| { | |
| "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": 4.0456725892457674, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -2.453125, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -26.25, | |
| "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": 11.977426170480863, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -26.625, | |
| "logps/rejected": -27.75, | |
| "loss": 0.6915, | |
| "rewards/accuracies": 0.0694444477558136, | |
| "rewards/chosen": -0.0001735687255859375, | |
| "rewards/margins": 0.00121307373046875, | |
| "rewards/rejected": -0.00139617919921875, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 15.706600018975474, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -26.5, | |
| "logps/rejected": -26.5, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": 0.0021820068359375, | |
| "rewards/margins": 0.00031280517578125, | |
| "rewards/rejected": 0.0018768310546875, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 5.751567257532746, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -26.125, | |
| "logps/rejected": -27.625, | |
| "loss": 0.6908, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.01904296875, | |
| "rewards/margins": 0.00421142578125, | |
| "rewards/rejected": 0.01483154296875, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 6.259469660517725, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -26.375, | |
| "logps/rejected": -25.625, | |
| "loss": 0.6894, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": 0.0244140625, | |
| "rewards/margins": 0.00016498565673828125, | |
| "rewards/rejected": 0.0242919921875, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 8.325294477210107, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -24.0, | |
| "logps/rejected": -24.25, | |
| "loss": 0.6879, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.1396484375, | |
| "rewards/margins": -0.0034332275390625, | |
| "rewards/rejected": 0.1435546875, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 1.35085346612318, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.21875, | |
| "logps/chosen": -24.5, | |
| "logps/rejected": -24.625, | |
| "loss": 0.6833, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": 0.150390625, | |
| "rewards/margins": 0.0286865234375, | |
| "rewards/rejected": 0.1220703125, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 9.495808902778645, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -27.0, | |
| "logps/rejected": -28.75, | |
| "loss": 0.6819, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -0.029052734375, | |
| "rewards/margins": 0.1015625, | |
| "rewards/rejected": -0.130859375, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 4.685855835475929, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -2.125, | |
| "logits/rejected": -2.15625, | |
| "logps/chosen": -26.5, | |
| "logps/rejected": -27.625, | |
| "loss": 0.6727, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.08447265625, | |
| "rewards/margins": 0.049072265625, | |
| "rewards/rejected": -0.1337890625, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 2.970279298363062, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -2.0625, | |
| "logits/rejected": -2.015625, | |
| "logps/chosen": -26.625, | |
| "logps/rejected": -27.625, | |
| "loss": 0.6571, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.0869140625, | |
| "rewards/margins": 0.158203125, | |
| "rewards/rejected": -0.2451171875, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 5.8567219411314095, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -2.0625, | |
| "logits/rejected": -2.125, | |
| "logps/chosen": -26.5, | |
| "logps/rejected": -28.5, | |
| "loss": 0.6824, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.05712890625, | |
| "rewards/margins": 0.1357421875, | |
| "rewards/rejected": -0.193359375, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 22.580144541109522, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -2.015625, | |
| "logits/rejected": -1.953125, | |
| "logps/chosen": -27.25, | |
| "logps/rejected": -27.5, | |
| "loss": 0.695, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.06982421875, | |
| "rewards/margins": 0.0859375, | |
| "rewards/rejected": -0.1552734375, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 10.875291592784304, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -1.8828125, | |
| "logits/rejected": -1.96875, | |
| "logps/chosen": -24.375, | |
| "logps/rejected": -26.625, | |
| "loss": 0.6694, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.10302734375, | |
| "rewards/margins": 0.134765625, | |
| "rewards/rejected": -0.03076171875, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 0.009510896766321154, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -2.078125, | |
| "logits/rejected": -2.078125, | |
| "logps/chosen": -28.0, | |
| "logps/rejected": -30.125, | |
| "loss": 0.6482, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.03515625, | |
| "rewards/margins": 0.283203125, | |
| "rewards/rejected": -0.2470703125, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 0.4374302395746067, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -1.9453125, | |
| "logits/rejected": -1.9609375, | |
| "logps/chosen": -26.375, | |
| "logps/rejected": -29.0, | |
| "loss": 0.6753, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.0238037109375, | |
| "rewards/margins": 0.2119140625, | |
| "rewards/rejected": -0.1884765625, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 6.5626177245998, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -2.0, | |
| "logits/rejected": -1.9921875, | |
| "logps/chosen": -25.75, | |
| "logps/rejected": -28.75, | |
| "loss": 0.6608, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.1103515625, | |
| "rewards/margins": 0.302734375, | |
| "rewards/rejected": -0.1923828125, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 14.632731675533192, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -2.0625, | |
| "logits/rejected": -2.0, | |
| "logps/chosen": -28.375, | |
| "logps/rejected": -28.875, | |
| "loss": 0.6794, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -0.10205078125, | |
| "rewards/margins": 0.13671875, | |
| "rewards/rejected": -0.23828125, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 12.431710703671124, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -1.9140625, | |
| "logits/rejected": -1.8671875, | |
| "logps/chosen": -26.625, | |
| "logps/rejected": -26.75, | |
| "loss": 0.6877, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.08935546875, | |
| "rewards/margins": 0.044189453125, | |
| "rewards/rejected": -0.1337890625, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 15.161032148442127, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -2.078125, | |
| "logits/rejected": -2.03125, | |
| "logps/chosen": -27.125, | |
| "logps/rejected": -28.125, | |
| "loss": 0.6599, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -0.03564453125, | |
| "rewards/margins": 0.1435546875, | |
| "rewards/rejected": -0.1796875, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 5.092871029243261, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -2.0, | |
| "logits/rejected": -2.046875, | |
| "logps/chosen": -25.75, | |
| "logps/rejected": -28.75, | |
| "loss": 0.6495, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.0177001953125, | |
| "rewards/margins": 0.208984375, | |
| "rewards/rejected": -0.19140625, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 2.5607191911586518, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -2.015625, | |
| "logits/rejected": -2.046875, | |
| "logps/chosen": -27.0, | |
| "logps/rejected": -29.25, | |
| "loss": 0.652, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -0.052978515625, | |
| "rewards/margins": 0.23828125, | |
| "rewards/rejected": -0.291015625, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 12.454055979699797, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -1.921875, | |
| "logits/rejected": -1.984375, | |
| "logps/chosen": -28.0, | |
| "logps/rejected": -30.625, | |
| "loss": 0.6657, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -0.2578125, | |
| "rewards/margins": 0.21484375, | |
| "rewards/rejected": -0.470703125, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 0.31981365864850264, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -1.9375, | |
| "logits/rejected": -1.921875, | |
| "logps/chosen": -29.125, | |
| "logps/rejected": -28.75, | |
| "loss": 0.6739, | |
| "rewards/accuracies": 0.012500000186264515, | |
| "rewards/chosen": -0.330078125, | |
| "rewards/margins": -0.00775146484375, | |
| "rewards/rejected": -0.322265625, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 11.545543036405014, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -2.015625, | |
| "logits/rejected": -1.9765625, | |
| "logps/chosen": -30.875, | |
| "logps/rejected": -32.75, | |
| "loss": 0.6743, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -0.376953125, | |
| "rewards/margins": 0.2275390625, | |
| "rewards/rejected": -0.6015625, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 1.7407594902100405, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -1.8671875, | |
| "logits/rejected": -1.84375, | |
| "logps/chosen": -29.625, | |
| "logps/rejected": -31.75, | |
| "loss": 0.6687, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -0.353515625, | |
| "rewards/margins": 0.2001953125, | |
| "rewards/rejected": -0.5546875, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 8.015592621613848, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -1.859375, | |
| "logits/rejected": -1.796875, | |
| "logps/chosen": -33.0, | |
| "logps/rejected": -33.0, | |
| "loss": 0.6549, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -0.6484375, | |
| "rewards/margins": 0.08544921875, | |
| "rewards/rejected": -0.734375, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 17.274773190103073, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -2.140625, | |
| "logits/rejected": -1.9921875, | |
| "logps/chosen": -31.0, | |
| "logps/rejected": -35.25, | |
| "loss": 0.6397, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -0.337890625, | |
| "rewards/margins": 0.5078125, | |
| "rewards/rejected": -0.84765625, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 11.6515546470372, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -1.7265625, | |
| "logits/rejected": -1.8203125, | |
| "logps/chosen": -28.25, | |
| "logps/rejected": -30.875, | |
| "loss": 0.6588, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -0.306640625, | |
| "rewards/margins": 0.162109375, | |
| "rewards/rejected": -0.46875, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 0.4544310442685568, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -30.25, | |
| "logps/rejected": -32.5, | |
| "loss": 0.6663, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -0.478515625, | |
| "rewards/margins": 0.1923828125, | |
| "rewards/rejected": -0.671875, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 0.5264502443577255, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.71875, | |
| "logps/chosen": -30.75, | |
| "logps/rejected": -32.5, | |
| "loss": 0.6458, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": -0.54296875, | |
| "rewards/margins": 0.12890625, | |
| "rewards/rejected": -0.671875, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 8.549601724926253, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -1.734375, | |
| "logits/rejected": -1.875, | |
| "logps/chosen": -29.375, | |
| "logps/rejected": -35.0, | |
| "loss": 0.6513, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -0.3515625, | |
| "rewards/margins": 0.453125, | |
| "rewards/rejected": -0.8046875, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 0.516543579681758, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -1.78125, | |
| "logits/rejected": -1.84375, | |
| "logps/chosen": -31.125, | |
| "logps/rejected": -36.5, | |
| "loss": 0.6654, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -0.37890625, | |
| "rewards/margins": 0.490234375, | |
| "rewards/rejected": -0.8671875, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -1.609375, | |
| "eval_logits/rejected": -1.7109375, | |
| "eval_logps/chosen": -27.625, | |
| "eval_logps/rejected": -32.0, | |
| "eval_loss": 0.6846874952316284, | |
| "eval_rewards/accuracies": 0.1071428582072258, | |
| "eval_rewards/chosen": -0.279296875, | |
| "eval_rewards/margins": 0.35546875, | |
| "eval_rewards/rejected": -0.63671875, | |
| "eval_runtime": 12.1377, | |
| "eval_samples_per_second": 16.478, | |
| "eval_steps_per_second": 0.577, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 2.3864188520261362, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -1.828125, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -33.75, | |
| "logps/rejected": -37.5, | |
| "loss": 0.6294, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -0.68359375, | |
| "rewards/margins": 0.421875, | |
| "rewards/rejected": -1.109375, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 3.965496848264155, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -1.7421875, | |
| "logits/rejected": -1.765625, | |
| "logps/chosen": -37.75, | |
| "logps/rejected": -45.0, | |
| "loss": 0.6225, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -1.1171875, | |
| "rewards/margins": 0.64453125, | |
| "rewards/rejected": -1.7578125, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 0.02899720766182005, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -38.0, | |
| "logps/rejected": -43.75, | |
| "loss": 0.6351, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -1.21875, | |
| "rewards/margins": 0.53515625, | |
| "rewards/rejected": -1.7578125, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 0.026441426744005847, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.7734375, | |
| "logps/chosen": -40.75, | |
| "logps/rejected": -49.25, | |
| "loss": 0.6348, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -1.4921875, | |
| "rewards/margins": 0.75, | |
| "rewards/rejected": -2.25, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 0.03574410260293087, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -1.7890625, | |
| "logits/rejected": -1.7890625, | |
| "logps/chosen": -46.0, | |
| "logps/rejected": -53.5, | |
| "loss": 0.6422, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -1.9140625, | |
| "rewards/margins": 0.61328125, | |
| "rewards/rejected": -2.53125, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 0.1835442118475042, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.546875, | |
| "logps/chosen": -50.0, | |
| "logps/rejected": -57.0, | |
| "loss": 0.6459, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -2.421875, | |
| "rewards/margins": 0.66015625, | |
| "rewards/rejected": -3.078125, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 4.312166836710742, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -53.0, | |
| "logps/rejected": -64.5, | |
| "loss": 0.6067, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": -2.65625, | |
| "rewards/margins": 1.0546875, | |
| "rewards/rejected": -3.703125, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 0.5769082265771389, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -1.5234375, | |
| "logits/rejected": -1.5703125, | |
| "logps/chosen": -50.75, | |
| "logps/rejected": -62.25, | |
| "loss": 0.632, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -2.484375, | |
| "rewards/margins": 1.09375, | |
| "rewards/rejected": -3.5625, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 1.0452071203702569, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -1.5234375, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -54.5, | |
| "logps/rejected": -63.0, | |
| "loss": 0.6097, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -2.703125, | |
| "rewards/margins": 0.9609375, | |
| "rewards/rejected": -3.671875, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 0.04970518243029372, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -1.265625, | |
| "logits/rejected": -1.2578125, | |
| "logps/chosen": -59.75, | |
| "logps/rejected": -61.0, | |
| "loss": 0.6496, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -3.40625, | |
| "rewards/margins": 0.1259765625, | |
| "rewards/rejected": -3.53125, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 4.264767440681185, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -1.2890625, | |
| "logits/rejected": -1.34375, | |
| "logps/chosen": -66.0, | |
| "logps/rejected": -74.0, | |
| "loss": 0.6317, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -3.9375, | |
| "rewards/margins": 0.73828125, | |
| "rewards/rejected": -4.6875, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 2.0262143632063334, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -1.203125, | |
| "logits/rejected": -1.34375, | |
| "logps/chosen": -60.5, | |
| "logps/rejected": -69.5, | |
| "loss": 0.6147, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -3.53125, | |
| "rewards/margins": 0.7421875, | |
| "rewards/rejected": -4.28125, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 26.831676259117806, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -1.2734375, | |
| "logits/rejected": -1.2734375, | |
| "logps/chosen": -65.0, | |
| "logps/rejected": -68.0, | |
| "loss": 0.6294, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -3.9375, | |
| "rewards/margins": 0.3046875, | |
| "rewards/rejected": -4.25, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 0.05025089218726005, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.546875, | |
| "logps/chosen": -63.25, | |
| "logps/rejected": -75.0, | |
| "loss": 0.6134, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -3.59375, | |
| "rewards/margins": 1.0625, | |
| "rewards/rejected": -4.65625, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 13.134253957976723, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -61.5, | |
| "logps/rejected": -78.5, | |
| "loss": 0.6079, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": -3.4375, | |
| "rewards/margins": 1.65625, | |
| "rewards/rejected": -5.09375, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 3.779870665578967, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -63.75, | |
| "logps/rejected": -78.0, | |
| "loss": 0.6183, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": -3.640625, | |
| "rewards/margins": 1.453125, | |
| "rewards/rejected": -5.09375, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 6.992742883163218, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.6015625, | |
| "logps/chosen": -50.5, | |
| "logps/rejected": -58.5, | |
| "loss": 0.619, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -2.53125, | |
| "rewards/margins": 0.75, | |
| "rewards/rejected": -3.296875, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 0.3993462915994891, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -52.75, | |
| "logps/rejected": -62.0, | |
| "loss": 0.6489, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -2.703125, | |
| "rewards/margins": 0.85546875, | |
| "rewards/rejected": -3.5625, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 39.70496334420096, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -1.3203125, | |
| "logits/rejected": -1.3125, | |
| "logps/chosen": -55.75, | |
| "logps/rejected": -58.75, | |
| "loss": 0.6345, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -3.0, | |
| "rewards/margins": 0.27734375, | |
| "rewards/rejected": -3.28125, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 4.389513296964734, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -1.34375, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -57.0, | |
| "logps/rejected": -68.0, | |
| "loss": 0.609, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -3.078125, | |
| "rewards/margins": 0.9140625, | |
| "rewards/rejected": -3.984375, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 45.74667388120851, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -1.2890625, | |
| "logits/rejected": -1.265625, | |
| "logps/chosen": -52.75, | |
| "logps/rejected": -60.5, | |
| "loss": 0.644, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -2.71875, | |
| "rewards/margins": 0.76953125, | |
| "rewards/rejected": -3.484375, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 0.11808620256612264, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -1.203125, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -66.0, | |
| "logps/rejected": -74.0, | |
| "loss": 0.6544, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -3.84375, | |
| "rewards/margins": 0.921875, | |
| "rewards/rejected": -4.75, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 43.20368695140686, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -1.15625, | |
| "logits/rejected": -0.99609375, | |
| "logps/chosen": -71.5, | |
| "logps/rejected": -76.5, | |
| "loss": 0.619, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -4.4375, | |
| "rewards/margins": 0.65625, | |
| "rewards/rejected": -5.09375, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 0.1237047158452992, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -68.0, | |
| "logps/rejected": -77.0, | |
| "loss": 0.6496, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -4.1875, | |
| "rewards/margins": 0.83984375, | |
| "rewards/rejected": -5.03125, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 2.362749678959825, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.125, | |
| "logps/chosen": -68.5, | |
| "logps/rejected": -74.0, | |
| "loss": 0.6285, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -4.3125, | |
| "rewards/margins": 0.51171875, | |
| "rewards/rejected": -4.8125, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 33.404592108022655, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -67.0, | |
| "logps/rejected": -70.5, | |
| "loss": 0.6661, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 0.353515625, | |
| "rewards/rejected": -4.46875, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 0.16477116623157045, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -1.125, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -74.5, | |
| "logps/rejected": -79.5, | |
| "loss": 0.6354, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -4.875, | |
| "rewards/margins": 0.55078125, | |
| "rewards/rejected": -5.40625, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 0.44059169502301365, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -1.140625, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -74.5, | |
| "logps/rejected": -79.0, | |
| "loss": 0.6175, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -4.875, | |
| "rewards/margins": 0.5078125, | |
| "rewards/rejected": -5.375, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 1.82923747322661, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -1.015625, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -78.0, | |
| "logps/rejected": -85.0, | |
| "loss": 0.6128, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -5.21875, | |
| "rewards/margins": 0.7265625, | |
| "rewards/rejected": -5.9375, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 0.12048853170410125, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -1.25, | |
| "logits/rejected": -1.21875, | |
| "logps/chosen": -78.5, | |
| "logps/rejected": -83.5, | |
| "loss": 0.633, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -5.0625, | |
| "rewards/margins": 0.64453125, | |
| "rewards/rejected": -5.71875, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 0.34148459203595716, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -1.125, | |
| "logits/rejected": -1.2265625, | |
| "logps/chosen": -73.5, | |
| "logps/rejected": -83.0, | |
| "loss": 0.6205, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 0.87109375, | |
| "rewards/rejected": -5.65625, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -1.171875, | |
| "eval_logits/rejected": -1.265625, | |
| "eval_logps/chosen": -73.0, | |
| "eval_logps/rejected": -80.0, | |
| "eval_loss": 0.7091405987739563, | |
| "eval_rewards/accuracies": 0.1071428582072258, | |
| "eval_rewards/chosen": -4.78125, | |
| "eval_rewards/margins": 0.62890625, | |
| "eval_rewards/rejected": -5.40625, | |
| "eval_runtime": 12.0597, | |
| "eval_samples_per_second": 16.584, | |
| "eval_steps_per_second": 0.58, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 0.10140143638014018, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -1.1796875, | |
| "logits/rejected": -1.2578125, | |
| "logps/chosen": -79.0, | |
| "logps/rejected": -85.0, | |
| "loss": 0.6348, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -5.34375, | |
| "rewards/margins": 0.546875, | |
| "rewards/rejected": -5.875, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 0.8245291368159866, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.296875, | |
| "logps/chosen": -77.0, | |
| "logps/rejected": -88.5, | |
| "loss": 0.611, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -4.9375, | |
| "rewards/margins": 1.25, | |
| "rewards/rejected": -6.15625, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 0.10712888479455715, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": -1.171875, | |
| "logits/rejected": -1.15625, | |
| "logps/chosen": -72.5, | |
| "logps/rejected": -78.0, | |
| "loss": 0.6271, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 0.55859375, | |
| "rewards/rejected": -5.3125, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 0.0657788228327599, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -85.0, | |
| "logps/rejected": -91.5, | |
| "loss": 0.6184, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -5.90625, | |
| "rewards/margins": 0.6484375, | |
| "rewards/rejected": -6.5625, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 0.04955247697639958, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -1.125, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -80.5, | |
| "logps/rejected": -91.0, | |
| "loss": 0.6175, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -5.4375, | |
| "rewards/margins": 1.0625, | |
| "rewards/rejected": -6.5, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 12.412215026265685, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -1.21875, | |
| "logits/rejected": -1.2890625, | |
| "logps/chosen": -81.5, | |
| "logps/rejected": -92.0, | |
| "loss": 0.6305, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": -5.4375, | |
| "rewards/margins": 0.99609375, | |
| "rewards/rejected": -6.4375, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 0.061456953130695664, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.0078125, | |
| "logps/chosen": -76.0, | |
| "logps/rejected": -90.0, | |
| "loss": 0.6098, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -5.03125, | |
| "rewards/margins": 1.4140625, | |
| "rewards/rejected": -6.4375, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 0.14188321138308624, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -79.5, | |
| "logps/rejected": -91.0, | |
| "loss": 0.6291, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -5.25, | |
| "rewards/margins": 1.2109375, | |
| "rewards/rejected": -6.46875, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 52.56243882325492, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -0.8828125, | |
| "logits/rejected": -0.9609375, | |
| "logps/chosen": -81.5, | |
| "logps/rejected": -93.5, | |
| "loss": 0.6206, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -5.625, | |
| "rewards/margins": 1.0859375, | |
| "rewards/rejected": -6.6875, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 0.06578735818761013, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -0.984375, | |
| "logits/rejected": -0.98046875, | |
| "logps/chosen": -78.5, | |
| "logps/rejected": -87.0, | |
| "loss": 0.635, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -5.25, | |
| "rewards/margins": 0.83984375, | |
| "rewards/rejected": -6.09375, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 0.6780286838032901, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -1.0, | |
| "logps/chosen": -77.5, | |
| "logps/rejected": -84.5, | |
| "loss": 0.6158, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -5.125, | |
| "rewards/margins": 0.72265625, | |
| "rewards/rejected": -5.84375, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 0.06689723594677775, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -0.9453125, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -73.0, | |
| "logps/rejected": -82.0, | |
| "loss": 0.5954, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 0.83203125, | |
| "rewards/rejected": -5.5625, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 0.11174742388085977, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": -1.1484375, | |
| "logits/rejected": -1.046875, | |
| "logps/chosen": -72.0, | |
| "logps/rejected": -81.5, | |
| "loss": 0.6171, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 1.0078125, | |
| "rewards/rejected": -5.53125, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 0.08477487268874864, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": -1.0859375, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -74.5, | |
| "logps/rejected": -95.0, | |
| "loss": 0.5923, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 1.953125, | |
| "rewards/rejected": -6.78125, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 0.05239889114118326, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.9765625, | |
| "logits/rejected": -1.3359375, | |
| "logps/chosen": -77.0, | |
| "logps/rejected": -96.5, | |
| "loss": 0.6119, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -5.125, | |
| "rewards/margins": 1.703125, | |
| "rewards/rejected": -6.84375, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 0.16332563379342213, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.1875, | |
| "logps/chosen": -73.0, | |
| "logps/rejected": -87.0, | |
| "loss": 0.614, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -4.71875, | |
| "rewards/margins": 1.2734375, | |
| "rewards/rejected": -6.0, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 0.342943270640063, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -0.97265625, | |
| "logps/chosen": -70.0, | |
| "logps/rejected": -81.5, | |
| "loss": 0.628, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -4.375, | |
| "rewards/margins": 1.2109375, | |
| "rewards/rejected": -5.59375, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 0.055314376786577586, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -75.0, | |
| "logps/rejected": -89.0, | |
| "loss": 0.6338, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -4.90625, | |
| "rewards/margins": 1.2734375, | |
| "rewards/rejected": -6.1875, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 0.04877004086710131, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -0.96484375, | |
| "logps/chosen": -69.5, | |
| "logps/rejected": -82.0, | |
| "loss": 0.6206, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 1.3359375, | |
| "rewards/rejected": -5.625, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 0.05125558393087492, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.1640625, | |
| "logps/chosen": -77.5, | |
| "logps/rejected": -98.0, | |
| "loss": 0.612, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -5.15625, | |
| "rewards/margins": 1.8828125, | |
| "rewards/rejected": -7.03125, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 0.06276234080739677, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.1171875, | |
| "logps/chosen": -69.5, | |
| "logps/rejected": -89.0, | |
| "loss": 0.6209, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -4.375, | |
| "rewards/margins": 1.78125, | |
| "rewards/rejected": -6.15625, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 0.23531553131275804, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -1.0234375, | |
| "logps/chosen": -74.5, | |
| "logps/rejected": -89.5, | |
| "loss": 0.6302, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -4.78125, | |
| "rewards/margins": 1.484375, | |
| "rewards/rejected": -6.25, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 0.4688889391540396, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -71.0, | |
| "logps/rejected": -84.5, | |
| "loss": 0.6162, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 1.3046875, | |
| "rewards/rejected": -5.78125, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 0.16709784182072895, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -0.90625, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -81.0, | |
| "logps/rejected": -92.0, | |
| "loss": 0.6119, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -5.53125, | |
| "rewards/margins": 1.0078125, | |
| "rewards/rejected": -6.5625, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 0.3793220360260412, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -1.0078125, | |
| "logits/rejected": -0.99609375, | |
| "logps/chosen": -82.5, | |
| "logps/rejected": -89.0, | |
| "loss": 0.6128, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": -5.625, | |
| "rewards/margins": 0.69921875, | |
| "rewards/rejected": -6.34375, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 0.05518027666658511, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -0.9609375, | |
| "logits/rejected": -1.0078125, | |
| "logps/chosen": -80.0, | |
| "logps/rejected": -90.0, | |
| "loss": 0.6333, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -5.46875, | |
| "rewards/margins": 0.9375, | |
| "rewards/rejected": -6.40625, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 0.10639054633224729, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -1.1484375, | |
| "logits/rejected": -1.015625, | |
| "logps/chosen": -82.0, | |
| "logps/rejected": -96.0, | |
| "loss": 0.6075, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -5.375, | |
| "rewards/margins": 1.5625, | |
| "rewards/rejected": -6.9375, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 0.05765927599486169, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -0.921875, | |
| "logits/rejected": -0.9609375, | |
| "logps/chosen": -78.0, | |
| "logps/rejected": -88.5, | |
| "loss": 0.6147, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": -5.28125, | |
| "rewards/margins": 1.03125, | |
| "rewards/rejected": -6.3125, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 0.06285322294345812, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -0.9453125, | |
| "logits/rejected": -0.96484375, | |
| "logps/chosen": -74.5, | |
| "logps/rejected": -87.0, | |
| "loss": 0.6118, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": -4.875, | |
| "rewards/margins": 1.2109375, | |
| "rewards/rejected": -6.09375, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 0.06222781535738933, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -1.1328125, | |
| "logits/rejected": -1.1484375, | |
| "logps/chosen": -76.5, | |
| "logps/rejected": -101.0, | |
| "loss": 0.5967, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": -5.0, | |
| "rewards/margins": 2.390625, | |
| "rewards/rejected": -7.375, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 0.17237229041303392, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": -1.171875, | |
| "logits/rejected": -0.9453125, | |
| "logps/chosen": -82.0, | |
| "logps/rejected": -89.0, | |
| "loss": 0.6408, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": -5.5, | |
| "rewards/margins": 0.828125, | |
| "rewards/rejected": -6.3125, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -0.921875, | |
| "eval_logits/rejected": -1.015625, | |
| "eval_logps/chosen": -78.0, | |
| "eval_logps/rejected": -85.5, | |
| "eval_loss": 0.7425781488418579, | |
| "eval_rewards/accuracies": 0.0892857164144516, | |
| "eval_rewards/chosen": -5.3125, | |
| "eval_rewards/margins": 0.6796875, | |
| "eval_rewards/rejected": -6.0, | |
| "eval_runtime": 15.1989, | |
| "eval_samples_per_second": 13.159, | |
| "eval_steps_per_second": 0.461, | |
| "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 | |
| } | |