| { | |
| "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": 42.83342291543744, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -0.3984375, | |
| "logits/rejected": -0.138671875, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -243.0, | |
| "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": 40.292381459319074, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -0.287109375, | |
| "logits/rejected": -0.2451171875, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.697, | |
| "rewards/accuracies": 0.1944444477558136, | |
| "rewards/chosen": -0.004180908203125, | |
| "rewards/margins": -0.0166015625, | |
| "rewards/rejected": 0.01251220703125, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 39.77411506737548, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -0.201171875, | |
| "logits/rejected": -0.2001953125, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.6827, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": -0.0050048828125, | |
| "rewards/margins": 0.03759765625, | |
| "rewards/rejected": -0.04248046875, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 36.58914456263147, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -0.341796875, | |
| "logits/rejected": -0.236328125, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.6503, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.048828125, | |
| "rewards/margins": 0.08251953125, | |
| "rewards/rejected": -0.1318359375, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 33.262544147493976, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -0.33203125, | |
| "logits/rejected": -0.22265625, | |
| "logps/chosen": -256.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.5671, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.111328125, | |
| "rewards/margins": 0.361328125, | |
| "rewards/rejected": -0.47265625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 34.58721127152837, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -0.25, | |
| "logits/rejected": -0.2041015625, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.4781, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.37890625, | |
| "rewards/margins": 0.6875, | |
| "rewards/rejected": -1.0625, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 28.7877935830883, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -0.2099609375, | |
| "logits/rejected": -0.146484375, | |
| "logps/chosen": -266.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.4472, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.8515625, | |
| "rewards/margins": 0.78515625, | |
| "rewards/rejected": -1.6328125, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 23.995799487313224, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -0.291015625, | |
| "logits/rejected": -0.2138671875, | |
| "logps/chosen": -255.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.349, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.494140625, | |
| "rewards/margins": 1.4453125, | |
| "rewards/rejected": -1.9453125, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 25.453610850800764, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -0.345703125, | |
| "logits/rejected": -0.248046875, | |
| "logps/chosen": -249.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.3588, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.51171875, | |
| "rewards/margins": 2.0, | |
| "rewards/rejected": -2.515625, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 27.015860960341136, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -0.265625, | |
| "logits/rejected": -0.2109375, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.3293, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -0.333984375, | |
| "rewards/margins": 2.359375, | |
| "rewards/rejected": -2.6875, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 24.07291733490132, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -0.251953125, | |
| "logits/rejected": -0.220703125, | |
| "logps/chosen": -256.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.3698, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.376953125, | |
| "rewards/margins": 1.921875, | |
| "rewards/rejected": -2.296875, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 20.6595462981731, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -0.310546875, | |
| "logits/rejected": -0.2109375, | |
| "logps/chosen": -256.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.3244, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.70703125, | |
| "rewards/margins": 2.03125, | |
| "rewards/rejected": -2.734375, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 17.508488770544073, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -0.33203125, | |
| "logits/rejected": -0.173828125, | |
| "logps/chosen": -250.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.2732, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.69140625, | |
| "rewards/margins": 2.28125, | |
| "rewards/rejected": -2.96875, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 27.880631705517214, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -0.255859375, | |
| "logits/rejected": -0.265625, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.3049, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.1875, | |
| "rewards/margins": 2.015625, | |
| "rewards/rejected": -3.203125, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 19.47719144816062, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -0.353515625, | |
| "logits/rejected": -0.3046875, | |
| "logps/chosen": -258.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.2261, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -1.03125, | |
| "rewards/margins": 2.671875, | |
| "rewards/rejected": -3.703125, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 11.758314734665644, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -0.3671875, | |
| "logits/rejected": -0.259765625, | |
| "logps/chosen": -255.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.2482, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.146484375, | |
| "rewards/margins": 3.09375, | |
| "rewards/rejected": -3.234375, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 29.065207349464256, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -0.34375, | |
| "logits/rejected": -0.271484375, | |
| "logps/chosen": -253.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.2434, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.2177734375, | |
| "rewards/margins": 2.59375, | |
| "rewards/rejected": -2.8125, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 20.887703509636495, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -0.28125, | |
| "logits/rejected": -0.259765625, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.2628, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -0.60546875, | |
| "rewards/margins": 2.90625, | |
| "rewards/rejected": -3.5, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 27.32113393451712, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -0.205078125, | |
| "logits/rejected": -0.185546875, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.2504, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.373046875, | |
| "rewards/margins": 2.8125, | |
| "rewards/rejected": -3.1875, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 31.86011708190834, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -0.3984375, | |
| "logits/rejected": -0.3671875, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.2232, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -1.015625, | |
| "rewards/margins": 3.0625, | |
| "rewards/rejected": -4.09375, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 20.86967149307628, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -0.318359375, | |
| "logits/rejected": -0.28515625, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.266, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -0.244140625, | |
| "rewards/margins": 3.96875, | |
| "rewards/rejected": -4.21875, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 13.926971252489295, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -0.318359375, | |
| "logits/rejected": -0.283203125, | |
| "logps/chosen": -256.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.2425, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.91015625, | |
| "rewards/margins": 3.265625, | |
| "rewards/rejected": -4.15625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 21.465057034897768, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -0.30078125, | |
| "logits/rejected": -0.298828125, | |
| "logps/chosen": -266.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.2302, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.015625, | |
| "rewards/margins": 3.40625, | |
| "rewards/rejected": -4.40625, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 21.21945214622638, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -0.4140625, | |
| "logits/rejected": -0.302734375, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.1859, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.0859375, | |
| "rewards/margins": 3.75, | |
| "rewards/rejected": -3.84375, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 28.907492068309367, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -0.310546875, | |
| "logits/rejected": -0.228515625, | |
| "logps/chosen": -255.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.1745, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.45703125, | |
| "rewards/margins": 3.9375, | |
| "rewards/rejected": -4.375, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 23.30973121947339, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -0.359375, | |
| "logits/rejected": -0.28125, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.2707, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.494140625, | |
| "rewards/margins": 3.65625, | |
| "rewards/rejected": -4.15625, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 23.69571503944608, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -0.357421875, | |
| "logits/rejected": -0.248046875, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.1933, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.2734375, | |
| "rewards/margins": 3.6875, | |
| "rewards/rejected": -3.96875, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 29.829129154488403, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -0.419921875, | |
| "logits/rejected": -0.39453125, | |
| "logps/chosen": -258.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.1997, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.796875, | |
| "rewards/margins": 3.5, | |
| "rewards/rejected": -4.3125, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 17.97288498350235, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -0.359375, | |
| "logits/rejected": -0.251953125, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.2377, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.0390625, | |
| "rewards/margins": 3.4375, | |
| "rewards/rejected": -4.46875, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 34.21932991148693, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -0.396484375, | |
| "logits/rejected": -0.34375, | |
| "logps/chosen": -266.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.1923, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.81640625, | |
| "rewards/margins": 3.328125, | |
| "rewards/rejected": -4.15625, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 38.79444116190329, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -0.330078125, | |
| "logits/rejected": -0.1689453125, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.1612, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -0.326171875, | |
| "rewards/margins": 3.765625, | |
| "rewards/rejected": -4.09375, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 18.981506275251785, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -0.41796875, | |
| "logits/rejected": -0.3515625, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.2426, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -1.421875, | |
| "rewards/margins": 3.203125, | |
| "rewards/rejected": -4.625, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -0.287109375, | |
| "eval_logits/rejected": -0.244140625, | |
| "eval_logps/chosen": -264.0, | |
| "eval_logps/rejected": -322.0, | |
| "eval_loss": 0.24867147207260132, | |
| "eval_rewards/accuracies": 0.8928571343421936, | |
| "eval_rewards/chosen": -0.9375, | |
| "eval_rewards/margins": 4.03125, | |
| "eval_rewards/rejected": -4.96875, | |
| "eval_runtime": 12.8353, | |
| "eval_samples_per_second": 15.348, | |
| "eval_steps_per_second": 0.545, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 25.00700651524855, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -0.341796875, | |
| "logits/rejected": -0.322265625, | |
| "logps/chosen": -255.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0898, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.6640625, | |
| "rewards/margins": 4.3125, | |
| "rewards/rejected": -5.0, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 14.978780635113292, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -0.4140625, | |
| "logits/rejected": -0.2275390625, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -316.0, | |
| "loss": 0.1135, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.9140625, | |
| "rewards/margins": 4.46875, | |
| "rewards/rejected": -5.375, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 10.90198802693596, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -0.265625, | |
| "logits/rejected": -0.240234375, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.1268, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.8828125, | |
| "rewards/margins": 3.890625, | |
| "rewards/rejected": -5.78125, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 8.227239532697844, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -0.373046875, | |
| "logits/rejected": -0.328125, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.1025, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.79296875, | |
| "rewards/margins": 4.59375, | |
| "rewards/rejected": -5.375, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 17.506218154474453, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -0.380859375, | |
| "logits/rejected": -0.298828125, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -316.0, | |
| "loss": 0.0969, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.6796875, | |
| "rewards/margins": 4.8125, | |
| "rewards/rejected": -5.5, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 13.952177817698395, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -0.41796875, | |
| "logits/rejected": -0.28125, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -328.0, | |
| "loss": 0.1174, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.7578125, | |
| "rewards/margins": 5.40625, | |
| "rewards/rejected": -6.15625, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 16.052371680620837, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -0.43359375, | |
| "logits/rejected": -0.328125, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -322.0, | |
| "loss": 0.1208, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.9453125, | |
| "rewards/margins": 4.625, | |
| "rewards/rejected": -5.5625, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 20.72199242370601, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -0.37890625, | |
| "logits/rejected": -0.212890625, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -328.0, | |
| "loss": 0.1105, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.109375, | |
| "rewards/margins": 5.28125, | |
| "rewards/rejected": -6.40625, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 13.139466416035857, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -0.400390625, | |
| "logits/rejected": -0.33203125, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -326.0, | |
| "loss": 0.1076, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.4375, | |
| "rewards/margins": 5.09375, | |
| "rewards/rejected": -6.53125, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 10.075029824744835, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -0.39453125, | |
| "logits/rejected": -0.26953125, | |
| "logps/chosen": -256.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.1034, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -0.515625, | |
| "rewards/margins": 4.34375, | |
| "rewards/rejected": -4.875, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 17.19960192431196, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -0.455078125, | |
| "logits/rejected": -0.30859375, | |
| "logps/chosen": -255.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.1215, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -0.71875, | |
| "rewards/margins": 4.78125, | |
| "rewards/rejected": -5.5, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 20.37753546784737, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -0.41796875, | |
| "logits/rejected": -0.3984375, | |
| "logps/chosen": -272.0, | |
| "logps/rejected": -326.0, | |
| "loss": 0.1057, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.421875, | |
| "rewards/margins": 4.875, | |
| "rewards/rejected": -6.28125, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 7.561218450127534, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -0.333984375, | |
| "logits/rejected": -0.259765625, | |
| "logps/chosen": -270.0, | |
| "logps/rejected": -324.0, | |
| "loss": 0.1042, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.546875, | |
| "rewards/margins": 4.40625, | |
| "rewards/rejected": -5.9375, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 9.798023352470738, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -0.359375, | |
| "logits/rejected": -0.267578125, | |
| "logps/chosen": -274.0, | |
| "logps/rejected": -332.0, | |
| "loss": 0.0847, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.625, | |
| "rewards/margins": 5.34375, | |
| "rewards/rejected": -6.96875, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 7.958325164856841, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -0.439453125, | |
| "logits/rejected": -0.2412109375, | |
| "logps/chosen": -258.0, | |
| "logps/rejected": -322.0, | |
| "loss": 0.1158, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -0.48046875, | |
| "rewards/margins": 5.46875, | |
| "rewards/rejected": -5.96875, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 15.835828220330061, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -0.3828125, | |
| "logits/rejected": -0.27734375, | |
| "logps/chosen": -258.0, | |
| "logps/rejected": -326.0, | |
| "loss": 0.0681, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.03125, | |
| "rewards/margins": 5.59375, | |
| "rewards/rejected": -6.65625, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 25.36228829636491, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -0.326171875, | |
| "logits/rejected": -0.21875, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -328.0, | |
| "loss": 0.1227, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.86328125, | |
| "rewards/margins": 5.3125, | |
| "rewards/rejected": -6.15625, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 8.57335808491846, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -0.423828125, | |
| "logits/rejected": -0.287109375, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.1359, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.84375, | |
| "rewards/margins": 4.75, | |
| "rewards/rejected": -5.59375, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 24.16008654516332, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -0.41796875, | |
| "logits/rejected": -0.34765625, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.0981, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.6640625, | |
| "rewards/margins": 5.1875, | |
| "rewards/rejected": -5.875, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 6.972376633136719, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -0.408203125, | |
| "logits/rejected": -0.30859375, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -322.0, | |
| "loss": 0.1303, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.078125, | |
| "rewards/margins": 5.09375, | |
| "rewards/rejected": -6.15625, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 18.844852297813603, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -0.330078125, | |
| "logits/rejected": -0.2080078125, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -328.0, | |
| "loss": 0.1214, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.125, | |
| "rewards/margins": 5.15625, | |
| "rewards/rejected": -6.28125, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 9.000512959166052, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -0.466796875, | |
| "logits/rejected": -0.326171875, | |
| "logps/chosen": -253.0, | |
| "logps/rejected": -324.0, | |
| "loss": 0.0701, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.498046875, | |
| "rewards/margins": 5.65625, | |
| "rewards/rejected": -6.125, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 16.792602690422132, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -0.40234375, | |
| "logits/rejected": -0.318359375, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.1153, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.80078125, | |
| "rewards/margins": 5.4375, | |
| "rewards/rejected": -6.21875, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 12.357523137887897, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -0.294921875, | |
| "logits/rejected": -0.2060546875, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -326.0, | |
| "loss": 0.0971, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.55078125, | |
| "rewards/margins": 5.5625, | |
| "rewards/rejected": -6.125, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 19.84116110881825, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -0.3828125, | |
| "logits/rejected": -0.298828125, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.0897, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.56640625, | |
| "rewards/margins": 5.1875, | |
| "rewards/rejected": -5.75, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 9.128386312996932, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -0.4296875, | |
| "logits/rejected": -0.419921875, | |
| "logps/chosen": -270.0, | |
| "logps/rejected": -324.0, | |
| "loss": 0.1074, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.71875, | |
| "rewards/margins": 4.4375, | |
| "rewards/rejected": -6.15625, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 21.846512169242597, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -0.439453125, | |
| "logits/rejected": -0.310546875, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -328.0, | |
| "loss": 0.1091, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -1.5234375, | |
| "rewards/margins": 5.0, | |
| "rewards/rejected": -6.53125, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 4.3074168032135765, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -0.416015625, | |
| "logits/rejected": -0.341796875, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -336.0, | |
| "loss": 0.0898, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.4609375, | |
| "rewards/margins": 5.53125, | |
| "rewards/rejected": -7.0, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 23.755214933049356, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -0.498046875, | |
| "logits/rejected": -0.2578125, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -328.0, | |
| "loss": 0.108, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.4296875, | |
| "rewards/margins": 5.28125, | |
| "rewards/rejected": -6.71875, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 20.987597780350296, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -0.43359375, | |
| "logits/rejected": -0.30078125, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -332.0, | |
| "loss": 0.1141, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.46875, | |
| "rewards/margins": 5.6875, | |
| "rewards/rejected": -7.15625, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 10.007848002910942, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -0.455078125, | |
| "logits/rejected": -0.333984375, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -322.0, | |
| "loss": 0.1053, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.0625, | |
| "rewards/margins": 5.46875, | |
| "rewards/rejected": -6.53125, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -0.337890625, | |
| "eval_logits/rejected": -0.275390625, | |
| "eval_logps/chosen": -278.0, | |
| "eval_logps/rejected": -344.0, | |
| "eval_loss": 0.24997025728225708, | |
| "eval_rewards/accuracies": 0.9285714030265808, | |
| "eval_rewards/chosen": -2.359375, | |
| "eval_rewards/margins": 4.9375, | |
| "eval_rewards/rejected": -7.3125, | |
| "eval_runtime": 12.6509, | |
| "eval_samples_per_second": 15.572, | |
| "eval_steps_per_second": 0.553, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 18.241699866247686, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -0.37109375, | |
| "logits/rejected": -0.376953125, | |
| "logps/chosen": -272.0, | |
| "logps/rejected": -334.0, | |
| "loss": 0.0797, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.7734375, | |
| "rewards/margins": 5.46875, | |
| "rewards/rejected": -7.25, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 7.402237455250574, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -0.302734375, | |
| "logits/rejected": -0.306640625, | |
| "logps/chosen": -276.0, | |
| "logps/rejected": -338.0, | |
| "loss": 0.0779, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.0625, | |
| "rewards/margins": 5.28125, | |
| "rewards/rejected": -7.34375, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 15.213558693968412, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": -0.4296875, | |
| "logits/rejected": -0.3125, | |
| "logps/chosen": -270.0, | |
| "logps/rejected": -334.0, | |
| "loss": 0.0986, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -1.4765625, | |
| "rewards/margins": 5.375, | |
| "rewards/rejected": -6.875, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 6.558513252601039, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -0.294921875, | |
| "logits/rejected": -0.23828125, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -330.0, | |
| "loss": 0.0463, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.98828125, | |
| "rewards/margins": 6.0, | |
| "rewards/rejected": -6.96875, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 2.7928900841374116, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -0.3984375, | |
| "logits/rejected": -0.3046875, | |
| "logps/chosen": -266.0, | |
| "logps/rejected": -336.0, | |
| "loss": 0.0371, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.4140625, | |
| "rewards/margins": 6.1875, | |
| "rewards/rejected": -7.59375, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 3.284910793509464, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -0.4765625, | |
| "logits/rejected": -0.30078125, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -336.0, | |
| "loss": 0.0587, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.78515625, | |
| "rewards/margins": 6.5625, | |
| "rewards/rejected": -7.34375, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 7.644866358289651, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -0.392578125, | |
| "logits/rejected": -0.3125, | |
| "logps/chosen": -270.0, | |
| "logps/rejected": -336.0, | |
| "loss": 0.0464, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.3203125, | |
| "rewards/margins": 6.125, | |
| "rewards/rejected": -7.4375, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 3.4166784290048615, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -0.4609375, | |
| "logits/rejected": -0.2578125, | |
| "logps/chosen": -253.0, | |
| "logps/rejected": -332.0, | |
| "loss": 0.0418, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.80078125, | |
| "rewards/margins": 6.375, | |
| "rewards/rejected": -7.1875, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 18.486708233609825, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -0.4921875, | |
| "logits/rejected": -0.361328125, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -342.0, | |
| "loss": 0.082, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.5234375, | |
| "rewards/margins": 6.34375, | |
| "rewards/rejected": -7.84375, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 7.094665874891602, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -0.4609375, | |
| "logits/rejected": -0.275390625, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -338.0, | |
| "loss": 0.0467, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.25, | |
| "rewards/margins": 6.5625, | |
| "rewards/rejected": -7.8125, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 12.579518491613028, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -0.46484375, | |
| "logits/rejected": -0.40625, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -336.0, | |
| "loss": 0.0498, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.625, | |
| "rewards/margins": 6.0, | |
| "rewards/rejected": -7.625, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 3.8020122383533996, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -0.396484375, | |
| "logits/rejected": -0.30078125, | |
| "logps/chosen": -272.0, | |
| "logps/rejected": -346.0, | |
| "loss": 0.053, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.7734375, | |
| "rewards/margins": 6.5625, | |
| "rewards/rejected": -8.3125, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 7.2662422494862255, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": -0.47265625, | |
| "logits/rejected": -0.361328125, | |
| "logps/chosen": -272.0, | |
| "logps/rejected": -354.0, | |
| "loss": 0.0467, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.96875, | |
| "rewards/margins": 6.5625, | |
| "rewards/rejected": -8.5625, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 9.44464858583, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": -0.353515625, | |
| "logits/rejected": -0.240234375, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -340.0, | |
| "loss": 0.0353, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.390625, | |
| "rewards/margins": 6.3125, | |
| "rewards/rejected": -7.71875, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 6.695861150174956, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.427734375, | |
| "logits/rejected": -0.28125, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -342.0, | |
| "loss": 0.0467, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.3984375, | |
| "rewards/margins": 6.3125, | |
| "rewards/rejected": -7.71875, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 5.8132707822864385, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -0.42578125, | |
| "logits/rejected": -0.384765625, | |
| "logps/chosen": -276.0, | |
| "logps/rejected": -348.0, | |
| "loss": 0.0304, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.21875, | |
| "rewards/margins": 6.375, | |
| "rewards/rejected": -8.625, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 31.743459957220768, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -0.34375, | |
| "logits/rejected": -0.337890625, | |
| "logps/chosen": -282.0, | |
| "logps/rejected": -356.0, | |
| "loss": 0.0535, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.328125, | |
| "rewards/margins": 6.25, | |
| "rewards/rejected": -8.625, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 14.397287073427687, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -0.4140625, | |
| "logits/rejected": -0.2265625, | |
| "logps/chosen": -266.0, | |
| "logps/rejected": -344.0, | |
| "loss": 0.041, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.15625, | |
| "rewards/margins": 6.59375, | |
| "rewards/rejected": -7.75, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 7.8843681357511155, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -0.43359375, | |
| "logits/rejected": -0.349609375, | |
| "logps/chosen": -266.0, | |
| "logps/rejected": -340.0, | |
| "loss": 0.0386, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.5625, | |
| "rewards/margins": 6.5, | |
| "rewards/rejected": -8.0625, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 21.116338835976237, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -0.4921875, | |
| "logits/rejected": -0.30078125, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -344.0, | |
| "loss": 0.0537, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.4140625, | |
| "rewards/margins": 6.53125, | |
| "rewards/rejected": -7.96875, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 8.528280731121223, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -0.4921875, | |
| "logits/rejected": -0.44140625, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -342.0, | |
| "loss": 0.0423, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.40625, | |
| "rewards/margins": 6.59375, | |
| "rewards/rejected": -8.0, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 23.00258608838155, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -0.435546875, | |
| "logits/rejected": -0.36328125, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -334.0, | |
| "loss": 0.0359, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.875, | |
| "rewards/margins": 6.59375, | |
| "rewards/rejected": -7.46875, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 10.971091921904996, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -0.439453125, | |
| "logits/rejected": -0.23828125, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -340.0, | |
| "loss": 0.0792, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.0625, | |
| "rewards/margins": 5.5, | |
| "rewards/rejected": -7.5625, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 23.47882603701124, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -0.44921875, | |
| "logits/rejected": -0.369140625, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -330.0, | |
| "loss": 0.0565, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.1015625, | |
| "rewards/margins": 5.84375, | |
| "rewards/rejected": -6.9375, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 18.207194355135897, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -0.54296875, | |
| "logits/rejected": -0.33984375, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -340.0, | |
| "loss": 0.0406, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.0390625, | |
| "rewards/margins": 6.6875, | |
| "rewards/rejected": -7.71875, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 3.414112525400886, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -0.353515625, | |
| "logits/rejected": -0.322265625, | |
| "logps/chosen": -276.0, | |
| "logps/rejected": -342.0, | |
| "loss": 0.0331, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.671875, | |
| "rewards/margins": 6.21875, | |
| "rewards/rejected": -7.875, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 3.645069513448636, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -0.41015625, | |
| "logits/rejected": -0.341796875, | |
| "logps/chosen": -258.0, | |
| "logps/rejected": -350.0, | |
| "loss": 0.0609, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.109375, | |
| "rewards/margins": 7.5, | |
| "rewards/rejected": -8.625, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 20.057840037106633, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -0.58203125, | |
| "logits/rejected": -0.296875, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -342.0, | |
| "loss": 0.0551, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.234375, | |
| "rewards/margins": 6.75, | |
| "rewards/rejected": -8.0, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 5.344958710263874, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -0.361328125, | |
| "logits/rejected": -0.228515625, | |
| "logps/chosen": -274.0, | |
| "logps/rejected": -348.0, | |
| "loss": 0.0422, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.9375, | |
| "rewards/margins": 6.375, | |
| "rewards/rejected": -8.3125, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 25.83356121662788, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -0.44140625, | |
| "logits/rejected": -0.33984375, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -340.0, | |
| "loss": 0.0397, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.265625, | |
| "rewards/margins": 6.4375, | |
| "rewards/rejected": -7.71875, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 11.37077001510433, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": -0.400390625, | |
| "logits/rejected": -0.31640625, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -346.0, | |
| "loss": 0.0539, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.6640625, | |
| "rewards/margins": 6.375, | |
| "rewards/rejected": -8.0625, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -0.35546875, | |
| "eval_logits/rejected": -0.283203125, | |
| "eval_logps/chosen": -274.0, | |
| "eval_logps/rejected": -350.0, | |
| "eval_loss": 0.2567913234233856, | |
| "eval_rewards/accuracies": 0.9285714030265808, | |
| "eval_rewards/chosen": -1.9375, | |
| "eval_rewards/margins": 5.9375, | |
| "eval_rewards/rejected": -7.875, | |
| "eval_runtime": 14.5298, | |
| "eval_samples_per_second": 13.558, | |
| "eval_steps_per_second": 0.482, | |
| "step": 939 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 939, | |
| "total_flos": 0.0, | |
| "train_loss": 0.1578371215044373, | |
| "train_runtime": 5401.3477, | |
| "train_samples_per_second": 5.554, | |
| "train_steps_per_second": 0.174 | |
| } | |
| ], | |
| "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 | |
| } | |