| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 939, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.003194888178913738, | |
| "grad_norm": 54.81250463683086, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -0.044189453125, | |
| "logits/rejected": -0.09228515625, | |
| "logps/chosen": -308.0, | |
| "logps/rejected": -134.0, | |
| "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": 53.95046919259797, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -0.3828125, | |
| "logits/rejected": -0.40234375, | |
| "logps/chosen": -320.0, | |
| "logps/rejected": -136.0, | |
| "loss": 0.6988, | |
| "rewards/accuracies": 0.3472222089767456, | |
| "rewards/chosen": 0.001068115234375, | |
| "rewards/margins": 0.00933837890625, | |
| "rewards/rejected": -0.00836181640625, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 49.585771413361975, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -0.390625, | |
| "logits/rejected": -0.41796875, | |
| "logps/chosen": -362.0, | |
| "logps/rejected": -144.0, | |
| "loss": 0.6667, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.02001953125, | |
| "rewards/margins": 0.05908203125, | |
| "rewards/rejected": -0.0390625, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 42.29466312552128, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -0.5625, | |
| "logits/rejected": -0.55078125, | |
| "logps/chosen": -350.0, | |
| "logps/rejected": -148.0, | |
| "loss": 0.5623, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.1181640625, | |
| "rewards/margins": 0.328125, | |
| "rewards/rejected": -0.2109375, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 30.439085013256786, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -0.294921875, | |
| "logits/rejected": -0.45703125, | |
| "logps/chosen": -350.0, | |
| "logps/rejected": -144.0, | |
| "loss": 0.411, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.2177734375, | |
| "rewards/margins": 0.73828125, | |
| "rewards/rejected": -0.5234375, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 12.799318586746557, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -0.3046875, | |
| "logits/rejected": -0.44921875, | |
| "logps/chosen": -290.0, | |
| "logps/rejected": -151.0, | |
| "loss": 0.213, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.251953125, | |
| "rewards/margins": 1.5234375, | |
| "rewards/rejected": -1.2734375, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 6.50898280557731, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -0.302734375, | |
| "logits/rejected": -0.5078125, | |
| "logps/chosen": -298.0, | |
| "logps/rejected": -170.0, | |
| "loss": 0.095, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.373046875, | |
| "rewards/margins": 2.8125, | |
| "rewards/rejected": -2.4375, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 6.930144113115375, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -0.23828125, | |
| "logits/rejected": -0.44140625, | |
| "logps/chosen": -288.0, | |
| "logps/rejected": -171.0, | |
| "loss": 0.0307, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.9765625, | |
| "rewards/margins": 4.90625, | |
| "rewards/rejected": -3.90625, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 4.074558805260814, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -0.275390625, | |
| "logits/rejected": -0.296875, | |
| "logps/chosen": -280.0, | |
| "logps/rejected": -188.0, | |
| "loss": 0.0156, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.296875, | |
| "rewards/margins": 6.1875, | |
| "rewards/rejected": -4.875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 3.3281914960233805, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -0.142578125, | |
| "logits/rejected": -0.31640625, | |
| "logps/chosen": -239.0, | |
| "logps/rejected": -189.0, | |
| "loss": 0.0115, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.46875, | |
| "rewards/margins": 7.21875, | |
| "rewards/rejected": -5.75, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 0.2013303942361548, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -0.169921875, | |
| "logits/rejected": -0.306640625, | |
| "logps/chosen": -338.0, | |
| "logps/rejected": -217.0, | |
| "loss": 0.0037, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.3984375, | |
| "rewards/margins": 8.25, | |
| "rewards/rejected": -6.84375, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 2.144766213627776, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -0.267578125, | |
| "logits/rejected": -0.2890625, | |
| "logps/chosen": -292.0, | |
| "logps/rejected": -216.0, | |
| "loss": 0.004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.265625, | |
| "rewards/margins": 8.6875, | |
| "rewards/rejected": -7.4375, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 0.1453071677708086, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -0.12890625, | |
| "logits/rejected": -0.20703125, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -201.0, | |
| "loss": 0.0032, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.453125, | |
| "rewards/margins": 9.0, | |
| "rewards/rejected": -7.5625, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 0.31129685645313604, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -0.125, | |
| "logits/rejected": -0.412109375, | |
| "logps/chosen": -304.0, | |
| "logps/rejected": -226.0, | |
| "loss": 0.0036, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.3125, | |
| "rewards/margins": 9.625, | |
| "rewards/rejected": -8.3125, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 1.4440341904660572, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -0.1650390625, | |
| "logits/rejected": -0.1845703125, | |
| "logps/chosen": -338.0, | |
| "logps/rejected": -222.0, | |
| "loss": 0.0025, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.25, | |
| "rewards/margins": 9.8125, | |
| "rewards/rejected": -8.625, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 0.25222347097127684, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -0.08935546875, | |
| "logits/rejected": -0.2470703125, | |
| "logps/chosen": -300.0, | |
| "logps/rejected": -224.0, | |
| "loss": 0.0062, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.640625, | |
| "rewards/margins": 10.3125, | |
| "rewards/rejected": -8.6875, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 0.04551534803700279, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -0.1279296875, | |
| "logits/rejected": -0.2314453125, | |
| "logps/chosen": -296.0, | |
| "logps/rejected": -240.0, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.6953125, | |
| "rewards/margins": 10.75, | |
| "rewards/rejected": -9.0625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 1.0539451781035367, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -0.1826171875, | |
| "logits/rejected": -0.201171875, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -232.0, | |
| "loss": 0.0027, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.3046875, | |
| "rewards/margins": 10.5625, | |
| "rewards/rejected": -9.25, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 0.008991111862560141, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -0.138671875, | |
| "logits/rejected": -0.318359375, | |
| "logps/chosen": -322.0, | |
| "logps/rejected": -237.0, | |
| "loss": 0.005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.2890625, | |
| "rewards/margins": 11.0, | |
| "rewards/rejected": -9.6875, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 0.040925930935537355, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -0.1015625, | |
| "logits/rejected": -0.259765625, | |
| "logps/chosen": -290.0, | |
| "logps/rejected": -240.0, | |
| "loss": 0.0017, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.953125, | |
| "rewards/margins": 11.4375, | |
| "rewards/rejected": -9.5, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 0.05836199428865348, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -0.150390625, | |
| "logits/rejected": -0.294921875, | |
| "logps/chosen": -340.0, | |
| "logps/rejected": -243.0, | |
| "loss": 0.0048, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.828125, | |
| "rewards/margins": 11.875, | |
| "rewards/rejected": -10.0625, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 0.009307515698978714, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -0.2470703125, | |
| "logits/rejected": -0.150390625, | |
| "logps/chosen": -300.0, | |
| "logps/rejected": -241.0, | |
| "loss": 0.0037, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.984375, | |
| "rewards/margins": 11.875, | |
| "rewards/rejected": -9.875, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 0.06477340975400131, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -0.287109375, | |
| "logits/rejected": -0.40234375, | |
| "logps/chosen": -348.0, | |
| "logps/rejected": -244.0, | |
| "loss": 0.0024, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.078125, | |
| "rewards/margins": 12.0625, | |
| "rewards/rejected": -10.0, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 0.12992102662346627, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -0.2060546875, | |
| "logits/rejected": -0.23046875, | |
| "logps/chosen": -270.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 12.1875, | |
| "rewards/rejected": -10.0, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 0.6536293524892515, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -0.1337890625, | |
| "logits/rejected": -0.25, | |
| "logps/chosen": -312.0, | |
| "logps/rejected": -245.0, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 12.5625, | |
| "rewards/rejected": -10.5625, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 0.3257530411474732, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -0.28125, | |
| "logits/rejected": -0.205078125, | |
| "logps/chosen": -316.0, | |
| "logps/rejected": -240.0, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.78125, | |
| "rewards/margins": 12.0625, | |
| "rewards/rejected": -10.3125, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 0.9726896506176986, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -0.166015625, | |
| "logits/rejected": -0.13671875, | |
| "logps/chosen": -302.0, | |
| "logps/rejected": -254.0, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.984375, | |
| "rewards/margins": 12.5625, | |
| "rewards/rejected": -10.5625, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 0.041715953039775905, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -0.15625, | |
| "logits/rejected": -0.263671875, | |
| "logps/chosen": -278.0, | |
| "logps/rejected": -246.0, | |
| "loss": 0.0046, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.03125, | |
| "rewards/margins": 12.25, | |
| "rewards/rejected": -10.25, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 1.3367486705088796, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -0.014404296875, | |
| "logits/rejected": -0.1171875, | |
| "logps/chosen": -272.0, | |
| "logps/rejected": -238.0, | |
| "loss": 0.0023, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.9296875, | |
| "rewards/margins": 12.125, | |
| "rewards/rejected": -10.25, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 0.02661420959172122, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -0.2236328125, | |
| "logits/rejected": -0.193359375, | |
| "logps/chosen": -320.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0016, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.8046875, | |
| "rewards/margins": 12.875, | |
| "rewards/rejected": -11.0625, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 0.2014648330486529, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -0.019775390625, | |
| "logits/rejected": -0.23046875, | |
| "logps/chosen": -302.0, | |
| "logps/rejected": -251.0, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.234375, | |
| "rewards/margins": 13.25, | |
| "rewards/rejected": -11.0625, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 0.02404893688543294, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -0.09814453125, | |
| "logits/rejected": -0.1494140625, | |
| "logps/chosen": -272.0, | |
| "logps/rejected": -247.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.984375, | |
| "rewards/margins": 13.3125, | |
| "rewards/rejected": -11.3125, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -0.0751953125, | |
| "eval_logits/rejected": -0.248046875, | |
| "eval_logps/chosen": -274.0, | |
| "eval_logps/rejected": -253.0, | |
| "eval_loss": 0.00043513119453564286, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": 2.015625, | |
| "eval_rewards/margins": 13.125, | |
| "eval_rewards/rejected": -11.0625, | |
| "eval_runtime": 18.0315, | |
| "eval_samples_per_second": 11.092, | |
| "eval_steps_per_second": 0.388, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 0.019269598789051393, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -0.058349609375, | |
| "logits/rejected": -0.1845703125, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -237.0, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.234375, | |
| "rewards/margins": 12.875, | |
| "rewards/rejected": -10.6875, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 0.040456088367899996, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -0.12890625, | |
| "logits/rejected": -0.2041015625, | |
| "logps/chosen": -296.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 13.5625, | |
| "rewards/rejected": -11.125, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 0.0007001885169442494, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -0.0771484375, | |
| "logits/rejected": -0.2080078125, | |
| "logps/chosen": -286.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.265625, | |
| "rewards/margins": 13.625, | |
| "rewards/rejected": -11.375, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 0.0007684508091218239, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -0.0712890625, | |
| "logits/rejected": -0.09326171875, | |
| "logps/chosen": -290.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0016, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.28125, | |
| "rewards/margins": 14.1875, | |
| "rewards/rejected": -11.9375, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 0.05722322451953868, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -0.02978515625, | |
| "logits/rejected": -0.228515625, | |
| "logps/chosen": -247.0, | |
| "logps/rejected": -249.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.828125, | |
| "rewards/margins": 13.25, | |
| "rewards/rejected": -11.4375, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 0.15262674343512006, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -0.1513671875, | |
| "logits/rejected": -0.1591796875, | |
| "logps/chosen": -352.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.953125, | |
| "rewards/margins": 13.625, | |
| "rewards/rejected": -11.6875, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 0.004924918876145697, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -0.039794921875, | |
| "logits/rejected": -0.1728515625, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -252.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 13.3125, | |
| "rewards/rejected": -11.0, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 0.020599626070146905, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -0.1796875, | |
| "logits/rejected": -0.2197265625, | |
| "logps/chosen": -324.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.9375, | |
| "rewards/margins": 13.5625, | |
| "rewards/rejected": -11.625, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 0.0028880458746553105, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -0.1201171875, | |
| "logits/rejected": -0.2197265625, | |
| "logps/chosen": -290.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.453125, | |
| "rewards/margins": 13.9375, | |
| "rewards/rejected": -11.5, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 0.004250020172600707, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -0.10107421875, | |
| "logits/rejected": -0.20703125, | |
| "logps/chosen": -292.0, | |
| "logps/rejected": -248.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 14.0, | |
| "rewards/rejected": -11.5625, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 0.00582186798031753, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -0.04736328125, | |
| "logits/rejected": -0.1376953125, | |
| "logps/chosen": -258.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 1.8125, | |
| "rewards/margins": 13.5, | |
| "rewards/rejected": -11.6875, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 0.014863585988386751, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -0.11474609375, | |
| "logits/rejected": -0.0615234375, | |
| "logps/chosen": -288.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 14.1875, | |
| "rewards/rejected": -11.875, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 0.5550627550740515, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -0.11474609375, | |
| "logits/rejected": -0.1455078125, | |
| "logps/chosen": -306.0, | |
| "logps/rejected": -248.0, | |
| "loss": 0.0034, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 14.0, | |
| "rewards/rejected": -11.625, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 0.040580102264137226, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -0.08935546875, | |
| "logits/rejected": -0.1552734375, | |
| "logps/chosen": -304.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.46875, | |
| "rewards/margins": 13.6875, | |
| "rewards/rejected": -11.25, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 0.008803992788275688, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -0.07080078125, | |
| "logits/rejected": -0.072265625, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.65625, | |
| "rewards/margins": 14.125, | |
| "rewards/rejected": -11.4375, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 0.013909569749826874, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -0.099609375, | |
| "logits/rejected": -0.1123046875, | |
| "logps/chosen": -250.0, | |
| "logps/rejected": -250.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.515625, | |
| "rewards/margins": 14.0625, | |
| "rewards/rejected": -11.5625, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 0.1150271151982318, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -0.115234375, | |
| "logits/rejected": -0.1396484375, | |
| "logps/chosen": -326.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.484375, | |
| "rewards/margins": 14.25, | |
| "rewards/rejected": -11.75, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 0.19996849325172428, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -0.1728515625, | |
| "logits/rejected": -0.2060546875, | |
| "logps/chosen": -318.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.5, | |
| "rewards/margins": 15.0, | |
| "rewards/rejected": -12.5, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 0.037794629543665755, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -0.1123046875, | |
| "logits/rejected": -0.091796875, | |
| "logps/chosen": -340.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.375, | |
| "rewards/margins": 14.125, | |
| "rewards/rejected": -11.75, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 0.015480400669133856, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -0.09033203125, | |
| "logits/rejected": -0.1357421875, | |
| "logps/chosen": -278.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 14.4375, | |
| "rewards/rejected": -11.875, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 0.008138448027567313, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -0.318359375, | |
| "logits/rejected": -0.2158203125, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -253.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.359375, | |
| "rewards/margins": 14.375, | |
| "rewards/rejected": -12.0, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 0.0014973167627475115, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -0.19140625, | |
| "logits/rejected": -0.314453125, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 14.5625, | |
| "rewards/rejected": -12.25, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 0.002044088145558404, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -0.0810546875, | |
| "logits/rejected": -0.061279296875, | |
| "logps/chosen": -328.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.21875, | |
| "rewards/margins": 14.625, | |
| "rewards/rejected": -12.4375, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 0.01037121413667403, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -0.08935546875, | |
| "logits/rejected": -0.14453125, | |
| "logps/chosen": -270.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 14.5625, | |
| "rewards/rejected": -12.125, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 0.19904568142685894, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -0.10498046875, | |
| "logits/rejected": -0.28515625, | |
| "logps/chosen": -346.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 15.0625, | |
| "rewards/rejected": -12.625, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 0.18769077833185094, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -0.1640625, | |
| "logits/rejected": -0.2314453125, | |
| "logps/chosen": -278.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 15.375, | |
| "rewards/rejected": -13.0625, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 0.0032252350206625866, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -0.1875, | |
| "logits/rejected": -0.1748046875, | |
| "logps/chosen": -298.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.265625, | |
| "rewards/margins": 14.4375, | |
| "rewards/rejected": -12.1875, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 0.05828221948192852, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -0.1171875, | |
| "logits/rejected": -0.2373046875, | |
| "logps/chosen": -348.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 14.625, | |
| "rewards/rejected": -12.3125, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 0.053966529810810196, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -0.11572265625, | |
| "logits/rejected": -0.1611328125, | |
| "logps/chosen": -282.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.765625, | |
| "rewards/margins": 15.25, | |
| "rewards/rejected": -12.5, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 0.0021271751622448315, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -0.078125, | |
| "logits/rejected": -0.11328125, | |
| "logps/chosen": -284.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 14.625, | |
| "rewards/rejected": -12.3125, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 0.13313793603466698, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -0.119140625, | |
| "logits/rejected": -0.271484375, | |
| "logps/chosen": -302.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0018, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 14.8125, | |
| "rewards/rejected": -12.1875, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -0.046142578125, | |
| "eval_logits/rejected": -0.20703125, | |
| "eval_logps/chosen": -272.0, | |
| "eval_logps/rejected": -264.0, | |
| "eval_loss": 0.000274659541901201, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": 2.375, | |
| "eval_rewards/margins": 14.625, | |
| "eval_rewards/rejected": -12.25, | |
| "eval_runtime": 18.5388, | |
| "eval_samples_per_second": 10.788, | |
| "eval_steps_per_second": 0.378, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 0.0023706227049919383, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -0.05029296875, | |
| "logits/rejected": -0.036865234375, | |
| "logps/chosen": -282.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 14.625, | |
| "rewards/rejected": -12.25, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 0.0014065175603586702, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -0.1767578125, | |
| "logits/rejected": -0.1953125, | |
| "logps/chosen": -318.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.71875, | |
| "rewards/margins": 15.375, | |
| "rewards/rejected": -12.625, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 0.00316511389636026, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": 0.08740234375, | |
| "logits/rejected": -0.024169921875, | |
| "logps/chosen": -288.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.421875, | |
| "rewards/margins": 14.9375, | |
| "rewards/rejected": -12.5, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 0.0038704769009460477, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -0.2158203125, | |
| "logits/rejected": -0.25390625, | |
| "logps/chosen": -346.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 15.25, | |
| "rewards/rejected": -12.625, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 0.023730689357337867, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -0.07373046875, | |
| "logits/rejected": -0.255859375, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.484375, | |
| "rewards/margins": 14.6875, | |
| "rewards/rejected": -12.25, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 0.011645737272854386, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -0.0625, | |
| "logits/rejected": -0.212890625, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 15.0625, | |
| "rewards/rejected": -12.6875, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 0.12674646282529275, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -0.158203125, | |
| "logits/rejected": -0.23046875, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.46875, | |
| "rewards/margins": 14.625, | |
| "rewards/rejected": -12.1875, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 0.05357739138258823, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -0.1396484375, | |
| "logits/rejected": -0.197265625, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -252.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.46875, | |
| "rewards/margins": 14.375, | |
| "rewards/rejected": -11.9375, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 0.024908852922529832, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -0.20703125, | |
| "logits/rejected": -0.20703125, | |
| "logps/chosen": -306.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.46875, | |
| "rewards/margins": 15.0625, | |
| "rewards/rejected": -12.625, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 0.002772591639076003, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -0.03173828125, | |
| "logits/rejected": -0.10107421875, | |
| "logps/chosen": -264.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.65625, | |
| "rewards/margins": 15.0, | |
| "rewards/rejected": -12.375, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 0.008649309371085316, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -0.10498046875, | |
| "logits/rejected": -0.0458984375, | |
| "logps/chosen": -340.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 15.8125, | |
| "rewards/rejected": -13.25, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 0.020922313798910405, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -0.1396484375, | |
| "logits/rejected": -0.12255859375, | |
| "logps/chosen": -296.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 15.75, | |
| "rewards/rejected": -13.125, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 0.3141667738750178, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": 0.0517578125, | |
| "logits/rejected": -0.2255859375, | |
| "logps/chosen": -238.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 15.1875, | |
| "rewards/rejected": -12.5625, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 0.4721965738839435, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": 0.034423828125, | |
| "logits/rejected": 0.014404296875, | |
| "logps/chosen": -282.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.671875, | |
| "rewards/margins": 15.625, | |
| "rewards/rejected": -13.0, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 0.0003171684755086141, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -0.1689453125, | |
| "logits/rejected": -0.038330078125, | |
| "logps/chosen": -292.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.359375, | |
| "rewards/margins": 15.625, | |
| "rewards/rejected": -13.25, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 0.004043689913700047, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -0.0267333984375, | |
| "logits/rejected": -0.1396484375, | |
| "logps/chosen": -284.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 14.875, | |
| "rewards/rejected": -12.625, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 0.003995150076889529, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -0.12109375, | |
| "logits/rejected": -0.1455078125, | |
| "logps/chosen": -314.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.09375, | |
| "rewards/margins": 15.375, | |
| "rewards/rejected": -13.25, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 0.00017060355418111367, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -0.10546875, | |
| "logits/rejected": -0.1796875, | |
| "logps/chosen": -278.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 15.4375, | |
| "rewards/rejected": -13.125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 0.006000422750795018, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -0.177734375, | |
| "logits/rejected": -0.2294921875, | |
| "logps/chosen": -302.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.484375, | |
| "rewards/margins": 15.75, | |
| "rewards/rejected": -13.3125, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 0.0022350286043502548, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -0.11181640625, | |
| "logits/rejected": -0.169921875, | |
| "logps/chosen": -318.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.53125, | |
| "rewards/margins": 15.625, | |
| "rewards/rejected": -13.125, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 0.0068241359959056724, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -0.09423828125, | |
| "logits/rejected": -0.125, | |
| "logps/chosen": -274.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 15.5625, | |
| "rewards/rejected": -13.1875, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 0.0033702221294609085, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -0.01226806640625, | |
| "logits/rejected": -0.0888671875, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.71875, | |
| "rewards/margins": 15.0, | |
| "rewards/rejected": -12.3125, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 0.02851003983496184, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -0.031005859375, | |
| "logits/rejected": -0.0281982421875, | |
| "logps/chosen": -276.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.609375, | |
| "rewards/margins": 15.5625, | |
| "rewards/rejected": -12.9375, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 2.39196902183789, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -0.06494140625, | |
| "logits/rejected": -0.07666015625, | |
| "logps/chosen": -308.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.703125, | |
| "rewards/margins": 15.3125, | |
| "rewards/rejected": -12.625, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 0.023936662776166576, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -0.10205078125, | |
| "logits/rejected": -0.1640625, | |
| "logps/chosen": -312.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.765625, | |
| "rewards/margins": 15.75, | |
| "rewards/rejected": -12.9375, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 0.013479278710533628, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -0.044921875, | |
| "logits/rejected": -0.13671875, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 15.5625, | |
| "rewards/rejected": -13.0, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 0.0009673808962921506, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -0.166015625, | |
| "logits/rejected": -0.052490234375, | |
| "logps/chosen": -340.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 15.75, | |
| "rewards/rejected": -13.125, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 0.00409502303119149, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -0.05322265625, | |
| "logits/rejected": -0.0791015625, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.28125, | |
| "rewards/margins": 15.1875, | |
| "rewards/rejected": -12.875, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 0.010091723690541614, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -0.01397705078125, | |
| "logits/rejected": -0.038330078125, | |
| "logps/chosen": -272.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.796875, | |
| "rewards/margins": 15.4375, | |
| "rewards/rejected": -12.625, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 0.24700675544339368, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -0.07080078125, | |
| "logits/rejected": -0.2314453125, | |
| "logps/chosen": -300.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 15.125, | |
| "rewards/rejected": -12.5625, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 0.0007465859295131363, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": 0.0390625, | |
| "logits/rejected": -0.033447265625, | |
| "logps/chosen": -286.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 15.3125, | |
| "rewards/rejected": -12.875, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -0.0235595703125, | |
| "eval_logits/rejected": -0.177734375, | |
| "eval_logps/chosen": -270.0, | |
| "eval_logps/rejected": -270.0, | |
| "eval_loss": 0.00022739515407010913, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": 2.4375, | |
| "eval_rewards/margins": 15.1875, | |
| "eval_rewards/rejected": -12.75, | |
| "eval_runtime": 18.1716, | |
| "eval_samples_per_second": 11.006, | |
| "eval_steps_per_second": 0.385, | |
| "step": 939 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 939, | |
| "total_flos": 0.0, | |
| "train_loss": 0.02960524848324539, | |
| "train_runtime": 6985.1453, | |
| "train_samples_per_second": 4.295, | |
| "train_steps_per_second": 0.134 | |
| } | |
| ], | |
| "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 | |
| } | |