| { | |
| "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": 54.30884971483562, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.59375, | |
| "logps/chosen": -116.0, | |
| "logps/rejected": -95.0, | |
| "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": 55.79365090335572, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -146.0, | |
| "logps/rejected": -110.0, | |
| "loss": 0.6897, | |
| "rewards/accuracies": 0.3888888955116272, | |
| "rewards/chosen": 0.00138092041015625, | |
| "rewards/margins": 0.0069580078125, | |
| "rewards/rejected": -0.00555419921875, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 43.39563938211016, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -172.0, | |
| "logps/rejected": -112.0, | |
| "loss": 0.6219, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.07373046875, | |
| "rewards/margins": 0.1513671875, | |
| "rewards/rejected": -0.0771484375, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 25.708981026181373, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -162.0, | |
| "logps/rejected": -118.0, | |
| "loss": 0.4138, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.306640625, | |
| "rewards/margins": 0.72265625, | |
| "rewards/rejected": -0.416015625, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 15.297617448707284, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -127.0, | |
| "logps/rejected": -129.0, | |
| "loss": 0.2096, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 0.59765625, | |
| "rewards/margins": 1.75, | |
| "rewards/rejected": -1.1484375, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 10.111799213403708, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -135.0, | |
| "loss": 0.09, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 0.91015625, | |
| "rewards/margins": 3.578125, | |
| "rewards/rejected": -2.65625, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 5.011861710134038, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -164.0, | |
| "logps/rejected": -152.0, | |
| "loss": 0.0777, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 0.98046875, | |
| "rewards/margins": 4.28125, | |
| "rewards/rejected": -3.296875, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 2.038654121141292, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -139.0, | |
| "logps/rejected": -150.0, | |
| "loss": 0.058, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 1.03125, | |
| "rewards/margins": 5.375, | |
| "rewards/rejected": -4.3125, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 0.8151022345347168, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -125.0, | |
| "logps/rejected": -169.0, | |
| "loss": 0.1014, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.5859375, | |
| "rewards/margins": 7.125, | |
| "rewards/rejected": -5.5625, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 14.044895624062985, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -131.0, | |
| "logps/rejected": -171.0, | |
| "loss": 0.1479, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.59375, | |
| "rewards/margins": 7.71875, | |
| "rewards/rejected": -6.125, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 8.035172138916089, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -148.0, | |
| "logps/rejected": -169.0, | |
| "loss": 0.0732, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.171875, | |
| "rewards/margins": 6.75, | |
| "rewards/rejected": -5.5625, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 6.405395884454611, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -138.0, | |
| "logps/rejected": -172.0, | |
| "loss": 0.0814, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.15625, | |
| "rewards/margins": 7.65625, | |
| "rewards/rejected": -6.5, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 4.8836636965287274, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -156.0, | |
| "logps/rejected": -176.0, | |
| "loss": 0.0543, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.1171875, | |
| "rewards/margins": 7.5625, | |
| "rewards/rejected": -6.4375, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 0.06385368374522883, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -134.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.0589, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.453125, | |
| "rewards/margins": 7.84375, | |
| "rewards/rejected": -6.40625, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 0.14249262103404972, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.71875, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -182.0, | |
| "loss": 0.1228, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.53125, | |
| "rewards/margins": 8.1875, | |
| "rewards/rejected": -6.6875, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 1.125258345768065, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -177.0, | |
| "loss": 0.09, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.5703125, | |
| "rewards/margins": 7.9375, | |
| "rewards/rejected": -6.375, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 12.78259443972307, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -164.0, | |
| "logps/rejected": -180.0, | |
| "loss": 0.0555, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.65625, | |
| "rewards/margins": 8.5, | |
| "rewards/rejected": -6.84375, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 24.274365993969184, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -138.0, | |
| "logps/rejected": -177.0, | |
| "loss": 0.0895, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.59375, | |
| "rewards/margins": 8.125, | |
| "rewards/rejected": -6.5, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 0.20380161572614333, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -178.0, | |
| "loss": 0.0873, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.9765625, | |
| "rewards/margins": 7.8125, | |
| "rewards/rejected": -5.8125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 16.482316832003793, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -158.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.0752, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 1.734375, | |
| "rewards/margins": 7.8125, | |
| "rewards/rejected": -6.0625, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 0.19644543626350036, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -133.0, | |
| "logps/rejected": -177.0, | |
| "loss": 0.0724, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.671875, | |
| "rewards/margins": 7.4375, | |
| "rewards/rejected": -5.78125, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 13.95704163752683, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.703125, | |
| "logps/chosen": -121.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.1151, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.6875, | |
| "rewards/margins": 7.59375, | |
| "rewards/rejected": -5.90625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 0.9512671990629835, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -157.0, | |
| "logps/rejected": -163.0, | |
| "loss": 0.0555, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.8359375, | |
| "rewards/margins": 6.5625, | |
| "rewards/rejected": -4.71875, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 10.125817972129635, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -1.6953125, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -161.0, | |
| "logps/rejected": -169.0, | |
| "loss": 0.0682, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.96875, | |
| "rewards/margins": 7.8125, | |
| "rewards/rejected": -5.84375, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 21.10318734954903, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -170.0, | |
| "loss": 0.0982, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.9609375, | |
| "rewards/margins": 7.8125, | |
| "rewards/rejected": -5.8125, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 0.9759357330496103, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.7265625, | |
| "logps/chosen": -165.0, | |
| "logps/rejected": -171.0, | |
| "loss": 0.0785, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 7.6875, | |
| "rewards/rejected": -5.6875, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 7.0236513754447545, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -1.6953125, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -159.0, | |
| "logps/rejected": -160.0, | |
| "loss": 0.0495, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.8359375, | |
| "rewards/margins": 7.15625, | |
| "rewards/rejected": -5.3125, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 2.049286158195271, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -138.0, | |
| "logps/rejected": -172.0, | |
| "loss": 0.0782, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.578125, | |
| "rewards/margins": 7.34375, | |
| "rewards/rejected": -5.78125, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 0.2795436901035824, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -124.0, | |
| "logps/rejected": -169.0, | |
| "loss": 0.0921, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.6328125, | |
| "rewards/margins": 7.625, | |
| "rewards/rejected": -6.0, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 9.508793778114665, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -140.0, | |
| "logps/rejected": -169.0, | |
| "loss": 0.0842, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.5078125, | |
| "rewards/margins": 7.09375, | |
| "rewards/rejected": -5.59375, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 10.523272369321887, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -1.6484375, | |
| "logits/rejected": -1.7421875, | |
| "logps/chosen": -151.0, | |
| "logps/rejected": -169.0, | |
| "loss": 0.0909, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.6875, | |
| "rewards/margins": 6.9375, | |
| "rewards/rejected": -5.25, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 5.122653578302724, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.71875, | |
| "logps/chosen": -152.0, | |
| "logps/rejected": -164.0, | |
| "loss": 0.1023, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.6796875, | |
| "rewards/margins": 7.125, | |
| "rewards/rejected": -5.4375, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -1.6484375, | |
| "eval_logits/rejected": -1.7109375, | |
| "eval_logps/chosen": -139.0, | |
| "eval_logps/rejected": -162.0, | |
| "eval_loss": 0.1479569971561432, | |
| "eval_rewards/accuracies": 0.9642857313156128, | |
| "eval_rewards/chosen": 1.71875, | |
| "eval_rewards/margins": 6.34375, | |
| "eval_rewards/rejected": -4.625, | |
| "eval_runtime": 15.5607, | |
| "eval_samples_per_second": 12.853, | |
| "eval_steps_per_second": 0.45, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 16.21457119609687, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -126.0, | |
| "logps/rejected": -156.0, | |
| "loss": 0.1181, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 1.6015625, | |
| "rewards/margins": 6.3125, | |
| "rewards/rejected": -4.71875, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 0.29904436862166595, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -1.703125, | |
| "logits/rejected": -1.7265625, | |
| "logps/chosen": -155.0, | |
| "logps/rejected": -170.0, | |
| "loss": 0.0205, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.734375, | |
| "rewards/margins": 6.3125, | |
| "rewards/rejected": -4.59375, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 13.679951332888173, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -145.0, | |
| "logps/rejected": -163.0, | |
| "loss": 0.0398, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.78125, | |
| "rewards/margins": 6.90625, | |
| "rewards/rejected": -5.125, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 11.9783902063343, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -111.0, | |
| "logps/rejected": -165.0, | |
| "loss": 0.0784, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.9609375, | |
| "rewards/margins": 7.34375, | |
| "rewards/rejected": -5.375, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 0.4679031202507738, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -130.0, | |
| "logps/rejected": -167.0, | |
| "loss": 0.0111, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.9140625, | |
| "rewards/margins": 7.15625, | |
| "rewards/rejected": -5.21875, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 0.20451574077109838, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -143.0, | |
| "logps/rejected": -168.0, | |
| "loss": 0.022, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.0625, | |
| "rewards/margins": 8.1875, | |
| "rewards/rejected": -6.125, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 10.598105422232411, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -1.6484375, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -116.5, | |
| "logps/rejected": -167.0, | |
| "loss": 0.0368, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.890625, | |
| "rewards/margins": 7.96875, | |
| "rewards/rejected": -6.09375, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 0.6399268248761598, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -138.0, | |
| "logps/rejected": -179.0, | |
| "loss": 0.0332, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.7265625, | |
| "rewards/margins": 7.4375, | |
| "rewards/rejected": -5.71875, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 1.7605838517748418, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -131.0, | |
| "logps/rejected": -170.0, | |
| "loss": 0.0362, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.9296875, | |
| "rewards/margins": 7.71875, | |
| "rewards/rejected": -5.8125, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 9.785775931105587, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.703125, | |
| "logps/chosen": -127.0, | |
| "logps/rejected": -166.0, | |
| "loss": 0.041, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.84375, | |
| "rewards/margins": 7.46875, | |
| "rewards/rejected": -5.625, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 6.0614459363104825, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -138.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.0306, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.8359375, | |
| "rewards/margins": 8.0, | |
| "rewards/rejected": -6.15625, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 6.761951499664274, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -1.6484375, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -140.0, | |
| "logps/rejected": -167.0, | |
| "loss": 0.0498, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 7.75, | |
| "rewards/rejected": -5.75, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 0.17536983369970263, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -150.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.0198, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.8359375, | |
| "rewards/margins": 7.96875, | |
| "rewards/rejected": -6.125, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 0.25285285279776293, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -140.0, | |
| "logps/rejected": -174.0, | |
| "loss": 0.0341, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.8671875, | |
| "rewards/margins": 8.125, | |
| "rewards/rejected": -6.28125, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 1.0751563884211028, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.7265625, | |
| "logps/chosen": -139.0, | |
| "logps/rejected": -176.0, | |
| "loss": 0.0062, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.1875, | |
| "rewards/margins": 8.75, | |
| "rewards/rejected": -6.59375, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 0.11127563987006106, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -142.0, | |
| "logps/rejected": -181.0, | |
| "loss": 0.0314, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.7265625, | |
| "rewards/margins": 8.9375, | |
| "rewards/rejected": -7.1875, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 6.3964315534774006, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.703125, | |
| "logps/chosen": -139.0, | |
| "logps/rejected": -184.0, | |
| "loss": 0.0347, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.78125, | |
| "rewards/margins": 8.9375, | |
| "rewards/rejected": -7.15625, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 11.099049063206122, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -135.0, | |
| "logps/rejected": -182.0, | |
| "loss": 0.0342, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.796875, | |
| "rewards/margins": 8.75, | |
| "rewards/rejected": -6.96875, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 25.72023107854396, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -147.0, | |
| "logps/rejected": -180.0, | |
| "loss": 0.1078, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.8515625, | |
| "rewards/margins": 8.4375, | |
| "rewards/rejected": -6.59375, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 10.898951400824425, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -131.0, | |
| "logps/rejected": -169.0, | |
| "loss": 0.048, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.078125, | |
| "rewards/margins": 7.96875, | |
| "rewards/rejected": -5.875, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 0.44841903505864433, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.6015625, | |
| "logps/chosen": -136.0, | |
| "logps/rejected": -175.0, | |
| "loss": 0.0194, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 2.34375, | |
| "rewards/margins": 8.375, | |
| "rewards/rejected": -6.03125, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 12.512633548125457, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -183.0, | |
| "loss": 0.0115, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.28125, | |
| "rewards/margins": 9.0625, | |
| "rewards/rejected": -6.8125, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 13.959766189558197, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.703125, | |
| "logps/chosen": -151.0, | |
| "logps/rejected": -183.0, | |
| "loss": 0.0304, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 9.0, | |
| "rewards/rejected": -6.5625, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 9.233719469042315, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -129.0, | |
| "logps/rejected": -180.0, | |
| "loss": 0.0313, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 2.0625, | |
| "rewards/margins": 8.625, | |
| "rewards/rejected": -6.5625, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 0.6258637700766689, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -152.0, | |
| "logps/rejected": -184.0, | |
| "loss": 0.0282, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.796875, | |
| "rewards/margins": 8.5625, | |
| "rewards/rejected": -6.8125, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 11.539213376888336, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.703125, | |
| "logps/chosen": -156.0, | |
| "logps/rejected": -186.0, | |
| "loss": 0.0412, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.578125, | |
| "rewards/margins": 8.375, | |
| "rewards/rejected": -6.8125, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 14.76772287276129, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.75, | |
| "logps/chosen": -127.5, | |
| "logps/rejected": -186.0, | |
| "loss": 0.0494, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.96875, | |
| "rewards/margins": 9.375, | |
| "rewards/rejected": -7.40625, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 0.1294660311123192, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.7265625, | |
| "logps/chosen": -113.5, | |
| "logps/rejected": -183.0, | |
| "loss": 0.0131, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 9.6875, | |
| "rewards/rejected": -7.59375, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 6.865783384868779, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -137.0, | |
| "logps/rejected": -176.0, | |
| "loss": 0.0681, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.828125, | |
| "rewards/margins": 8.625, | |
| "rewards/rejected": -6.78125, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 0.08359605869340851, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -1.6484375, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -185.0, | |
| "loss": 0.0534, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 2.09375, | |
| "rewards/margins": 9.25, | |
| "rewards/rejected": -7.125, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 0.3421157310137346, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -155.0, | |
| "logps/rejected": -179.0, | |
| "loss": 0.0236, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.9140625, | |
| "rewards/margins": 8.75, | |
| "rewards/rejected": -6.84375, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -1.625, | |
| "eval_logits/rejected": -1.6796875, | |
| "eval_logps/chosen": -137.0, | |
| "eval_logps/rejected": -179.0, | |
| "eval_loss": 0.17432327568531036, | |
| "eval_rewards/accuracies": 0.9642857313156128, | |
| "eval_rewards/chosen": 1.984375, | |
| "eval_rewards/margins": 8.3125, | |
| "eval_rewards/rejected": -6.34375, | |
| "eval_runtime": 15.7689, | |
| "eval_samples_per_second": 12.683, | |
| "eval_steps_per_second": 0.444, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 0.6498565946867418, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -147.0, | |
| "logps/rejected": -182.0, | |
| "loss": 0.0192, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 9.4375, | |
| "rewards/rejected": -7.09375, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 0.3293295305702466, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -149.0, | |
| "logps/rejected": -184.0, | |
| "loss": 0.0076, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 9.0625, | |
| "rewards/rejected": -6.9375, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 0.03540134173988831, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.703125, | |
| "logps/chosen": -168.0, | |
| "logps/rejected": -186.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 9.9375, | |
| "rewards/rejected": -7.59375, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 16.091702934411547, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -1.609375, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -139.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.0426, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 9.125, | |
| "rewards/rejected": -7.0, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 0.2098027250760498, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -149.0, | |
| "logps/rejected": -178.0, | |
| "loss": 0.013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.84375, | |
| "rewards/margins": 8.375, | |
| "rewards/rejected": -6.5625, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 0.5339899857614889, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -1.609375, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -123.5, | |
| "logps/rejected": -182.0, | |
| "loss": 0.0247, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.8203125, | |
| "rewards/margins": 8.5, | |
| "rewards/rejected": -6.6875, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 0.16146396901518884, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -173.0, | |
| "logps/rejected": -185.0, | |
| "loss": 0.0159, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.9453125, | |
| "rewards/margins": 8.3125, | |
| "rewards/rejected": -6.375, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 0.07042850207697388, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -137.0, | |
| "logps/rejected": -180.0, | |
| "loss": 0.0176, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.984375, | |
| "rewards/margins": 9.0, | |
| "rewards/rejected": -7.0, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 0.1655932469804816, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -149.0, | |
| "logps/rejected": -180.0, | |
| "loss": 0.0023, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 8.75, | |
| "rewards/rejected": -6.625, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 2.1568103368975344, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -117.0, | |
| "logps/rejected": -181.0, | |
| "loss": 0.0671, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 1.703125, | |
| "rewards/margins": 8.6875, | |
| "rewards/rejected": -6.9375, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 0.314943777613801, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -156.0, | |
| "logps/rejected": -181.0, | |
| "loss": 0.0022, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.09375, | |
| "rewards/margins": 9.0, | |
| "rewards/rejected": -6.9375, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 0.1717048855680156, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -163.0, | |
| "logps/rejected": -185.0, | |
| "loss": 0.0042, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.203125, | |
| "rewards/margins": 9.1875, | |
| "rewards/rejected": -7.0, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 0.13036139910530478, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -128.0, | |
| "logps/rejected": -190.0, | |
| "loss": 0.0172, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.90625, | |
| "rewards/margins": 8.6875, | |
| "rewards/rejected": -6.78125, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 2.55842893808289, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -146.0, | |
| "logps/rejected": -193.0, | |
| "loss": 0.0212, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.515625, | |
| "rewards/margins": 8.375, | |
| "rewards/rejected": -6.875, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 0.09940252022628729, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -1.6484375, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -132.0, | |
| "logps/rejected": -187.0, | |
| "loss": 0.0017, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 9.25, | |
| "rewards/rejected": -7.15625, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 17.356505869545202, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -131.0, | |
| "logps/rejected": -182.0, | |
| "loss": 0.0192, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 9.375, | |
| "rewards/rejected": -7.15625, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 0.593428314749854, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -141.0, | |
| "logps/rejected": -186.0, | |
| "loss": 0.021, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.359375, | |
| "rewards/margins": 9.5, | |
| "rewards/rejected": -7.15625, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 11.900912740835807, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -132.0, | |
| "logps/rejected": -186.0, | |
| "loss": 0.0351, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 9.5625, | |
| "rewards/rejected": -7.34375, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 0.06771001113976342, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -109.0, | |
| "logps/rejected": -177.0, | |
| "loss": 0.014, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.8046875, | |
| "rewards/margins": 8.75, | |
| "rewards/rejected": -6.96875, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 0.19065837416735965, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -118.5, | |
| "logps/rejected": -178.0, | |
| "loss": 0.0135, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.03125, | |
| "rewards/margins": 8.6875, | |
| "rewards/rejected": -6.65625, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 1.1870544618648544, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -1.6484375, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -141.0, | |
| "logps/rejected": -187.0, | |
| "loss": 0.0074, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.96875, | |
| "rewards/margins": 9.1875, | |
| "rewards/rejected": -7.25, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 1.9119353584457313, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -130.0, | |
| "logps/rejected": -183.0, | |
| "loss": 0.0046, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.125, | |
| "rewards/margins": 9.5, | |
| "rewards/rejected": -7.375, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 0.10117790082745341, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -121.5, | |
| "logps/rejected": -177.0, | |
| "loss": 0.0396, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.15625, | |
| "rewards/margins": 8.625, | |
| "rewards/rejected": -6.4375, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 7.809237170202005, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -105.5, | |
| "logps/rejected": -180.0, | |
| "loss": 0.0254, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.8203125, | |
| "rewards/margins": 8.75, | |
| "rewards/rejected": -6.90625, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 0.09331472530413627, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -143.0, | |
| "logps/rejected": -183.0, | |
| "loss": 0.0163, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.9453125, | |
| "rewards/margins": 9.0, | |
| "rewards/rejected": -7.0625, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 5.639594850245226, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -147.0, | |
| "logps/rejected": -179.0, | |
| "loss": 0.0034, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.9609375, | |
| "rewards/margins": 8.9375, | |
| "rewards/rejected": -6.9375, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 0.20860276541522338, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -143.0, | |
| "logps/rejected": -182.0, | |
| "loss": 0.0047, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.265625, | |
| "rewards/margins": 9.625, | |
| "rewards/rejected": -7.375, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 6.794488064302702, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -155.0, | |
| "logps/rejected": -189.0, | |
| "loss": 0.011, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.21875, | |
| "rewards/margins": 9.4375, | |
| "rewards/rejected": -7.1875, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 0.9529687426961281, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -146.0, | |
| "logps/rejected": -179.0, | |
| "loss": 0.0325, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.09375, | |
| "rewards/margins": 9.125, | |
| "rewards/rejected": -7.03125, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 0.122655595633226, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -132.0, | |
| "logps/rejected": -184.0, | |
| "loss": 0.0072, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.9296875, | |
| "rewards/margins": 9.3125, | |
| "rewards/rejected": -7.40625, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 0.46195117646435746, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -156.0, | |
| "logps/rejected": -189.0, | |
| "loss": 0.0125, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 9.0625, | |
| "rewards/rejected": -7.0625, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -1.6171875, | |
| "eval_logits/rejected": -1.671875, | |
| "eval_logps/chosen": -138.0, | |
| "eval_logps/rejected": -181.0, | |
| "eval_loss": 0.1789417862892151, | |
| "eval_rewards/accuracies": 0.9642857313156128, | |
| "eval_rewards/chosen": 1.84375, | |
| "eval_rewards/margins": 8.375, | |
| "eval_rewards/rejected": -6.5, | |
| "eval_runtime": 18.4526, | |
| "eval_samples_per_second": 10.839, | |
| "eval_steps_per_second": 0.379, | |
| "step": 939 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 939, | |
| "total_flos": 0.0, | |
| "train_loss": 0.06296250589104634, | |
| "train_runtime": 5854.5161, | |
| "train_samples_per_second": 5.118, | |
| "train_steps_per_second": 0.16 | |
| } | |
| ], | |
| "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 | |
| } | |