| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 100, | |
| "global_step": 391, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0025575447570332483, | |
| "grad_norm": 39.26094609355056, | |
| "learning_rate": 2.5e-09, | |
| "logits/chosen": -4.633832931518555, | |
| "logits/rejected": -4.853580951690674, | |
| "logps/chosen": -233.06472778320312, | |
| "logps/rejected": -207.33529663085938, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02557544757033248, | |
| "grad_norm": 36.68935602174721, | |
| "learning_rate": 2.5e-08, | |
| "logits/chosen": -4.338743686676025, | |
| "logits/rejected": -4.649880886077881, | |
| "logps/chosen": -262.4483642578125, | |
| "logps/rejected": -221.99293518066406, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4791666567325592, | |
| "rewards/chosen": 0.00011811345757450908, | |
| "rewards/margins": 0.0013624767307192087, | |
| "rewards/rejected": -0.0012443631421774626, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05115089514066496, | |
| "grad_norm": 39.59560101486009, | |
| "learning_rate": 5e-08, | |
| "logits/chosen": -4.518132209777832, | |
| "logits/rejected": -4.7573723793029785, | |
| "logps/chosen": -265.6820373535156, | |
| "logps/rejected": -224.56332397460938, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": 0.0007334979018196464, | |
| "rewards/margins": 0.0019747079350054264, | |
| "rewards/rejected": -0.00124121003318578, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.07672634271099744, | |
| "grad_norm": 42.07292039672945, | |
| "learning_rate": 7.5e-08, | |
| "logits/chosen": -4.586892127990723, | |
| "logits/rejected": -4.772247314453125, | |
| "logps/chosen": -254.6095428466797, | |
| "logps/rejected": -221.1687469482422, | |
| "loss": 0.6881, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.004800780676305294, | |
| "rewards/margins": 0.01116220187395811, | |
| "rewards/rejected": -0.006361422594636679, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.10230179028132992, | |
| "grad_norm": 39.99985818605015, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -4.643338680267334, | |
| "logits/rejected": -4.734442234039307, | |
| "logps/chosen": -250.8112030029297, | |
| "logps/rejected": -227.5035858154297, | |
| "loss": 0.674, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": 0.015374640934169292, | |
| "rewards/margins": 0.04068901389837265, | |
| "rewards/rejected": -0.025314366444945335, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1278772378516624, | |
| "grad_norm": 39.42008970309411, | |
| "learning_rate": 9.979985922607475e-08, | |
| "logits/chosen": -4.586745262145996, | |
| "logits/rejected": -4.8019514083862305, | |
| "logps/chosen": -266.86578369140625, | |
| "logps/rejected": -236.32333374023438, | |
| "loss": 0.6472, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.025944575667381287, | |
| "rewards/margins": 0.09520022571086884, | |
| "rewards/rejected": -0.06925566494464874, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.1534526854219949, | |
| "grad_norm": 38.00390022566866, | |
| "learning_rate": 9.92010391574745e-08, | |
| "logits/chosen": -4.757911205291748, | |
| "logits/rejected": -4.887364387512207, | |
| "logps/chosen": -239.2329559326172, | |
| "logps/rejected": -240.6675262451172, | |
| "loss": 0.5963, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -0.0015811750199645758, | |
| "rewards/margins": 0.20014917850494385, | |
| "rewards/rejected": -0.20173034071922302, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.17902813299232737, | |
| "grad_norm": 40.739477874841256, | |
| "learning_rate": 9.820833372667812e-08, | |
| "logits/chosen": -4.700837135314941, | |
| "logits/rejected": -4.890389442443848, | |
| "logps/chosen": -262.3451843261719, | |
| "logps/rejected": -265.8101501464844, | |
| "loss": 0.552, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.08908917009830475, | |
| "rewards/margins": 0.345054566860199, | |
| "rewards/rejected": -0.4341437816619873, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.20460358056265984, | |
| "grad_norm": 39.05798828996261, | |
| "learning_rate": 9.682969016701356e-08, | |
| "logits/chosen": -4.702755451202393, | |
| "logits/rejected": -4.89903450012207, | |
| "logps/chosen": -271.28289794921875, | |
| "logps/rejected": -292.86474609375, | |
| "loss": 0.4978, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.11036349833011627, | |
| "rewards/margins": 0.46904292702674866, | |
| "rewards/rejected": -0.5794063806533813, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.23017902813299232, | |
| "grad_norm": 35.899018808141314, | |
| "learning_rate": 9.507614539004081e-08, | |
| "logits/chosen": -4.802427768707275, | |
| "logits/rejected": -5.013784408569336, | |
| "logps/chosen": -254.51138305664062, | |
| "logps/rejected": -277.52117919921875, | |
| "loss": 0.4653, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.007465996779501438, | |
| "rewards/margins": 0.6835082769393921, | |
| "rewards/rejected": -0.6909742951393127, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.2557544757033248, | |
| "grad_norm": 36.72256884048717, | |
| "learning_rate": 9.296173762811083e-08, | |
| "logits/chosen": -4.685734748840332, | |
| "logits/rejected": -4.969820976257324, | |
| "logps/chosen": -261.9112243652344, | |
| "logps/rejected": -306.5802001953125, | |
| "loss": 0.4622, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -0.03183584660291672, | |
| "rewards/margins": 0.7450669407844543, | |
| "rewards/rejected": -0.7769027948379517, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2557544757033248, | |
| "eval_logits/chosen": -4.788325309753418, | |
| "eval_logits/rejected": -5.006863594055176, | |
| "eval_logps/chosen": -439.2997131347656, | |
| "eval_logps/rejected": -570.1284790039062, | |
| "eval_loss": 0.7195349335670471, | |
| "eval_rewards/accuracies": 0.48828125, | |
| "eval_rewards/chosen": -0.4882526695728302, | |
| "eval_rewards/margins": 0.029155578464269638, | |
| "eval_rewards/rejected": -0.5174082517623901, | |
| "eval_runtime": 98.6046, | |
| "eval_samples_per_second": 20.283, | |
| "eval_steps_per_second": 0.325, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2813299232736573, | |
| "grad_norm": 37.92952007987077, | |
| "learning_rate": 9.050339404945832e-08, | |
| "logits/chosen": -4.800631999969482, | |
| "logits/rejected": -5.014934539794922, | |
| "logps/chosen": -261.18560791015625, | |
| "logps/rejected": -321.3470458984375, | |
| "loss": 0.4274, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.1468946933746338, | |
| "rewards/margins": 0.8339446783065796, | |
| "rewards/rejected": -0.9808392524719238, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.3069053708439898, | |
| "grad_norm": 34.34594445177828, | |
| "learning_rate": 8.77207952455395e-08, | |
| "logits/chosen": -4.794532775878906, | |
| "logits/rejected": -5.045430660247803, | |
| "logps/chosen": -288.1357727050781, | |
| "logps/rejected": -340.7710876464844, | |
| "loss": 0.4231, | |
| "rewards/accuracies": 0.8687499761581421, | |
| "rewards/chosen": -0.13442620635032654, | |
| "rewards/margins": 0.9264079928398132, | |
| "rewards/rejected": -1.0608341693878174, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.33248081841432225, | |
| "grad_norm": 38.8387240251847, | |
| "learning_rate": 8.463621767547997e-08, | |
| "logits/chosen": -4.894143104553223, | |
| "logits/rejected": -5.190768718719482, | |
| "logps/chosen": -281.547607421875, | |
| "logps/rejected": -352.4320983886719, | |
| "loss": 0.4055, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -0.25050926208496094, | |
| "rewards/margins": 1.063452959060669, | |
| "rewards/rejected": -1.3139623403549194, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.35805626598465473, | |
| "grad_norm": 36.960561396839225, | |
| "learning_rate": 8.127435532896387e-08, | |
| "logits/chosen": -4.948976993560791, | |
| "logits/rejected": -5.229872703552246, | |
| "logps/chosen": -320.14996337890625, | |
| "logps/rejected": -386.9942932128906, | |
| "loss": 0.392, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.3797687590122223, | |
| "rewards/margins": 1.1361901760101318, | |
| "rewards/rejected": -1.5159590244293213, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.3836317135549872, | |
| "grad_norm": 43.84353984691604, | |
| "learning_rate": 7.766212203526569e-08, | |
| "logits/chosen": -5.017346382141113, | |
| "logits/rejected": -5.299422740936279, | |
| "logps/chosen": -287.28350830078125, | |
| "logps/rejected": -373.42730712890625, | |
| "loss": 0.3891, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.374920517206192, | |
| "rewards/margins": 1.0954115390777588, | |
| "rewards/rejected": -1.4703319072723389, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.4092071611253197, | |
| "grad_norm": 45.6486792527098, | |
| "learning_rate": 7.382843600106538e-08, | |
| "logits/chosen": -5.09747314453125, | |
| "logits/rejected": -5.334885120391846, | |
| "logps/chosen": -294.9914855957031, | |
| "logps/rejected": -389.2686462402344, | |
| "loss": 0.3586, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -0.47372356057167053, | |
| "rewards/margins": 1.2067290544509888, | |
| "rewards/rejected": -1.680452585220337, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.43478260869565216, | |
| "grad_norm": 41.55934089834119, | |
| "learning_rate": 6.980398830195784e-08, | |
| "logits/chosen": -5.022572040557861, | |
| "logits/rejected": -5.321537494659424, | |
| "logps/chosen": -307.09521484375, | |
| "logps/rejected": -427.2223205566406, | |
| "loss": 0.3397, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.5031875967979431, | |
| "rewards/margins": 1.4829540252685547, | |
| "rewards/rejected": -1.9861418008804321, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.46035805626598464, | |
| "grad_norm": 37.76415174175888, | |
| "learning_rate": 6.562099718102787e-08, | |
| "logits/chosen": -5.190648078918457, | |
| "logits/rejected": -5.442112922668457, | |
| "logps/chosen": -293.28350830078125, | |
| "logps/rejected": -403.9653625488281, | |
| "loss": 0.3551, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -0.5512182116508484, | |
| "rewards/margins": 1.491098165512085, | |
| "rewards/rejected": -2.042316198348999, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.4859335038363171, | |
| "grad_norm": 52.26242886674056, | |
| "learning_rate": 6.131295012148612e-08, | |
| "logits/chosen": -5.1211113929748535, | |
| "logits/rejected": -5.269497871398926, | |
| "logps/chosen": -328.1192321777344, | |
| "logps/rejected": -461.3627014160156, | |
| "loss": 0.3605, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -0.7219178080558777, | |
| "rewards/margins": 1.4698314666748047, | |
| "rewards/rejected": -2.1917495727539062, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.5115089514066496, | |
| "grad_norm": 41.210687347051376, | |
| "learning_rate": 5.691433575823665e-08, | |
| "logits/chosen": -5.160891056060791, | |
| "logits/rejected": -5.3507490158081055, | |
| "logps/chosen": -323.77532958984375, | |
| "logps/rejected": -448.09271240234375, | |
| "loss": 0.3618, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.7529298067092896, | |
| "rewards/margins": 1.5445276498794556, | |
| "rewards/rejected": -2.297457456588745, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5115089514066496, | |
| "eval_logits/chosen": -5.146201133728027, | |
| "eval_logits/rejected": -5.404381275177002, | |
| "eval_logps/chosen": -551.28759765625, | |
| "eval_logps/rejected": -696.0471801757812, | |
| "eval_loss": 0.7717023491859436, | |
| "eval_rewards/accuracies": 0.5703125, | |
| "eval_rewards/chosen": -1.6081316471099854, | |
| "eval_rewards/margins": 0.16846325993537903, | |
| "eval_rewards/rejected": -1.7765947580337524, | |
| "eval_runtime": 98.5146, | |
| "eval_samples_per_second": 20.302, | |
| "eval_steps_per_second": 0.325, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5370843989769821, | |
| "grad_norm": 33.990804012478485, | |
| "learning_rate": 5.2460367774593905e-08, | |
| "logits/chosen": -5.241884708404541, | |
| "logits/rejected": -5.514366626739502, | |
| "logps/chosen": -325.63385009765625, | |
| "logps/rejected": -494.1783142089844, | |
| "loss": 0.3361, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -0.631729245185852, | |
| "rewards/margins": 1.9169912338256836, | |
| "rewards/rejected": -2.548720359802246, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.5626598465473146, | |
| "grad_norm": 43.799668582141386, | |
| "learning_rate": 4.798670299452925e-08, | |
| "logits/chosen": -5.0720014572143555, | |
| "logits/rejected": -5.483579635620117, | |
| "logps/chosen": -326.8290100097656, | |
| "logps/rejected": -476.4143981933594, | |
| "loss": 0.3361, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.6208838224411011, | |
| "rewards/margins": 1.8470569849014282, | |
| "rewards/rejected": -2.4679408073425293, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.5882352941176471, | |
| "grad_norm": 49.054323066624825, | |
| "learning_rate": 4.3529155927297226e-08, | |
| "logits/chosen": -5.175291538238525, | |
| "logits/rejected": -5.545914649963379, | |
| "logps/chosen": -346.18603515625, | |
| "logps/rejected": -467.39093017578125, | |
| "loss": 0.3482, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.8635517954826355, | |
| "rewards/margins": 1.513214111328125, | |
| "rewards/rejected": -2.376765727996826, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.6138107416879796, | |
| "grad_norm": 39.76260402422977, | |
| "learning_rate": 3.9123412049691636e-08, | |
| "logits/chosen": -5.179227828979492, | |
| "logits/rejected": -5.498720645904541, | |
| "logps/chosen": -356.7921142578125, | |
| "logps/rejected": -494.8655700683594, | |
| "loss": 0.3289, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -0.8680986166000366, | |
| "rewards/margins": 1.7818403244018555, | |
| "rewards/rejected": -2.6499390602111816, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.639386189258312, | |
| "grad_norm": 39.31547537418584, | |
| "learning_rate": 3.480474212128766e-08, | |
| "logits/chosen": -5.349845886230469, | |
| "logits/rejected": -5.674462795257568, | |
| "logps/chosen": -343.0057067871094, | |
| "logps/rejected": -444.6160583496094, | |
| "loss": 0.3376, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.9290262460708618, | |
| "rewards/margins": 1.3633143901824951, | |
| "rewards/rejected": -2.2923407554626465, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.6649616368286445, | |
| "grad_norm": 47.933967621919784, | |
| "learning_rate": 3.060771981975726e-08, | |
| "logits/chosen": -5.2227373123168945, | |
| "logits/rejected": -5.566103935241699, | |
| "logps/chosen": -334.39306640625, | |
| "logps/rejected": -502.48211669921875, | |
| "loss": 0.3271, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -0.8672464489936829, | |
| "rewards/margins": 1.90872323513031, | |
| "rewards/rejected": -2.7759692668914795, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.690537084398977, | |
| "grad_norm": 37.918454908956036, | |
| "learning_rate": 2.6565944956764818e-08, | |
| "logits/chosen": -5.3946356773376465, | |
| "logits/rejected": -5.655900478363037, | |
| "logps/chosen": -349.11785888671875, | |
| "logps/rejected": -490.94720458984375, | |
| "loss": 0.3334, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.8713266253471375, | |
| "rewards/margins": 1.8250868320465088, | |
| "rewards/rejected": -2.696413516998291, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.7161125319693095, | |
| "grad_norm": 43.57355633058329, | |
| "learning_rate": 2.2711774490274766e-08, | |
| "logits/chosen": -5.313577651977539, | |
| "logits/rejected": -5.556331157684326, | |
| "logps/chosen": -347.73529052734375, | |
| "logps/rejected": -519.9879150390625, | |
| "loss": 0.3132, | |
| "rewards/accuracies": 0.8687499761581421, | |
| "rewards/chosen": -0.8589982986450195, | |
| "rewards/margins": 1.8729617595672607, | |
| "rewards/rejected": -2.731959819793701, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.7416879795396419, | |
| "grad_norm": 48.474190706284276, | |
| "learning_rate": 1.9076063486687256e-08, | |
| "logits/chosen": -5.153687477111816, | |
| "logits/rejected": -5.569068908691406, | |
| "logps/chosen": -342.547607421875, | |
| "logps/rejected": -473.8719787597656, | |
| "loss": 0.3082, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -0.7416984438896179, | |
| "rewards/margins": 1.8063793182373047, | |
| "rewards/rejected": -2.5480778217315674, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.7672634271099744, | |
| "grad_norm": 46.47610707831878, | |
| "learning_rate": 1.5687918106563324e-08, | |
| "logits/chosen": -5.316671371459961, | |
| "logits/rejected": -5.562603950500488, | |
| "logps/chosen": -339.9733581542969, | |
| "logps/rejected": -511.3289489746094, | |
| "loss": 0.3249, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.8634950518608093, | |
| "rewards/margins": 2.052011251449585, | |
| "rewards/rejected": -2.915506601333618, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.7672634271099744, | |
| "eval_logits/chosen": -5.345163345336914, | |
| "eval_logits/rejected": -5.603174209594727, | |
| "eval_logps/chosen": -585.470458984375, | |
| "eval_logps/rejected": -737.188720703125, | |
| "eval_loss": 0.7740334272384644, | |
| "eval_rewards/accuracies": 0.56640625, | |
| "eval_rewards/chosen": -1.9499597549438477, | |
| "eval_rewards/margins": 0.23805105686187744, | |
| "eval_rewards/rejected": -2.1880106925964355, | |
| "eval_runtime": 98.564, | |
| "eval_samples_per_second": 20.291, | |
| "eval_steps_per_second": 0.325, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.7928388746803069, | |
| "grad_norm": 43.72352052570602, | |
| "learning_rate": 1.257446259144494e-08, | |
| "logits/chosen": -5.234416961669922, | |
| "logits/rejected": -5.61405611038208, | |
| "logps/chosen": -347.01263427734375, | |
| "logps/rejected": -529.2141723632812, | |
| "loss": 0.3198, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.8649846911430359, | |
| "rewards/margins": 2.1693644523620605, | |
| "rewards/rejected": -3.034349203109741, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.8184143222506394, | |
| "grad_norm": 49.6074266455155, | |
| "learning_rate": 9.760622117187234e-09, | |
| "logits/chosen": -5.37489128112793, | |
| "logits/rejected": -5.706605434417725, | |
| "logps/chosen": -341.7244567871094, | |
| "logps/rejected": -499.4188537597656, | |
| "loss": 0.3337, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.9388681650161743, | |
| "rewards/margins": 1.9423179626464844, | |
| "rewards/rejected": -2.881186008453369, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.8439897698209718, | |
| "grad_norm": 47.67492578137724, | |
| "learning_rate": 7.2689232521989885e-09, | |
| "logits/chosen": -5.29935884475708, | |
| "logits/rejected": -5.635709762573242, | |
| "logps/chosen": -363.49591064453125, | |
| "logps/rejected": -517.4185180664062, | |
| "loss": 0.314, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.9602580070495605, | |
| "rewards/margins": 1.8374437093734741, | |
| "rewards/rejected": -2.797701597213745, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.8695652173913043, | |
| "grad_norm": 39.135188403156114, | |
| "learning_rate": 5.119313618049309e-09, | |
| "logits/chosen": -5.324515342712402, | |
| "logits/rejected": -5.7237324714660645, | |
| "logps/chosen": -366.15264892578125, | |
| "logps/rejected": -471.0101623535156, | |
| "loss": 0.3135, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -0.9129904508590698, | |
| "rewards/margins": 1.6364881992340088, | |
| "rewards/rejected": -2.549478530883789, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.8951406649616368, | |
| "grad_norm": 43.15700039763216, | |
| "learning_rate": 3.3290021961708158e-09, | |
| "logits/chosen": -5.358675956726074, | |
| "logits/rejected": -5.501515865325928, | |
| "logps/chosen": -351.61016845703125, | |
| "logps/rejected": -492.70294189453125, | |
| "loss": 0.3421, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -1.0095672607421875, | |
| "rewards/margins": 1.6046628952026367, | |
| "rewards/rejected": -2.614229917526245, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.9207161125319693, | |
| "grad_norm": 43.21141833066524, | |
| "learning_rate": 1.9123215591052013e-09, | |
| "logits/chosen": -5.308594703674316, | |
| "logits/rejected": -5.525307655334473, | |
| "logps/chosen": -355.9210510253906, | |
| "logps/rejected": -509.302001953125, | |
| "loss": 0.3229, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.9471023678779602, | |
| "rewards/margins": 1.8037946224212646, | |
| "rewards/rejected": -2.75089693069458, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.9462915601023018, | |
| "grad_norm": 45.460982961010096, | |
| "learning_rate": 8.806131292167618e-10, | |
| "logits/chosen": -5.335926055908203, | |
| "logits/rejected": -5.556703567504883, | |
| "logps/chosen": -349.836181640625, | |
| "logps/rejected": -495.7314453125, | |
| "loss": 0.3405, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.957531750202179, | |
| "rewards/margins": 1.595903754234314, | |
| "rewards/rejected": -2.5534353256225586, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.9718670076726342, | |
| "grad_norm": 44.49515613968937, | |
| "learning_rate": 2.4213638345040867e-10, | |
| "logits/chosen": -5.496818542480469, | |
| "logits/rejected": -5.771291255950928, | |
| "logps/chosen": -356.3369445800781, | |
| "logps/rejected": -501.03448486328125, | |
| "loss": 0.3269, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.9707748293876648, | |
| "rewards/margins": 1.7586256265640259, | |
| "rewards/rejected": -2.729400634765625, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.9974424552429667, | |
| "grad_norm": 54.69003383274636, | |
| "learning_rate": 2.0027310073833516e-12, | |
| "logits/chosen": -5.4842376708984375, | |
| "logits/rejected": -5.72286319732666, | |
| "logps/chosen": -359.2212829589844, | |
| "logps/rejected": -514.3115234375, | |
| "loss": 0.3144, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.9742454290390015, | |
| "rewards/margins": 1.8987795114517212, | |
| "rewards/rejected": -2.8730247020721436, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 391, | |
| "total_flos": 0.0, | |
| "train_loss": 0.4100414538932273, | |
| "train_runtime": 6205.0725, | |
| "train_samples_per_second": 8.058, | |
| "train_steps_per_second": 0.063 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 391, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |