| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.9904153354632586, | |
| "eval_steps": 500, | |
| "global_step": 468, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.006389776357827476, | |
| "grad_norm": 29.368945511033267, | |
| "learning_rate": 1.0638297872340425e-08, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.28125, | |
| "logps/chosen": -145.0, | |
| "logps/rejected": -130.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 28.719594921405, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.34375, | |
| "logps/chosen": -145.0, | |
| "logps/rejected": -129.0, | |
| "loss": 0.687, | |
| "rewards/accuracies": 0.3055555522441864, | |
| "rewards/chosen": 0.002960205078125, | |
| "rewards/margins": 0.0189208984375, | |
| "rewards/rejected": -0.0159912109375, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 22.494308989969756, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.390625, | |
| "logps/chosen": -146.0, | |
| "logps/rejected": -134.0, | |
| "loss": 0.5984, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 0.00970458984375, | |
| "rewards/margins": 0.203125, | |
| "rewards/rejected": -0.193359375, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 9.449890990720535, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.296875, | |
| "logps/chosen": -140.0, | |
| "logps/rejected": -136.0, | |
| "loss": 0.3229, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.025634765625, | |
| "rewards/margins": 1.0625, | |
| "rewards/rejected": -1.0390625, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 0.38707031057010677, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -149.0, | |
| "logps/rejected": -169.0, | |
| "loss": 0.0577, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.2578125, | |
| "rewards/margins": 3.671875, | |
| "rewards/rejected": -3.9375, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 0.20241899573369684, | |
| "learning_rate": 4.96437054631829e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -147.0, | |
| "logps/rejected": -203.0, | |
| "loss": 0.0033, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.66015625, | |
| "rewards/margins": 6.875, | |
| "rewards/rejected": -7.53125, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 0.3039285596197402, | |
| "learning_rate": 4.845605700712589e-07, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -221.0, | |
| "loss": 0.0025, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7265625, | |
| "rewards/margins": 8.0, | |
| "rewards/rejected": -8.75, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 0.362275931660039, | |
| "learning_rate": 4.7268408551068883e-07, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -151.0, | |
| "logps/rejected": -222.0, | |
| "loss": 0.0018, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.8515625, | |
| "rewards/margins": 8.375, | |
| "rewards/rejected": -9.25, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 0.2868857164394034, | |
| "learning_rate": 4.6080760095011875e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -231.0, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.734375, | |
| "rewards/margins": 9.3125, | |
| "rewards/rejected": -10.0625, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 0.15240372298676685, | |
| "learning_rate": 4.4893111638954866e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -148.0, | |
| "logps/rejected": -230.0, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.73046875, | |
| "rewards/margins": 9.5, | |
| "rewards/rejected": -10.25, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 0.02392673493900165, | |
| "learning_rate": 4.3705463182897863e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.3828125, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -235.0, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7578125, | |
| "rewards/margins": 9.75, | |
| "rewards/rejected": -10.5625, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 0.09175324835066521, | |
| "learning_rate": 4.251781472684085e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -150.0, | |
| "logps/rejected": -234.0, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.83203125, | |
| "rewards/margins": 10.0625, | |
| "rewards/rejected": -10.875, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 0.06184587725303002, | |
| "learning_rate": 4.1330166270783846e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -145.0, | |
| "logps/rejected": -240.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.89453125, | |
| "rewards/margins": 10.375, | |
| "rewards/rejected": -11.25, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 0.024649453162613934, | |
| "learning_rate": 4.0142517814726837e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -150.0, | |
| "logps/rejected": -242.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.859375, | |
| "rewards/margins": 10.625, | |
| "rewards/rejected": -11.5, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 0.022941440244503526, | |
| "learning_rate": 3.8954869358669834e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -171.0, | |
| "logps/rejected": -242.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.046875, | |
| "rewards/margins": 10.375, | |
| "rewards/rejected": -11.4375, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 0.2502811167205942, | |
| "learning_rate": 3.7767220902612825e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -160.0, | |
| "logps/rejected": -250.0, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.984375, | |
| "rewards/margins": 10.8125, | |
| "rewards/rejected": -11.75, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.9968051118210862, | |
| "eval_logits/chosen": -1.4921875, | |
| "eval_logits/rejected": -1.46875, | |
| "eval_logps/chosen": -148.0, | |
| "eval_logps/rejected": -250.0, | |
| "eval_loss": 7.716297841398045e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -0.90625, | |
| "eval_rewards/margins": 10.9375, | |
| "eval_rewards/rejected": -11.8125, | |
| "eval_runtime": 11.6151, | |
| "eval_samples_per_second": 17.219, | |
| "eval_steps_per_second": 0.603, | |
| "step": 156 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 0.05686403061946404, | |
| "learning_rate": 3.6579572446555817e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -174.0, | |
| "logps/rejected": -249.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.95703125, | |
| "rewards/margins": 10.875, | |
| "rewards/rejected": -11.8125, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 0.025647512662174864, | |
| "learning_rate": 3.5391923990498813e-07, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -151.0, | |
| "logps/rejected": -253.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.890625, | |
| "rewards/margins": 11.4375, | |
| "rewards/rejected": -12.3125, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 0.10600901838990953, | |
| "learning_rate": 3.42042755344418e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -160.0, | |
| "logps/rejected": -249.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.94140625, | |
| "rewards/margins": 11.25, | |
| "rewards/rejected": -12.25, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 0.13119830498432972, | |
| "learning_rate": 3.3016627078384796e-07, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -146.0, | |
| "logps/rejected": -251.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.88671875, | |
| "rewards/margins": 11.4375, | |
| "rewards/rejected": -12.3125, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 0.030239526412913494, | |
| "learning_rate": 3.182897862232779e-07, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -134.0, | |
| "logps/rejected": -252.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.9453125, | |
| "rewards/margins": 11.6875, | |
| "rewards/rejected": -12.625, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 0.0080485091857848, | |
| "learning_rate": 3.0641330166270784e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.97265625, | |
| "rewards/margins": 11.8125, | |
| "rewards/rejected": -12.8125, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 0.01503760669759659, | |
| "learning_rate": 2.9453681710213776e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.890625, | |
| "rewards/margins": 11.8125, | |
| "rewards/rejected": -12.6875, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 0.01745595116822664, | |
| "learning_rate": 2.8266033254156767e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -152.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.73828125, | |
| "rewards/margins": 12.0, | |
| "rewards/rejected": -12.75, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 0.010639490905132864, | |
| "learning_rate": 2.7078384798099764e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -152.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.8359375, | |
| "rewards/margins": 12.125, | |
| "rewards/rejected": -13.0, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 0.021325021450464145, | |
| "learning_rate": 2.589073634204275e-07, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.80859375, | |
| "rewards/margins": 11.6875, | |
| "rewards/rejected": -12.5, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 0.03402727752531963, | |
| "learning_rate": 2.4703087885985747e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -143.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7265625, | |
| "rewards/margins": 11.6875, | |
| "rewards/rejected": -12.4375, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 0.01735332750787422, | |
| "learning_rate": 2.351543942992874e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -165.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.80859375, | |
| "rewards/margins": 12.1875, | |
| "rewards/rejected": -13.0, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 0.02817904426646656, | |
| "learning_rate": 2.2327790973871732e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -158.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7734375, | |
| "rewards/margins": 12.375, | |
| "rewards/rejected": -13.125, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 0.026646622883463183, | |
| "learning_rate": 2.1140142517814726e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.90234375, | |
| "rewards/margins": 11.9375, | |
| "rewards/rejected": -12.875, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 0.005464018856443713, | |
| "learning_rate": 1.9952494061757718e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -155.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.296875, | |
| "rewards/margins": 12.3125, | |
| "rewards/rejected": -13.625, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 0.01607250092110866, | |
| "learning_rate": 1.876484560570071e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -152.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.34375, | |
| "rewards/margins": 12.5625, | |
| "rewards/rejected": -13.875, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -1.5078125, | |
| "eval_logits/rejected": -1.515625, | |
| "eval_logps/chosen": -151.0, | |
| "eval_logps/rejected": -266.0, | |
| "eval_loss": 1.9190907551092096e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -1.1640625, | |
| "eval_rewards/margins": 12.3125, | |
| "eval_rewards/rejected": -13.4375, | |
| "eval_runtime": 11.5575, | |
| "eval_samples_per_second": 17.305, | |
| "eval_steps_per_second": 0.606, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 0.01361688937176726, | |
| "learning_rate": 1.7577197149643706e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -157.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.203125, | |
| "rewards/margins": 12.6875, | |
| "rewards/rejected": -13.9375, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 0.05503143614977478, | |
| "learning_rate": 1.6389548693586697e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -162.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1796875, | |
| "rewards/margins": 12.375, | |
| "rewards/rejected": -13.5625, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 0.010269985783571838, | |
| "learning_rate": 1.520190023752969e-07, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -152.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.078125, | |
| "rewards/margins": 12.75, | |
| "rewards/rejected": -13.8125, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 0.0683919347457657, | |
| "learning_rate": 1.4014251781472683e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -156.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.09375, | |
| "rewards/margins": 12.75, | |
| "rewards/rejected": -13.8125, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 0.015461126488895941, | |
| "learning_rate": 1.2826603325415677e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.546875, | |
| "logps/chosen": -146.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1015625, | |
| "rewards/margins": 12.625, | |
| "rewards/rejected": -13.75, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 0.004666909345169454, | |
| "learning_rate": 1.163895486935867e-07, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -160.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1328125, | |
| "rewards/margins": 12.8125, | |
| "rewards/rejected": -13.9375, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 0.021768817881518163, | |
| "learning_rate": 1.0451306413301662e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -157.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1484375, | |
| "rewards/margins": 12.8125, | |
| "rewards/rejected": -14.0, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 0.006712968496189408, | |
| "learning_rate": 9.263657957244655e-08, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -150.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.046875, | |
| "rewards/margins": 13.125, | |
| "rewards/rejected": -14.125, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 0.03157363956222121, | |
| "learning_rate": 8.076009501187649e-08, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -160.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.0625, | |
| "rewards/margins": 13.0625, | |
| "rewards/rejected": -14.125, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 0.012508580641767728, | |
| "learning_rate": 6.88836104513064e-08, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -151.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.03125, | |
| "rewards/margins": 12.9375, | |
| "rewards/rejected": -13.9375, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 0.014858628369684351, | |
| "learning_rate": 5.700712589073634e-08, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -159.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.171875, | |
| "rewards/margins": 12.8125, | |
| "rewards/rejected": -14.0, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 0.00119033851952325, | |
| "learning_rate": 4.5130641330166267e-08, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -165.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.0703125, | |
| "rewards/margins": 13.25, | |
| "rewards/rejected": -14.3125, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 0.01000781639917085, | |
| "learning_rate": 3.32541567695962e-08, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -142.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.0234375, | |
| "rewards/margins": 13.375, | |
| "rewards/rejected": -14.375, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 0.032333252949445446, | |
| "learning_rate": 2.1377672209026125e-08, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -159.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.078125, | |
| "rewards/margins": 13.3125, | |
| "rewards/rejected": -14.375, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 0.0060308968638626, | |
| "learning_rate": 9.501187648456057e-09, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.0390625, | |
| "rewards/margins": 13.3125, | |
| "rewards/rejected": -14.375, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 2.9904153354632586, | |
| "eval_logits/chosen": -1.5078125, | |
| "eval_logits/rejected": -1.5234375, | |
| "eval_logps/chosen": -150.0, | |
| "eval_logps/rejected": -270.0, | |
| "eval_loss": 1.3302266779646743e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -1.09375, | |
| "eval_rewards/margins": 12.8125, | |
| "eval_rewards/rejected": -13.9375, | |
| "eval_runtime": 14.4196, | |
| "eval_samples_per_second": 13.87, | |
| "eval_steps_per_second": 0.485, | |
| "step": 468 | |
| }, | |
| { | |
| "epoch": 2.9904153354632586, | |
| "step": 468, | |
| "total_flos": 0.0, | |
| "train_loss": 0.036010837255037814, | |
| "train_runtime": 4360.4368, | |
| "train_samples_per_second": 6.88, | |
| "train_steps_per_second": 0.107 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 468, | |
| "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 | |
| } | |