| { | |
| "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": 33.83219308939002, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -256.0, | |
| "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": 34.295600964581766, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.5546875, | |
| "logps/chosen": -238.0, | |
| "logps/rejected": -247.0, | |
| "loss": 0.6968, | |
| "rewards/accuracies": 0.1805555522441864, | |
| "rewards/chosen": -0.002777099609375, | |
| "rewards/margins": -0.002777099609375, | |
| "rewards/rejected": 1.3589859008789062e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 35.06874635564751, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -240.0, | |
| "logps/rejected": -254.0, | |
| "loss": 0.6888, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -0.0024871826171875, | |
| "rewards/margins": 0.0150146484375, | |
| "rewards/rejected": -0.017578125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 32.66967705371805, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -241.0, | |
| "logps/rejected": -247.0, | |
| "loss": 0.6846, | |
| "rewards/accuracies": 0.4000000059604645, | |
| "rewards/chosen": -0.0250244140625, | |
| "rewards/margins": 0.0162353515625, | |
| "rewards/rejected": -0.041259765625, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 33.3498861828891, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -234.0, | |
| "logps/rejected": -246.0, | |
| "loss": 0.6682, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.027587890625, | |
| "rewards/margins": 0.07763671875, | |
| "rewards/rejected": -0.10498046875, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 32.8066635955503, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.5625, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -252.0, | |
| "loss": 0.6564, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.050048828125, | |
| "rewards/margins": 0.15234375, | |
| "rewards/rejected": -0.203125, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 31.151319250680135, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -1.390625, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -254.0, | |
| "loss": 0.5997, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.1787109375, | |
| "rewards/margins": 0.259765625, | |
| "rewards/rejected": -0.4375, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 32.261295417505444, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -253.0, | |
| "loss": 0.6068, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.357421875, | |
| "rewards/margins": 0.2314453125, | |
| "rewards/rejected": -0.5859375, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 29.607136046645028, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.5773, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.55078125, | |
| "rewards/margins": 0.25, | |
| "rewards/rejected": -0.80078125, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 29.91290745142416, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.390625, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.5479, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.416015625, | |
| "rewards/margins": 0.546875, | |
| "rewards/rejected": -0.9609375, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 26.305448444999005, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.4975, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.48828125, | |
| "rewards/margins": 0.609375, | |
| "rewards/rejected": -1.09375, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 32.72166445591288, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -249.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.5314, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.71875, | |
| "rewards/margins": 0.6484375, | |
| "rewards/rejected": -1.3671875, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 30.27365733119751, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -247.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.5104, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.5, | |
| "rewards/margins": 0.64453125, | |
| "rewards/rejected": -1.140625, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 28.430091683563568, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -247.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.4292, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.7109375, | |
| "rewards/margins": 0.9375, | |
| "rewards/rejected": -1.6484375, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 26.4562874890803, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.4294, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.458984375, | |
| "rewards/margins": 1.140625, | |
| "rewards/rejected": -1.6015625, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 28.469599730279953, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.4552, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.9921875, | |
| "rewards/margins": 0.86328125, | |
| "rewards/rejected": -1.859375, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 24.793876978222805, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.390625, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.437, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.83984375, | |
| "rewards/margins": 1.0078125, | |
| "rewards/rejected": -1.8515625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 31.65572470155131, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.4342, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.8125, | |
| "rewards/margins": 1.2890625, | |
| "rewards/rejected": -2.09375, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 23.03742604789954, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.3242, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.484375, | |
| "rewards/margins": 1.59375, | |
| "rewards/rejected": -2.078125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 25.359399398545317, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.3737, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.2421875, | |
| "rewards/margins": 1.375, | |
| "rewards/rejected": -2.609375, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 24.81415089513789, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.3879, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.73828125, | |
| "rewards/margins": 1.6875, | |
| "rewards/rejected": -2.4375, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 25.835490969114602, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.3863, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.125, | |
| "rewards/margins": 1.5, | |
| "rewards/rejected": -2.625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 36.89571558397856, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.3658, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.83984375, | |
| "rewards/margins": 1.3359375, | |
| "rewards/rejected": -2.171875, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 29.264392415424865, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -253.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.3908, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.046875, | |
| "rewards/margins": 1.4296875, | |
| "rewards/rejected": -2.484375, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 21.579221137355532, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.3577, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.89453125, | |
| "rewards/margins": 1.6484375, | |
| "rewards/rejected": -2.546875, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 29.086956776688627, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -241.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.3608, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.58984375, | |
| "rewards/margins": 1.90625, | |
| "rewards/rejected": -2.5, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 16.121334472586927, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.344, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.95703125, | |
| "rewards/margins": 1.8671875, | |
| "rewards/rejected": -2.828125, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 27.16809374599435, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.3484, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.82421875, | |
| "rewards/margins": 1.9140625, | |
| "rewards/rejected": -2.734375, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 17.824726123830576, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -253.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.3234, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -1.125, | |
| "rewards/margins": 1.65625, | |
| "rewards/rejected": -2.78125, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 20.459767771162962, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.3206, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.87109375, | |
| "rewards/margins": 1.90625, | |
| "rewards/rejected": -2.765625, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 27.174346183527405, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -249.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.3199, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.8125, | |
| "rewards/margins": 2.140625, | |
| "rewards/rejected": -2.953125, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 25.180996364480237, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -1.4296875, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -247.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.3335, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.78515625, | |
| "rewards/margins": 1.984375, | |
| "rewards/rejected": -2.765625, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -1.3203125, | |
| "eval_logits/rejected": -1.3828125, | |
| "eval_logps/chosen": -251.0, | |
| "eval_logps/rejected": -280.0, | |
| "eval_loss": 0.3685546815395355, | |
| "eval_rewards/accuracies": 0.75, | |
| "eval_rewards/chosen": -1.0703125, | |
| "eval_rewards/margins": 1.890625, | |
| "eval_rewards/rejected": -2.96875, | |
| "eval_runtime": 11.107, | |
| "eval_samples_per_second": 18.007, | |
| "eval_steps_per_second": 0.63, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 22.992187694664622, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.2266, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.78515625, | |
| "rewards/margins": 2.796875, | |
| "rewards/rejected": -3.59375, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 21.541180860944344, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -249.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.1895, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.0546875, | |
| "rewards/margins": 2.765625, | |
| "rewards/rejected": -3.8125, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 15.23008721372514, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.1719, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.96484375, | |
| "rewards/margins": 3.078125, | |
| "rewards/rejected": -4.0625, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 24.68250321450141, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -255.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.218, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.2421875, | |
| "rewards/margins": 2.9375, | |
| "rewards/rejected": -4.1875, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 24.1341079399757, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.1583, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.5078125, | |
| "rewards/margins": 3.296875, | |
| "rewards/rejected": -3.8125, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 19.227045061905727, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.1652, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.7109375, | |
| "rewards/margins": 3.046875, | |
| "rewards/rejected": -3.75, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 26.621847691836305, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.1974, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.91015625, | |
| "rewards/margins": 3.15625, | |
| "rewards/rejected": -4.0625, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 13.381405496459287, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.1562, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -0.60546875, | |
| "rewards/margins": 3.15625, | |
| "rewards/rejected": -3.765625, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 22.66314472539665, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.390625, | |
| "logps/chosen": -247.0, | |
| "logps/rejected": -294.0, | |
| "loss": 0.1676, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -1.1171875, | |
| "rewards/margins": 3.28125, | |
| "rewards/rejected": -4.40625, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 21.158315786539344, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -1.4140625, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -250.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.1362, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -0.8046875, | |
| "rewards/margins": 3.453125, | |
| "rewards/rejected": -4.25, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 21.52447791309729, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -250.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.1837, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -0.6796875, | |
| "rewards/margins": 3.171875, | |
| "rewards/rejected": -3.84375, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 17.74412482948413, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.1336, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.625, | |
| "rewards/margins": 3.5625, | |
| "rewards/rejected": -4.1875, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 27.789758242475365, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -294.0, | |
| "loss": 0.1467, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.1875, | |
| "rewards/margins": 3.53125, | |
| "rewards/rejected": -4.71875, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 14.190845365082243, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -249.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.142, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.234375, | |
| "rewards/margins": 3.640625, | |
| "rewards/rejected": -4.875, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 14.191633300976386, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -247.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.1435, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.8046875, | |
| "rewards/margins": 3.84375, | |
| "rewards/rejected": -4.65625, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 12.282091047977525, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -249.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.1215, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.03125, | |
| "rewards/margins": 4.15625, | |
| "rewards/rejected": -5.1875, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 17.774657190397015, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -249.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.1482, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.75, | |
| "rewards/margins": 3.515625, | |
| "rewards/rejected": -4.28125, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 12.548718190832725, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.1153, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.765625, | |
| "rewards/margins": 4.25, | |
| "rewards/rejected": -5.0, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 21.03007948633376, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.328125, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.1462, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.69140625, | |
| "rewards/margins": 3.578125, | |
| "rewards/rejected": -4.28125, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 22.21936248599112, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -253.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.1452, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -0.76953125, | |
| "rewards/margins": 3.96875, | |
| "rewards/rejected": -4.75, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 21.144140431346198, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -239.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.1311, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.208984375, | |
| "rewards/margins": 3.96875, | |
| "rewards/rejected": -4.1875, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 18.47380195962148, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.1292, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -0.69140625, | |
| "rewards/margins": 3.671875, | |
| "rewards/rejected": -4.375, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 20.91408268731602, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -244.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.1426, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -0.703125, | |
| "rewards/margins": 3.390625, | |
| "rewards/rejected": -4.09375, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 16.824757445565048, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -239.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.1063, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.3984375, | |
| "rewards/margins": 4.4375, | |
| "rewards/rejected": -4.84375, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 25.287519792001007, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.1261, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.1328125, | |
| "rewards/margins": 4.15625, | |
| "rewards/rejected": -5.28125, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 27.218111317330546, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.390625, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.1666, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.349609375, | |
| "rewards/margins": 4.8125, | |
| "rewards/rejected": -5.15625, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 23.1627237247012, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -250.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.1231, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.76953125, | |
| "rewards/margins": 4.25, | |
| "rewards/rejected": -5.03125, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 12.216445362218424, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -1.4140625, | |
| "logits/rejected": -1.3359375, | |
| "logps/chosen": -247.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.1365, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.50390625, | |
| "rewards/margins": 4.0625, | |
| "rewards/rejected": -4.5625, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 21.535957401095253, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -253.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.1603, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.87109375, | |
| "rewards/margins": 4.09375, | |
| "rewards/rejected": -4.96875, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 6.766720668443772, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.0994, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -0.427734375, | |
| "rewards/margins": 4.0625, | |
| "rewards/rejected": -4.46875, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 18.394899397655916, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.1398, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.8125, | |
| "rewards/margins": 4.46875, | |
| "rewards/rejected": -5.28125, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -1.3125, | |
| "eval_logits/rejected": -1.3515625, | |
| "eval_logps/chosen": -253.0, | |
| "eval_logps/rejected": -298.0, | |
| "eval_loss": 0.26944825053215027, | |
| "eval_rewards/accuracies": 0.8571428656578064, | |
| "eval_rewards/chosen": -1.296875, | |
| "eval_rewards/margins": 3.546875, | |
| "eval_rewards/rejected": -4.84375, | |
| "eval_runtime": 11.2185, | |
| "eval_samples_per_second": 17.828, | |
| "eval_steps_per_second": 0.624, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 9.008859978395401, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -1.5234375, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.089, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.3359375, | |
| "rewards/margins": 4.25, | |
| "rewards/rejected": -4.59375, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 6.394226306430514, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.0568, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6015625, | |
| "rewards/margins": 5.5, | |
| "rewards/rejected": -6.125, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 3.079234171886653, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -310.0, | |
| "loss": 0.0519, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.71875, | |
| "rewards/margins": 4.9375, | |
| "rewards/rejected": -5.6875, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 6.272950983225102, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.0804, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.8125, | |
| "rewards/margins": 4.59375, | |
| "rewards/rejected": -5.40625, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 8.988487131576974, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.34375, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.0496, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.49609375, | |
| "rewards/margins": 4.75, | |
| "rewards/rejected": -5.25, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 8.159838760463082, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -250.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.0578, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.87109375, | |
| "rewards/margins": 4.96875, | |
| "rewards/rejected": -5.84375, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 17.597694183277287, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.0673, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.07421875, | |
| "rewards/margins": 5.875, | |
| "rewards/rejected": -5.9375, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 19.792820570766843, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -1.375, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0702, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7421875, | |
| "rewards/margins": 5.15625, | |
| "rewards/rejected": -5.90625, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 8.600381448035137, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -240.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.0539, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.1923828125, | |
| "rewards/margins": 5.75, | |
| "rewards/rejected": -5.96875, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 17.41941039012227, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.069, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.61328125, | |
| "rewards/margins": 5.3125, | |
| "rewards/rejected": -5.9375, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 25.037192351858263, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -244.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0497, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.8125, | |
| "rewards/margins": 5.34375, | |
| "rewards/rejected": -6.15625, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 6.1257474979571045, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0463, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.59375, | |
| "rewards/margins": 5.78125, | |
| "rewards/rejected": -6.375, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 9.957701932091602, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -250.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0614, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1171875, | |
| "rewards/margins": 5.28125, | |
| "rewards/rejected": -6.40625, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 8.151894348408907, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": -1.390625, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0503, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.078125, | |
| "rewards/margins": 5.15625, | |
| "rewards/rejected": -6.25, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 7.245862853607653, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.0485, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1640625, | |
| "rewards/margins": 5.0, | |
| "rewards/rejected": -6.1875, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 18.973643902153352, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.3359375, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -310.0, | |
| "loss": 0.0523, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.640625, | |
| "rewards/margins": 5.59375, | |
| "rewards/rejected": -6.25, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 15.56196898970075, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.0726, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.392578125, | |
| "rewards/margins": 5.0625, | |
| "rewards/rejected": -5.46875, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 7.434354083474048, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.3671875, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0518, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7578125, | |
| "rewards/margins": 5.53125, | |
| "rewards/rejected": -6.28125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 6.043379670574249, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.0828, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.453125, | |
| "rewards/margins": 5.09375, | |
| "rewards/rejected": -5.53125, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 9.788784681363204, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.3671875, | |
| "logps/chosen": -250.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.0491, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6875, | |
| "rewards/margins": 5.25, | |
| "rewards/rejected": -5.9375, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 21.30703661009818, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -255.0, | |
| "logps/rejected": -310.0, | |
| "loss": 0.0517, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.3515625, | |
| "rewards/margins": 4.71875, | |
| "rewards/rejected": -6.09375, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 9.373420801368813, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -250.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.059, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.03125, | |
| "rewards/margins": 5.21875, | |
| "rewards/rejected": -6.25, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 13.613909483856304, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -310.0, | |
| "loss": 0.0567, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.75390625, | |
| "rewards/margins": 6.0, | |
| "rewards/rejected": -6.75, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 15.430662024682558, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -1.390625, | |
| "logits/rejected": -1.2734375, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -316.0, | |
| "loss": 0.0413, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.0390625, | |
| "rewards/margins": 5.71875, | |
| "rewards/rejected": -6.75, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 15.722766184638564, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0437, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.875, | |
| "rewards/margins": 5.0, | |
| "rewards/rejected": -5.875, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 14.977181098784836, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -238.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.046, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.384765625, | |
| "rewards/margins": 5.1875, | |
| "rewards/rejected": -5.59375, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 26.584208659094074, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0499, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.546875, | |
| "rewards/margins": 5.5625, | |
| "rewards/rejected": -6.09375, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 7.397188016373842, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.0416, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.015625, | |
| "rewards/margins": 5.5, | |
| "rewards/rejected": -6.5, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 4.16555751994792, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.3828125, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -316.0, | |
| "loss": 0.0393, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.85546875, | |
| "rewards/margins": 6.0625, | |
| "rewards/rejected": -6.90625, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 18.370934081127842, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0439, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.046875, | |
| "rewards/margins": 5.21875, | |
| "rewards/rejected": -6.25, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 9.265481933726917, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -255.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0392, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.171875, | |
| "rewards/margins": 5.3125, | |
| "rewards/rejected": -6.46875, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -1.3046875, | |
| "eval_logits/rejected": -1.34375, | |
| "eval_logps/chosen": -260.0, | |
| "eval_logps/rejected": -310.0, | |
| "eval_loss": 0.29661378264427185, | |
| "eval_rewards/accuracies": 0.8214285969734192, | |
| "eval_rewards/chosen": -1.9375, | |
| "eval_rewards/margins": 3.984375, | |
| "eval_rewards/rejected": -5.90625, | |
| "eval_runtime": 13.7446, | |
| "eval_samples_per_second": 14.551, | |
| "eval_steps_per_second": 0.509, | |
| "step": 939 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 939, | |
| "total_flos": 0.0, | |
| "train_loss": 0.22188479552507653, | |
| "train_runtime": 4635.7821, | |
| "train_samples_per_second": 6.468, | |
| "train_steps_per_second": 0.203 | |
| } | |
| ], | |
| "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 | |
| } | |