| { | |
| "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": 280.8462782751874, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -3.109375, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -396.0, | |
| "logps/rejected": -104.5, | |
| "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": 193.61448145669556, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -3.109375, | |
| "logits/rejected": -3.234375, | |
| "logps/chosen": -362.0, | |
| "logps/rejected": -123.5, | |
| "loss": 0.6821, | |
| "rewards/accuracies": 0.4166666567325592, | |
| "rewards/chosen": 0.0003490447998046875, | |
| "rewards/margins": 0.035888671875, | |
| "rewards/rejected": -0.035400390625, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 191.21117575266513, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -2.96875, | |
| "logits/rejected": -3.0625, | |
| "logps/chosen": -418.0, | |
| "logps/rejected": -126.0, | |
| "loss": 0.5132, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.0888671875, | |
| "rewards/margins": 0.423828125, | |
| "rewards/rejected": -0.333984375, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 10.716448104729054, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -3.0625, | |
| "logits/rejected": -3.046875, | |
| "logps/chosen": -382.0, | |
| "logps/rejected": -149.0, | |
| "loss": 0.1531, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.365234375, | |
| "rewards/margins": 2.234375, | |
| "rewards/rejected": -1.875, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 2.469465641467824, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -2.9375, | |
| "logits/rejected": -3.03125, | |
| "logps/chosen": -382.0, | |
| "logps/rejected": -163.0, | |
| "loss": 0.0145, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.2578125, | |
| "rewards/margins": 5.3125, | |
| "rewards/rejected": -4.0625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 0.378024187208137, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -2.921875, | |
| "logits/rejected": -2.921875, | |
| "logps/chosen": -318.0, | |
| "logps/rejected": -187.0, | |
| "loss": 0.0073, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.1328125, | |
| "rewards/margins": 7.25, | |
| "rewards/rejected": -6.125, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 0.14287121298206507, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -2.8125, | |
| "logits/rejected": -2.84375, | |
| "logps/chosen": -338.0, | |
| "logps/rejected": -201.0, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.6796875, | |
| "rewards/margins": 9.125, | |
| "rewards/rejected": -7.46875, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 0.2437445873125328, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -2.859375, | |
| "logits/rejected": -2.78125, | |
| "logps/chosen": -316.0, | |
| "logps/rejected": -208.0, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.5390625, | |
| "rewards/margins": 10.5625, | |
| "rewards/rejected": -9.0625, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 0.3817280043216981, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -2.703125, | |
| "logits/rejected": -2.578125, | |
| "logps/chosen": -360.0, | |
| "logps/rejected": -214.0, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.0, | |
| "rewards/margins": 11.5, | |
| "rewards/rejected": -9.5, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 0.011934078569469323, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.578125, | |
| "logps/chosen": -272.0, | |
| "logps/rejected": -216.0, | |
| "loss": 0.0073, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.640625, | |
| "rewards/margins": 11.5625, | |
| "rewards/rejected": -9.9375, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 0.10330250583125518, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -2.75, | |
| "logits/rejected": -2.71875, | |
| "logps/chosen": -402.0, | |
| "logps/rejected": -227.0, | |
| "loss": 0.0037, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.8671875, | |
| "rewards/margins": 11.9375, | |
| "rewards/rejected": -10.0625, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 0.04492212094130509, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -2.71875, | |
| "logits/rejected": -2.640625, | |
| "logps/chosen": -322.0, | |
| "logps/rejected": -238.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.046875, | |
| "rewards/margins": 12.5, | |
| "rewards/rejected": -11.5, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 0.008955090901863088, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -2.515625, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -270.0, | |
| "logps/rejected": -248.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.171875, | |
| "rewards/margins": 14.375, | |
| "rewards/rejected": -13.1875, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 0.4806132060039022, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -2.578125, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -364.0, | |
| "logps/rejected": -245.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.59375, | |
| "rewards/margins": 13.6875, | |
| "rewards/rejected": -12.125, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 0.11448825443399689, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -2.640625, | |
| "logits/rejected": -2.515625, | |
| "logps/chosen": -404.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.5859375, | |
| "rewards/margins": 15.5625, | |
| "rewards/rejected": -14.0, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 0.045880492246870935, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -2.75, | |
| "logits/rejected": -2.703125, | |
| "logps/chosen": -354.0, | |
| "logps/rejected": -250.0, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.8671875, | |
| "rewards/margins": 14.5625, | |
| "rewards/rejected": -12.6875, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 0.003833084071827597, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -2.796875, | |
| "logits/rejected": -2.609375, | |
| "logps/chosen": -348.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.625, | |
| "rewards/margins": 15.5625, | |
| "rewards/rejected": -14.0, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 0.010411651571998708, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -2.8125, | |
| "logits/rejected": -2.609375, | |
| "logps/chosen": -326.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.4296875, | |
| "rewards/margins": 15.5625, | |
| "rewards/rejected": -14.125, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 0.0029715259908925555, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -2.796875, | |
| "logits/rejected": -2.578125, | |
| "logps/chosen": -354.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.84375, | |
| "rewards/margins": 16.125, | |
| "rewards/rejected": -14.3125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 0.015891116030461273, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -2.671875, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -362.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.921875, | |
| "rewards/margins": 17.5, | |
| "rewards/rejected": -14.625, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 0.0157480530740379, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -2.765625, | |
| "logits/rejected": -2.6875, | |
| "logps/chosen": -388.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0026, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.859375, | |
| "rewards/margins": 17.25, | |
| "rewards/rejected": -14.375, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 0.0006495551759343999, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -2.84375, | |
| "logits/rejected": -2.625, | |
| "logps/chosen": -348.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 16.75, | |
| "rewards/rejected": -14.4375, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 0.0055130402632981934, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -2.75, | |
| "logits/rejected": -2.625, | |
| "logps/chosen": -386.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 17.375, | |
| "rewards/rejected": -14.9375, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 0.003053645252154487, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -2.828125, | |
| "logits/rejected": -2.5625, | |
| "logps/chosen": -312.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.34375, | |
| "rewards/margins": 17.25, | |
| "rewards/rejected": -14.875, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 0.0017309994643395468, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -2.75, | |
| "logits/rejected": -2.71875, | |
| "logps/chosen": -352.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 16.75, | |
| "rewards/rejected": -14.375, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 0.006736967050209572, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -2.84375, | |
| "logits/rejected": -2.6875, | |
| "logps/chosen": -352.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.171875, | |
| "rewards/margins": 16.625, | |
| "rewards/rejected": -15.4375, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 0.22749571813014804, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -2.859375, | |
| "logits/rejected": -2.625, | |
| "logps/chosen": -346.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.03125, | |
| "rewards/margins": 17.125, | |
| "rewards/rejected": -15.125, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 0.40533640282802086, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -2.796875, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -320.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.28125, | |
| "rewards/margins": 17.625, | |
| "rewards/rejected": -15.375, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 1.6419213417764489, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -318.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.875, | |
| "rewards/margins": 17.875, | |
| "rewards/rejected": -15.9375, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 0.0021810663315091007, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -2.828125, | |
| "logits/rejected": -2.65625, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.09375, | |
| "rewards/margins": 17.375, | |
| "rewards/rejected": -15.25, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 0.002542401923963208, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -2.734375, | |
| "logits/rejected": -2.65625, | |
| "logps/chosen": -352.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.671875, | |
| "rewards/margins": 18.375, | |
| "rewards/rejected": -15.75, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 0.0017692339924974662, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -2.640625, | |
| "logits/rejected": -2.5625, | |
| "logps/chosen": -324.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.484375, | |
| "rewards/margins": 19.25, | |
| "rewards/rejected": -16.75, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -2.84375, | |
| "eval_logits/rejected": -2.65625, | |
| "eval_logps/chosen": -302.0, | |
| "eval_logps/rejected": -274.0, | |
| "eval_loss": 3.4866854548454285e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": 2.359375, | |
| "eval_rewards/margins": 17.125, | |
| "eval_rewards/rejected": -14.8125, | |
| "eval_runtime": 19.1357, | |
| "eval_samples_per_second": 10.452, | |
| "eval_steps_per_second": 0.366, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 0.005478609557178483, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -2.78125, | |
| "logits/rejected": -2.5625, | |
| "logps/chosen": -296.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.421875, | |
| "rewards/margins": 19.125, | |
| "rewards/rejected": -16.75, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 0.013737635995063696, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -2.8125, | |
| "logits/rejected": -2.65625, | |
| "logps/chosen": -324.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 18.875, | |
| "rewards/rejected": -16.25, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 0.00019349035033339429, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -2.765625, | |
| "logits/rejected": -2.5625, | |
| "logps/chosen": -322.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 19.5, | |
| "rewards/rejected": -16.875, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 0.0005813199789647137, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -2.828125, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -320.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.609375, | |
| "rewards/margins": 18.375, | |
| "rewards/rejected": -15.8125, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 0.006146815019397078, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -2.78125, | |
| "logits/rejected": -2.671875, | |
| "logps/chosen": -280.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.84375, | |
| "rewards/margins": 18.625, | |
| "rewards/rejected": -16.75, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 0.001769493435224336, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -2.59375, | |
| "logits/rejected": -2.640625, | |
| "logps/chosen": -418.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.96875, | |
| "rewards/margins": 19.125, | |
| "rewards/rejected": -16.125, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 0.00289437017638318, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -2.734375, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -304.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.75, | |
| "rewards/margins": 19.625, | |
| "rewards/rejected": -16.875, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 0.002541615596439298, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -2.765625, | |
| "logits/rejected": -2.6875, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 19.125, | |
| "rewards/rejected": -16.5, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 0.0004182225706778624, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -2.6875, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -354.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.765625, | |
| "rewards/margins": 19.75, | |
| "rewards/rejected": -17.0, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 0.0058634982116960045, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -2.75, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -332.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 19.25, | |
| "rewards/rejected": -17.0, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 0.0005877666510578203, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -2.65625, | |
| "logits/rejected": -2.578125, | |
| "logps/chosen": -300.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.28125, | |
| "rewards/margins": 18.25, | |
| "rewards/rejected": -16.0, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 7.688662094360472e-05, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -2.703125, | |
| "logits/rejected": -2.578125, | |
| "logps/chosen": -336.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.953125, | |
| "rewards/margins": 19.875, | |
| "rewards/rejected": -16.875, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 0.0013188626087899836, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -2.671875, | |
| "logits/rejected": -2.5, | |
| "logps/chosen": -362.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.65625, | |
| "rewards/margins": 19.375, | |
| "rewards/rejected": -16.625, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 0.000560100668346436, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -2.734375, | |
| "logits/rejected": -2.46875, | |
| "logps/chosen": -340.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.859375, | |
| "rewards/margins": 18.625, | |
| "rewards/rejected": -15.75, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 0.0004914338037975381, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -2.71875, | |
| "logits/rejected": -2.578125, | |
| "logps/chosen": -298.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 18.5, | |
| "rewards/rejected": -16.125, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 0.0014282123670870752, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -2.671875, | |
| "logits/rejected": -2.5, | |
| "logps/chosen": -298.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.75, | |
| "rewards/margins": 20.25, | |
| "rewards/rejected": -17.5, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 0.0011868778955687227, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -2.734375, | |
| "logits/rejected": -2.578125, | |
| "logps/chosen": -370.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 19.0, | |
| "rewards/rejected": -16.375, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 0.00017059860211962754, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -2.6875, | |
| "logits/rejected": -2.5, | |
| "logps/chosen": -362.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.546875, | |
| "rewards/margins": 20.375, | |
| "rewards/rejected": -17.875, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 0.000533897211000123, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.46875, | |
| "logps/chosen": -410.0, | |
| "logps/rejected": -294.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.84375, | |
| "rewards/margins": 19.125, | |
| "rewards/rejected": -16.375, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 0.00016451813084725905, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -2.59375, | |
| "logits/rejected": -2.5625, | |
| "logps/chosen": -330.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.65625, | |
| "rewards/margins": 20.75, | |
| "rewards/rejected": -18.125, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 0.0022106764918951973, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -2.703125, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -294.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.6171875, | |
| "rewards/margins": 20.125, | |
| "rewards/rejected": -18.5, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 0.004548328476876588, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -2.65625, | |
| "logits/rejected": -2.5625, | |
| "logps/chosen": -300.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.15625, | |
| "rewards/margins": 19.25, | |
| "rewards/rejected": -17.0, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 0.0012387511087695242, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -2.71875, | |
| "logits/rejected": -2.46875, | |
| "logps/chosen": -380.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.609375, | |
| "rewards/margins": 20.625, | |
| "rewards/rejected": -18.0, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 0.00011744334330627099, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -2.703125, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -328.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.6875, | |
| "rewards/margins": 20.0, | |
| "rewards/rejected": -17.375, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 0.0003087811407451836, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -2.765625, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -390.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 20.5, | |
| "rewards/rejected": -18.0, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 0.0004052547360073496, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -2.703125, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -314.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.265625, | |
| "rewards/margins": 20.0, | |
| "rewards/rejected": -17.75, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 0.0014702518762848597, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -2.6875, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -342.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.53125, | |
| "rewards/margins": 19.875, | |
| "rewards/rejected": -17.25, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 0.0003324653250062881, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -2.71875, | |
| "logits/rejected": -2.515625, | |
| "logps/chosen": -390.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.625, | |
| "rewards/margins": 19.875, | |
| "rewards/rejected": -17.25, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 0.0010880215884730468, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -2.546875, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -328.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.484375, | |
| "rewards/margins": 20.375, | |
| "rewards/rejected": -18.0, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 0.3586647424788816, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -2.5625, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -332.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 20.5, | |
| "rewards/rejected": -18.25, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 0.0008306417375702101, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -2.8125, | |
| "logits/rejected": -2.5, | |
| "logps/chosen": -342.0, | |
| "logps/rejected": -310.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.7734375, | |
| "rewards/margins": 20.625, | |
| "rewards/rejected": -18.75, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -2.75, | |
| "eval_logits/rejected": -2.53125, | |
| "eval_logps/chosen": -304.0, | |
| "eval_logps/rejected": -292.0, | |
| "eval_loss": 2.103640690620523e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": 2.109375, | |
| "eval_rewards/margins": 18.875, | |
| "eval_rewards/rejected": -16.75, | |
| "eval_runtime": 19.4457, | |
| "eval_samples_per_second": 10.285, | |
| "eval_steps_per_second": 0.36, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 0.00010554475208693832, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -2.734375, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -350.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 21.0, | |
| "rewards/rejected": -18.625, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 0.0013682912627571986, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -2.640625, | |
| "logits/rejected": -2.453125, | |
| "logps/chosen": -362.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 20.5, | |
| "rewards/rejected": -17.875, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 0.0016285347343782553, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": -2.53125, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -336.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.984375, | |
| "rewards/margins": 21.875, | |
| "rewards/rejected": -18.875, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 0.00037335332803088234, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -2.59375, | |
| "logits/rejected": -2.515625, | |
| "logps/chosen": -390.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.1875, | |
| "rewards/margins": 19.75, | |
| "rewards/rejected": -17.625, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 0.005423332011391937, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -2.671875, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -308.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.75, | |
| "rewards/margins": 19.5, | |
| "rewards/rejected": -17.875, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 0.0029105677835870505, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -2.75, | |
| "logits/rejected": -2.515625, | |
| "logps/chosen": -310.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.7421875, | |
| "rewards/margins": 20.375, | |
| "rewards/rejected": -18.625, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 0.020037162303156882, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -2.609375, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -296.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.703125, | |
| "rewards/margins": 20.75, | |
| "rewards/rejected": -19.125, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 0.0002419171006255769, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -2.609375, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -298.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.9609375, | |
| "rewards/margins": 20.875, | |
| "rewards/rejected": -19.0, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 0.0004291807828884535, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -2.671875, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -348.0, | |
| "logps/rejected": -310.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 20.625, | |
| "rewards/rejected": -18.375, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 0.00034178613830759455, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -2.5625, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -316.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.796875, | |
| "rewards/margins": 21.0, | |
| "rewards/rejected": -18.125, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 0.0007654424868007646, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -2.65625, | |
| "logits/rejected": -2.578125, | |
| "logps/chosen": -396.0, | |
| "logps/rejected": -322.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.5, | |
| "rewards/margins": 22.25, | |
| "rewards/rejected": -19.75, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 0.0012292244990740038, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -2.71875, | |
| "logits/rejected": -2.5, | |
| "logps/chosen": -346.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 20.75, | |
| "rewards/rejected": -18.25, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 6.35629759218304e-05, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -290.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.125, | |
| "rewards/margins": 21.5, | |
| "rewards/rejected": -19.375, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 0.0012458510016512745, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": -2.5, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.609375, | |
| "rewards/margins": 21.0, | |
| "rewards/rejected": -18.375, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 9.565857408200781e-05, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -2.640625, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -352.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.703125, | |
| "rewards/margins": 20.625, | |
| "rewards/rejected": -18.0, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 0.0003579063195204582, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -2.640625, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -366.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 20.875, | |
| "rewards/rejected": -18.625, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 0.0006374161403948482, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -2.609375, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.53125, | |
| "rewards/margins": 20.375, | |
| "rewards/rejected": -17.875, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 0.00012985600570182622, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -2.578125, | |
| "logits/rejected": -2.21875, | |
| "logps/chosen": -328.0, | |
| "logps/rejected": -310.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.796875, | |
| "rewards/margins": 21.0, | |
| "rewards/rejected": -19.25, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 0.0012142032429552098, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -2.53125, | |
| "logits/rejected": -2.5, | |
| "logps/chosen": -352.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 21.125, | |
| "rewards/rejected": -18.875, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 0.00032345170115246773, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -2.640625, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -350.0, | |
| "logps/rejected": -318.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.46875, | |
| "rewards/margins": 21.5, | |
| "rewards/rejected": -19.125, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 0.00016853438104820527, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -2.65625, | |
| "logits/rejected": -2.46875, | |
| "logps/chosen": -312.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.03125, | |
| "rewards/margins": 20.25, | |
| "rewards/rejected": -18.25, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 0.0010929596646721339, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -280.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.34375, | |
| "rewards/margins": 21.5, | |
| "rewards/rejected": -19.125, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 0.0007238125685981078, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -2.671875, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -322.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.984375, | |
| "rewards/margins": 21.375, | |
| "rewards/rejected": -19.375, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 0.000992936673131661, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -2.5, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -366.0, | |
| "logps/rejected": -322.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 22.875, | |
| "rewards/rejected": -20.25, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 0.00017936737590329918, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -2.65625, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -364.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.71875, | |
| "rewards/margins": 22.375, | |
| "rewards/rejected": -19.625, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 0.004580709768232918, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -2.609375, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -258.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.1875, | |
| "rewards/margins": 21.0, | |
| "rewards/rejected": -18.75, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 9.549896107098972e-05, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -2.5625, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -396.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.5, | |
| "rewards/margins": 22.125, | |
| "rewards/rejected": -19.625, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 0.008331840196421716, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -2.71875, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -276.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 21.5, | |
| "rewards/rejected": -19.25, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 0.00011991166733590545, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -2.53125, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -336.0, | |
| "logps/rejected": -310.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.28125, | |
| "rewards/margins": 21.375, | |
| "rewards/rejected": -19.125, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 0.00024048757621631005, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -2.6875, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -348.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.28125, | |
| "rewards/margins": 21.875, | |
| "rewards/rejected": -19.625, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 0.00012022583726080308, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": -2.765625, | |
| "logits/rejected": -2.5, | |
| "logps/chosen": -324.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 20.375, | |
| "rewards/rejected": -18.125, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -2.75, | |
| "eval_logits/rejected": -2.515625, | |
| "eval_logps/chosen": -304.0, | |
| "eval_logps/rejected": -298.0, | |
| "eval_loss": 2.0297271476010792e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": 2.109375, | |
| "eval_rewards/margins": 19.5, | |
| "eval_rewards/rejected": -17.375, | |
| "eval_runtime": 19.8217, | |
| "eval_samples_per_second": 10.09, | |
| "eval_steps_per_second": 0.353, | |
| "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 | |
| } | |