| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.47426924660354053, | |
| "eval_steps": 500, | |
| "global_step": 36, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.013174145738987238, | |
| "grad_norm": 0.7230468392372131, | |
| "learning_rate": 1.25e-07, | |
| "logits/chosen": 9.993395805358887, | |
| "logits/rejected": 10.444890022277832, | |
| "logps/chosen": -0.8203260898590088, | |
| "logps/rejected": -0.8579692840576172, | |
| "loss": 1.4012, | |
| "rewards/accuracies": 0.546875, | |
| "rewards/chosen": -1.6406521797180176, | |
| "rewards/margins": 0.07528629899024963, | |
| "rewards/rejected": -1.7159385681152344, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.026348291477974475, | |
| "grad_norm": 1.0001572370529175, | |
| "learning_rate": 2.5e-07, | |
| "logits/chosen": 10.52372932434082, | |
| "logits/rejected": 10.788372993469238, | |
| "logps/chosen": -0.8189243078231812, | |
| "logps/rejected": -0.9178226590156555, | |
| "loss": 1.2834, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -1.6378486156463623, | |
| "rewards/margins": 0.19779649376869202, | |
| "rewards/rejected": -1.835645318031311, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.03952243721696171, | |
| "grad_norm": 0.979387640953064, | |
| "learning_rate": 3.75e-07, | |
| "logits/chosen": 10.042590141296387, | |
| "logits/rejected": 10.521858215332031, | |
| "logps/chosen": -0.917183518409729, | |
| "logps/rejected": -0.9847703576087952, | |
| "loss": 1.3664, | |
| "rewards/accuracies": 0.5546875, | |
| "rewards/chosen": -1.834367036819458, | |
| "rewards/margins": 0.13517364859580994, | |
| "rewards/rejected": -1.9695407152175903, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.05269658295594895, | |
| "grad_norm": 1.867851972579956, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": 10.141695022583008, | |
| "logits/rejected": 10.483091354370117, | |
| "logps/chosen": -0.8512520790100098, | |
| "logps/rejected": -0.8865377902984619, | |
| "loss": 1.4018, | |
| "rewards/accuracies": 0.546875, | |
| "rewards/chosen": -1.7025041580200195, | |
| "rewards/margins": 0.0705714151263237, | |
| "rewards/rejected": -1.7730755805969238, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.06587072869493618, | |
| "grad_norm": 0.5915809273719788, | |
| "learning_rate": 6.249999999999999e-07, | |
| "logits/chosen": 10.462024688720703, | |
| "logits/rejected": 11.114532470703125, | |
| "logps/chosen": -0.7721583247184753, | |
| "logps/rejected": -0.7985925078392029, | |
| "loss": 1.3829, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -1.5443166494369507, | |
| "rewards/margins": 0.052868396043777466, | |
| "rewards/rejected": -1.5971850156784058, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.07904487443392343, | |
| "grad_norm": 0.6252888441085815, | |
| "learning_rate": 7.5e-07, | |
| "logits/chosen": 10.718672752380371, | |
| "logits/rejected": 11.112823486328125, | |
| "logps/chosen": -0.8559192419052124, | |
| "logps/rejected": -0.9305768013000488, | |
| "loss": 1.3618, | |
| "rewards/accuracies": 0.5859375, | |
| "rewards/chosen": -1.7118384838104248, | |
| "rewards/margins": 0.1493152379989624, | |
| "rewards/rejected": -1.8611536026000977, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.09221902017291066, | |
| "grad_norm": 0.7208986878395081, | |
| "learning_rate": 8.75e-07, | |
| "logits/chosen": 9.864439010620117, | |
| "logits/rejected": 10.09216022491455, | |
| "logps/chosen": -0.7553099393844604, | |
| "logps/rejected": -0.8098596930503845, | |
| "loss": 1.3613, | |
| "rewards/accuracies": 0.5859375, | |
| "rewards/chosen": -1.510619878768921, | |
| "rewards/margins": 0.10909969359636307, | |
| "rewards/rejected": -1.619719386100769, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.1053931659118979, | |
| "grad_norm": 1.6331408023834229, | |
| "learning_rate": 1e-06, | |
| "logits/chosen": 10.65585994720459, | |
| "logits/rejected": 11.107338905334473, | |
| "logps/chosen": -0.774090051651001, | |
| "logps/rejected": -0.8291972279548645, | |
| "loss": 1.3462, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -1.548180103302002, | |
| "rewards/margins": 0.11021438986063004, | |
| "rewards/rejected": -1.658394455909729, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.11856731165088513, | |
| "grad_norm": 1.0860488414764404, | |
| "learning_rate": 9.994504457428556e-07, | |
| "logits/chosen": 10.446334838867188, | |
| "logits/rejected": 10.82742691040039, | |
| "logps/chosen": -0.938445508480072, | |
| "logps/rejected": -0.8894600868225098, | |
| "loss": 1.5125, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -1.876891016960144, | |
| "rewards/margins": -0.09797099232673645, | |
| "rewards/rejected": -1.7789201736450195, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.13174145738987236, | |
| "grad_norm": 1.3078101873397827, | |
| "learning_rate": 9.97802991010949e-07, | |
| "logits/chosen": 10.292729377746582, | |
| "logits/rejected": 10.629972457885742, | |
| "logps/chosen": -0.8660728931427002, | |
| "logps/rejected": -0.881781280040741, | |
| "loss": 1.4654, | |
| "rewards/accuracies": 0.5390625, | |
| "rewards/chosen": -1.7321457862854004, | |
| "rewards/margins": 0.0314166434109211, | |
| "rewards/rejected": -1.763562560081482, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.14491560312885962, | |
| "grad_norm": 1.0685713291168213, | |
| "learning_rate": 9.950612572673255e-07, | |
| "logits/chosen": 10.143664360046387, | |
| "logits/rejected": 10.639056205749512, | |
| "logps/chosen": -0.8788573145866394, | |
| "logps/rejected": -0.9421562552452087, | |
| "loss": 1.3638, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -1.7577146291732788, | |
| "rewards/margins": 0.12659788131713867, | |
| "rewards/rejected": -1.8843125104904175, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.15808974886784685, | |
| "grad_norm": 0.9801668524742126, | |
| "learning_rate": 9.912312714377879e-07, | |
| "logits/chosen": 10.19462776184082, | |
| "logits/rejected": 10.269098281860352, | |
| "logps/chosen": -0.8250212669372559, | |
| "logps/rejected": -0.8715258836746216, | |
| "loss": 1.3649, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -1.6500425338745117, | |
| "rewards/margins": 0.09300932288169861, | |
| "rewards/rejected": -1.7430517673492432, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.17126389460683408, | |
| "grad_norm": 0.9345762729644775, | |
| "learning_rate": 9.863214526624063e-07, | |
| "logits/chosen": 9.898378372192383, | |
| "logits/rejected": 10.37912368774414, | |
| "logps/chosen": -0.8251993656158447, | |
| "logps/rejected": -0.9702485799789429, | |
| "loss": 1.3106, | |
| "rewards/accuracies": 0.5546875, | |
| "rewards/chosen": -1.6503987312316895, | |
| "rewards/margins": 0.2900983989238739, | |
| "rewards/rejected": -1.9404971599578857, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.1844380403458213, | |
| "grad_norm": 0.6539536118507385, | |
| "learning_rate": 9.8034259378842e-07, | |
| "logits/chosen": 10.518888473510742, | |
| "logits/rejected": 11.198535919189453, | |
| "logps/chosen": -0.7980682253837585, | |
| "logps/rejected": -0.8631402254104614, | |
| "loss": 1.3561, | |
| "rewards/accuracies": 0.609375, | |
| "rewards/chosen": -1.596136450767517, | |
| "rewards/margins": 0.13014405965805054, | |
| "rewards/rejected": -1.7262804508209229, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.19761218608480857, | |
| "grad_norm": 0.3341844975948334, | |
| "learning_rate": 9.73307837645217e-07, | |
| "logits/chosen": 10.087669372558594, | |
| "logits/rejected": 10.595584869384766, | |
| "logps/chosen": -0.7868794798851013, | |
| "logps/rejected": -0.8722506165504456, | |
| "loss": 1.3174, | |
| "rewards/accuracies": 0.5546875, | |
| "rewards/chosen": -1.5737589597702026, | |
| "rewards/margins": 0.17074236273765564, | |
| "rewards/rejected": -1.7445012331008911, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.2107863318237958, | |
| "grad_norm": 1.1852020025253296, | |
| "learning_rate": 9.652326481535433e-07, | |
| "logits/chosen": 10.743181228637695, | |
| "logits/rejected": 11.05751895904541, | |
| "logps/chosen": -0.791871190071106, | |
| "logps/rejected": -0.8441797494888306, | |
| "loss": 1.3868, | |
| "rewards/accuracies": 0.5078125, | |
| "rewards/chosen": -1.583742380142212, | |
| "rewards/margins": 0.1046171486377716, | |
| "rewards/rejected": -1.6883594989776611, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.22396047756278303, | |
| "grad_norm": 0.5402158498764038, | |
| "learning_rate": 9.561347763324483e-07, | |
| "logits/chosen": 10.244048118591309, | |
| "logits/rejected": 10.602608680725098, | |
| "logps/chosen": -0.8015426993370056, | |
| "logps/rejected": -0.823470950126648, | |
| "loss": 1.4002, | |
| "rewards/accuracies": 0.5546875, | |
| "rewards/chosen": -1.6030853986740112, | |
| "rewards/margins": 0.04385652765631676, | |
| "rewards/rejected": -1.646941900253296, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.23713462330177026, | |
| "grad_norm": 1.1475958824157715, | |
| "learning_rate": 9.460342212786932e-07, | |
| "logits/chosen": 10.25997543334961, | |
| "logits/rejected": 10.875801086425781, | |
| "logps/chosen": -0.9126402735710144, | |
| "logps/rejected": -0.8353321552276611, | |
| "loss": 1.5624, | |
| "rewards/accuracies": 0.5390625, | |
| "rewards/chosen": -1.8252805471420288, | |
| "rewards/margins": -0.1546163707971573, | |
| "rewards/rejected": -1.6706643104553223, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.2503087690407575, | |
| "grad_norm": 0.7070759534835815, | |
| "learning_rate": 9.349531862043951e-07, | |
| "logits/chosen": 10.557324409484863, | |
| "logits/rejected": 10.653444290161133, | |
| "logps/chosen": -0.8390055894851685, | |
| "logps/rejected": -0.8851607441902161, | |
| "loss": 1.3597, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -1.678011178970337, | |
| "rewards/margins": 0.09231019765138626, | |
| "rewards/rejected": -1.7703214883804321, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.2634829147797447, | |
| "grad_norm": 1.4036712646484375, | |
| "learning_rate": 9.229160296295487e-07, | |
| "logits/chosen": 10.47518253326416, | |
| "logits/rejected": 11.045392990112305, | |
| "logps/chosen": -0.861572265625, | |
| "logps/rejected": -0.8943375945091248, | |
| "loss": 1.4097, | |
| "rewards/accuracies": 0.578125, | |
| "rewards/chosen": -1.72314453125, | |
| "rewards/margins": 0.06553083658218384, | |
| "rewards/rejected": -1.7886751890182495, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.276657060518732, | |
| "grad_norm": 0.537728488445282, | |
| "learning_rate": 9.099492118367122e-07, | |
| "logits/chosen": 10.20445728302002, | |
| "logits/rejected": 10.791345596313477, | |
| "logps/chosen": -0.8058649301528931, | |
| "logps/rejected": -0.9033350944519043, | |
| "loss": 1.2939, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -1.6117298603057861, | |
| "rewards/margins": 0.19494034349918365, | |
| "rewards/rejected": -1.8066701889038086, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.28983120625771924, | |
| "grad_norm": 1.7850818634033203, | |
| "learning_rate": 8.960812367055646e-07, | |
| "logits/chosen": 10.360614776611328, | |
| "logits/rejected": 10.701253890991211, | |
| "logps/chosen": -0.8701018691062927, | |
| "logps/rejected": -0.8806692957878113, | |
| "loss": 1.4251, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -1.7402037382125854, | |
| "rewards/margins": 0.021134957671165466, | |
| "rewards/rejected": -1.7613385915756226, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.3030053519967065, | |
| "grad_norm": 0.34783315658569336, | |
| "learning_rate": 8.813425890551909e-07, | |
| "logits/chosen": 10.343542098999023, | |
| "logits/rejected": 10.988011360168457, | |
| "logps/chosen": -0.8861572742462158, | |
| "logps/rejected": -0.8626466989517212, | |
| "loss": 1.4816, | |
| "rewards/accuracies": 0.546875, | |
| "rewards/chosen": -1.7723145484924316, | |
| "rewards/margins": -0.04702116921544075, | |
| "rewards/rejected": -1.7252933979034424, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.3161794977356937, | |
| "grad_norm": 0.642213761806488, | |
| "learning_rate": 8.657656676318345e-07, | |
| "logits/chosen": 10.253456115722656, | |
| "logits/rejected": 10.651442527770996, | |
| "logps/chosen": -0.8050903677940369, | |
| "logps/rejected": -0.8499569296836853, | |
| "loss": 1.3893, | |
| "rewards/accuracies": 0.5078125, | |
| "rewards/chosen": -1.6101807355880737, | |
| "rewards/margins": 0.08973318338394165, | |
| "rewards/rejected": -1.6999138593673706, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.32935364347468093, | |
| "grad_norm": 2.578787088394165, | |
| "learning_rate": 8.493847138894208e-07, | |
| "logits/chosen": 10.33389949798584, | |
| "logits/rejected": 10.824821472167969, | |
| "logps/chosen": -0.7404808402061462, | |
| "logps/rejected": -0.7938650846481323, | |
| "loss": 1.3623, | |
| "rewards/accuracies": 0.578125, | |
| "rewards/chosen": -1.4809616804122925, | |
| "rewards/margins": 0.10676843672990799, | |
| "rewards/rejected": -1.5877301692962646, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.34252778921366817, | |
| "grad_norm": 0.6688835620880127, | |
| "learning_rate": 8.322357367194108e-07, | |
| "logits/chosen": 10.620444297790527, | |
| "logits/rejected": 10.987481117248535, | |
| "logps/chosen": -0.81562340259552, | |
| "logps/rejected": -0.8406739234924316, | |
| "loss": 1.4056, | |
| "rewards/accuracies": 0.5078125, | |
| "rewards/chosen": -1.63124680519104, | |
| "rewards/margins": 0.050100862979888916, | |
| "rewards/rejected": -1.6813478469848633, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.3557019349526554, | |
| "grad_norm": 1.578236699104309, | |
| "learning_rate": 8.143564332954425e-07, | |
| "logits/chosen": 10.240825653076172, | |
| "logits/rejected": 10.97463321685791, | |
| "logps/chosen": -0.86244797706604, | |
| "logps/rejected": -0.852953314781189, | |
| "loss": 1.4416, | |
| "rewards/accuracies": 0.4921875, | |
| "rewards/chosen": -1.72489595413208, | |
| "rewards/margins": -0.01898936554789543, | |
| "rewards/rejected": -1.705906629562378, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.3688760806916426, | |
| "grad_norm": 0.8747299313545227, | |
| "learning_rate": 7.957861062067612e-07, | |
| "logits/chosen": 9.665928840637207, | |
| "logits/rejected": 10.360410690307617, | |
| "logps/chosen": -0.7768715620040894, | |
| "logps/rejected": -0.8124402761459351, | |
| "loss": 1.3916, | |
| "rewards/accuracies": 0.5703125, | |
| "rewards/chosen": -1.5537431240081787, | |
| "rewards/margins": 0.07113761454820633, | |
| "rewards/rejected": -1.6248805522918701, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.3820502264306299, | |
| "grad_norm": 0.8407192826271057, | |
| "learning_rate": 7.765655770625996e-07, | |
| "logits/chosen": 10.551313400268555, | |
| "logits/rejected": 10.729636192321777, | |
| "logps/chosen": -0.747046709060669, | |
| "logps/rejected": -0.7931259870529175, | |
| "loss": 1.3627, | |
| "rewards/accuracies": 0.640625, | |
| "rewards/chosen": -1.494093418121338, | |
| "rewards/margins": 0.09215845912694931, | |
| "rewards/rejected": -1.586251974105835, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.39522437216961714, | |
| "grad_norm": 0.650399923324585, | |
| "learning_rate": 7.567370967574209e-07, | |
| "logits/chosen": 10.303435325622559, | |
| "logits/rejected": 11.132701873779297, | |
| "logps/chosen": -0.8180528283119202, | |
| "logps/rejected": -0.8380974531173706, | |
| "loss": 1.4193, | |
| "rewards/accuracies": 0.515625, | |
| "rewards/chosen": -1.6361056566238403, | |
| "rewards/margins": 0.04008916765451431, | |
| "rewards/rejected": -1.6761949062347412, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.4083985179086044, | |
| "grad_norm": 0.8736916780471802, | |
| "learning_rate": 7.363442525942826e-07, | |
| "logits/chosen": 10.033166885375977, | |
| "logits/rejected": 10.649832725524902, | |
| "logps/chosen": -0.7705740332603455, | |
| "logps/rejected": -0.8067623376846313, | |
| "loss": 1.3867, | |
| "rewards/accuracies": 0.5546875, | |
| "rewards/chosen": -1.541148066520691, | |
| "rewards/margins": 0.07237657904624939, | |
| "rewards/rejected": -1.6135246753692627, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.4215726636475916, | |
| "grad_norm": 0.2823183536529541, | |
| "learning_rate": 7.154318724704851e-07, | |
| "logits/chosen": 10.305078506469727, | |
| "logits/rejected": 10.909933090209961, | |
| "logps/chosen": -0.7408677339553833, | |
| "logps/rejected": -0.7292467355728149, | |
| "loss": 1.4487, | |
| "rewards/accuracies": 0.5078125, | |
| "rewards/chosen": -1.4817354679107666, | |
| "rewards/margins": -0.02324201911687851, | |
| "rewards/rejected": -1.4584934711456299, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.43474680938657884, | |
| "grad_norm": 0.9221538305282593, | |
| "learning_rate": 6.940459263361248e-07, | |
| "logits/chosen": 10.10401725769043, | |
| "logits/rejected": 10.791908264160156, | |
| "logps/chosen": -0.8303573131561279, | |
| "logps/rejected": -0.7867410182952881, | |
| "loss": 1.5131, | |
| "rewards/accuracies": 0.484375, | |
| "rewards/chosen": -1.6607146263122559, | |
| "rewards/margins": -0.08723282814025879, | |
| "rewards/rejected": -1.5734820365905762, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.44792095512556607, | |
| "grad_norm": 0.7228991389274597, | |
| "learning_rate": 6.722334251421664e-07, | |
| "logits/chosen": 9.976701736450195, | |
| "logits/rejected": 10.480062484741211, | |
| "logps/chosen": -0.923545241355896, | |
| "logps/rejected": -0.9711360335350037, | |
| "loss": 1.4215, | |
| "rewards/accuracies": 0.5859375, | |
| "rewards/chosen": -1.847090482711792, | |
| "rewards/margins": 0.09518149495124817, | |
| "rewards/rejected": -1.9422720670700073, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.4610951008645533, | |
| "grad_norm": 2.438734769821167, | |
| "learning_rate": 6.500423175001703e-07, | |
| "logits/chosen": 10.54722785949707, | |
| "logits/rejected": 10.914085388183594, | |
| "logps/chosen": -0.8670656681060791, | |
| "logps/rejected": -0.9043451547622681, | |
| "loss": 1.4441, | |
| "rewards/accuracies": 0.5703125, | |
| "rewards/chosen": -1.7341313362121582, | |
| "rewards/margins": 0.07455900311470032, | |
| "rewards/rejected": -1.8086903095245361, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.47426924660354053, | |
| "grad_norm": 0.5184939503669739, | |
| "learning_rate": 6.275213842808382e-07, | |
| "logits/chosen": 10.422931671142578, | |
| "logits/rejected": 10.859081268310547, | |
| "logps/chosen": -0.7457296848297119, | |
| "logps/rejected": -0.7759172320365906, | |
| "loss": 1.385, | |
| "rewards/accuracies": 0.515625, | |
| "rewards/chosen": -1.4914593696594238, | |
| "rewards/margins": 0.0603751465678215, | |
| "rewards/rejected": -1.5518344640731812, | |
| "step": 36 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 75, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 12, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |