| { | |
| "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": 0.0010465724751439038, | |
| "grad_norm": 4.695332597526886, | |
| "learning_rate": 5.208333333333333e-09, | |
| "logits/chosen": -0.7187488675117493, | |
| "logits/rejected": -0.5983389019966125, | |
| "logps/chosen": -433.40936279296875, | |
| "logps/rejected": -346.82391357421875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.010465724751439037, | |
| "grad_norm": 4.094353723689003, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -0.4951174557209015, | |
| "logits/rejected": -0.4637991189956665, | |
| "logps/chosen": -283.0612487792969, | |
| "logps/rejected": -261.8544921875, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.3888888955116272, | |
| "rewards/chosen": -0.00028089413535781205, | |
| "rewards/margins": -0.000883792235981673, | |
| "rewards/rejected": 0.0006028979551047087, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.020931449502878074, | |
| "grad_norm": 4.493448151438676, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -0.5218612551689148, | |
| "logits/rejected": -0.4537959694862366, | |
| "logps/chosen": -301.08050537109375, | |
| "logps/rejected": -288.3440856933594, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.0001712679659249261, | |
| "rewards/margins": 0.00038912202580831945, | |
| "rewards/rejected": -0.00021785404533147812, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03139717425431711, | |
| "grad_norm": 4.560411028573212, | |
| "learning_rate": 1.5624999999999999e-07, | |
| "logits/chosen": -0.5537956953048706, | |
| "logits/rejected": -0.47262096405029297, | |
| "logps/chosen": -287.0794372558594, | |
| "logps/rejected": -264.6256103515625, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.0005557264084927738, | |
| "rewards/margins": -0.00028174123144708574, | |
| "rewards/rejected": 0.0008374677272513509, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04186289900575615, | |
| "grad_norm": 4.747451905252561, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -0.5157141089439392, | |
| "logits/rejected": -0.42310914397239685, | |
| "logps/chosen": -324.68865966796875, | |
| "logps/rejected": -281.6961364746094, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.004000083543360233, | |
| "rewards/margins": 0.0023048361763358116, | |
| "rewards/rejected": 0.0016952479491010308, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.052328623757195186, | |
| "grad_norm": 4.1236572047382545, | |
| "learning_rate": 2.604166666666667e-07, | |
| "logits/chosen": -0.4842318594455719, | |
| "logits/rejected": -0.4082712233066559, | |
| "logps/chosen": -241.1922607421875, | |
| "logps/rejected": -204.27723693847656, | |
| "loss": 0.6916, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.009572440758347511, | |
| "rewards/margins": 0.004367954097688198, | |
| "rewards/rejected": 0.0052044871263206005, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06279434850863422, | |
| "grad_norm": 4.38318666746926, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -0.5186334252357483, | |
| "logits/rejected": -0.47192978858947754, | |
| "logps/chosen": -322.8270568847656, | |
| "logps/rejected": -279.4931640625, | |
| "loss": 0.6901, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.019420895725488663, | |
| "rewards/margins": 0.006516781635582447, | |
| "rewards/rejected": 0.012904113158583641, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07326007326007326, | |
| "grad_norm": 4.392009777368473, | |
| "learning_rate": 3.645833333333333e-07, | |
| "logits/chosen": -0.44262728095054626, | |
| "logits/rejected": -0.4167408049106598, | |
| "logps/chosen": -241.4150848388672, | |
| "logps/rejected": -265.60296630859375, | |
| "loss": 0.6873, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.02704036608338356, | |
| "rewards/margins": 0.014237403869628906, | |
| "rewards/rejected": 0.01280296128243208, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.0837257980115123, | |
| "grad_norm": 4.432632376020607, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -0.5181559920310974, | |
| "logits/rejected": -0.4063476622104645, | |
| "logps/chosen": -259.34405517578125, | |
| "logps/rejected": -240.4481658935547, | |
| "loss": 0.683, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.032934751361608505, | |
| "rewards/margins": 0.019226713106036186, | |
| "rewards/rejected": 0.013708041980862617, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.09419152276295134, | |
| "grad_norm": 3.7960885338069263, | |
| "learning_rate": 4.6874999999999996e-07, | |
| "logits/chosen": -0.4854390025138855, | |
| "logits/rejected": -0.44603580236434937, | |
| "logps/chosen": -280.05841064453125, | |
| "logps/rejected": -260.02008056640625, | |
| "loss": 0.6793, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.03177185729146004, | |
| "rewards/margins": 0.019615832716226578, | |
| "rewards/rejected": 0.01215602271258831, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "grad_norm": 5.321767060392544, | |
| "learning_rate": 4.999732492681437e-07, | |
| "logits/chosen": -0.5290640592575073, | |
| "logits/rejected": -0.4432452321052551, | |
| "logps/chosen": -280.30877685546875, | |
| "logps/rejected": -274.52349853515625, | |
| "loss": 0.6703, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.03181055933237076, | |
| "rewards/margins": 0.04814142733812332, | |
| "rewards/rejected": -0.016330868005752563, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "eval_logits/chosen": -0.5812728404998779, | |
| "eval_logits/rejected": -0.4756637215614319, | |
| "eval_logps/chosen": -285.9060974121094, | |
| "eval_logps/rejected": -265.83782958984375, | |
| "eval_loss": 0.6694991588592529, | |
| "eval_rewards/accuracies": 0.682539701461792, | |
| "eval_rewards/chosen": 0.017283162102103233, | |
| "eval_rewards/margins": 0.05808327719569206, | |
| "eval_rewards/rejected": -0.04080010578036308, | |
| "eval_runtime": 227.6437, | |
| "eval_samples_per_second": 8.786, | |
| "eval_steps_per_second": 0.277, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1151229722658294, | |
| "grad_norm": 4.6587263741397384, | |
| "learning_rate": 4.996723692767926e-07, | |
| "logits/chosen": -0.5552923083305359, | |
| "logits/rejected": -0.463242769241333, | |
| "logps/chosen": -263.10443115234375, | |
| "logps/rejected": -231.2692413330078, | |
| "loss": 0.6685, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.011513032019138336, | |
| "rewards/margins": 0.08094070851802826, | |
| "rewards/rejected": -0.06942768394947052, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12558869701726844, | |
| "grad_norm": 5.15038375709292, | |
| "learning_rate": 4.990375746213598e-07, | |
| "logits/chosen": -0.5475479960441589, | |
| "logits/rejected": -0.5249751210212708, | |
| "logps/chosen": -238.50537109375, | |
| "logps/rejected": -264.0198974609375, | |
| "loss": 0.6508, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.0012080561136826873, | |
| "rewards/margins": 0.08254374563694, | |
| "rewards/rejected": -0.08375179767608643, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.1360544217687075, | |
| "grad_norm": 5.6036300426343715, | |
| "learning_rate": 4.980697142834314e-07, | |
| "logits/chosen": -0.6076304316520691, | |
| "logits/rejected": -0.5554308295249939, | |
| "logps/chosen": -297.47943115234375, | |
| "logps/rejected": -291.3108825683594, | |
| "loss": 0.6389, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.10223034769296646, | |
| "rewards/margins": 0.1436336785554886, | |
| "rewards/rejected": -0.24586400389671326, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.14652014652014653, | |
| "grad_norm": 6.407440966129605, | |
| "learning_rate": 4.967700826904229e-07, | |
| "logits/chosen": -0.576185405254364, | |
| "logits/rejected": -0.5214563608169556, | |
| "logps/chosen": -312.6526794433594, | |
| "logps/rejected": -317.4207763671875, | |
| "loss": 0.6255, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.17939871549606323, | |
| "rewards/margins": 0.13575369119644165, | |
| "rewards/rejected": -0.3151523768901825, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15698587127158556, | |
| "grad_norm": 6.101819441628303, | |
| "learning_rate": 4.951404179843962e-07, | |
| "logits/chosen": -0.7110940217971802, | |
| "logits/rejected": -0.5994465947151184, | |
| "logps/chosen": -330.38751220703125, | |
| "logps/rejected": -281.0096435546875, | |
| "loss": 0.6202, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.14434388279914856, | |
| "rewards/margins": 0.2144981324672699, | |
| "rewards/rejected": -0.35884204506874084, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.1674515960230246, | |
| "grad_norm": 20.68712383660321, | |
| "learning_rate": 4.931828996974498e-07, | |
| "logits/chosen": -0.7509492635726929, | |
| "logits/rejected": -0.571489155292511, | |
| "logps/chosen": -354.3675231933594, | |
| "logps/rejected": -311.22052001953125, | |
| "loss": 0.5982, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.2555951774120331, | |
| "rewards/margins": 0.3567826747894287, | |
| "rewards/rejected": -0.6123778223991394, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.17791732077446362, | |
| "grad_norm": 8.080191404091543, | |
| "learning_rate": 4.909001458367866e-07, | |
| "logits/chosen": -0.6748760938644409, | |
| "logits/rejected": -0.5670477747917175, | |
| "logps/chosen": -308.5774841308594, | |
| "logps/rejected": -334.3050842285156, | |
| "loss": 0.5899, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.4368668496608734, | |
| "rewards/margins": 0.32547345757484436, | |
| "rewards/rejected": -0.7623404264450073, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.18838304552590268, | |
| "grad_norm": 11.93364822141488, | |
| "learning_rate": 4.882952093833627e-07, | |
| "logits/chosen": -0.7000716924667358, | |
| "logits/rejected": -0.6841859221458435, | |
| "logps/chosen": -266.9092712402344, | |
| "logps/rejected": -318.55902099609375, | |
| "loss": 0.5811, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.33145153522491455, | |
| "rewards/margins": 0.372114360332489, | |
| "rewards/rejected": -0.7035658955574036, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.1988487702773417, | |
| "grad_norm": 8.64087542458168, | |
| "learning_rate": 4.853715742087946e-07, | |
| "logits/chosen": -0.6523844003677368, | |
| "logits/rejected": -0.6216700673103333, | |
| "logps/chosen": -304.2262268066406, | |
| "logps/rejected": -387.6643981933594, | |
| "loss": 0.5856, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.47744685411453247, | |
| "rewards/margins": 0.3826759457588196, | |
| "rewards/rejected": -0.860122799873352, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "grad_norm": 12.162378152563793, | |
| "learning_rate": 4.821331504159906e-07, | |
| "logits/chosen": -0.8054048418998718, | |
| "logits/rejected": -0.660961925983429, | |
| "logps/chosen": -370.180419921875, | |
| "logps/rejected": -326.85565185546875, | |
| "loss": 0.5922, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.3945849537849426, | |
| "rewards/margins": 0.4013177454471588, | |
| "rewards/rejected": -0.7959026098251343, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "eval_logits/chosen": -0.74112468957901, | |
| "eval_logits/rejected": -0.6443408131599426, | |
| "eval_logps/chosen": -333.7917175292969, | |
| "eval_logps/rejected": -346.79608154296875, | |
| "eval_loss": 0.5902364253997803, | |
| "eval_rewards/accuracies": 0.7063491940498352, | |
| "eval_rewards/chosen": -0.46157318353652954, | |
| "eval_rewards/margins": 0.388809472322464, | |
| "eval_rewards/rejected": -0.8503828048706055, | |
| "eval_runtime": 226.0253, | |
| "eval_samples_per_second": 8.849, | |
| "eval_steps_per_second": 0.279, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21978021978021978, | |
| "grad_norm": 9.47218509665613, | |
| "learning_rate": 4.785842691097342e-07, | |
| "logits/chosen": -0.6910367012023926, | |
| "logits/rejected": -0.6332982182502747, | |
| "logps/chosen": -335.70611572265625, | |
| "logps/rejected": -382.6346740722656, | |
| "loss": 0.5955, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.5499948859214783, | |
| "rewards/margins": 0.3104594051837921, | |
| "rewards/rejected": -0.860454261302948, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.2302459445316588, | |
| "grad_norm": 9.803517786698613, | |
| "learning_rate": 4.7472967660421603e-07, | |
| "logits/chosen": -0.7180417776107788, | |
| "logits/rejected": -0.621491014957428, | |
| "logps/chosen": -331.8623046875, | |
| "logps/rejected": -331.8077087402344, | |
| "loss": 0.5901, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.46668997406959534, | |
| "rewards/margins": 0.48732930421829224, | |
| "rewards/rejected": -0.9540191888809204, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.24071166928309787, | |
| "grad_norm": 10.560915230537866, | |
| "learning_rate": 4.705745280752585e-07, | |
| "logits/chosen": -0.5524531006813049, | |
| "logits/rejected": -0.5356574058532715, | |
| "logps/chosen": -321.02392578125, | |
| "logps/rejected": -364.1357116699219, | |
| "loss": 0.578, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.621025800704956, | |
| "rewards/margins": 0.4221344590187073, | |
| "rewards/rejected": -1.0431602001190186, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.25117739403453687, | |
| "grad_norm": 9.336666465482518, | |
| "learning_rate": 4.6612438066572555e-07, | |
| "logits/chosen": -0.5498469471931458, | |
| "logits/rejected": -0.5183900594711304, | |
| "logps/chosen": -326.112060546875, | |
| "logps/rejected": -366.96673583984375, | |
| "loss": 0.5612, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.6049496531486511, | |
| "rewards/margins": 0.49857011437416077, | |
| "rewards/rejected": -1.1035196781158447, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.2616431187859759, | |
| "grad_norm": 14.340514917233914, | |
| "learning_rate": 4.6138518605333664e-07, | |
| "logits/chosen": -0.5425869226455688, | |
| "logits/rejected": -0.5251628160476685, | |
| "logps/chosen": -321.1498107910156, | |
| "logps/rejected": -371.6676330566406, | |
| "loss": 0.5388, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.5140849351882935, | |
| "rewards/margins": 0.5452972650527954, | |
| "rewards/rejected": -1.0593822002410889, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.272108843537415, | |
| "grad_norm": 10.300342307961989, | |
| "learning_rate": 4.5636328249082514e-07, | |
| "logits/chosen": -0.6172841787338257, | |
| "logits/rejected": -0.45737963914871216, | |
| "logps/chosen": -340.3819274902344, | |
| "logps/rejected": -368.44586181640625, | |
| "loss": 0.5476, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.6206396222114563, | |
| "rewards/margins": 0.3971046805381775, | |
| "rewards/rejected": -1.0177444219589233, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.282574568288854, | |
| "grad_norm": 11.855640878189652, | |
| "learning_rate": 4.510653863290871e-07, | |
| "logits/chosen": -0.5061743259429932, | |
| "logits/rejected": -0.3946554660797119, | |
| "logps/chosen": -368.46331787109375, | |
| "logps/rejected": -373.1842346191406, | |
| "loss": 0.5366, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.6316210627555847, | |
| "rewards/margins": 0.5877398252487183, | |
| "rewards/rejected": -1.2193609476089478, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.29304029304029305, | |
| "grad_norm": 13.86402294982031, | |
| "learning_rate": 4.4549858303465737e-07, | |
| "logits/chosen": -0.4837842881679535, | |
| "logits/rejected": -0.3030511736869812, | |
| "logps/chosen": -394.0540771484375, | |
| "logps/rejected": -411.68902587890625, | |
| "loss": 0.567, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.7208216786384583, | |
| "rewards/margins": 0.6727169752120972, | |
| "rewards/rejected": -1.3935387134552002, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.3035060177917321, | |
| "grad_norm": 14.705610202375755, | |
| "learning_rate": 4.396703177135261e-07, | |
| "logits/chosen": -0.5852281451225281, | |
| "logits/rejected": -0.36015063524246216, | |
| "logps/chosen": -368.44146728515625, | |
| "logps/rejected": -360.4360046386719, | |
| "loss": 0.5489, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.6637938022613525, | |
| "rewards/margins": 0.5562569499015808, | |
| "rewards/rejected": -1.2200508117675781, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "grad_norm": 14.361033433220218, | |
| "learning_rate": 4.335883851539693e-07, | |
| "logits/chosen": -0.36654001474380493, | |
| "logits/rejected": -0.317962646484375, | |
| "logps/chosen": -305.54656982421875, | |
| "logps/rejected": -349.1251525878906, | |
| "loss": 0.5592, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.5999817848205566, | |
| "rewards/margins": 0.45001593232154846, | |
| "rewards/rejected": -1.0499978065490723, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "eval_logits/chosen": -0.43302780389785767, | |
| "eval_logits/rejected": -0.2678911089897156, | |
| "eval_logps/chosen": -349.07769775390625, | |
| "eval_logps/rejected": -383.3018493652344, | |
| "eval_loss": 0.5462030172348022, | |
| "eval_rewards/accuracies": 0.7420634627342224, | |
| "eval_rewards/chosen": -0.6144329309463501, | |
| "eval_rewards/margins": 0.6010070443153381, | |
| "eval_rewards/rejected": -1.2154401540756226, | |
| "eval_runtime": 227.0069, | |
| "eval_samples_per_second": 8.81, | |
| "eval_steps_per_second": 0.278, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32443746729461015, | |
| "grad_norm": 12.59126443650798, | |
| "learning_rate": 4.272609194017105e-07, | |
| "logits/chosen": -0.5242592096328735, | |
| "logits/rejected": -0.22999849915504456, | |
| "logps/chosen": -345.8105163574219, | |
| "logps/rejected": -334.08111572265625, | |
| "loss": 0.5479, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.6000981330871582, | |
| "rewards/margins": 0.637532114982605, | |
| "rewards/rejected": -1.2376301288604736, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.3349031920460492, | |
| "grad_norm": 11.732525849526553, | |
| "learning_rate": 4.2069638288135547e-07, | |
| "logits/chosen": -0.34528014063835144, | |
| "logits/rejected": -0.2457645833492279, | |
| "logps/chosen": -330.71527099609375, | |
| "logps/rejected": -412.20953369140625, | |
| "loss": 0.55, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.695503294467926, | |
| "rewards/margins": 0.5903774499893188, | |
| "rewards/rejected": -1.2858808040618896, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.3453689167974882, | |
| "grad_norm": 13.250553221571959, | |
| "learning_rate": 4.139035550786494e-07, | |
| "logits/chosen": -0.21062567830085754, | |
| "logits/rejected": -0.08340445160865784, | |
| "logps/chosen": -331.62542724609375, | |
| "logps/rejected": -363.3686828613281, | |
| "loss": 0.5412, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.7190684676170349, | |
| "rewards/margins": 0.5235351920127869, | |
| "rewards/rejected": -1.2426036596298218, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.35583464154892724, | |
| "grad_norm": 12.344702305971902, | |
| "learning_rate": 4.0689152079869306e-07, | |
| "logits/chosen": -0.05835915356874466, | |
| "logits/rejected": 0.08525937795639038, | |
| "logps/chosen": -335.3782958984375, | |
| "logps/rejected": -347.57257080078125, | |
| "loss": 0.5462, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.8475249409675598, | |
| "rewards/margins": 0.44163185358047485, | |
| "rewards/rejected": -1.2891566753387451, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.3663003663003663, | |
| "grad_norm": 13.72798924306665, | |
| "learning_rate": 3.99669658015821e-07, | |
| "logits/chosen": -0.2664807438850403, | |
| "logits/rejected": -0.0909152552485466, | |
| "logps/chosen": -316.62506103515625, | |
| "logps/rejected": -341.7252197265625, | |
| "loss": 0.535, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.6483460664749146, | |
| "rewards/margins": 0.47694116830825806, | |
| "rewards/rejected": -1.1252872943878174, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.37676609105180536, | |
| "grad_norm": 16.67040726492694, | |
| "learning_rate": 3.92247625331392e-07, | |
| "logits/chosen": -0.3472316861152649, | |
| "logits/rejected": -0.20045724511146545, | |
| "logps/chosen": -332.7682800292969, | |
| "logps/rejected": -366.4549865722656, | |
| "loss": 0.5055, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.5464685559272766, | |
| "rewards/margins": 0.5930379629135132, | |
| "rewards/rejected": -1.1395065784454346, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.3872318158032444, | |
| "grad_norm": 14.93675802203948, | |
| "learning_rate": 3.846353490562664e-07, | |
| "logits/chosen": -0.325096994638443, | |
| "logits/rejected": -0.2412107288837433, | |
| "logps/chosen": -333.1510925292969, | |
| "logps/rejected": -368.9770202636719, | |
| "loss": 0.5258, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.7361698746681213, | |
| "rewards/margins": 0.49931269884109497, | |
| "rewards/rejected": -1.2354824542999268, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.3976975405546834, | |
| "grad_norm": 13.251455539085676, | |
| "learning_rate": 3.768430099352445e-07, | |
| "logits/chosen": -0.21928231418132782, | |
| "logits/rejected": -0.08141541481018066, | |
| "logps/chosen": -347.12762451171875, | |
| "logps/rejected": -409.29571533203125, | |
| "loss": 0.5477, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.8551673889160156, | |
| "rewards/margins": 0.5591238737106323, | |
| "rewards/rejected": -1.4142911434173584, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.40816326530612246, | |
| "grad_norm": 12.09475227958707, | |
| "learning_rate": 3.6888102953122304e-07, | |
| "logits/chosen": -0.19851084053516388, | |
| "logits/rejected": -0.08052431792020798, | |
| "logps/chosen": -308.36724853515625, | |
| "logps/rejected": -357.89349365234375, | |
| "loss": 0.542, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.7273977398872375, | |
| "rewards/margins": 0.501800537109375, | |
| "rewards/rejected": -1.2291982173919678, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "grad_norm": 27.84407880505837, | |
| "learning_rate": 3.607600562872785e-07, | |
| "logits/chosen": -0.20401215553283691, | |
| "logits/rejected": -0.21550790965557098, | |
| "logps/chosen": -336.02923583984375, | |
| "logps/rejected": -415.23760986328125, | |
| "loss": 0.5461, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.8249620199203491, | |
| "rewards/margins": 0.3732803463935852, | |
| "rewards/rejected": -1.1982421875, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "eval_logits/chosen": -0.24119864404201508, | |
| "eval_logits/rejected": -0.010025140829384327, | |
| "eval_logps/chosen": -357.9294738769531, | |
| "eval_logps/rejected": -397.44207763671875, | |
| "eval_loss": 0.5322944521903992, | |
| "eval_rewards/accuracies": 0.738095223903656, | |
| "eval_rewards/chosen": -0.7029505372047424, | |
| "eval_rewards/margins": 0.6538920998573303, | |
| "eval_rewards/rejected": -1.3568426370620728, | |
| "eval_runtime": 230.0486, | |
| "eval_samples_per_second": 8.694, | |
| "eval_steps_per_second": 0.274, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4290947148090005, | |
| "grad_norm": 14.555012611239015, | |
| "learning_rate": 3.5249095128531856e-07, | |
| "logits/chosen": -0.054116807878017426, | |
| "logits/rejected": 0.22516992688179016, | |
| "logps/chosen": -371.7142639160156, | |
| "logps/rejected": -408.3119201660156, | |
| "loss": 0.5431, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.9307588338851929, | |
| "rewards/margins": 0.6702965497970581, | |
| "rewards/rejected": -1.6010555028915405, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.43956043956043955, | |
| "grad_norm": 13.358004942776654, | |
| "learning_rate": 3.4408477372034736e-07, | |
| "logits/chosen": -0.11378423124551773, | |
| "logits/rejected": 0.2778463363647461, | |
| "logps/chosen": -429.87109375, | |
| "logps/rejected": -423.0447692871094, | |
| "loss": 0.5372, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.0181143283843994, | |
| "rewards/margins": 0.6527779698371887, | |
| "rewards/rejected": -1.670892357826233, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.4500261643118786, | |
| "grad_norm": 13.83059515121346, | |
| "learning_rate": 3.3555276610977276e-07, | |
| "logits/chosen": 0.1387651413679123, | |
| "logits/rejected": 0.3719004988670349, | |
| "logps/chosen": -305.9029235839844, | |
| "logps/rejected": -406.22869873046875, | |
| "loss": 0.5252, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.9015353918075562, | |
| "rewards/margins": 0.7788639664649963, | |
| "rewards/rejected": -1.6803992986679077, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.4604918890633176, | |
| "grad_norm": 16.016907625597792, | |
| "learning_rate": 3.269063392575352e-07, | |
| "logits/chosen": 0.27102214097976685, | |
| "logits/rejected": 0.3719123899936676, | |
| "logps/chosen": -328.58782958984375, | |
| "logps/rejected": -388.52410888671875, | |
| "loss": 0.5343, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.8982616662979126, | |
| "rewards/margins": 0.6498668789863586, | |
| "rewards/rejected": -1.5481284856796265, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.47095761381475665, | |
| "grad_norm": 11.748927642696774, | |
| "learning_rate": 3.1815705699316964e-07, | |
| "logits/chosen": 0.14817403256893158, | |
| "logits/rejected": 0.3451501727104187, | |
| "logps/chosen": -369.4537658691406, | |
| "logps/rejected": -422.6024475097656, | |
| "loss": 0.5277, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.9175182580947876, | |
| "rewards/margins": 0.7517377138137817, | |
| "rewards/rejected": -1.6692559719085693, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.48142333856619574, | |
| "grad_norm": 12.97699616095586, | |
| "learning_rate": 3.0931662070620794e-07, | |
| "logits/chosen": -0.02564959228038788, | |
| "logits/rejected": 0.29681330919265747, | |
| "logps/chosen": -292.01348876953125, | |
| "logps/rejected": -352.1918029785156, | |
| "loss": 0.5503, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.6568984389305115, | |
| "rewards/margins": 0.8044195175170898, | |
| "rewards/rejected": -1.461318016052246, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.49188906331763477, | |
| "grad_norm": 13.884696801765326, | |
| "learning_rate": 3.003968536966078e-07, | |
| "logits/chosen": 0.07226122915744781, | |
| "logits/rejected": 0.24507398903369904, | |
| "logps/chosen": -342.1966247558594, | |
| "logps/rejected": -380.1782531738281, | |
| "loss": 0.4913, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.6603950262069702, | |
| "rewards/margins": 0.6150242686271667, | |
| "rewards/rejected": -1.2754193544387817, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5023547880690737, | |
| "grad_norm": 16.314360967355586, | |
| "learning_rate": 2.9140968536213693e-07, | |
| "logits/chosen": 0.1461777538061142, | |
| "logits/rejected": 0.5193515419960022, | |
| "logps/chosen": -365.03143310546875, | |
| "logps/rejected": -450.1629333496094, | |
| "loss": 0.5062, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.841938853263855, | |
| "rewards/margins": 0.9499468803405762, | |
| "rewards/rejected": -1.7918857336044312, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.5128205128205128, | |
| "grad_norm": 17.255358445672606, | |
| "learning_rate": 2.823671352438608e-07, | |
| "logits/chosen": 0.15712898969650269, | |
| "logits/rejected": 0.6440854072570801, | |
| "logps/chosen": -392.81866455078125, | |
| "logps/rejected": -430.689208984375, | |
| "loss": 0.5112, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.8158360719680786, | |
| "rewards/margins": 0.96565181016922, | |
| "rewards/rejected": -1.7814878225326538, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "grad_norm": 12.135159172710598, | |
| "learning_rate": 2.73281296951072e-07, | |
| "logits/chosen": 0.16401800513267517, | |
| "logits/rejected": 0.4998152256011963, | |
| "logps/chosen": -322.4615783691406, | |
| "logps/rejected": -393.34197998046875, | |
| "loss": 0.5211, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.8438663482666016, | |
| "rewards/margins": 0.9155359268188477, | |
| "rewards/rejected": -1.7594020366668701, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "eval_logits/chosen": 0.24328011274337769, | |
| "eval_logits/rejected": 0.5345854759216309, | |
| "eval_logps/chosen": -396.3761901855469, | |
| "eval_logps/rejected": -449.1320495605469, | |
| "eval_loss": 0.5214627981185913, | |
| "eval_rewards/accuracies": 0.7341269850730896, | |
| "eval_rewards/chosen": -1.087417721748352, | |
| "eval_rewards/margins": 0.7863243818283081, | |
| "eval_rewards/rejected": -1.8737419843673706, | |
| "eval_runtime": 226.3881, | |
| "eval_samples_per_second": 8.834, | |
| "eval_steps_per_second": 0.278, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.533751962323391, | |
| "grad_norm": 13.608551630251748, | |
| "learning_rate": 2.641643219871597e-07, | |
| "logits/chosen": 0.27356186509132385, | |
| "logits/rejected": 0.4992881417274475, | |
| "logps/chosen": -343.2596435546875, | |
| "logps/rejected": -415.26275634765625, | |
| "loss": 0.4984, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.923366367816925, | |
| "rewards/margins": 0.8507606387138367, | |
| "rewards/rejected": -1.7741270065307617, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.54421768707483, | |
| "grad_norm": 14.648106425594214, | |
| "learning_rate": 2.550284034980507e-07, | |
| "logits/chosen": 0.24416789412498474, | |
| "logits/rejected": 0.3334105610847473, | |
| "logps/chosen": -345.0733642578125, | |
| "logps/rejected": -388.39544677734375, | |
| "loss": 0.5276, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.8896657824516296, | |
| "rewards/margins": 0.5996683835983276, | |
| "rewards/rejected": -1.489334225654602, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.554683411826269, | |
| "grad_norm": 18.62263649476354, | |
| "learning_rate": 2.4588575996495794e-07, | |
| "logits/chosen": 0.11800795793533325, | |
| "logits/rejected": 0.39625564217567444, | |
| "logps/chosen": -364.69830322265625, | |
| "logps/rejected": -451.4327697753906, | |
| "loss": 0.5108, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.6179436445236206, | |
| "rewards/margins": 1.0892958641052246, | |
| "rewards/rejected": -1.7072395086288452, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.565149136577708, | |
| "grad_norm": 14.08705064458867, | |
| "learning_rate": 2.367486188632446e-07, | |
| "logits/chosen": 0.32275527715682983, | |
| "logits/rejected": 0.5255266427993774, | |
| "logps/chosen": -370.3330383300781, | |
| "logps/rejected": -402.99365234375, | |
| "loss": 0.5189, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.8699628114700317, | |
| "rewards/margins": 0.6969233155250549, | |
| "rewards/rejected": -1.5668861865997314, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.5756148613291471, | |
| "grad_norm": 17.14042985699126, | |
| "learning_rate": 2.276292003092593e-07, | |
| "logits/chosen": 0.22727811336517334, | |
| "logits/rejected": 0.5902983546257019, | |
| "logps/chosen": -366.0690002441406, | |
| "logps/rejected": -373.06427001953125, | |
| "loss": 0.514, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.8379491567611694, | |
| "rewards/margins": 0.5861997008323669, | |
| "rewards/rejected": -1.4241489171981812, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5860805860805861, | |
| "grad_norm": 23.49625757400348, | |
| "learning_rate": 2.185397007170141e-07, | |
| "logits/chosen": 0.4517914354801178, | |
| "logits/rejected": 0.6837216019630432, | |
| "logps/chosen": -288.7613525390625, | |
| "logps/rejected": -369.98590087890625, | |
| "loss": 0.5293, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.694735050201416, | |
| "rewards/margins": 0.7958725690841675, | |
| "rewards/rejected": -1.490607500076294, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.5965463108320251, | |
| "grad_norm": 17.265899168005756, | |
| "learning_rate": 2.094922764865619e-07, | |
| "logits/chosen": 0.43373093008995056, | |
| "logits/rejected": 0.621323823928833, | |
| "logps/chosen": -347.9602966308594, | |
| "logps/rejected": -448.27166748046875, | |
| "loss": 0.5007, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.0440847873687744, | |
| "rewards/margins": 0.6639469265937805, | |
| "rewards/rejected": -1.7080316543579102, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.6070120355834642, | |
| "grad_norm": 20.271007557729554, | |
| "learning_rate": 2.0049902774588797e-07, | |
| "logits/chosen": 0.3310932517051697, | |
| "logits/rejected": 0.6366420984268188, | |
| "logps/chosen": -326.4140625, | |
| "logps/rejected": -378.14251708984375, | |
| "loss": 0.5319, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.8602741956710815, | |
| "rewards/margins": 0.7902485728263855, | |
| "rewards/rejected": -1.6505229473114014, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.6174777603349032, | |
| "grad_norm": 14.597787676388892, | |
| "learning_rate": 1.9157198216806238e-07, | |
| "logits/chosen": 0.1060560941696167, | |
| "logits/rejected": 0.4027911126613617, | |
| "logps/chosen": -391.3057861328125, | |
| "logps/rejected": -447.2630310058594, | |
| "loss": 0.5146, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.9476065635681152, | |
| "rewards/margins": 0.8027095794677734, | |
| "rewards/rejected": -1.7503160238265991, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "grad_norm": 13.72232592119506, | |
| "learning_rate": 1.8272307888529274e-07, | |
| "logits/chosen": 0.25353509187698364, | |
| "logits/rejected": 0.5968682169914246, | |
| "logps/chosen": -425.5113830566406, | |
| "logps/rejected": -419.0316467285156, | |
| "loss": 0.4932, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.7882476449012756, | |
| "rewards/margins": 0.640188992023468, | |
| "rewards/rejected": -1.4284366369247437, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "eval_logits/chosen": 0.12464320659637451, | |
| "eval_logits/rejected": 0.42349082231521606, | |
| "eval_logps/chosen": -360.2087707519531, | |
| "eval_logps/rejected": -411.3826904296875, | |
| "eval_loss": 0.5179768800735474, | |
| "eval_rewards/accuracies": 0.7539682388305664, | |
| "eval_rewards/chosen": -0.7257437109947205, | |
| "eval_rewards/margins": 0.7705051302909851, | |
| "eval_rewards/rejected": -1.4962489604949951, | |
| "eval_runtime": 228.2238, | |
| "eval_samples_per_second": 8.763, | |
| "eval_steps_per_second": 0.276, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6384092098377813, | |
| "grad_norm": 16.319608597107443, | |
| "learning_rate": 1.7396415252139288e-07, | |
| "logits/chosen": 0.15204815566539764, | |
| "logits/rejected": 0.42100468277931213, | |
| "logps/chosen": -377.43377685546875, | |
| "logps/rejected": -418.968994140625, | |
| "loss": 0.5296, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.7400127649307251, | |
| "rewards/margins": 0.7931601405143738, | |
| "rewards/rejected": -1.533172845840454, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.6488749345892203, | |
| "grad_norm": 14.946981838762534, | |
| "learning_rate": 1.6530691736402316e-07, | |
| "logits/chosen": 0.19925448298454285, | |
| "logits/rejected": 0.5180339813232422, | |
| "logps/chosen": -383.7248229980469, | |
| "logps/rejected": -406.205322265625, | |
| "loss": 0.483, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.8672540783882141, | |
| "rewards/margins": 0.9261838793754578, | |
| "rewards/rejected": -1.7934379577636719, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.6593406593406593, | |
| "grad_norm": 17.582461648265248, | |
| "learning_rate": 1.5676295169786864e-07, | |
| "logits/chosen": 0.29265791177749634, | |
| "logits/rejected": 0.5982542634010315, | |
| "logps/chosen": -355.6764221191406, | |
| "logps/rejected": -397.57403564453125, | |
| "loss": 0.4936, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.9091702699661255, | |
| "rewards/margins": 0.7706948518753052, | |
| "rewards/rejected": -1.6798651218414307, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.6698063840920984, | |
| "grad_norm": 16.068638596661998, | |
| "learning_rate": 1.483436823197092e-07, | |
| "logits/chosen": 0.3285290002822876, | |
| "logits/rejected": 0.7033597230911255, | |
| "logps/chosen": -322.0961608886719, | |
| "logps/rejected": -381.0562438964844, | |
| "loss": 0.4991, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.8638845682144165, | |
| "rewards/margins": 0.8404154777526855, | |
| "rewards/rejected": -1.7043001651763916, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.6802721088435374, | |
| "grad_norm": 14.977459971087322, | |
| "learning_rate": 1.4006036925609243e-07, | |
| "logits/chosen": 0.31715768575668335, | |
| "logits/rejected": 0.4414668679237366, | |
| "logps/chosen": -351.3278503417969, | |
| "logps/rejected": -434.75799560546875, | |
| "loss": 0.523, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.7902597188949585, | |
| "rewards/margins": 0.7662769556045532, | |
| "rewards/rejected": -1.5565365552902222, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6907378335949764, | |
| "grad_norm": 18.437373509306592, | |
| "learning_rate": 1.319240907040458e-07, | |
| "logits/chosen": 0.04035192355513573, | |
| "logits/rejected": 0.22156131267547607, | |
| "logps/chosen": -339.9199523925781, | |
| "logps/rejected": -384.633544921875, | |
| "loss": 0.512, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.7180854678153992, | |
| "rewards/margins": 0.6694477200508118, | |
| "rewards/rejected": -1.3875333070755005, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.7012035583464155, | |
| "grad_norm": 18.725340795468888, | |
| "learning_rate": 1.239457282149695e-07, | |
| "logits/chosen": 0.2705515921115875, | |
| "logits/rejected": 0.5743520259857178, | |
| "logps/chosen": -367.09326171875, | |
| "logps/rejected": -425.185546875, | |
| "loss": 0.4988, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.9605787992477417, | |
| "rewards/margins": 0.8168357610702515, | |
| "rewards/rejected": -1.7774145603179932, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.7116692830978545, | |
| "grad_norm": 15.807433822852937, | |
| "learning_rate": 1.1613595214152711e-07, | |
| "logits/chosen": 0.4535256326198578, | |
| "logits/rejected": 0.6553749442100525, | |
| "logps/chosen": -414.3898010253906, | |
| "logps/rejected": -449.7381286621094, | |
| "loss": 0.5116, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.0415313243865967, | |
| "rewards/margins": 0.6825998425483704, | |
| "rewards/rejected": -1.7241313457489014, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.7221350078492935, | |
| "grad_norm": 12.622124400991629, | |
| "learning_rate": 1.0850520736699362e-07, | |
| "logits/chosen": 0.19162841141223907, | |
| "logits/rejected": 0.6098980903625488, | |
| "logps/chosen": -341.79766845703125, | |
| "logps/rejected": -394.4463806152344, | |
| "loss": 0.4674, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.813568115234375, | |
| "rewards/margins": 0.8326117396354675, | |
| "rewards/rejected": -1.6461797952651978, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "grad_norm": 14.178960044408889, | |
| "learning_rate": 1.0106369933615042e-07, | |
| "logits/chosen": 0.4778602123260498, | |
| "logits/rejected": 0.7655819654464722, | |
| "logps/chosen": -350.98883056640625, | |
| "logps/rejected": -405.68756103515625, | |
| "loss": 0.4891, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.9857735633850098, | |
| "rewards/margins": 0.8198670148849487, | |
| "rewards/rejected": -1.805640459060669, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "eval_logits/chosen": 0.3793373703956604, | |
| "eval_logits/rejected": 0.7265903353691101, | |
| "eval_logps/chosen": -383.8190002441406, | |
| "eval_logps/rejected": -441.8805847167969, | |
| "eval_loss": 0.5096952319145203, | |
| "eval_rewards/accuracies": 0.7579365372657776, | |
| "eval_rewards/chosen": -0.9618459939956665, | |
| "eval_rewards/margins": 0.839381217956543, | |
| "eval_rewards/rejected": -1.801227331161499, | |
| "eval_runtime": 227.351, | |
| "eval_samples_per_second": 8.797, | |
| "eval_steps_per_second": 0.277, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7430664573521716, | |
| "grad_norm": 19.16673815297702, | |
| "learning_rate": 9.382138040640714e-08, | |
| "logits/chosen": 0.48337316513061523, | |
| "logits/rejected": 0.8561640977859497, | |
| "logps/chosen": -375.72723388671875, | |
| "logps/rejected": -441.17657470703125, | |
| "loss": 0.5328, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.00464928150177, | |
| "rewards/margins": 0.9693723917007446, | |
| "rewards/rejected": -1.9740216732025146, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.7535321821036107, | |
| "grad_norm": 15.335356540441524, | |
| "learning_rate": 8.678793653740632e-08, | |
| "logits/chosen": 0.19264373183250427, | |
| "logits/rejected": 0.5776667594909668, | |
| "logps/chosen": -418.45806884765625, | |
| "logps/rejected": -453.5069274902344, | |
| "loss": 0.4996, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.7923117876052856, | |
| "rewards/margins": 0.863797664642334, | |
| "rewards/rejected": -1.6561095714569092, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.7639979068550498, | |
| "grad_norm": 19.284158091087093, | |
| "learning_rate": 7.997277433690983e-08, | |
| "logits/chosen": 0.3528473973274231, | |
| "logits/rejected": 0.6614034175872803, | |
| "logps/chosen": -321.63165283203125, | |
| "logps/rejected": -384.52557373046875, | |
| "loss": 0.5141, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8356879949569702, | |
| "rewards/margins": 0.7253657579421997, | |
| "rewards/rejected": -1.5610538721084595, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.7744636316064888, | |
| "grad_norm": 13.11426429788781, | |
| "learning_rate": 7.338500848029602e-08, | |
| "logits/chosen": 0.3238237798213959, | |
| "logits/rejected": 0.7854956388473511, | |
| "logps/chosen": -352.84112548828125, | |
| "logps/rejected": -388.8981018066406, | |
| "loss": 0.5039, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.754696786403656, | |
| "rewards/margins": 0.8794612884521484, | |
| "rewards/rejected": -1.6341578960418701, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.7849293563579278, | |
| "grad_norm": 13.51761512540701, | |
| "learning_rate": 6.70334495204884e-08, | |
| "logits/chosen": 0.6912227869033813, | |
| "logits/rejected": 0.9092944264411926, | |
| "logps/chosen": -344.513916015625, | |
| "logps/rejected": -462.43536376953125, | |
| "loss": 0.5054, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.0211327075958252, | |
| "rewards/margins": 1.0520156621932983, | |
| "rewards/rejected": -2.073148250579834, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7953950811093669, | |
| "grad_norm": 13.59497749011639, | |
| "learning_rate": 6.092659210462231e-08, | |
| "logits/chosen": 0.677158772945404, | |
| "logits/rejected": 0.942747950553894, | |
| "logps/chosen": -345.63848876953125, | |
| "logps/rejected": -470.269775390625, | |
| "loss": 0.5058, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.0826094150543213, | |
| "rewards/margins": 1.0310587882995605, | |
| "rewards/rejected": -2.113668203353882, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.8058608058608059, | |
| "grad_norm": 11.18434909589235, | |
| "learning_rate": 5.507260361320737e-08, | |
| "logits/chosen": 0.4663829207420349, | |
| "logits/rejected": 0.6142369508743286, | |
| "logps/chosen": -447.09613037109375, | |
| "logps/rejected": -531.8663330078125, | |
| "loss": 0.4845, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.1781532764434814, | |
| "rewards/margins": 0.7445544004440308, | |
| "rewards/rejected": -1.9227077960968018, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.8163265306122449, | |
| "grad_norm": 20.85850537945954, | |
| "learning_rate": 4.947931323697982e-08, | |
| "logits/chosen": 0.3884517252445221, | |
| "logits/rejected": 0.8345146179199219, | |
| "logps/chosen": -403.45867919921875, | |
| "logps/rejected": -429.47064208984375, | |
| "loss": 0.5072, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.74444979429245, | |
| "rewards/margins": 1.009735107421875, | |
| "rewards/rejected": -1.7541850805282593, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.826792255363684, | |
| "grad_norm": 13.990147261782901, | |
| "learning_rate": 4.415420150605398e-08, | |
| "logits/chosen": 0.5786755084991455, | |
| "logits/rejected": 0.9725602865219116, | |
| "logps/chosen": -360.92657470703125, | |
| "logps/rejected": -401.515380859375, | |
| "loss": 0.5104, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.9632360339164734, | |
| "rewards/margins": 0.7449162006378174, | |
| "rewards/rejected": -1.708152174949646, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "grad_norm": 16.235900872693833, | |
| "learning_rate": 3.9104390285376374e-08, | |
| "logits/chosen": 0.5984079837799072, | |
| "logits/rejected": 0.9596022367477417, | |
| "logps/chosen": -344.7926330566406, | |
| "logps/rejected": -415.468505859375, | |
| "loss": 0.5052, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.9547444581985474, | |
| "rewards/margins": 0.812140166759491, | |
| "rewards/rejected": -1.766884446144104, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "eval_logits/chosen": 0.45483413338661194, | |
| "eval_logits/rejected": 0.8224205374717712, | |
| "eval_logps/chosen": -380.42578125, | |
| "eval_logps/rejected": -441.0578308105469, | |
| "eval_loss": 0.5066840052604675, | |
| "eval_rewards/accuracies": 0.7539682388305664, | |
| "eval_rewards/chosen": -0.9279137253761292, | |
| "eval_rewards/margins": 0.8650867342948914, | |
| "eval_rewards/rejected": -1.7930004596710205, | |
| "eval_runtime": 229.2471, | |
| "eval_samples_per_second": 8.724, | |
| "eval_steps_per_second": 0.275, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.847723704866562, | |
| "grad_norm": 12.324929368398774, | |
| "learning_rate": 3.433663324986208e-08, | |
| "logits/chosen": 0.3393861651420593, | |
| "logits/rejected": 0.9386290311813354, | |
| "logps/chosen": -431.0074157714844, | |
| "logps/rejected": -448.7564392089844, | |
| "loss": 0.5038, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.036725640296936, | |
| "rewards/margins": 0.8281686902046204, | |
| "rewards/rejected": -1.8648942708969116, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.858189429618001, | |
| "grad_norm": 18.594511256107943, | |
| "learning_rate": 2.9857306851953897e-08, | |
| "logits/chosen": 0.40348076820373535, | |
| "logits/rejected": 0.633434534072876, | |
| "logps/chosen": -367.580810546875, | |
| "logps/rejected": -448.944580078125, | |
| "loss": 0.4981, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.8985809087753296, | |
| "rewards/margins": 1.0237659215927124, | |
| "rewards/rejected": -1.922347068786621, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.8686551543694401, | |
| "grad_norm": 13.517371830628358, | |
| "learning_rate": 2.567240179368185e-08, | |
| "logits/chosen": 0.6697582006454468, | |
| "logits/rejected": 0.8797575235366821, | |
| "logps/chosen": -338.66162109375, | |
| "logps/rejected": -448.2554626464844, | |
| "loss": 0.505, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.0065336227416992, | |
| "rewards/margins": 0.8832134008407593, | |
| "rewards/rejected": -1.889746904373169, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.8791208791208791, | |
| "grad_norm": 11.944300741165499, | |
| "learning_rate": 2.1787515014630357e-08, | |
| "logits/chosen": 0.6317728757858276, | |
| "logits/rejected": 0.991460919380188, | |
| "logps/chosen": -384.9150085449219, | |
| "logps/rejected": -464.59326171875, | |
| "loss": 0.4855, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.089877724647522, | |
| "rewards/margins": 0.8564338684082031, | |
| "rewards/rejected": -1.946311593055725, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.8895866038723181, | |
| "grad_norm": 16.102405155905764, | |
| "learning_rate": 1.820784220652766e-08, | |
| "logits/chosen": 0.6404603719711304, | |
| "logits/rejected": 0.9315497279167175, | |
| "logps/chosen": -365.2972106933594, | |
| "logps/rejected": -447.9342346191406, | |
| "loss": 0.474, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.9282188415527344, | |
| "rewards/margins": 1.0321100950241089, | |
| "rewards/rejected": -1.9603290557861328, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9000523286237572, | |
| "grad_norm": 14.744690323569634, | |
| "learning_rate": 1.4938170864468636e-08, | |
| "logits/chosen": 0.46648722887039185, | |
| "logits/rejected": 0.9038979411125183, | |
| "logps/chosen": -401.80963134765625, | |
| "logps/rejected": -458.4877014160156, | |
| "loss": 0.4834, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.0028564929962158, | |
| "rewards/margins": 0.8283039927482605, | |
| "rewards/rejected": -1.8311609029769897, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.9105180533751962, | |
| "grad_norm": 16.179311044599356, | |
| "learning_rate": 1.1982873884064465e-08, | |
| "logits/chosen": 0.43730059266090393, | |
| "logits/rejected": 0.8710809946060181, | |
| "logps/chosen": -334.9444885253906, | |
| "logps/rejected": -371.6940612792969, | |
| "loss": 0.4911, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.9459562301635742, | |
| "rewards/margins": 0.7860196232795715, | |
| "rewards/rejected": -1.7319759130477905, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.9209837781266352, | |
| "grad_norm": 14.166797085920381, | |
| "learning_rate": 9.345903713082304e-09, | |
| "logits/chosen": 0.4502927362918854, | |
| "logits/rejected": 0.6843646764755249, | |
| "logps/chosen": -367.3271179199219, | |
| "logps/rejected": -403.81610107421875, | |
| "loss": 0.5062, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.8845589756965637, | |
| "rewards/margins": 0.7547686696052551, | |
| "rewards/rejected": -1.6393276453018188, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.9314495028780743, | |
| "grad_norm": 13.843768410288419, | |
| "learning_rate": 7.030787065396865e-09, | |
| "logits/chosen": 0.6794019937515259, | |
| "logits/rejected": 1.035869836807251, | |
| "logps/chosen": -337.71124267578125, | |
| "logps/rejected": -409.55108642578125, | |
| "loss": 0.5119, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.9029817581176758, | |
| "rewards/margins": 0.996401309967041, | |
| "rewards/rejected": -1.8993831872940063, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "grad_norm": 18.40121909693918, | |
| "learning_rate": 5.04062020432286e-09, | |
| "logits/chosen": 0.5905700325965881, | |
| "logits/rejected": 0.9023973345756531, | |
| "logps/chosen": -391.4457702636719, | |
| "logps/rejected": -482.791015625, | |
| "loss": 0.4946, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.083058476448059, | |
| "rewards/margins": 0.8590753674507141, | |
| "rewards/rejected": -1.9421336650848389, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "eval_logits/chosen": 0.48986998200416565, | |
| "eval_logits/rejected": 0.865327775478363, | |
| "eval_logps/chosen": -382.1980285644531, | |
| "eval_logps/rejected": -444.3302307128906, | |
| "eval_loss": 0.5059770941734314, | |
| "eval_rewards/accuracies": 0.7579365372657776, | |
| "eval_rewards/chosen": -0.945636510848999, | |
| "eval_rewards/margins": 0.880087673664093, | |
| "eval_rewards/rejected": -1.8257242441177368, | |
| "eval_runtime": 227.0574, | |
| "eval_samples_per_second": 8.808, | |
| "eval_steps_per_second": 0.277, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9523809523809523, | |
| "grad_norm": 16.77062501745593, | |
| "learning_rate": 3.3780648016376866e-09, | |
| "logits/chosen": 0.5361930131912231, | |
| "logits/rejected": 0.99513179063797, | |
| "logps/chosen": -326.2569885253906, | |
| "logps/rejected": -383.8536071777344, | |
| "loss": 0.5026, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.8974063992500305, | |
| "rewards/margins": 0.9173014760017395, | |
| "rewards/rejected": -1.8147079944610596, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.9628466771323915, | |
| "grad_norm": 26.414125387862228, | |
| "learning_rate": 2.0453443778310766e-09, | |
| "logits/chosen": 0.4901656210422516, | |
| "logits/rejected": 0.9345698356628418, | |
| "logps/chosen": -391.57965087890625, | |
| "logps/rejected": -420.29833984375, | |
| "loss": 0.5179, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.931963324546814, | |
| "rewards/margins": 0.8185266256332397, | |
| "rewards/rejected": -1.7504901885986328, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.9733124018838305, | |
| "grad_norm": 16.660540365959708, | |
| "learning_rate": 1.0442413283435758e-09, | |
| "logits/chosen": 0.4114384651184082, | |
| "logits/rejected": 0.9512575268745422, | |
| "logps/chosen": -420.1183166503906, | |
| "logps/rejected": -420.1231384277344, | |
| "loss": 0.4762, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8815310597419739, | |
| "rewards/margins": 0.7737995386123657, | |
| "rewards/rejected": -1.6553304195404053, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.9837781266352695, | |
| "grad_norm": 15.455558961505814, | |
| "learning_rate": 3.760945397705828e-10, | |
| "logits/chosen": 0.6433460116386414, | |
| "logits/rejected": 0.7288548350334167, | |
| "logps/chosen": -335.63751220703125, | |
| "logps/rejected": -465.32647705078125, | |
| "loss": 0.4908, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.8627341985702515, | |
| "rewards/margins": 0.863608181476593, | |
| "rewards/rejected": -1.7263424396514893, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.9942438513867086, | |
| "grad_norm": 15.993090751409031, | |
| "learning_rate": 4.17975992204056e-11, | |
| "logits/chosen": 0.3498649597167969, | |
| "logits/rejected": 0.7463272213935852, | |
| "logps/chosen": -401.7789001464844, | |
| "logps/rejected": -454.7509765625, | |
| "loss": 0.4996, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.9310733675956726, | |
| "rewards/margins": 0.8322589993476868, | |
| "rewards/rejected": -1.7633323669433594, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.9994767137624281, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5461677596207064, | |
| "train_runtime": 19976.9989, | |
| "train_samples_per_second": 3.06, | |
| "train_steps_per_second": 0.048 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 955, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |