| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984301412872841, | |
| "eval_steps": 100, | |
| "global_step": 477, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -2.5889174938201904, | |
| "logits/rejected": -2.4813222885131836, | |
| "logps/chosen": -289.8450622558594, | |
| "logps/rejected": -264.9564514160156, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.0416666666666667e-06, | |
| "logits/chosen": -2.388092517852783, | |
| "logits/rejected": -2.4257497787475586, | |
| "logps/chosen": -260.3330078125, | |
| "logps/rejected": -219.36460876464844, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.5277777910232544, | |
| "rewards/chosen": 0.006709899753332138, | |
| "rewards/margins": 0.001456400495953858, | |
| "rewards/rejected": 0.005253499373793602, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.0833333333333334e-06, | |
| "logits/chosen": -2.4777274131774902, | |
| "logits/rejected": -2.4627153873443604, | |
| "logps/chosen": -269.85845947265625, | |
| "logps/rejected": -241.0397186279297, | |
| "loss": 0.6883, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": 0.036987464874982834, | |
| "rewards/margins": 0.010376101359724998, | |
| "rewards/rejected": 0.026611363515257835, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.125e-06, | |
| "logits/chosen": -2.4684128761291504, | |
| "logits/rejected": -2.4569170475006104, | |
| "logps/chosen": -248.72634887695312, | |
| "logps/rejected": -245.45999145507812, | |
| "loss": 0.6814, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.04978276044130325, | |
| "rewards/margins": 0.02046312391757965, | |
| "rewards/rejected": 0.029319632798433304, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.166666666666667e-06, | |
| "logits/chosen": -2.350592613220215, | |
| "logits/rejected": -2.3819007873535156, | |
| "logps/chosen": -286.8092041015625, | |
| "logps/rejected": -258.41522216796875, | |
| "loss": 0.6638, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.04874344915151596, | |
| "rewards/margins": 0.06839489936828613, | |
| "rewards/rejected": -0.019651446491479874, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.999731868769027e-06, | |
| "logits/chosen": -2.371358871459961, | |
| "logits/rejected": -2.3549230098724365, | |
| "logps/chosen": -286.4510192871094, | |
| "logps/rejected": -292.4720153808594, | |
| "loss": 0.6532, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.040428485721349716, | |
| "rewards/margins": 0.11083565652370453, | |
| "rewards/rejected": -0.15126413106918335, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.9903533134293035e-06, | |
| "logits/chosen": -2.3133959770202637, | |
| "logits/rejected": -2.2497916221618652, | |
| "logps/chosen": -266.2447204589844, | |
| "logps/rejected": -250.51513671875, | |
| "loss": 0.6381, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.177861288189888, | |
| "rewards/margins": 0.15977302193641663, | |
| "rewards/rejected": -0.33763426542282104, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.967625656594782e-06, | |
| "logits/chosen": -2.359027624130249, | |
| "logits/rejected": -2.3221073150634766, | |
| "logps/chosen": -293.37640380859375, | |
| "logps/rejected": -272.8565979003906, | |
| "loss": 0.6222, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.28658825159072876, | |
| "rewards/margins": 0.1545184701681137, | |
| "rewards/rejected": -0.4411067068576813, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.93167072587771e-06, | |
| "logits/chosen": -2.2765626907348633, | |
| "logits/rejected": -2.2555174827575684, | |
| "logps/chosen": -295.1297607421875, | |
| "logps/rejected": -313.0910339355469, | |
| "loss": 0.6196, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.2227451503276825, | |
| "rewards/margins": 0.2126256674528122, | |
| "rewards/rejected": -0.43537086248397827, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.882681251368549e-06, | |
| "logits/chosen": -2.245281219482422, | |
| "logits/rejected": -2.227741241455078, | |
| "logps/chosen": -269.5973205566406, | |
| "logps/rejected": -291.8226318359375, | |
| "loss": 0.5989, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.3465212285518646, | |
| "rewards/margins": 0.26312902569770813, | |
| "rewards/rejected": -0.6096502542495728, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.8209198325401815e-06, | |
| "logits/chosen": -2.2300758361816406, | |
| "logits/rejected": -2.1660382747650146, | |
| "logps/chosen": -315.1083068847656, | |
| "logps/rejected": -333.6819763183594, | |
| "loss": 0.5985, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.5296173691749573, | |
| "rewards/margins": 0.28275665640830994, | |
| "rewards/rejected": -0.8123741149902344, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": -1.999224066734314, | |
| "eval_logits/rejected": -2.0060811042785645, | |
| "eval_logps/chosen": -312.6618347167969, | |
| "eval_logps/rejected": -347.36639404296875, | |
| "eval_loss": 0.6166529059410095, | |
| "eval_rewards/accuracies": 0.703125, | |
| "eval_rewards/chosen": -0.6621690392494202, | |
| "eval_rewards/margins": 0.3358937203884125, | |
| "eval_rewards/rejected": -0.9980627298355103, | |
| "eval_runtime": 73.6215, | |
| "eval_samples_per_second": 27.166, | |
| "eval_steps_per_second": 0.435, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.746717530629565e-06, | |
| "logits/chosen": -2.0477728843688965, | |
| "logits/rejected": -2.079465389251709, | |
| "logps/chosen": -331.4808349609375, | |
| "logps/rejected": -322.7523193359375, | |
| "loss": 0.603, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.47515854239463806, | |
| "rewards/margins": 0.4070638120174408, | |
| "rewards/rejected": -0.8822224736213684, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.660472094042121e-06, | |
| "logits/chosen": -1.9152177572250366, | |
| "logits/rejected": -1.9233732223510742, | |
| "logps/chosen": -306.1752014160156, | |
| "logps/rejected": -293.25958251953125, | |
| "loss": 0.5842, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.44953522086143494, | |
| "rewards/margins": 0.3475509285926819, | |
| "rewards/rejected": -0.7970861196517944, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.5626458262912745e-06, | |
| "logits/chosen": -1.70361328125, | |
| "logits/rejected": -1.6728603839874268, | |
| "logps/chosen": -318.77362060546875, | |
| "logps/rejected": -358.8970642089844, | |
| "loss": 0.5741, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.5537757873535156, | |
| "rewards/margins": 0.4414924681186676, | |
| "rewards/rejected": -0.9952683448791504, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.453763107901676e-06, | |
| "logits/chosen": -1.389970302581787, | |
| "logits/rejected": -1.3908889293670654, | |
| "logps/chosen": -365.573974609375, | |
| "logps/rejected": -389.08026123046875, | |
| "loss": 0.5673, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.8941513895988464, | |
| "rewards/margins": 0.5348166823387146, | |
| "rewards/rejected": -1.4289681911468506, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.33440758555951e-06, | |
| "logits/chosen": -1.2829688787460327, | |
| "logits/rejected": -1.1666593551635742, | |
| "logps/chosen": -331.8448181152344, | |
| "logps/rejected": -347.62164306640625, | |
| "loss": 0.5587, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.8844560384750366, | |
| "rewards/margins": 0.41986608505249023, | |
| "rewards/rejected": -1.304322361946106, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.205219043576955e-06, | |
| "logits/chosen": -0.9949747323989868, | |
| "logits/rejected": -0.5801770687103271, | |
| "logps/chosen": -321.4789123535156, | |
| "logps/rejected": -394.52911376953125, | |
| "loss": 0.5409, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.5671305060386658, | |
| "rewards/margins": 0.7361670732498169, | |
| "rewards/rejected": -1.3032976388931274, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.066889974440757e-06, | |
| "logits/chosen": -0.5876813530921936, | |
| "logits/rejected": -0.3649616539478302, | |
| "logps/chosen": -293.5026550292969, | |
| "logps/rejected": -311.59423828125, | |
| "loss": 0.5916, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.5926370024681091, | |
| "rewards/margins": 0.5672849416732788, | |
| "rewards/rejected": -1.1599220037460327, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.92016186682789e-06, | |
| "logits/chosen": -0.8265250325202942, | |
| "logits/rejected": -0.7326392531394958, | |
| "logps/chosen": -324.5079650878906, | |
| "logps/rejected": -388.0706481933594, | |
| "loss": 0.5533, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.7355499863624573, | |
| "rewards/margins": 0.6323290467262268, | |
| "rewards/rejected": -1.3678789138793945, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.7658212309857576e-06, | |
| "logits/chosen": -0.7887569665908813, | |
| "logits/rejected": -0.5450983643531799, | |
| "logps/chosen": -341.79412841796875, | |
| "logps/rejected": -386.400146484375, | |
| "loss": 0.5368, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.7941572666168213, | |
| "rewards/margins": 0.5156592130661011, | |
| "rewards/rejected": -1.309816598892212, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.604695382782159e-06, | |
| "logits/chosen": -0.2513345181941986, | |
| "logits/rejected": -0.26383644342422485, | |
| "logps/chosen": -360.3321228027344, | |
| "logps/rejected": -372.03326416015625, | |
| "loss": 0.5302, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.8325196504592896, | |
| "rewards/margins": 0.6745045185089111, | |
| "rewards/rejected": -1.5070240497589111, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": 0.4000808000564575, | |
| "eval_logits/rejected": 0.31158187985420227, | |
| "eval_logps/chosen": -334.02044677734375, | |
| "eval_logps/rejected": -407.42919921875, | |
| "eval_loss": 0.54950350522995, | |
| "eval_rewards/accuracies": 0.74609375, | |
| "eval_rewards/chosen": -0.8757554292678833, | |
| "eval_rewards/margins": 0.7229353189468384, | |
| "eval_rewards/rejected": -1.5986907482147217, | |
| "eval_runtime": 72.2579, | |
| "eval_samples_per_second": 27.679, | |
| "eval_steps_per_second": 0.443, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.437648009023905e-06, | |
| "logits/chosen": -0.355979859828949, | |
| "logits/rejected": -0.13378120958805084, | |
| "logps/chosen": -354.21832275390625, | |
| "logps/rejected": -387.26171875, | |
| "loss": 0.5542, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.7991288304328918, | |
| "rewards/margins": 0.5489404201507568, | |
| "rewards/rejected": -1.348069190979004, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.265574537815398e-06, | |
| "logits/chosen": -0.2700692415237427, | |
| "logits/rejected": -0.08166289329528809, | |
| "logps/chosen": -290.8014831542969, | |
| "logps/rejected": -332.3136901855469, | |
| "loss": 0.569, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.608197033405304, | |
| "rewards/margins": 0.507323145866394, | |
| "rewards/rejected": -1.1155202388763428, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.089397338773569e-06, | |
| "logits/chosen": -0.06517831236124039, | |
| "logits/rejected": 0.004061543848365545, | |
| "logps/chosen": -317.51495361328125, | |
| "logps/rejected": -371.9200439453125, | |
| "loss": 0.5576, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.7464595437049866, | |
| "rewards/margins": 0.5365883111953735, | |
| "rewards/rejected": -1.2830479145050049, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.9100607788275547e-06, | |
| "logits/chosen": -0.13817472755908966, | |
| "logits/rejected": -0.04675758630037308, | |
| "logps/chosen": -334.0376281738281, | |
| "logps/rejected": -356.2500915527344, | |
| "loss": 0.5592, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.908129096031189, | |
| "rewards/margins": 0.4950867295265198, | |
| "rewards/rejected": -1.403215765953064, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.72852616010567e-06, | |
| "logits/chosen": -0.0858619436621666, | |
| "logits/rejected": 0.23496215045452118, | |
| "logps/chosen": -347.7973327636719, | |
| "logps/rejected": -415.5159606933594, | |
| "loss": 0.539, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.9283957481384277, | |
| "rewards/margins": 0.7460072636604309, | |
| "rewards/rejected": -1.6744029521942139, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.5457665670441937e-06, | |
| "logits/chosen": 0.35799893736839294, | |
| "logits/rejected": 0.7192636728286743, | |
| "logps/chosen": -364.86138916015625, | |
| "logps/rejected": -400.7781066894531, | |
| "loss": 0.5156, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.0014389753341675, | |
| "rewards/margins": 0.67198646068573, | |
| "rewards/rejected": -1.673425316810608, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.3627616503391813e-06, | |
| "logits/chosen": 0.7610759735107422, | |
| "logits/rejected": 0.6359414458274841, | |
| "logps/chosen": -394.61541748046875, | |
| "logps/rejected": -451.08563232421875, | |
| "loss": 0.5183, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.1308982372283936, | |
| "rewards/margins": 0.7238563299179077, | |
| "rewards/rejected": -1.8547546863555908, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.1804923757009885e-06, | |
| "logits/chosen": 1.2029451131820679, | |
| "logits/rejected": 1.3649919033050537, | |
| "logps/chosen": -380.61932373046875, | |
| "logps/rejected": -396.6190490722656, | |
| "loss": 0.5485, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.2345253229141235, | |
| "rewards/margins": 0.6180087327957153, | |
| "rewards/rejected": -1.8525340557098389, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 1.9999357655598894e-06, | |
| "logits/chosen": 0.7018269300460815, | |
| "logits/rejected": 0.5022421479225159, | |
| "logps/chosen": -329.9044189453125, | |
| "logps/rejected": -380.2184143066406, | |
| "loss": 0.5464, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.9144694209098816, | |
| "rewards/margins": 0.6528812050819397, | |
| "rewards/rejected": -1.5673506259918213, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.8220596619089576e-06, | |
| "logits/chosen": 0.12925171852111816, | |
| "logits/rejected": 0.4913701117038727, | |
| "logps/chosen": -334.60748291015625, | |
| "logps/rejected": -398.9501647949219, | |
| "loss": 0.533, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.7113243341445923, | |
| "rewards/margins": 0.6202197670936584, | |
| "rewards/rejected": -1.3315439224243164, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": 0.6809147000312805, | |
| "eval_logits/rejected": 0.5716233849525452, | |
| "eval_logps/chosen": -327.86053466796875, | |
| "eval_logps/rejected": -399.1313171386719, | |
| "eval_loss": 0.5384255051612854, | |
| "eval_rewards/accuracies": 0.76171875, | |
| "eval_rewards/chosen": -0.8141559362411499, | |
| "eval_rewards/margins": 0.7015555500984192, | |
| "eval_rewards/rejected": -1.5157114267349243, | |
| "eval_runtime": 72.3068, | |
| "eval_samples_per_second": 27.66, | |
| "eval_steps_per_second": 0.443, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.647817538357072e-06, | |
| "logits/chosen": 0.2949855327606201, | |
| "logits/rejected": 0.6500759124755859, | |
| "logps/chosen": -364.088134765625, | |
| "logps/rejected": -378.37872314453125, | |
| "loss": 0.5199, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.0496848821640015, | |
| "rewards/margins": 0.6011860966682434, | |
| "rewards/rejected": -1.6508712768554688, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.4781433892011132e-06, | |
| "logits/chosen": 1.0437233448028564, | |
| "logits/rejected": 0.9824169278144836, | |
| "logps/chosen": -368.3329772949219, | |
| "logps/rejected": -443.03125, | |
| "loss": 0.5141, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.1911169290542603, | |
| "rewards/margins": 0.9159032702445984, | |
| "rewards/rejected": -2.107020139694214, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3139467229135999e-06, | |
| "logits/chosen": 0.43222084641456604, | |
| "logits/rejected": 0.5403100252151489, | |
| "logps/chosen": -406.03656005859375, | |
| "logps/rejected": -415.65667724609375, | |
| "loss": 0.5332, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.136765718460083, | |
| "rewards/margins": 0.6447176337242126, | |
| "rewards/rejected": -1.7814832925796509, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.1561076868822756e-06, | |
| "logits/chosen": 0.783063530921936, | |
| "logits/rejected": 0.8185106515884399, | |
| "logps/chosen": -352.68560791015625, | |
| "logps/rejected": -389.525390625, | |
| "loss": 0.5327, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.0253931283950806, | |
| "rewards/margins": 0.48898547887802124, | |
| "rewards/rejected": -1.514378547668457, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0054723495346484e-06, | |
| "logits/chosen": 0.8168965578079224, | |
| "logits/rejected": 0.7234944105148315, | |
| "logps/chosen": -320.1710510253906, | |
| "logps/rejected": -381.6952209472656, | |
| "loss": 0.5137, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.005994200706482, | |
| "rewards/margins": 0.8249969482421875, | |
| "rewards/rejected": -1.8309911489486694, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.628481651367876e-07, | |
| "logits/chosen": 0.688895046710968, | |
| "logits/rejected": 1.0163129568099976, | |
| "logps/chosen": -380.68157958984375, | |
| "logps/rejected": -419.84234619140625, | |
| "loss": 0.5228, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.1236674785614014, | |
| "rewards/margins": 0.698872447013855, | |
| "rewards/rejected": -1.8225399255752563, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.289996455765749e-07, | |
| "logits/chosen": 0.6197929382324219, | |
| "logits/rejected": 0.7506722807884216, | |
| "logps/chosen": -420.45263671875, | |
| "logps/rejected": -426.29107666015625, | |
| "loss": 0.5353, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.1160409450531006, | |
| "rewards/margins": 0.6991242170333862, | |
| "rewards/rejected": -1.8151648044586182, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 6.046442623320145e-07, | |
| "logits/chosen": 0.5390521287918091, | |
| "logits/rejected": 0.5552736520767212, | |
| "logps/chosen": -365.60577392578125, | |
| "logps/rejected": -421.53564453125, | |
| "loss": 0.5076, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.0088932514190674, | |
| "rewards/margins": 0.6299774050712585, | |
| "rewards/rejected": -1.6388708353042603, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.904486005914027e-07, | |
| "logits/chosen": 0.7070841789245605, | |
| "logits/rejected": 0.8798855543136597, | |
| "logps/chosen": -369.456787109375, | |
| "logps/rejected": -406.3539123535156, | |
| "loss": 0.5334, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.0536937713623047, | |
| "rewards/margins": 0.5137845277786255, | |
| "rewards/rejected": -1.5674784183502197, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.8702478614051353e-07, | |
| "logits/chosen": 0.5031794905662537, | |
| "logits/rejected": 0.47197189927101135, | |
| "logps/chosen": -409.0977478027344, | |
| "logps/rejected": -434.04693603515625, | |
| "loss": 0.518, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -1.1779779195785522, | |
| "rewards/margins": 0.5962620973587036, | |
| "rewards/rejected": -1.7742401361465454, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": 1.195478081703186, | |
| "eval_logits/rejected": 1.1053041219711304, | |
| "eval_logps/chosen": -351.9891662597656, | |
| "eval_logps/rejected": -432.54376220703125, | |
| "eval_loss": 0.5275784730911255, | |
| "eval_rewards/accuracies": 0.75, | |
| "eval_rewards/chosen": -1.0554425716400146, | |
| "eval_rewards/margins": 0.7943933010101318, | |
| "eval_rewards/rejected": -1.849835753440857, | |
| "eval_runtime": 72.3032, | |
| "eval_samples_per_second": 27.661, | |
| "eval_steps_per_second": 0.443, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 2.9492720416985004e-07, | |
| "logits/chosen": 0.8437727093696594, | |
| "logits/rejected": 0.800951361656189, | |
| "logps/chosen": -389.89471435546875, | |
| "logps/rejected": -402.897705078125, | |
| "loss": 0.536, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.1922531127929688, | |
| "rewards/margins": 0.5932148694992065, | |
| "rewards/rejected": -1.7854681015014648, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.1464952759020857e-07, | |
| "logits/chosen": 0.8531819581985474, | |
| "logits/rejected": 0.7011052966117859, | |
| "logps/chosen": -365.4083557128906, | |
| "logps/rejected": -429.77392578125, | |
| "loss": 0.5293, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.1621034145355225, | |
| "rewards/margins": 0.5474345684051514, | |
| "rewards/rejected": -1.7095378637313843, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.4662207078575685e-07, | |
| "logits/chosen": 0.9001744389533997, | |
| "logits/rejected": 0.7595891356468201, | |
| "logps/chosen": -376.25653076171875, | |
| "logps/rejected": -414.7420959472656, | |
| "loss": 0.5296, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.0551135540008545, | |
| "rewards/margins": 0.7705513834953308, | |
| "rewards/rejected": -1.825664758682251, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 9.120948298936422e-08, | |
| "logits/chosen": 0.6783544421195984, | |
| "logits/rejected": 0.9758931994438171, | |
| "logps/chosen": -374.05902099609375, | |
| "logps/rejected": -429.77825927734375, | |
| "loss": 0.5249, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.049117088317871, | |
| "rewards/margins": 0.7034090161323547, | |
| "rewards/rejected": -1.752526044845581, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 4.870879364444109e-08, | |
| "logits/chosen": 0.8478315472602844, | |
| "logits/rejected": 0.7550174593925476, | |
| "logps/chosen": -369.22027587890625, | |
| "logps/rejected": -430.88873291015625, | |
| "loss": 0.5406, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.088853120803833, | |
| "rewards/margins": 0.6529593467712402, | |
| "rewards/rejected": -1.7418124675750732, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 1.93478202307823e-08, | |
| "logits/chosen": 0.6956031918525696, | |
| "logits/rejected": 0.911687970161438, | |
| "logps/chosen": -368.9820251464844, | |
| "logps/rejected": -417.018310546875, | |
| "loss": 0.516, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -1.0993984937667847, | |
| "rewards/margins": 0.6686335206031799, | |
| "rewards/rejected": -1.7680320739746094, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 3.283947088983663e-09, | |
| "logits/chosen": 0.7104119658470154, | |
| "logits/rejected": 0.7404820322990417, | |
| "logps/chosen": -359.8403625488281, | |
| "logps/rejected": -435.29803466796875, | |
| "loss": 0.509, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.0412371158599854, | |
| "rewards/margins": 0.8249354362487793, | |
| "rewards/rejected": -1.8661725521087646, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 477, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5624207920498319, | |
| "train_runtime": 4926.1363, | |
| "train_samples_per_second": 12.41, | |
| "train_steps_per_second": 0.097 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 477, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |