "loss": 0.6513, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -0.3671875, | |
| "rewards/margins": 0.1513671875, | |
| "rewards/rejected": -0.51953125, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 4.349346321646269, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.7265625, | |
| "logps/chosen": -28.875, | |
| "logps/rejected": -33.25, | |
| "loss": 0.6461, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -0.318359375, | |
| "rewards/margins": 0.416015625, | |
| "rewards/rejected": -0.734375, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 0.3972672117279365, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -27.25, | |
| "logps/rejected": -29.75, | |
| "loss": 0.6523, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.10986328125, | |
| "rewards/margins": 0.302734375, | |
| "rewards/rejected": -0.4140625, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 0.4119772243153088, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -30.125, | |
| "logps/rejected": -36.5, | |
| "loss": 0.6497, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -0.369140625, | |
| "rewards/margins": 0.625, | |
| "rewards/rejected": -0.9921875, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 1.0151292168029609, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -33.25, | |
| "logps/rejected": -37.0, | |
| "loss": 0.6692, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.419921875, | |
| "rewards/margins": 0.494140625, | |
| "rewards/rejected": -0.9140625, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -1.40625, | |
| "eval_logits/rejected": -1.46875, | |
| "eval_logps/chosen": -33.0, | |
| "eval_logps/rejected": -33.5, | |
| "eval_loss": 0.6996874809265137, | |
| "eval_rewards/accuracies": 0.0357142873108387, | |
| "eval_rewards/chosen": -0.78125, | |
| "eval_rewards/margins": 0.01226806640625, | |
| "eval_rewards/rejected": -0.79296875, | |
| "eval_runtime": 12.161, | |
| "eval_samples_per_second": 16.446, | |
| "eval_steps_per_second": 0.576, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 7.202635825304889, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -34.75, | |
| "logps/rejected": -42.5, | |
| "loss": 0.6235, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -0.89453125, | |
| "rewards/margins": 0.74609375, | |
| "rewards/rejected": -1.640625, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 1.275580126538154, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -31.125, | |
| "logps/rejected": -41.75, | |
| "loss": 0.6151, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -0.455078125, | |
| "rewards/margins": 0.8984375, | |
| "rewards/rejected": -1.359375, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 0.019832018851623313, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -1.3125, | |
| "logits/rejected": -1.296875, | |
| "logps/chosen": -31.375, | |
| "logps/rejected": -37.0, | |
| "loss": 0.6486, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -0.54296875, | |
| "rewards/margins": 0.5546875, | |
| "rewards/rejected": -1.09375, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 0.03839229395328657, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -32.75, | |
| "logps/rejected": -42.0, | |
| "loss": 0.6342, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -0.458984375, | |
| "rewards/margins": 1.015625, | |
| "rewards/rejected": -1.4765625, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 0.06878809997710522, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.5546875, | |
| "logps/chosen": -38.0, | |
| "logps/rejected": -47.75, | |
| "loss": 0.6545, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -1.09375, | |
| "rewards/margins": 0.77734375, | |
| "rewards/rejected": -1.875, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 1.1695819275941122, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -1.296875, | |
| "logits/rejected": -1.3046875, | |
| "logps/chosen": -42.0, | |
| "logps/rejected": -50.0, | |
| "loss": 0.6375, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -1.6015625, | |
| "rewards/margins": 0.765625, | |
| "rewards/rejected": -2.359375, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 21.891055398219233, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.3828125, | |
| "logps/chosen": -47.0, | |
| "logps/rejected": -59.5, | |
| "loss": 0.6187, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": -1.9453125, | |
| "rewards/margins": 1.2890625, | |
| "rewards/rejected": -3.234375, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 1.1116911096255169, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -42.25, | |
| "logps/rejected": -49.25, | |
| "loss": 0.6397, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -1.6015625, | |
| "rewards/margins": 0.72265625, | |
| "rewards/rejected": -2.328125, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 0.08927777959520299, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.578125, | |
| "logps/chosen": -53.0, | |
| "logps/rejected": -62.5, | |
| "loss": 0.6009, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -2.53125, | |
| "rewards/margins": 1.0546875, | |
| "rewards/rejected": -3.59375, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 0.24547463712332385, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -57.25, | |
| "logps/rejected": -60.0, | |
| "loss": 0.6563, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": -3.125, | |
| "rewards/margins": 0.279296875, | |
| "rewards/rejected": -3.40625, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 0.3724631921085405, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -56.5, | |
| "logps/rejected": -67.5, | |
| "loss": 0.6086, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -3.078125, | |
| "rewards/margins": 0.96875, | |
| "rewards/rejected": -4.03125, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 3.7873334439757507, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -51.5, | |
| "logps/rejected": -58.25, | |
| "loss": 0.6094, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -2.609375, | |
| "rewards/margins": 0.59375, | |
| "rewards/rejected": -3.21875, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 0.15146281496723069, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -52.75, | |
| "logps/rejected": -59.5, | |
| "loss": 0.6366, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -2.734375, | |
| "rewards/margins": 0.62109375, | |
| "rewards/rejected": -3.34375, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 0.05373739227024865, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -52.5, | |
| "logps/rejected": -61.25, | |
| "loss": 0.6361, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -2.546875, | |
| "rewards/margins": 0.9140625, | |
| "rewards/rejected": -3.453125, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 2.1108258783173444, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -1.609375, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -55.25, | |
| "logps/rejected": -68.5, | |
| "loss": 0.6239, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -2.875, | |
| "rewards/margins": 1.2421875, | |
| "rewards/rejected": -4.09375, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 16.825032338501476, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -1.3359375, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -70.5, | |
| "logps/rejected": -82.0, | |
| "loss": 0.6223, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -4.375, | |
| "rewards/margins": 1.1015625, | |
| "rewards/rejected": -5.46875, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 0.9352320114366829, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -1.1640625, | |
| "logits/rejected": -1.2578125, | |
| "logps/chosen": -61.0, | |
| "logps/rejected": -68.0, | |
| "loss": 0.6176, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -3.578125, | |
| "rewards/margins": 0.65234375, | |
| "rewards/rejected": -4.25, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 0.13140397685816607, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -1.234375, | |
| "logits/rejected": -1.203125, | |
| "logps/chosen": -57.25, | |
| "logps/rejected": -68.0, | |
| "loss": 0.6313, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -3.109375, | |
| "rewards/margins": 1.078125, | |
| "rewards/rejected": -4.1875, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 2.857678302528537, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -0.94921875, | |
| "logits/rejected": -0.98828125, | |
| "logps/chosen": -65.0, | |
| "logps/rejected": -75.0, | |
| "loss": 0.6415, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -3.9375, | |
| "rewards/margins": 0.9296875, | |
| "rewards/rejected": -4.84375, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 6.27783518417277, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -1.125, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -62.75, | |
| "logps/rejected": -72.0, | |
| "loss": 0.606, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -3.609375, | |
| "rewards/margins": 0.97265625, | |
| "rewards/rejected": -4.59375, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 46.466393549381415, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -0.98046875, | |
| "logps/chosen": -63.75, | |
| "logps/rejected": -69.5, | |
| "loss": 0.632, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -3.78125, | |
| "rewards/margins": 0.65234375, | |
| "rewards/rejected": -4.4375, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 0.06164881019856861, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -0.921875, | |
| "logits/rejected": -0.97265625, | |
| "logps/chosen": -72.0, | |
| "logps/rejected": -82.0, | |
| "loss": 0.6158, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 0.93359375, | |
| "rewards/rejected": -5.5, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 18.95472182535036, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -0.8828125, | |
| "logits/rejected": -0.76953125, | |
| "logps/chosen": -73.5, | |
| "logps/rejected": -77.0, | |
| "loss": 0.6165, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 0.43359375, | |
| "rewards/rejected": -5.1875, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 4.937928022035688, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -0.80859375, | |
| "logits/rejected": -0.80078125, | |
| "logps/chosen": -74.5, | |
| "logps/rejected": -84.5, | |
| "loss": 0.6299, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -4.875, | |
| "rewards/margins": 0.9453125, | |
| "rewards/rejected": -5.8125, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 0.052775064664981906, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -0.625, | |
| "logits/rejected": -0.63671875, | |
| "logps/chosen": -76.0, | |
| "logps/rejected": -81.0, | |
| "loss": 0.6185, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -5.09375, | |
| "rewards/margins": 0.451171875, | |
| "rewards/rejected": -5.5625, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 0.05881650618601094, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -0.76953125, | |
| "logits/rejected": -0.8671875, | |
| "logps/chosen": -69.0, | |
| "logps/rejected": -82.0, | |
| "loss": 0.6275, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -4.3125, | |
| "rewards/margins": 1.1640625, | |
| "rewards/rejected": -5.5, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 0.09080364117757252, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -0.78125, | |
| "logits/rejected": -0.7421875, | |
| "logps/chosen": -70.5, | |
| "logps/rejected": -76.5, | |
| "loss": 0.6017, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -4.4375, | |
| "rewards/margins": 0.65625, | |
| "rewards/rejected": -5.09375, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 88.56747771687814, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -0.6953125, | |
| "logits/rejected": -0.7109375, | |
| "logps/chosen": -79.5, | |
| "logps/rejected": -86.0, | |
| "loss": 0.6423, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -5.34375, | |
| "rewards/margins": 0.63671875, | |
| "rewards/rejected": -6.0, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 0.09697018729820749, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -0.9765625, | |
| "logps/chosen": -65.0, | |
| "logps/rejected": -73.5, | |
| "loss": 0.6195, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -3.859375, | |
| "rewards/margins": 0.8984375, | |
| "rewards/rejected": -4.75, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 60.05545053119793, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.1953125, | |
| "logps/chosen": -62.25, | |
| "logps/rejected": -75.5, | |
| "loss": 0.6418, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -3.640625, | |
| "rewards/margins": 1.0859375, | |
| "rewards/rejected": -4.71875, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 3.014858955881416, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.171875, | |
| "logps/chosen": -61.0, | |
| "logps/rejected": -71.0, | |
| "loss": 0.6314, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -3.5625, | |
| "rewards/margins": 0.8671875, | |
| "rewards/rejected": -4.4375, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -1.078125, | |
| "eval_logits/rejected": -1.109375, | |
| "eval_logps/chosen": -63.25, | |
| "eval_logps/rejected": -63.75, | |
| "eval_loss": 0.7364062666893005, | |
| "eval_rewards/accuracies": 0.0535714291036129, | |
| "eval_rewards/chosen": -3.8125, | |
| "eval_rewards/margins": 0.035888671875, | |
| "eval_rewards/rejected": -3.84375, | |
| "eval_runtime": 12.1644, | |
| "eval_samples_per_second": 16.441, | |
| "eval_steps_per_second": 0.575, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 0.08332814736176879, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -0.96875, | |
| "logits/rejected": -1.1171875, | |
| "logps/chosen": -74.5, | |
| "logps/rejected": -83.0, | |
| "loss": 0.6226, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -4.84375, | |
| "rewards/margins": 0.671875, | |
| "rewards/rejected": -5.53125, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 0.164985938182988, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -1.1796875, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -67.5, | |
| "logps/rejected": -82.0, | |
| "loss": 0.6191, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -4.03125, | |
| "rewards/margins": 1.484375, | |
| "rewards/rejected": -5.53125, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 0.07809236634230764, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.203125, | |
| "logps/chosen": -57.75, | |
| "logps/rejected": -70.0, | |
| "loss": 0.632, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -3.25, | |
| "rewards/margins": 0.9921875, | |
| "rewards/rejected": -4.25, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 0.06516317067237166, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -0.7265625, | |
| "logits/rejected": -0.72265625, | |
| "logps/chosen": -76.5, | |
| "logps/rejected": -81.0, | |
| "loss": 0.6186, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -5.03125, | |
| "rewards/margins": 0.5078125, | |
| "rewards/rejected": -5.5625, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 0.051643150499330034, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -0.76171875, | |
| "logits/rejected": -0.6796875, | |
| "logps/chosen": -79.5, | |
| "logps/rejected": -91.5, | |
| "loss": 0.6134, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -5.28125, | |
| "rewards/margins": 1.25, | |
| "rewards/rejected": -6.53125, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 0.08025556916050038, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -0.8671875, | |
| "logits/rejected": -0.84765625, | |
| "logps/chosen": -79.5, | |
| "logps/rejected": -90.5, | |
| "loss": 0.6419, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -5.1875, | |
| "rewards/margins": 1.125, | |
| "rewards/rejected": -6.28125, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 0.0576214364801174, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -0.7109375, | |
| "logits/rejected": -0.7265625, | |
| "logps/chosen": -73.5, | |
| "logps/rejected": -88.5, | |
| "loss": 0.6011, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 1.4375, | |
| "rewards/rejected": -6.25, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 0.07023261300538382, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -0.703125, | |
| "logits/rejected": -0.7265625, | |
| "logps/chosen": -80.5, | |
| "logps/rejected": -93.5, | |
| "loss": 0.6226, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -5.40625, | |
| "rewards/margins": 1.2109375, | |
| "rewards/rejected": -6.625, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 2.0705799687047497, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -0.53515625, | |
| "logits/rejected": -0.69921875, | |
| "logps/chosen": -82.5, | |
| "logps/rejected": -97.5, | |
| "loss": 0.6126, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -5.75, | |
| "rewards/margins": 1.2421875, | |
| "rewards/rejected": -7.0, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 0.05267573804141954, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -0.5390625, | |
| "logits/rejected": -0.416015625, | |
| "logps/chosen": -80.0, | |
| "logps/rejected": -91.5, | |
| "loss": 0.6208, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -5.3125, | |
| "rewards/margins": 1.2578125, | |
| "rewards/rejected": -6.59375, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 0.06157412465679981, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -0.333984375, | |
| "logits/rejected": -0.330078125, | |
| "logps/chosen": -93.0, | |
| "logps/rejected": -100.0, | |
| "loss": 0.6077, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -6.6875, | |
| "rewards/margins": 0.69921875, | |
| "rewards/rejected": -7.375, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 0.5344640945556689, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -0.36328125, | |
| "logits/rejected": -0.359375, | |
| "logps/chosen": -84.0, | |
| "logps/rejected": -92.0, | |
| "loss": 0.6148, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -5.875, | |
| "rewards/margins": 0.75390625, | |
| "rewards/rejected": -6.625, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 0.062456246205116854, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": -0.5625, | |
| "logits/rejected": -0.388671875, | |
| "logps/chosen": -87.0, | |
| "logps/rejected": -99.0, | |
| "loss": 0.6118, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -5.90625, | |
| "rewards/margins": 1.40625, | |
| "rewards/rejected": -7.3125, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 0.04969846393608146, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": -0.40625, | |
| "logits/rejected": -0.3515625, | |
| "logps/chosen": -86.0, | |
| "logps/rejected": -106.5, | |
| "loss": 0.6032, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -5.96875, | |
| "rewards/margins": 2.015625, | |
| "rewards/rejected": -8.0, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 0.07639092395372227, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.357421875, | |
| "logits/rejected": -0.5, | |
| "logps/chosen": -84.5, | |
| "logps/rejected": -107.5, | |
| "loss": 0.6055, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -5.875, | |
| "rewards/margins": 2.1875, | |
| "rewards/rejected": -8.0625, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 0.07447444348205871, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -0.5078125, | |
| "logits/rejected": -0.61328125, | |
| "logps/chosen": -81.0, | |
| "logps/rejected": -103.5, | |
| "loss": 0.6054, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": -5.46875, | |
| "rewards/margins": 2.09375, | |
| "rewards/rejected": -7.5625, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 0.10040200439902261, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -0.578125, | |
| "logits/rejected": -0.482421875, | |
| "logps/chosen": -74.0, | |
| "logps/rejected": -89.0, | |
| "loss": 0.6253, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 1.453125, | |
| "rewards/rejected": -6.28125, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 0.23663153627221573, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -0.482421875, | |
| "logits/rejected": -0.5, | |
| "logps/chosen": -76.5, | |
| "logps/rejected": -93.5, | |
| "loss": 0.5994, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -5.09375, | |
| "rewards/margins": 1.6796875, | |
| "rewards/rejected": -6.78125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 1.1904000627005848, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -0.61328125, | |
| "logits/rejected": -0.55078125, | |
| "logps/chosen": -81.0, | |
| "logps/rejected": -97.0, | |
| "loss": 0.6114, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -5.5, | |
| "rewards/margins": 1.5546875, | |
| "rewards/rejected": -7.0625, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 0.06181041676045537, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -0.5625, | |
| "logits/rejected": -0.58203125, | |
| "logps/chosen": -85.5, | |
| "logps/rejected": -99.5, | |
| "loss": 0.6229, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -5.78125, | |
| "rewards/margins": 1.421875, | |
| "rewards/rejected": -7.21875, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 0.06412244004564162, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -0.68359375, | |
| "logits/rejected": -0.57421875, | |
| "logps/chosen": -73.0, | |
| "logps/rejected": -86.0, | |
| "loss": 0.6204, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -4.59375, | |
| "rewards/margins": 1.375, | |
| "rewards/rejected": -5.96875, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 18.114477819614837, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -0.71875, | |
| "logits/rejected": -0.54296875, | |
| "logps/chosen": -87.5, | |
| "logps/rejected": -125.5, | |
| "loss": 0.6208, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -5.875, | |
| "rewards/margins": 3.96875, | |
| "rewards/rejected": -9.8125, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 0.0580429780200346, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -0.5078125, | |
| "logits/rejected": -0.494140625, | |
| "logps/chosen": -84.0, | |
| "logps/rejected": -91.0, | |
| "loss": 0.6254, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -5.8125, | |
| "rewards/margins": 0.6640625, | |
| "rewards/rejected": -6.5, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 0.07862204388717929, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -0.52734375, | |
| "logits/rejected": -0.5390625, | |
| "logps/chosen": -82.0, | |
| "logps/rejected": -100.0, | |
| "loss": 0.5989, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -5.59375, | |
| "rewards/margins": 1.6875, | |
| "rewards/rejected": -7.28125, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 1.3694461818497943, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -0.546875, | |
| "logits/rejected": -0.5234375, | |
| "logps/chosen": -84.5, | |
| "logps/rejected": -93.5, | |
| "loss": 0.6144, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -5.84375, | |
| "rewards/margins": 0.87890625, | |
| "rewards/rejected": -6.71875, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 0.09117311902195645, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -0.478515625, | |
| "logits/rejected": -0.44140625, | |
| "logps/chosen": -85.0, | |
| "logps/rejected": -95.5, | |
| "loss": 0.6142, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -5.9375, | |
| "rewards/margins": 1.046875, | |
| "rewards/rejected": -6.96875, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 6.5275431311324645, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -0.5234375, | |
| "logits/rejected": -0.416015625, | |
| "logps/chosen": -92.0, | |
| "logps/rejected": -106.0, | |
| "loss": 0.617, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -6.4375, | |
| "rewards/margins": 1.484375, | |
| "rewards/rejected": -7.9375, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 0.06494665152999182, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -0.349609375, | |
| "logits/rejected": -0.37890625, | |
| "logps/chosen": -84.0, | |
| "logps/rejected": -97.0, | |
| "loss": 0.5968, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -5.90625, | |
| "rewards/margins": 1.2421875, | |
| "rewards/rejected": -7.15625, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 0.10041685712641052, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -0.392578125, | |
| "logits/rejected": -0.32421875, | |
| "logps/chosen": -82.5, | |
| "logps/rejected": -97.0, | |
| "loss": 0.601, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -5.6875, | |
| "rewards/margins": 1.4140625, | |
| "rewards/rejected": -7.09375, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 0.07270087893217783, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -0.53515625, | |
| "logits/rejected": -0.474609375, | |
| "logps/chosen": -81.5, | |
| "logps/rejected": -101.5, | |
| "loss": 0.6098, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -5.4375, | |
| "rewards/margins": 2.015625, | |
| "rewards/rejected": -7.46875, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 0.05599532699973568, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": -0.376953125, | |
| "logits/rejected": -0.314453125, | |
| "logps/chosen": -88.5, | |
| "logps/rejected": -98.0, | |
| "loss": 0.6397, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -6.25, | |
| "rewards/margins": 0.94921875, | |
| "rewards/rejected": -7.1875, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -0.375, | |
| "eval_logits/rejected": -0.40625, | |
| "eval_logps/chosen": -85.0, | |
| "eval_logps/rejected": -85.5, | |
| "eval_loss": 0.7239843606948853, | |
| "eval_rewards/accuracies": 0.0357142873108387, | |
| "eval_rewards/chosen": -5.96875, | |
| "eval_rewards/margins": 0.046142578125, | |
| "eval_rewards/rejected": -6.0, | |
| "eval_runtime": 15.3339, | |
| "eval_samples_per_second": 13.043, | |
| "eval_steps_per_second": 0.457, | |
| "step": 939 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 939, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6368570703541001, | |
| "train_runtime": 4786.5679, | |
| "train_samples_per_second": 6.268, | |
| "train_steps_per_second": 0.196 | |
| } | |
| ], | |
| "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 | |
| } | |