| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9994767137624281, | |
| "eval_steps": 100, | |
| "global_step": 955, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -2.4445395469665527, | |
| "logits/rejected": -2.388306140899658, | |
| "logps/chosen": -221.89984130859375, | |
| "logps/rejected": -197.11672973632812, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 5.208333333333334e-07, | |
| "logits/chosen": -2.385066509246826, | |
| "logits/rejected": -2.237809181213379, | |
| "logps/chosen": -315.8031005859375, | |
| "logps/rejected": -248.2730255126953, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.4444444477558136, | |
| "rewards/chosen": 8.269523277704138e-06, | |
| "rewards/margins": -0.00012648198753595352, | |
| "rewards/rejected": 0.00013475156447384506, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.0416666666666667e-06, | |
| "logits/chosen": -2.166614294052124, | |
| "logits/rejected": -2.138037919998169, | |
| "logps/chosen": -230.2421112060547, | |
| "logps/rejected": -215.0667724609375, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.00011211812670808285, | |
| "rewards/margins": -0.00017659642617218196, | |
| "rewards/rejected": 6.44782921881415e-05, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.5625e-06, | |
| "logits/chosen": -2.359757900238037, | |
| "logits/rejected": -2.239750623703003, | |
| "logps/chosen": -287.16546630859375, | |
| "logps/rejected": -250.483154296875, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": 0.0006519377348013222, | |
| "rewards/margins": -0.000464016105979681, | |
| "rewards/rejected": 0.0011159538989886642, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.0833333333333334e-06, | |
| "logits/chosen": -2.332831859588623, | |
| "logits/rejected": -2.2222561836242676, | |
| "logps/chosen": -268.5550231933594, | |
| "logps/rejected": -245.136474609375, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.0009799805702641606, | |
| "rewards/margins": 0.00043143361108377576, | |
| "rewards/rejected": 0.0005485471338033676, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.604166666666667e-06, | |
| "logits/chosen": -2.3112823963165283, | |
| "logits/rejected": -2.315413236618042, | |
| "logps/chosen": -252.92160034179688, | |
| "logps/rejected": -255.1580352783203, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.0017378001939505339, | |
| "rewards/margins": 0.000829376163892448, | |
| "rewards/rejected": 0.0009084242628887296, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.125e-06, | |
| "logits/chosen": -2.2834067344665527, | |
| "logits/rejected": -2.2284858226776123, | |
| "logps/chosen": -261.4151306152344, | |
| "logps/rejected": -252.23080444335938, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.0034083526115864515, | |
| "rewards/margins": 0.0015303167747333646, | |
| "rewards/rejected": 0.0018780359532684088, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 3.6458333333333333e-06, | |
| "logits/chosen": -2.319131374359131, | |
| "logits/rejected": -2.1937270164489746, | |
| "logps/chosen": -305.9224548339844, | |
| "logps/rejected": -275.6637268066406, | |
| "loss": 0.6905, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.010511571541428566, | |
| "rewards/margins": 0.006470891647040844, | |
| "rewards/rejected": 0.004040678963065147, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.166666666666667e-06, | |
| "logits/chosen": -2.155569553375244, | |
| "logits/rejected": -1.9963390827178955, | |
| "logps/chosen": -284.39007568359375, | |
| "logps/rejected": -246.76455688476562, | |
| "loss": 0.6876, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.014133408665657043, | |
| "rewards/margins": 0.012079447507858276, | |
| "rewards/rejected": 0.002053960459306836, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 4.6875000000000004e-06, | |
| "logits/chosen": -2.2905330657958984, | |
| "logits/rejected": -2.254070520401001, | |
| "logps/chosen": -289.9971008300781, | |
| "logps/rejected": -273.9203796386719, | |
| "loss": 0.6847, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.032837510108947754, | |
| "rewards/margins": 0.019895007833838463, | |
| "rewards/rejected": 0.012942495755851269, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.9997324926814375e-06, | |
| "logits/chosen": -2.243922710418701, | |
| "logits/rejected": -2.2560582160949707, | |
| "logps/chosen": -273.7684326171875, | |
| "logps/rejected": -287.9248962402344, | |
| "loss": 0.6812, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.03909907117486, | |
| "rewards/margins": 0.024433141574263573, | |
| "rewards/rejected": 0.014665926806628704, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "eval_logits/chosen": -2.251671075820923, | |
| "eval_logits/rejected": -2.194241762161255, | |
| "eval_logps/chosen": -245.63218688964844, | |
| "eval_logps/rejected": -250.59286499023438, | |
| "eval_loss": 0.6787430644035339, | |
| "eval_rewards/accuracies": 0.69921875, | |
| "eval_rewards/chosen": 0.04521488770842552, | |
| "eval_rewards/margins": 0.033204685896635056, | |
| "eval_rewards/rejected": 0.012010199017822742, | |
| "eval_runtime": 100.4664, | |
| "eval_samples_per_second": 19.907, | |
| "eval_steps_per_second": 0.319, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 4.996723692767927e-06, | |
| "logits/chosen": -2.2775979042053223, | |
| "logits/rejected": -2.168391704559326, | |
| "logps/chosen": -244.2776641845703, | |
| "logps/rejected": -219.0230712890625, | |
| "loss": 0.6751, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.03759222850203514, | |
| "rewards/margins": 0.03267201781272888, | |
| "rewards/rejected": 0.004920212086290121, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.9903757462135984e-06, | |
| "logits/chosen": -2.20393705368042, | |
| "logits/rejected": -2.0801148414611816, | |
| "logps/chosen": -254.13204956054688, | |
| "logps/rejected": -222.89602661132812, | |
| "loss": 0.6697, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.03248171508312225, | |
| "rewards/margins": 0.04965170845389366, | |
| "rewards/rejected": -0.017169995233416557, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 4.980697142834315e-06, | |
| "logits/chosen": -2.3278605937957764, | |
| "logits/rejected": -2.1921324729919434, | |
| "logps/chosen": -306.6998596191406, | |
| "logps/rejected": -257.21942138671875, | |
| "loss": 0.661, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.0004948576679453254, | |
| "rewards/margins": 0.06153837591409683, | |
| "rewards/rejected": -0.062033236026763916, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.967700826904229e-06, | |
| "logits/chosen": -2.1966867446899414, | |
| "logits/rejected": -2.082227945327759, | |
| "logps/chosen": -234.10623168945312, | |
| "logps/rejected": -225.2598114013672, | |
| "loss": 0.649, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.027638202533125877, | |
| "rewards/margins": 0.09491725265979767, | |
| "rewards/rejected": -0.1225554570555687, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.951404179843963e-06, | |
| "logits/chosen": -2.0880074501037598, | |
| "logits/rejected": -2.0883960723876953, | |
| "logps/chosen": -279.95965576171875, | |
| "logps/rejected": -290.29327392578125, | |
| "loss": 0.6478, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.09183233976364136, | |
| "rewards/margins": 0.10491160303354263, | |
| "rewards/rejected": -0.1967439353466034, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.931828996974498e-06, | |
| "logits/chosen": -2.1437039375305176, | |
| "logits/rejected": -2.035672187805176, | |
| "logps/chosen": -293.6642761230469, | |
| "logps/rejected": -308.59246826171875, | |
| "loss": 0.6161, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.10774645954370499, | |
| "rewards/margins": 0.21509592235088348, | |
| "rewards/rejected": -0.32284238934516907, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.909001458367867e-06, | |
| "logits/chosen": -2.1584107875823975, | |
| "logits/rejected": -1.9964697360992432, | |
| "logps/chosen": -278.0213928222656, | |
| "logps/rejected": -249.1472930908203, | |
| "loss": 0.6207, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.1264086812734604, | |
| "rewards/margins": 0.18005268275737762, | |
| "rewards/rejected": -0.306461364030838, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.882952093833628e-06, | |
| "logits/chosen": -2.049844741821289, | |
| "logits/rejected": -2.0112996101379395, | |
| "logps/chosen": -238.4535369873047, | |
| "logps/rejected": -298.28375244140625, | |
| "loss": 0.6092, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.26498228311538696, | |
| "rewards/margins": 0.2150886058807373, | |
| "rewards/rejected": -0.4800708293914795, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.853715742087947e-06, | |
| "logits/chosen": -2.2220816612243652, | |
| "logits/rejected": -2.0675864219665527, | |
| "logps/chosen": -294.275390625, | |
| "logps/rejected": -279.9860534667969, | |
| "loss": 0.6071, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.271901398897171, | |
| "rewards/margins": 0.23620745539665222, | |
| "rewards/rejected": -0.5081088542938232, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.821331504159906e-06, | |
| "logits/chosen": -2.120783567428589, | |
| "logits/rejected": -2.0360219478607178, | |
| "logps/chosen": -284.9825134277344, | |
| "logps/rejected": -325.1084899902344, | |
| "loss": 0.6066, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.18728381395339966, | |
| "rewards/margins": 0.21014633774757385, | |
| "rewards/rejected": -0.3974301517009735, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": -2.0610172748565674, | |
| "eval_logits/rejected": -1.9906072616577148, | |
| "eval_logps/chosen": -273.18548583984375, | |
| "eval_logps/rejected": -301.99749755859375, | |
| "eval_loss": 0.6150946021080017, | |
| "eval_rewards/accuracies": 0.69921875, | |
| "eval_rewards/chosen": -0.23031830787658691, | |
| "eval_rewards/margins": 0.2717178463935852, | |
| "eval_rewards/rejected": -0.5020360946655273, | |
| "eval_runtime": 100.7938, | |
| "eval_samples_per_second": 19.842, | |
| "eval_steps_per_second": 0.317, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.7858426910973435e-06, | |
| "logits/chosen": -2.1503536701202393, | |
| "logits/rejected": -1.9159533977508545, | |
| "logps/chosen": -337.6061706542969, | |
| "logps/rejected": -290.7955322265625, | |
| "loss": 0.6126, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.21691349148750305, | |
| "rewards/margins": 0.3816342353820801, | |
| "rewards/rejected": -0.5985476970672607, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.747296766042161e-06, | |
| "logits/chosen": -2.0444750785827637, | |
| "logits/rejected": -1.9728294610977173, | |
| "logps/chosen": -281.8607177734375, | |
| "logps/rejected": -295.66534423828125, | |
| "loss": 0.6085, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.19859151542186737, | |
| "rewards/margins": 0.28593915700912476, | |
| "rewards/rejected": -0.4845306873321533, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.705745280752586e-06, | |
| "logits/chosen": -2.0718815326690674, | |
| "logits/rejected": -1.9338264465332031, | |
| "logps/chosen": -306.795166015625, | |
| "logps/rejected": -283.50030517578125, | |
| "loss": 0.6077, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.2848009765148163, | |
| "rewards/margins": 0.24246558547019958, | |
| "rewards/rejected": -0.5272665619850159, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.661243806657256e-06, | |
| "logits/chosen": -1.994541883468628, | |
| "logits/rejected": -1.8617355823516846, | |
| "logps/chosen": -278.3886413574219, | |
| "logps/rejected": -256.62603759765625, | |
| "loss": 0.6076, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.27843061089515686, | |
| "rewards/margins": 0.25517329573631287, | |
| "rewards/rejected": -0.5336039066314697, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.613851860533367e-06, | |
| "logits/chosen": -1.980902910232544, | |
| "logits/rejected": -1.8700807094573975, | |
| "logps/chosen": -309.05877685546875, | |
| "logps/rejected": -351.293212890625, | |
| "loss": 0.6002, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.37613645195961, | |
| "rewards/margins": 0.3849337697029114, | |
| "rewards/rejected": -0.761070191860199, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.563632824908252e-06, | |
| "logits/chosen": -1.9446996450424194, | |
| "logits/rejected": -1.862592339515686, | |
| "logps/chosen": -309.58544921875, | |
| "logps/rejected": -325.975341796875, | |
| "loss": 0.5955, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.4704383909702301, | |
| "rewards/margins": 0.2729704976081848, | |
| "rewards/rejected": -0.7434087991714478, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 4.510653863290871e-06, | |
| "logits/chosen": -2.056185483932495, | |
| "logits/rejected": -1.9320430755615234, | |
| "logps/chosen": -311.8960876464844, | |
| "logps/rejected": -307.3726501464844, | |
| "loss": 0.5843, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.35895800590515137, | |
| "rewards/margins": 0.3666152358055115, | |
| "rewards/rejected": -0.7255732417106628, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.454985830346574e-06, | |
| "logits/chosen": -2.163935422897339, | |
| "logits/rejected": -2.0811800956726074, | |
| "logps/chosen": -313.5486145019531, | |
| "logps/rejected": -329.7764587402344, | |
| "loss": 0.595, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.3059764504432678, | |
| "rewards/margins": 0.3349177837371826, | |
| "rewards/rejected": -0.6408942937850952, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.396703177135262e-06, | |
| "logits/chosen": -2.090761423110962, | |
| "logits/rejected": -2.0192935466766357, | |
| "logps/chosen": -313.79547119140625, | |
| "logps/rejected": -292.3644104003906, | |
| "loss": 0.5862, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.42916345596313477, | |
| "rewards/margins": 0.19854620099067688, | |
| "rewards/rejected": -0.6277095675468445, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.335883851539693e-06, | |
| "logits/chosen": -2.0020194053649902, | |
| "logits/rejected": -1.8026950359344482, | |
| "logps/chosen": -264.9991455078125, | |
| "logps/rejected": -290.67193603515625, | |
| "loss": 0.5711, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.41096872091293335, | |
| "rewards/margins": 0.36055171489715576, | |
| "rewards/rejected": -0.7715204358100891, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "eval_logits/chosen": -2.0222675800323486, | |
| "eval_logits/rejected": -1.9416760206222534, | |
| "eval_logps/chosen": -294.5666198730469, | |
| "eval_logps/rejected": -336.92279052734375, | |
| "eval_loss": 0.5926566123962402, | |
| "eval_rewards/accuracies": 0.71875, | |
| "eval_rewards/chosen": -0.4441298246383667, | |
| "eval_rewards/margins": 0.40715909004211426, | |
| "eval_rewards/rejected": -0.851288914680481, | |
| "eval_runtime": 100.4732, | |
| "eval_samples_per_second": 19.906, | |
| "eval_steps_per_second": 0.318, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.2726091940171055e-06, | |
| "logits/chosen": -1.9914287328720093, | |
| "logits/rejected": -1.8461036682128906, | |
| "logps/chosen": -303.94287109375, | |
| "logps/rejected": -359.288330078125, | |
| "loss": 0.5712, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.4550764560699463, | |
| "rewards/margins": 0.5030550956726074, | |
| "rewards/rejected": -0.9581316113471985, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.206963828813555e-06, | |
| "logits/chosen": -1.9920743703842163, | |
| "logits/rejected": -1.8297895193099976, | |
| "logps/chosen": -330.64007568359375, | |
| "logps/rejected": -369.685302734375, | |
| "loss": 0.5752, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.5053039789199829, | |
| "rewards/margins": 0.44878053665161133, | |
| "rewards/rejected": -0.9540846943855286, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.139035550786495e-06, | |
| "logits/chosen": -1.977081298828125, | |
| "logits/rejected": -1.7652740478515625, | |
| "logps/chosen": -318.32421875, | |
| "logps/rejected": -294.5540466308594, | |
| "loss": 0.6065, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.5532656311988831, | |
| "rewards/margins": 0.37579816579818726, | |
| "rewards/rejected": -0.9290636777877808, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.068915207986931e-06, | |
| "logits/chosen": -1.8924455642700195, | |
| "logits/rejected": -1.7873607873916626, | |
| "logps/chosen": -267.74505615234375, | |
| "logps/rejected": -279.3656921386719, | |
| "loss": 0.6171, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.5295882225036621, | |
| "rewards/margins": 0.28365662693977356, | |
| "rewards/rejected": -0.8132449388504028, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 3.996696580158211e-06, | |
| "logits/chosen": -1.8571460247039795, | |
| "logits/rejected": -1.9016132354736328, | |
| "logps/chosen": -282.2213134765625, | |
| "logps/rejected": -360.13616943359375, | |
| "loss": 0.5854, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.5124384164810181, | |
| "rewards/margins": 0.5215515494346619, | |
| "rewards/rejected": -1.0339899063110352, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.922476253313921e-06, | |
| "logits/chosen": -2.048137664794922, | |
| "logits/rejected": -1.8757222890853882, | |
| "logps/chosen": -334.61102294921875, | |
| "logps/rejected": -348.38677978515625, | |
| "loss": 0.577, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.5598105192184448, | |
| "rewards/margins": 0.36239486932754517, | |
| "rewards/rejected": -0.9222054481506348, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 3.846353490562664e-06, | |
| "logits/chosen": -1.8339828252792358, | |
| "logits/rejected": -1.7505193948745728, | |
| "logps/chosen": -286.8439025878906, | |
| "logps/rejected": -329.70751953125, | |
| "loss": 0.5692, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.46078333258628845, | |
| "rewards/margins": 0.2924962639808655, | |
| "rewards/rejected": -0.7532795667648315, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.768430099352445e-06, | |
| "logits/chosen": -1.9885179996490479, | |
| "logits/rejected": -1.9461723566055298, | |
| "logps/chosen": -340.37432861328125, | |
| "logps/rejected": -348.7174377441406, | |
| "loss": 0.5628, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.5099226236343384, | |
| "rewards/margins": 0.3366420865058899, | |
| "rewards/rejected": -0.846564769744873, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 3.6888102953122307e-06, | |
| "logits/chosen": -1.7645511627197266, | |
| "logits/rejected": -1.5222413539886475, | |
| "logps/chosen": -293.2809753417969, | |
| "logps/rejected": -302.76483154296875, | |
| "loss": 0.586, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.5093228220939636, | |
| "rewards/margins": 0.5271816253662109, | |
| "rewards/rejected": -1.0365045070648193, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.607600562872785e-06, | |
| "logits/chosen": -2.1015868186950684, | |
| "logits/rejected": -1.896120309829712, | |
| "logps/chosen": -386.51507568359375, | |
| "logps/rejected": -355.5737609863281, | |
| "loss": 0.557, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.6686216592788696, | |
| "rewards/margins": 0.3796016275882721, | |
| "rewards/rejected": -1.0482232570648193, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": -1.8364356756210327, | |
| "eval_logits/rejected": -1.743373155593872, | |
| "eval_logps/chosen": -309.73779296875, | |
| "eval_logps/rejected": -359.1117248535156, | |
| "eval_loss": 0.5816810727119446, | |
| "eval_rewards/accuracies": 0.72265625, | |
| "eval_rewards/chosen": -0.5958413481712341, | |
| "eval_rewards/margins": 0.47733715176582336, | |
| "eval_rewards/rejected": -1.0731785297393799, | |
| "eval_runtime": 100.2581, | |
| "eval_samples_per_second": 19.949, | |
| "eval_steps_per_second": 0.319, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.5249095128531863e-06, | |
| "logits/chosen": -2.0195209980010986, | |
| "logits/rejected": -1.954064130783081, | |
| "logps/chosen": -382.4500732421875, | |
| "logps/rejected": -396.9842834472656, | |
| "loss": 0.5825, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.507409393787384, | |
| "rewards/margins": 0.45680397748947144, | |
| "rewards/rejected": -0.964213490486145, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.4408477372034743e-06, | |
| "logits/chosen": -1.6744167804718018, | |
| "logits/rejected": -1.5238146781921387, | |
| "logps/chosen": -277.4457702636719, | |
| "logps/rejected": -303.87213134765625, | |
| "loss": 0.5778, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.5154739022254944, | |
| "rewards/margins": 0.3824000358581543, | |
| "rewards/rejected": -0.8978738784790039, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.355527661097728e-06, | |
| "logits/chosen": -1.8267688751220703, | |
| "logits/rejected": -1.8169567584991455, | |
| "logps/chosen": -271.2919616699219, | |
| "logps/rejected": -293.12933349609375, | |
| "loss": 0.5999, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.4637935757637024, | |
| "rewards/margins": 0.3348856568336487, | |
| "rewards/rejected": -0.7986792325973511, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.269063392575352e-06, | |
| "logits/chosen": -1.7343038320541382, | |
| "logits/rejected": -1.7039591073989868, | |
| "logps/chosen": -298.5570983886719, | |
| "logps/rejected": -324.7281799316406, | |
| "loss": 0.5865, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.56922447681427, | |
| "rewards/margins": 0.3157515227794647, | |
| "rewards/rejected": -0.8849760293960571, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.181570569931697e-06, | |
| "logits/chosen": -1.3839043378829956, | |
| "logits/rejected": -1.3726271390914917, | |
| "logps/chosen": -274.1092529296875, | |
| "logps/rejected": -318.86492919921875, | |
| "loss": 0.578, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.5215469598770142, | |
| "rewards/margins": 0.42288708686828613, | |
| "rewards/rejected": -0.9444340467453003, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.09316620706208e-06, | |
| "logits/chosen": -1.7264198064804077, | |
| "logits/rejected": -1.616563081741333, | |
| "logps/chosen": -324.797119140625, | |
| "logps/rejected": -361.42303466796875, | |
| "loss": 0.585, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.5468829274177551, | |
| "rewards/margins": 0.3461124002933502, | |
| "rewards/rejected": -0.8929953575134277, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 3.0039685369660785e-06, | |
| "logits/chosen": -1.6359647512435913, | |
| "logits/rejected": -1.3307011127471924, | |
| "logps/chosen": -346.19561767578125, | |
| "logps/rejected": -339.08978271484375, | |
| "loss": 0.5644, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.6865732669830322, | |
| "rewards/margins": 0.44049325585365295, | |
| "rewards/rejected": -1.1270664930343628, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.91409685362137e-06, | |
| "logits/chosen": -1.1181867122650146, | |
| "logits/rejected": -0.9776461720466614, | |
| "logps/chosen": -293.79425048828125, | |
| "logps/rejected": -316.5558166503906, | |
| "loss": 0.6066, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.7594456076622009, | |
| "rewards/margins": 0.25457096099853516, | |
| "rewards/rejected": -1.0140166282653809, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 2.8236713524386085e-06, | |
| "logits/chosen": -1.3658815622329712, | |
| "logits/rejected": -1.1311851739883423, | |
| "logps/chosen": -310.53594970703125, | |
| "logps/rejected": -355.07257080078125, | |
| "loss": 0.5688, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.6269619464874268, | |
| "rewards/margins": 0.4374374449253082, | |
| "rewards/rejected": -1.0643993616104126, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.7328129695107205e-06, | |
| "logits/chosen": -1.2549781799316406, | |
| "logits/rejected": -0.9512443542480469, | |
| "logps/chosen": -343.2566833496094, | |
| "logps/rejected": -383.65936279296875, | |
| "loss": 0.5703, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.7175710201263428, | |
| "rewards/margins": 0.569162905216217, | |
| "rewards/rejected": -1.286733865737915, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "eval_logits/chosen": -0.9966566562652588, | |
| "eval_logits/rejected": -0.8466635942459106, | |
| "eval_logps/chosen": -322.3067932128906, | |
| "eval_logps/rejected": -375.8402099609375, | |
| "eval_loss": 0.5679102540016174, | |
| "eval_rewards/accuracies": 0.7265625, | |
| "eval_rewards/chosen": -0.7215311527252197, | |
| "eval_rewards/margins": 0.5189324617385864, | |
| "eval_rewards/rejected": -1.2404634952545166, | |
| "eval_runtime": 100.2462, | |
| "eval_samples_per_second": 19.951, | |
| "eval_steps_per_second": 0.319, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 2.641643219871597e-06, | |
| "logits/chosen": -1.133755087852478, | |
| "logits/rejected": -0.6763060092926025, | |
| "logps/chosen": -343.71722412109375, | |
| "logps/rejected": -342.73590087890625, | |
| "loss": 0.5469, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.820611834526062, | |
| "rewards/margins": 0.41005903482437134, | |
| "rewards/rejected": -1.2306709289550781, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.5502840349805074e-06, | |
| "logits/chosen": -1.0033422708511353, | |
| "logits/rejected": -0.8227261304855347, | |
| "logps/chosen": -342.0159606933594, | |
| "logps/rejected": -367.8998718261719, | |
| "loss": 0.5467, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.6553717851638794, | |
| "rewards/margins": 0.48232603073120117, | |
| "rewards/rejected": -1.1376978158950806, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2.4588575996495797e-06, | |
| "logits/chosen": -1.015627145767212, | |
| "logits/rejected": -0.7538983225822449, | |
| "logps/chosen": -377.20703125, | |
| "logps/rejected": -394.6551513671875, | |
| "loss": 0.5437, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.8781946301460266, | |
| "rewards/margins": 0.4648224413394928, | |
| "rewards/rejected": -1.3430172204971313, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.367486188632446e-06, | |
| "logits/chosen": -1.0752605199813843, | |
| "logits/rejected": -0.7782305479049683, | |
| "logps/chosen": -355.3316955566406, | |
| "logps/rejected": -417.9510803222656, | |
| "loss": 0.5496, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.738643229007721, | |
| "rewards/margins": 0.645391583442688, | |
| "rewards/rejected": -1.3840347528457642, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.276292003092593e-06, | |
| "logits/chosen": -0.9742870330810547, | |
| "logits/rejected": -0.5593458414077759, | |
| "logps/chosen": -347.0143127441406, | |
| "logps/rejected": -346.98065185546875, | |
| "loss": 0.5747, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.7598128914833069, | |
| "rewards/margins": 0.5269237756729126, | |
| "rewards/rejected": -1.2867367267608643, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.1853970071701415e-06, | |
| "logits/chosen": -0.7387961149215698, | |
| "logits/rejected": -0.4743451476097107, | |
| "logps/chosen": -338.78662109375, | |
| "logps/rejected": -341.47320556640625, | |
| "loss": 0.554, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.8780034780502319, | |
| "rewards/margins": 0.41334277391433716, | |
| "rewards/rejected": -1.2913460731506348, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.0949227648656194e-06, | |
| "logits/chosen": -0.6523474454879761, | |
| "logits/rejected": -0.389020174741745, | |
| "logps/chosen": -314.07220458984375, | |
| "logps/rejected": -362.8746643066406, | |
| "loss": 0.5822, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.8458138704299927, | |
| "rewards/margins": 0.4794779419898987, | |
| "rewards/rejected": -1.3252918720245361, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 2.00499027745888e-06, | |
| "logits/chosen": -0.516838788986206, | |
| "logits/rejected": -0.14539854228496552, | |
| "logps/chosen": -334.552001953125, | |
| "logps/rejected": -365.7865295410156, | |
| "loss": 0.5619, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.7920180559158325, | |
| "rewards/margins": 0.6189525723457336, | |
| "rewards/rejected": -1.410970687866211, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.915719821680624e-06, | |
| "logits/chosen": -0.4588368833065033, | |
| "logits/rejected": -0.17827937006950378, | |
| "logps/chosen": -302.94464111328125, | |
| "logps/rejected": -334.8900146484375, | |
| "loss": 0.5691, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.7475640773773193, | |
| "rewards/margins": 0.3807729184627533, | |
| "rewards/rejected": -1.1283371448516846, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.8272307888529276e-06, | |
| "logits/chosen": -0.7035941481590271, | |
| "logits/rejected": -0.5019673109054565, | |
| "logps/chosen": -361.04827880859375, | |
| "logps/rejected": -436.74560546875, | |
| "loss": 0.5498, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.5611416101455688, | |
| "rewards/margins": 0.6350690126419067, | |
| "rewards/rejected": -1.1962106227874756, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": -0.44625866413116455, | |
| "eval_logits/rejected": -0.25102561712265015, | |
| "eval_logps/chosen": -320.17938232421875, | |
| "eval_logps/rejected": -380.2698974609375, | |
| "eval_loss": 0.5582141280174255, | |
| "eval_rewards/accuracies": 0.7578125, | |
| "eval_rewards/chosen": -0.7002571821212769, | |
| "eval_rewards/margins": 0.5845023393630981, | |
| "eval_rewards/rejected": -1.284759521484375, | |
| "eval_runtime": 100.0013, | |
| "eval_samples_per_second": 20.0, | |
| "eval_steps_per_second": 0.32, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.739641525213929e-06, | |
| "logits/chosen": -0.5911251306533813, | |
| "logits/rejected": -0.18869177997112274, | |
| "logps/chosen": -325.91033935546875, | |
| "logps/rejected": -326.54095458984375, | |
| "loss": 0.5425, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.7186974287033081, | |
| "rewards/margins": 0.4664137363433838, | |
| "rewards/rejected": -1.185111165046692, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.6530691736402317e-06, | |
| "logits/chosen": -0.8488653898239136, | |
| "logits/rejected": -0.5028330087661743, | |
| "logps/chosen": -356.1060485839844, | |
| "logps/rejected": -355.0657958984375, | |
| "loss": 0.5542, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.8343976140022278, | |
| "rewards/margins": 0.4692300260066986, | |
| "rewards/rejected": -1.303627610206604, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1.5676295169786864e-06, | |
| "logits/chosen": -0.2444450557231903, | |
| "logits/rejected": 0.13049665093421936, | |
| "logps/chosen": -350.13702392578125, | |
| "logps/rejected": -413.64398193359375, | |
| "loss": 0.5379, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.8719974756240845, | |
| "rewards/margins": 0.7344092130661011, | |
| "rewards/rejected": -1.606406807899475, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.4834368231970922e-06, | |
| "logits/chosen": -0.38939914107322693, | |
| "logits/rejected": 0.05538179352879524, | |
| "logps/chosen": -322.5318298339844, | |
| "logps/rejected": -368.46343994140625, | |
| "loss": 0.5401, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.7830222845077515, | |
| "rewards/margins": 0.7185503244400024, | |
| "rewards/rejected": -1.501572847366333, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.4006036925609245e-06, | |
| "logits/chosen": -0.8077508807182312, | |
| "logits/rejected": -0.2668865919113159, | |
| "logps/chosen": -380.40252685546875, | |
| "logps/rejected": -398.7056579589844, | |
| "loss": 0.5383, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.8857347369194031, | |
| "rewards/margins": 0.5326262712478638, | |
| "rewards/rejected": -1.4183608293533325, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3192409070404582e-06, | |
| "logits/chosen": -0.5546111464500427, | |
| "logits/rejected": -0.17088347673416138, | |
| "logps/chosen": -398.8465881347656, | |
| "logps/rejected": -376.5204162597656, | |
| "loss": 0.554, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.9672527313232422, | |
| "rewards/margins": 0.5284374952316284, | |
| "rewards/rejected": -1.4956903457641602, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.2394572821496953e-06, | |
| "logits/chosen": -0.39746540784835815, | |
| "logits/rejected": -0.14205250144004822, | |
| "logps/chosen": -365.32281494140625, | |
| "logps/rejected": -412.72650146484375, | |
| "loss": 0.5385, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.8937069177627563, | |
| "rewards/margins": 0.5911084413528442, | |
| "rewards/rejected": -1.4848153591156006, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.1613595214152713e-06, | |
| "logits/chosen": -0.23002564907073975, | |
| "logits/rejected": -0.18054810166358948, | |
| "logps/chosen": -314.38275146484375, | |
| "logps/rejected": -324.4413146972656, | |
| "loss": 0.5657, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.8053790330886841, | |
| "rewards/margins": 0.24868443608283997, | |
| "rewards/rejected": -1.054063320159912, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1.0850520736699362e-06, | |
| "logits/chosen": -0.35535928606987, | |
| "logits/rejected": 0.08797712624073029, | |
| "logps/chosen": -279.0433349609375, | |
| "logps/rejected": -323.66754150390625, | |
| "loss": 0.5362, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.672313928604126, | |
| "rewards/margins": 0.6532977223396301, | |
| "rewards/rejected": -1.3256117105484009, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0106369933615043e-06, | |
| "logits/chosen": -0.46180295944213867, | |
| "logits/rejected": 0.07596547156572342, | |
| "logps/chosen": -314.94793701171875, | |
| "logps/rejected": -355.2444152832031, | |
| "loss": 0.5279, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.8018286824226379, | |
| "rewards/margins": 0.6314573884010315, | |
| "rewards/rejected": -1.4332860708236694, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "eval_logits/chosen": -0.198812335729599, | |
| "eval_logits/rejected": 0.014529339037835598, | |
| "eval_logps/chosen": -334.1552734375, | |
| "eval_logps/rejected": -400.80816650390625, | |
| "eval_loss": 0.5489765405654907, | |
| "eval_rewards/accuracies": 0.75, | |
| "eval_rewards/chosen": -0.8400164246559143, | |
| "eval_rewards/margins": 0.650126576423645, | |
| "eval_rewards/rejected": -1.4901429414749146, | |
| "eval_runtime": 99.961, | |
| "eval_samples_per_second": 20.008, | |
| "eval_steps_per_second": 0.32, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 9.382138040640714e-07, | |
| "logits/chosen": -0.22463122010231018, | |
| "logits/rejected": 0.20351561903953552, | |
| "logps/chosen": -352.7989196777344, | |
| "logps/rejected": -359.4533386230469, | |
| "loss": 0.546, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.9452872276306152, | |
| "rewards/margins": 0.49813446402549744, | |
| "rewards/rejected": -1.443421721458435, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.678793653740633e-07, | |
| "logits/chosen": -0.256120890378952, | |
| "logits/rejected": 0.14627498388290405, | |
| "logps/chosen": -387.221435546875, | |
| "logps/rejected": -435.5755310058594, | |
| "loss": 0.5552, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.9933661222457886, | |
| "rewards/margins": 0.663570761680603, | |
| "rewards/rejected": -1.6569368839263916, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 7.997277433690984e-07, | |
| "logits/chosen": -0.48140692710876465, | |
| "logits/rejected": 0.015109086409211159, | |
| "logps/chosen": -395.16815185546875, | |
| "logps/rejected": -368.5047302246094, | |
| "loss": 0.5629, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.8996418714523315, | |
| "rewards/margins": 0.48376113176345825, | |
| "rewards/rejected": -1.3834030628204346, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.338500848029603e-07, | |
| "logits/chosen": -0.3346256911754608, | |
| "logits/rejected": 0.1683504283428192, | |
| "logps/chosen": -410.95672607421875, | |
| "logps/rejected": -427.506591796875, | |
| "loss": 0.5523, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.8928782343864441, | |
| "rewards/margins": 0.6983044147491455, | |
| "rewards/rejected": -1.5911829471588135, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 6.70334495204884e-07, | |
| "logits/chosen": -0.5112616419792175, | |
| "logits/rejected": -0.16424962878227234, | |
| "logps/chosen": -346.962890625, | |
| "logps/rejected": -423.4779357910156, | |
| "loss": 0.5376, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.8174928426742554, | |
| "rewards/margins": 0.6463109850883484, | |
| "rewards/rejected": -1.4638036489486694, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 6.092659210462232e-07, | |
| "logits/chosen": -0.33304744958877563, | |
| "logits/rejected": -0.1934703141450882, | |
| "logps/chosen": -360.96124267578125, | |
| "logps/rejected": -388.1688537597656, | |
| "loss": 0.5219, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8719509840011597, | |
| "rewards/margins": 0.5291545987129211, | |
| "rewards/rejected": -1.4011056423187256, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 5.507260361320738e-07, | |
| "logits/chosen": -0.3076072931289673, | |
| "logits/rejected": -0.03598857671022415, | |
| "logps/chosen": -417.52679443359375, | |
| "logps/rejected": -437.1398010253906, | |
| "loss": 0.5557, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.8676769137382507, | |
| "rewards/margins": 0.46329760551452637, | |
| "rewards/rejected": -1.3309745788574219, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.947931323697983e-07, | |
| "logits/chosen": 0.10995174944400787, | |
| "logits/rejected": 0.25233790278434753, | |
| "logps/chosen": -298.9948425292969, | |
| "logps/rejected": -337.920166015625, | |
| "loss": 0.5422, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.9167734980583191, | |
| "rewards/margins": 0.42617493867874146, | |
| "rewards/rejected": -1.3429481983184814, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.4154201506053985e-07, | |
| "logits/chosen": -0.15819688141345978, | |
| "logits/rejected": -0.1625043898820877, | |
| "logps/chosen": -366.3085632324219, | |
| "logps/rejected": -413.33343505859375, | |
| "loss": 0.5391, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.967660129070282, | |
| "rewards/margins": 0.4664042592048645, | |
| "rewards/rejected": -1.434064507484436, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.910439028537638e-07, | |
| "logits/chosen": -0.2914949953556061, | |
| "logits/rejected": 0.14831864833831787, | |
| "logps/chosen": -412.54827880859375, | |
| "logps/rejected": -403.07647705078125, | |
| "loss": 0.5264, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.9057921171188354, | |
| "rewards/margins": 0.6066713333129883, | |
| "rewards/rejected": -1.5124633312225342, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": -0.15486477315425873, | |
| "eval_logits/rejected": 0.06035802513360977, | |
| "eval_logps/chosen": -336.28326416015625, | |
| "eval_logps/rejected": -404.0750732421875, | |
| "eval_loss": 0.5475199818611145, | |
| "eval_rewards/accuracies": 0.74609375, | |
| "eval_rewards/chosen": -0.8612961769104004, | |
| "eval_rewards/margins": 0.6615157127380371, | |
| "eval_rewards/rejected": -1.5228118896484375, | |
| "eval_runtime": 100.0077, | |
| "eval_samples_per_second": 19.998, | |
| "eval_steps_per_second": 0.32, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 3.4336633249862084e-07, | |
| "logits/chosen": -0.3513889014720917, | |
| "logits/rejected": 0.16097551584243774, | |
| "logps/chosen": -367.8076477050781, | |
| "logps/rejected": -368.91473388671875, | |
| "loss": 0.5401, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -1.0032860040664673, | |
| "rewards/margins": 0.39093494415283203, | |
| "rewards/rejected": -1.3942210674285889, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 2.98573068519539e-07, | |
| "logits/chosen": -0.27493131160736084, | |
| "logits/rejected": 0.1548646241426468, | |
| "logps/chosen": -365.89007568359375, | |
| "logps/rejected": -377.8901672363281, | |
| "loss": 0.5636, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.8509780764579773, | |
| "rewards/margins": 0.6635745167732239, | |
| "rewards/rejected": -1.5145527124404907, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 2.5672401793681854e-07, | |
| "logits/chosen": -0.1434798389673233, | |
| "logits/rejected": -0.3032146990299225, | |
| "logps/chosen": -335.52349853515625, | |
| "logps/rejected": -418.47979736328125, | |
| "loss": 0.5174, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.9104019999504089, | |
| "rewards/margins": 0.5144548416137695, | |
| "rewards/rejected": -1.4248569011688232, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.178751501463036e-07, | |
| "logits/chosen": -0.28891128301620483, | |
| "logits/rejected": -0.015888934955000877, | |
| "logps/chosen": -357.6486511230469, | |
| "logps/rejected": -387.8254089355469, | |
| "loss": 0.5675, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.9722326993942261, | |
| "rewards/margins": 0.40555062890052795, | |
| "rewards/rejected": -1.3777834177017212, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 1.820784220652766e-07, | |
| "logits/chosen": -0.457050621509552, | |
| "logits/rejected": -0.08179013431072235, | |
| "logps/chosen": -365.02935791015625, | |
| "logps/rejected": -377.90484619140625, | |
| "loss": 0.5464, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.8087032437324524, | |
| "rewards/margins": 0.6943231821060181, | |
| "rewards/rejected": -1.5030266046524048, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.4938170864468636e-07, | |
| "logits/chosen": -0.2849724292755127, | |
| "logits/rejected": -0.0881032794713974, | |
| "logps/chosen": -353.9926452636719, | |
| "logps/rejected": -402.39886474609375, | |
| "loss": 0.551, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.8892512321472168, | |
| "rewards/margins": 0.6870072484016418, | |
| "rewards/rejected": -1.5762584209442139, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.1982873884064466e-07, | |
| "logits/chosen": -0.2418334186077118, | |
| "logits/rejected": -0.05195746570825577, | |
| "logps/chosen": -314.6041564941406, | |
| "logps/rejected": -390.5324401855469, | |
| "loss": 0.5528, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.6465216279029846, | |
| "rewards/margins": 0.8042505383491516, | |
| "rewards/rejected": -1.4507721662521362, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 9.345903713082305e-08, | |
| "logits/chosen": -0.3975564241409302, | |
| "logits/rejected": -0.0711708664894104, | |
| "logps/chosen": -384.4748229980469, | |
| "logps/rejected": -408.9742126464844, | |
| "loss": 0.5344, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.9061468243598938, | |
| "rewards/margins": 0.4584124684333801, | |
| "rewards/rejected": -1.3645591735839844, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 7.030787065396866e-08, | |
| "logits/chosen": -0.34069642424583435, | |
| "logits/rejected": -0.19395461678504944, | |
| "logps/chosen": -320.58380126953125, | |
| "logps/rejected": -402.9022216796875, | |
| "loss": 0.5508, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.795324981212616, | |
| "rewards/margins": 0.6657498478889465, | |
| "rewards/rejected": -1.461074948310852, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 5.0406202043228604e-08, | |
| "logits/chosen": -0.0745534598827362, | |
| "logits/rejected": 0.35236138105392456, | |
| "logps/chosen": -382.5845642089844, | |
| "logps/rejected": -400.34735107421875, | |
| "loss": 0.5639, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.9554711580276489, | |
| "rewards/margins": 0.41256728768348694, | |
| "rewards/rejected": -1.3680384159088135, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "eval_logits/chosen": -0.14664390683174133, | |
| "eval_logits/rejected": 0.07037673145532608, | |
| "eval_logps/chosen": -336.4347839355469, | |
| "eval_logps/rejected": -404.4688415527344, | |
| "eval_loss": 0.5475045442581177, | |
| "eval_rewards/accuracies": 0.7421875, | |
| "eval_rewards/chosen": -0.8628111481666565, | |
| "eval_rewards/margins": 0.6639385223388672, | |
| "eval_rewards/rejected": -1.5267497301101685, | |
| "eval_runtime": 99.9101, | |
| "eval_samples_per_second": 20.018, | |
| "eval_steps_per_second": 0.32, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 3.378064801637687e-08, | |
| "logits/chosen": -0.18566574156284332, | |
| "logits/rejected": 0.08027581125497818, | |
| "logps/chosen": -338.36224365234375, | |
| "logps/rejected": -387.63946533203125, | |
| "loss": 0.5317, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.8497447967529297, | |
| "rewards/margins": 0.5736583471298218, | |
| "rewards/rejected": -1.423403263092041, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.0453443778310766e-08, | |
| "logits/chosen": -0.29515573382377625, | |
| "logits/rejected": 0.24348752200603485, | |
| "logps/chosen": -364.8608703613281, | |
| "logps/rejected": -390.37200927734375, | |
| "loss": 0.5321, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.9079607725143433, | |
| "rewards/margins": 0.5625349283218384, | |
| "rewards/rejected": -1.470495581626892, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 1.0442413283435759e-08, | |
| "logits/chosen": -0.34043288230895996, | |
| "logits/rejected": -0.010392585769295692, | |
| "logps/chosen": -288.91424560546875, | |
| "logps/rejected": -375.74945068359375, | |
| "loss": 0.51, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.7094988226890564, | |
| "rewards/margins": 0.8463606834411621, | |
| "rewards/rejected": -1.5558595657348633, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 3.760945397705828e-09, | |
| "logits/chosen": -0.4917779862880707, | |
| "logits/rejected": -0.014173048548400402, | |
| "logps/chosen": -391.59149169921875, | |
| "logps/rejected": -432.360107421875, | |
| "loss": 0.5395, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.9179250597953796, | |
| "rewards/margins": 0.5526332855224609, | |
| "rewards/rejected": -1.4705584049224854, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 4.1797599220405605e-10, | |
| "logits/chosen": -0.1975761204957962, | |
| "logits/rejected": 0.10586023330688477, | |
| "logps/chosen": -334.07745361328125, | |
| "logps/rejected": -401.03680419921875, | |
| "loss": 0.5484, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.9642227292060852, | |
| "rewards/margins": 0.754417359828949, | |
| "rewards/rejected": -1.7186399698257446, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 0.583915277301329, | |
| "train_runtime": 6210.8046, | |
| "train_samples_per_second": 9.843, | |
| "train_steps_per_second": 0.154 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 955, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |