| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 0.44742729306487694, |
| "eval_steps": 300, |
| "global_step": 100, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.0044742729306487695, |
| "grad_norm": 4.521431922912598, |
| "learning_rate": 8.695652173913044e-06, |
| "logits/chosen": -0.37893733382225037, |
| "logits/rejected": -0.4597730040550232, |
| "logps/chosen": -11.001541137695312, |
| "logps/rejected": -17.05784034729004, |
| "loss": 1.3586950302124023, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6655479669570923, |
| "rewards/accuracies": 0.0, |
| "rewards/chosen": 0.0, |
| "rewards/margins": 0.0, |
| "rewards/rejected": 0.0, |
| "step": 1, |
| "train_speed(iter/s)": 0.005374 |
| }, |
| { |
| "epoch": 0.008948545861297539, |
| "grad_norm": 3.470855951309204, |
| "learning_rate": 1.739130434782609e-05, |
| "logits/chosen": -0.3633067309856415, |
| "logits/rejected": -0.4392029941082001, |
| "logps/chosen": -7.973292350769043, |
| "logps/rejected": -13.90955924987793, |
| "loss": 1.2437297105789185, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5505824685096741, |
| "rewards/accuracies": 0.0, |
| "rewards/chosen": 0.0, |
| "rewards/margins": 0.0, |
| "rewards/rejected": 0.0, |
| "step": 2, |
| "train_speed(iter/s)": 0.005501 |
| }, |
| { |
| "epoch": 0.013422818791946308, |
| "grad_norm": 5.51271390914917, |
| "learning_rate": 2.608695652173913e-05, |
| "logits/chosen": -0.26578277349472046, |
| "logits/rejected": -0.45589232444763184, |
| "logps/chosen": -9.600592613220215, |
| "logps/rejected": -22.355058670043945, |
| "loss": 1.392338514328003, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6997992396354675, |
| "rewards/accuracies": 0.453125, |
| "rewards/chosen": 4.354381235316396e-05, |
| "rewards/margins": 0.0013819998130202293, |
| "rewards/rejected": -0.001338455593213439, |
| "step": 3, |
| "train_speed(iter/s)": 0.005521 |
| }, |
| { |
| "epoch": 0.017897091722595078, |
| "grad_norm": 6.5741682052612305, |
| "learning_rate": 3.478260869565218e-05, |
| "logits/chosen": -0.40403854846954346, |
| "logits/rejected": -0.5402908325195312, |
| "logps/chosen": -6.373080253601074, |
| "logps/rejected": -16.791393280029297, |
| "loss": 1.222740888595581, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5450983047485352, |
| "rewards/accuracies": 0.765625, |
| "rewards/chosen": 0.010452846065163612, |
| "rewards/margins": 0.031955912709236145, |
| "rewards/rejected": -0.021503068506717682, |
| "step": 4, |
| "train_speed(iter/s)": 0.005551 |
| }, |
| { |
| "epoch": 0.02237136465324385, |
| "grad_norm": 3.046576738357544, |
| "learning_rate": 4.347826086956522e-05, |
| "logits/chosen": -0.3187038004398346, |
| "logits/rejected": -0.4441537857055664, |
| "logps/chosen": -9.070188522338867, |
| "logps/rejected": -17.255901336669922, |
| "loss": 1.2276911735534668, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5655408501625061, |
| "rewards/accuracies": 0.625, |
| "rewards/chosen": -0.007649715524166822, |
| "rewards/margins": 0.07260076701641083, |
| "rewards/rejected": -0.08025049418210983, |
| "step": 5, |
| "train_speed(iter/s)": 0.005557 |
| }, |
| { |
| "epoch": 0.026845637583892617, |
| "grad_norm": 4.846445560455322, |
| "learning_rate": 5.217391304347826e-05, |
| "logits/chosen": -0.30176016688346863, |
| "logits/rejected": -0.45638489723205566, |
| "logps/chosen": -8.63841438293457, |
| "logps/rejected": -18.97266960144043, |
| "loss": 1.1484646797180176, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5371410250663757, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": -0.024880727753043175, |
| "rewards/margins": 0.23170879483222961, |
| "rewards/rejected": -0.25658950209617615, |
| "step": 6, |
| "train_speed(iter/s)": 0.005568 |
| }, |
| { |
| "epoch": 0.03131991051454139, |
| "grad_norm": 7.562114715576172, |
| "learning_rate": 6.086956521739131e-05, |
| "logits/chosen": -0.3937205672264099, |
| "logits/rejected": -0.48791196942329407, |
| "logps/chosen": -10.315498352050781, |
| "logps/rejected": -20.046728134155273, |
| "loss": 1.3093780279159546, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6873191595077515, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": -0.14371810853481293, |
| "rewards/margins": 0.2797008454799652, |
| "rewards/rejected": -0.42341893911361694, |
| "step": 7, |
| "train_speed(iter/s)": 0.005577 |
| }, |
| { |
| "epoch": 0.035794183445190156, |
| "grad_norm": 7.459152698516846, |
| "learning_rate": 6.956521739130436e-05, |
| "logits/chosen": -0.3540106415748596, |
| "logits/rejected": -0.43836480379104614, |
| "logps/chosen": -8.758136749267578, |
| "logps/rejected": -18.294340133666992, |
| "loss": 1.3215807676315308, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6817622184753418, |
| "rewards/accuracies": 0.640625, |
| "rewards/chosen": -0.19053705036640167, |
| "rewards/margins": 0.2214164435863495, |
| "rewards/rejected": -0.41195347905158997, |
| "step": 8, |
| "train_speed(iter/s)": 0.005585 |
| }, |
| { |
| "epoch": 0.040268456375838924, |
| "grad_norm": 3.0581202507019043, |
| "learning_rate": 7.82608695652174e-05, |
| "logits/chosen": -0.3466319143772125, |
| "logits/rejected": -0.4212740361690521, |
| "logps/chosen": -10.387115478515625, |
| "logps/rejected": -20.363929748535156, |
| "loss": 1.1448516845703125, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5626436471939087, |
| "rewards/accuracies": 0.703125, |
| "rewards/chosen": -0.06380077451467514, |
| "rewards/margins": 0.3404861092567444, |
| "rewards/rejected": -0.40428686141967773, |
| "step": 9, |
| "train_speed(iter/s)": 0.005583 |
| }, |
| { |
| "epoch": 0.0447427293064877, |
| "grad_norm": 3.9353108406066895, |
| "learning_rate": 8.695652173913044e-05, |
| "logits/chosen": -0.4188399314880371, |
| "logits/rejected": -0.5007613301277161, |
| "logps/chosen": -9.096799850463867, |
| "logps/rejected": -18.05609130859375, |
| "loss": 1.2927519083023071, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6729815006256104, |
| "rewards/accuracies": 0.625, |
| "rewards/chosen": -0.060015156865119934, |
| "rewards/margins": 0.20498159527778625, |
| "rewards/rejected": -0.264996737241745, |
| "step": 10, |
| "train_speed(iter/s)": 0.00559 |
| }, |
| { |
| "epoch": 0.049217002237136466, |
| "grad_norm": 2.2350685596466064, |
| "learning_rate": 9.565217391304348e-05, |
| "logits/chosen": -0.41371461749076843, |
| "logits/rejected": -0.4829203188419342, |
| "logps/chosen": -7.282920837402344, |
| "logps/rejected": -15.015157699584961, |
| "loss": 1.167080044746399, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5672391653060913, |
| "rewards/accuracies": 0.765625, |
| "rewards/chosen": 0.07586651295423508, |
| "rewards/margins": 0.2256288230419159, |
| "rewards/rejected": -0.14976228773593903, |
| "step": 11, |
| "train_speed(iter/s)": 0.005596 |
| }, |
| { |
| "epoch": 0.053691275167785234, |
| "grad_norm": 2.6733052730560303, |
| "learning_rate": 0.00010434782608695653, |
| "logits/chosen": -0.38637417554855347, |
| "logits/rejected": -0.45029905438423157, |
| "logps/chosen": -11.042367935180664, |
| "logps/rejected": -16.67047882080078, |
| "loss": 1.2763197422027588, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6531897783279419, |
| "rewards/accuracies": 0.671875, |
| "rewards/chosen": 0.06360332667827606, |
| "rewards/margins": 0.17586028575897217, |
| "rewards/rejected": -0.1122569665312767, |
| "step": 12, |
| "train_speed(iter/s)": 0.005598 |
| }, |
| { |
| "epoch": 0.058165548098434, |
| "grad_norm": 1.9472479820251465, |
| "learning_rate": 0.00011304347826086956, |
| "logits/chosen": -0.4195230007171631, |
| "logits/rejected": -0.4734322428703308, |
| "logps/chosen": -7.205465316772461, |
| "logps/rejected": -13.222125053405762, |
| "loss": 1.1376935243606567, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4900382459163666, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.018632210791110992, |
| "rewards/margins": 0.11014734208583832, |
| "rewards/rejected": -0.09151513874530792, |
| "step": 13, |
| "train_speed(iter/s)": 0.005601 |
| }, |
| { |
| "epoch": 0.06263982102908278, |
| "grad_norm": 1.828997015953064, |
| "learning_rate": 0.00012173913043478263, |
| "logits/chosen": -0.399485319852829, |
| "logits/rejected": -0.4682197570800781, |
| "logps/chosen": -8.005407333374023, |
| "logps/rejected": -16.06272315979004, |
| "loss": 1.1697134971618652, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5445835590362549, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.0030609495006501675, |
| "rewards/margins": 0.16438230872154236, |
| "rewards/rejected": -0.16132137179374695, |
| "step": 14, |
| "train_speed(iter/s)": 0.005604 |
| }, |
| { |
| "epoch": 0.06711409395973154, |
| "grad_norm": 2.1229512691497803, |
| "learning_rate": 0.00013043478260869567, |
| "logits/chosen": -0.3835911750793457, |
| "logits/rejected": -0.4592927396297455, |
| "logps/chosen": -6.499156475067139, |
| "logps/rejected": -15.707530975341797, |
| "loss": 1.0722893476486206, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.503006637096405, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.06762155145406723, |
| "rewards/margins": 0.31041574478149414, |
| "rewards/rejected": -0.24279415607452393, |
| "step": 15, |
| "train_speed(iter/s)": 0.005608 |
| }, |
| { |
| "epoch": 0.07158836689038031, |
| "grad_norm": 4.802647113800049, |
| "learning_rate": 0.0001391304347826087, |
| "logits/chosen": -0.3306896388530731, |
| "logits/rejected": -0.41153809428215027, |
| "logps/chosen": -10.432010650634766, |
| "logps/rejected": -21.58324432373047, |
| "loss": 1.3572790622711182, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.7762681245803833, |
| "rewards/accuracies": 0.671875, |
| "rewards/chosen": -0.06817664206027985, |
| "rewards/margins": 0.363118052482605, |
| "rewards/rejected": -0.4312947690486908, |
| "step": 16, |
| "train_speed(iter/s)": 0.005608 |
| }, |
| { |
| "epoch": 0.07606263982102908, |
| "grad_norm": 4.122814178466797, |
| "learning_rate": 0.00014782608695652173, |
| "logits/chosen": -0.32593876123428345, |
| "logits/rejected": -0.4176383316516876, |
| "logps/chosen": -9.068134307861328, |
| "logps/rejected": -21.460546493530273, |
| "loss": 1.1545155048370361, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5955549478530884, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.0028565023094415665, |
| "rewards/margins": 0.3992985486984253, |
| "rewards/rejected": -0.3964419960975647, |
| "step": 17, |
| "train_speed(iter/s)": 0.00561 |
| }, |
| { |
| "epoch": 0.08053691275167785, |
| "grad_norm": 1.9508388042449951, |
| "learning_rate": 0.0001565217391304348, |
| "logits/chosen": -0.3363662660121918, |
| "logits/rejected": -0.40299132466316223, |
| "logps/chosen": -7.306541919708252, |
| "logps/rejected": -15.145895004272461, |
| "loss": 1.1673789024353027, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5968848466873169, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.13542409241199493, |
| "rewards/margins": 0.3165813684463501, |
| "rewards/rejected": -0.18115724623203278, |
| "step": 18, |
| "train_speed(iter/s)": 0.005611 |
| }, |
| { |
| "epoch": 0.08501118568232663, |
| "grad_norm": 4.076289176940918, |
| "learning_rate": 0.00016521739130434784, |
| "logits/chosen": -0.33677053451538086, |
| "logits/rejected": -0.42119455337524414, |
| "logps/chosen": -6.263703346252441, |
| "logps/rejected": -17.470884323120117, |
| "loss": 0.9588873982429504, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.42532750964164734, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.17247594892978668, |
| "rewards/margins": 0.41344738006591797, |
| "rewards/rejected": -0.24097144603729248, |
| "step": 19, |
| "train_speed(iter/s)": 0.005613 |
| }, |
| { |
| "epoch": 0.0894854586129754, |
| "grad_norm": 2.8480069637298584, |
| "learning_rate": 0.00017391304347826088, |
| "logits/chosen": -0.27281832695007324, |
| "logits/rejected": -0.3382987976074219, |
| "logps/chosen": -8.123124122619629, |
| "logps/rejected": -17.66653060913086, |
| "loss": 1.1668881177902222, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5924651622772217, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": -0.00576989259570837, |
| "rewards/margins": 0.35684970021247864, |
| "rewards/rejected": -0.36261963844299316, |
| "step": 20, |
| "train_speed(iter/s)": 0.005615 |
| }, |
| { |
| "epoch": 0.09395973154362416, |
| "grad_norm": 2.664872646331787, |
| "learning_rate": 0.00018260869565217392, |
| "logits/chosen": -0.2556913197040558, |
| "logits/rejected": -0.3474789261817932, |
| "logps/chosen": -5.220917701721191, |
| "logps/rejected": -23.547605514526367, |
| "loss": 0.9417474865913391, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4541093111038208, |
| "rewards/accuracies": 0.796875, |
| "rewards/chosen": 0.043988656252622604, |
| "rewards/margins": 0.6349199414253235, |
| "rewards/rejected": -0.5909312963485718, |
| "step": 21, |
| "train_speed(iter/s)": 0.005615 |
| }, |
| { |
| "epoch": 0.09843400447427293, |
| "grad_norm": 2.325854539871216, |
| "learning_rate": 0.00019130434782608697, |
| "logits/chosen": -0.2636699676513672, |
| "logits/rejected": -0.33711570501327515, |
| "logps/chosen": -8.976820945739746, |
| "logps/rejected": -19.829130172729492, |
| "loss": 1.0729048252105713, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.595395028591156, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.09026271104812622, |
| "rewards/margins": 0.623466432094574, |
| "rewards/rejected": -0.5332037210464478, |
| "step": 22, |
| "train_speed(iter/s)": 0.005617 |
| }, |
| { |
| "epoch": 0.1029082774049217, |
| "grad_norm": 2.684507369995117, |
| "learning_rate": 0.0002, |
| "logits/chosen": -0.24679847061634064, |
| "logits/rejected": -0.3378849923610687, |
| "logps/chosen": -7.14666223526001, |
| "logps/rejected": -23.644922256469727, |
| "loss": 0.9377666115760803, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.427379846572876, |
| "rewards/accuracies": 0.765625, |
| "rewards/chosen": -0.0773518830537796, |
| "rewards/margins": 0.6581345796585083, |
| "rewards/rejected": -0.7354865074157715, |
| "step": 23, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.10738255033557047, |
| "grad_norm": 2.6381781101226807, |
| "learning_rate": 0.00019999724204599747, |
| "logits/chosen": -0.246232807636261, |
| "logits/rejected": -0.3716464936733246, |
| "logps/chosen": -6.111550331115723, |
| "logps/rejected": -24.672321319580078, |
| "loss": 0.9850844740867615, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5655074119567871, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.027247514575719833, |
| "rewards/margins": 0.8617139458656311, |
| "rewards/rejected": -0.8344664573669434, |
| "step": 24, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.11185682326621924, |
| "grad_norm": 2.5178840160369873, |
| "learning_rate": 0.00019998896833611603, |
| "logits/chosen": -0.23005428910255432, |
| "logits/rejected": -0.3447636365890503, |
| "logps/chosen": -8.92835807800293, |
| "logps/rejected": -25.173357009887695, |
| "loss": 1.1135735511779785, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5982357263565063, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": -0.07063186913728714, |
| "rewards/margins": 0.6180722117424011, |
| "rewards/rejected": -0.6887040138244629, |
| "step": 25, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.116331096196868, |
| "grad_norm": 1.9486284255981445, |
| "learning_rate": 0.0001999751793267259, |
| "logits/chosen": -0.17476646602153778, |
| "logits/rejected": -0.28910771012306213, |
| "logps/chosen": -6.982065200805664, |
| "logps/rejected": -25.235271453857422, |
| "loss": 0.9616149067878723, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4814175069332123, |
| "rewards/accuracies": 0.828125, |
| "rewards/chosen": 0.017612511292099953, |
| "rewards/margins": 0.7644082307815552, |
| "rewards/rejected": -0.7467955350875854, |
| "step": 26, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.12080536912751678, |
| "grad_norm": 2.024102210998535, |
| "learning_rate": 0.0001999558757784162, |
| "logits/chosen": -0.22182828187942505, |
| "logits/rejected": -0.33474183082580566, |
| "logps/chosen": -8.283013343811035, |
| "logps/rejected": -23.36388397216797, |
| "loss": 1.071007251739502, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5651244521141052, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": -0.005644451826810837, |
| "rewards/margins": 0.549856424331665, |
| "rewards/rejected": -0.5555009245872498, |
| "step": 27, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.12527964205816555, |
| "grad_norm": 2.334916830062866, |
| "learning_rate": 0.0001999310587559529, |
| "logits/chosen": -0.1745743602514267, |
| "logits/rejected": -0.28861111402511597, |
| "logps/chosen": -7.148108959197998, |
| "logps/rejected": -22.50945472717285, |
| "loss": 1.0254650115966797, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5092177391052246, |
| "rewards/accuracies": 0.828125, |
| "rewards/chosen": 0.024923313409090042, |
| "rewards/margins": 0.6763253211975098, |
| "rewards/rejected": -0.6514019966125488, |
| "step": 28, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.1297539149888143, |
| "grad_norm": 4.392460823059082, |
| "learning_rate": 0.00019990072962822007, |
| "logits/chosen": -0.1607450246810913, |
| "logits/rejected": -0.30258849263191223, |
| "logps/chosen": -8.330341339111328, |
| "logps/rejected": -23.302955627441406, |
| "loss": 1.1018993854522705, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6234543919563293, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.1296255737543106, |
| "rewards/margins": 0.7057831883430481, |
| "rewards/rejected": -0.5761575102806091, |
| "step": 29, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.1342281879194631, |
| "grad_norm": 1.9557713270187378, |
| "learning_rate": 0.00019986489006814452, |
| "logits/chosen": -0.15406475961208344, |
| "logits/rejected": -0.32140052318573, |
| "logps/chosen": -6.338605880737305, |
| "logps/rejected": -27.934860229492188, |
| "loss": 0.8741156458854675, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.43781277537345886, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.005505967885255814, |
| "rewards/margins": 0.9152470827102661, |
| "rewards/rejected": -0.9097411632537842, |
| "step": 30, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.13870246085011187, |
| "grad_norm": 3.570540189743042, |
| "learning_rate": 0.00019982354205260347, |
| "logits/chosen": -0.13595342636108398, |
| "logits/rejected": -0.24140222370624542, |
| "logps/chosen": -9.938725471496582, |
| "logps/rejected": -26.17535972595215, |
| "loss": 1.1157177686691284, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6451507806777954, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": -0.06947211921215057, |
| "rewards/margins": 0.9398424029350281, |
| "rewards/rejected": -1.0093145370483398, |
| "step": 31, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.14317673378076062, |
| "grad_norm": 2.263465404510498, |
| "learning_rate": 0.00019977668786231534, |
| "logits/chosen": -0.13489112257957458, |
| "logits/rejected": -0.2546093165874481, |
| "logps/chosen": -9.485723495483398, |
| "logps/rejected": -30.659465789794922, |
| "loss": 1.1024084091186523, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.7035719752311707, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.094739630818367, |
| "rewards/margins": 1.1513433456420898, |
| "rewards/rejected": -1.0566036701202393, |
| "step": 32, |
| "train_speed(iter/s)": 0.005617 |
| }, |
| { |
| "epoch": 0.1476510067114094, |
| "grad_norm": 4.003927230834961, |
| "learning_rate": 0.00019972433008171416, |
| "logits/chosen": -0.2770022451877594, |
| "logits/rejected": -0.3445180356502533, |
| "logps/chosen": -7.854065895080566, |
| "logps/rejected": -17.47032356262207, |
| "loss": 1.1709215641021729, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.635529637336731, |
| "rewards/accuracies": 0.765625, |
| "rewards/chosen": 0.03749306499958038, |
| "rewards/margins": 0.6043045520782471, |
| "rewards/rejected": -0.5668113827705383, |
| "step": 33, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.15212527964205816, |
| "grad_norm": 2.930697202682495, |
| "learning_rate": 0.00019966647159880703, |
| "logits/chosen": -0.20737791061401367, |
| "logits/rejected": -0.3068307638168335, |
| "logps/chosen": -5.924644470214844, |
| "logps/rejected": -20.47597885131836, |
| "loss": 0.9971151947975159, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4416566789150238, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.04966810718178749, |
| "rewards/margins": 0.48412013053894043, |
| "rewards/rejected": -0.4344520568847656, |
| "step": 34, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.15659955257270694, |
| "grad_norm": 2.292167901992798, |
| "learning_rate": 0.00019960311560501454, |
| "logits/chosen": -0.20404092967510223, |
| "logits/rejected": -0.2819429636001587, |
| "logps/chosen": -9.679859161376953, |
| "logps/rejected": -22.75543785095215, |
| "loss": 1.1032215356826782, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6184647679328918, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.07092233002185822, |
| "rewards/margins": 0.8317320942878723, |
| "rewards/rejected": -0.7608097791671753, |
| "step": 35, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.1610738255033557, |
| "grad_norm": 1.5171147584915161, |
| "learning_rate": 0.0001995342655949951, |
| "logits/chosen": -0.18677955865859985, |
| "logits/rejected": -0.30433592200279236, |
| "logps/chosen": -7.317014217376709, |
| "logps/rejected": -23.58348274230957, |
| "loss": 1.0045303106307983, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5375936031341553, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.13633473217487335, |
| "rewards/margins": 0.7853459119796753, |
| "rewards/rejected": -0.6490110754966736, |
| "step": 36, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.16554809843400448, |
| "grad_norm": 1.8760979175567627, |
| "learning_rate": 0.00019945992536645187, |
| "logits/chosen": -0.17943330109119415, |
| "logits/rejected": -0.27492690086364746, |
| "logps/chosen": -9.596693992614746, |
| "logps/rejected": -25.057254791259766, |
| "loss": 1.0518603324890137, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5895128846168518, |
| "rewards/accuracies": 0.859375, |
| "rewards/chosen": 4.468485713005066e-06, |
| "rewards/margins": 0.8271102905273438, |
| "rewards/rejected": -0.827105700969696, |
| "step": 37, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.17002237136465326, |
| "grad_norm": 2.1765875816345215, |
| "learning_rate": 0.0001993800990199235, |
| "logits/chosen": -0.17796888947486877, |
| "logits/rejected": -0.27494707703590393, |
| "logps/chosen": -8.854700088500977, |
| "logps/rejected": -22.589916229248047, |
| "loss": 1.1736024618148804, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6680553555488586, |
| "rewards/accuracies": 0.796875, |
| "rewards/chosen": 0.027395736426115036, |
| "rewards/margins": 0.702995777130127, |
| "rewards/rejected": -0.6756000518798828, |
| "step": 38, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.174496644295302, |
| "grad_norm": 2.0398788452148438, |
| "learning_rate": 0.0001992947909585578, |
| "logits/chosen": -0.15950845181941986, |
| "logits/rejected": -0.285598486661911, |
| "logps/chosen": -7.358231544494629, |
| "logps/rejected": -23.962072372436523, |
| "loss": 0.9235795736312866, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5008021593093872, |
| "rewards/accuracies": 0.890625, |
| "rewards/chosen": 0.1876223385334015, |
| "rewards/margins": 0.9989414215087891, |
| "rewards/rejected": -0.8113190531730652, |
| "step": 39, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.1789709172259508, |
| "grad_norm": 2.137016773223877, |
| "learning_rate": 0.000199204005887869, |
| "logits/chosen": -0.18114963173866272, |
| "logits/rejected": -0.23074296116828918, |
| "logps/chosen": -9.713254928588867, |
| "logps/rejected": -15.0142822265625, |
| "loss": 1.1612474918365479, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5743556618690491, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.08143825083971024, |
| "rewards/margins": 0.3889071047306061, |
| "rewards/rejected": -0.30746886134147644, |
| "step": 40, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.18344519015659955, |
| "grad_norm": 2.5089111328125, |
| "learning_rate": 0.000199107748815478, |
| "logits/chosen": -0.17036814987659454, |
| "logits/rejected": -0.25284841656684875, |
| "logps/chosen": -7.5790181159973145, |
| "logps/rejected": -22.757585525512695, |
| "loss": 0.9598883986473083, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5350608229637146, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.10258667171001434, |
| "rewards/margins": 0.982707679271698, |
| "rewards/rejected": -0.8801209926605225, |
| "step": 41, |
| "train_speed(iter/s)": 0.005622 |
| }, |
| { |
| "epoch": 0.18791946308724833, |
| "grad_norm": 2.425978660583496, |
| "learning_rate": 0.00019900602505083648, |
| "logits/chosen": -0.17056672275066376, |
| "logits/rejected": -0.24402382969856262, |
| "logps/chosen": -9.11720085144043, |
| "logps/rejected": -26.321962356567383, |
| "loss": 0.9588575959205627, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5199980735778809, |
| "rewards/accuracies": 0.796875, |
| "rewards/chosen": -0.02633209154009819, |
| "rewards/margins": 1.148821234703064, |
| "rewards/rejected": -1.1751532554626465, |
| "step": 42, |
| "train_speed(iter/s)": 0.005622 |
| }, |
| { |
| "epoch": 0.19239373601789708, |
| "grad_norm": 4.055136203765869, |
| "learning_rate": 0.0001988988402049336, |
| "logits/chosen": -0.06654427945613861, |
| "logits/rejected": -0.19404850900173187, |
| "logps/chosen": -10.976348876953125, |
| "logps/rejected": -35.80363464355469, |
| "loss": 1.205069661140442, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.7504922151565552, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": -0.11243688315153122, |
| "rewards/margins": 1.5014084577560425, |
| "rewards/rejected": -1.6138453483581543, |
| "step": 43, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.19686800894854586, |
| "grad_norm": 2.9351325035095215, |
| "learning_rate": 0.00019878620018998696, |
| "logits/chosen": -0.21978600323200226, |
| "logits/rejected": -0.32882797718048096, |
| "logps/chosen": -6.976373195648193, |
| "logps/rejected": -30.215131759643555, |
| "loss": 0.9490640759468079, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5300356149673462, |
| "rewards/accuracies": 0.828125, |
| "rewards/chosen": -0.11029693484306335, |
| "rewards/margins": 1.1996657848358154, |
| "rewards/rejected": -1.3099626302719116, |
| "step": 44, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.20134228187919462, |
| "grad_norm": 2.6874403953552246, |
| "learning_rate": 0.00019866811121911607, |
| "logits/chosen": -0.1936551034450531, |
| "logits/rejected": -0.284149706363678, |
| "logps/chosen": -6.9035515785217285, |
| "logps/rejected": -22.284351348876953, |
| "loss": 0.8805959820747375, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4068413972854614, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.07986758649349213, |
| "rewards/margins": 0.8386067152023315, |
| "rewards/rejected": -0.7587391138076782, |
| "step": 45, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.2058165548098434, |
| "grad_norm": 1.7893315553665161, |
| "learning_rate": 0.000198544579806, |
| "logits/chosen": -0.20961534976959229, |
| "logits/rejected": -0.33521464467048645, |
| "logps/chosen": -7.002056121826172, |
| "logps/rejected": -25.93248748779297, |
| "loss": 0.9666993618011475, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4805544912815094, |
| "rewards/accuracies": 0.796875, |
| "rewards/chosen": 0.07972076535224915, |
| "rewards/margins": 0.8979825377464294, |
| "rewards/rejected": -0.8182618021965027, |
| "step": 46, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.21029082774049218, |
| "grad_norm": 2.2033731937408447, |
| "learning_rate": 0.0001984156127645178, |
| "logits/chosen": -0.21569889783859253, |
| "logits/rejected": -0.38949355483055115, |
| "logps/chosen": -5.476630687713623, |
| "logps/rejected": -27.572505950927734, |
| "loss": 0.8785698413848877, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.42979833483695984, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.17229261994361877, |
| "rewards/margins": 1.044838547706604, |
| "rewards/rejected": -0.8725459575653076, |
| "step": 47, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.21476510067114093, |
| "grad_norm": 2.1163032054901123, |
| "learning_rate": 0.00019828121720837286, |
| "logits/chosen": -0.34833359718322754, |
| "logits/rejected": -0.41568437218666077, |
| "logps/chosen": -7.711272239685059, |
| "logps/rejected": -25.60678482055664, |
| "loss": 0.9890228509902954, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.548301100730896, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.07654333114624023, |
| "rewards/margins": 1.2576009035110474, |
| "rewards/rejected": -1.1810575723648071, |
| "step": 48, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.21923937360178972, |
| "grad_norm": 1.987210988998413, |
| "learning_rate": 0.00019814140055070042, |
| "logits/chosen": -0.2835751473903656, |
| "logits/rejected": -0.38263481855392456, |
| "logps/chosen": -8.624626159667969, |
| "logps/rejected": -25.244403839111328, |
| "loss": 0.9971789121627808, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5687417387962341, |
| "rewards/accuracies": 0.859375, |
| "rewards/chosen": 0.09697108715772629, |
| "rewards/margins": 1.0362679958343506, |
| "rewards/rejected": -0.9392969608306885, |
| "step": 49, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.22371364653243847, |
| "grad_norm": 1.9501866102218628, |
| "learning_rate": 0.0001979961705036587, |
| "logits/chosen": -0.38833191990852356, |
| "logits/rejected": -0.4256289303302765, |
| "logps/chosen": -8.22580623626709, |
| "logps/rejected": -23.093332290649414, |
| "loss": 0.9298551082611084, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4849992096424103, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.04563959687948227, |
| "rewards/margins": 1.0999000072479248, |
| "rewards/rejected": -1.054260492324829, |
| "step": 50, |
| "train_speed(iter/s)": 0.005622 |
| }, |
| { |
| "epoch": 0.22818791946308725, |
| "grad_norm": 2.496234893798828, |
| "learning_rate": 0.00019784553507800349, |
| "logits/chosen": -0.2710403501987457, |
| "logits/rejected": -0.41735631227493286, |
| "logps/chosen": -8.894371032714844, |
| "logps/rejected": -28.586936950683594, |
| "loss": 1.1125377416610718, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6875916719436646, |
| "rewards/accuracies": 0.796875, |
| "rewards/chosen": 0.10749680548906326, |
| "rewards/margins": 1.2817683219909668, |
| "rewards/rejected": -1.174271583557129, |
| "step": 51, |
| "train_speed(iter/s)": 0.005611 |
| }, |
| { |
| "epoch": 0.232662192393736, |
| "grad_norm": 3.274833917617798, |
| "learning_rate": 0.00019768950258264623, |
| "logits/chosen": -0.39167118072509766, |
| "logits/rejected": -0.4531523883342743, |
| "logps/chosen": -9.231929779052734, |
| "logps/rejected": -22.90334701538086, |
| "loss": 1.1200237274169922, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6103911399841309, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": -0.16369852423667908, |
| "rewards/margins": 0.9332659244537354, |
| "rewards/rejected": -1.0969644784927368, |
| "step": 52, |
| "train_speed(iter/s)": 0.005612 |
| }, |
| { |
| "epoch": 0.2371364653243848, |
| "grad_norm": 2.9111413955688477, |
| "learning_rate": 0.0001975280816241959, |
| "logits/chosen": -0.3340080976486206, |
| "logits/rejected": -0.3975221514701843, |
| "logps/chosen": -11.631303787231445, |
| "logps/rejected": -26.26183319091797, |
| "loss": 1.1326872110366821, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6245000958442688, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": -0.032076627016067505, |
| "rewards/margins": 0.9003015160560608, |
| "rewards/rejected": -0.9323781132698059, |
| "step": 53, |
| "train_speed(iter/s)": 0.005612 |
| }, |
| { |
| "epoch": 0.24161073825503357, |
| "grad_norm": 2.7828478813171387, |
| "learning_rate": 0.00019736128110648407, |
| "logits/chosen": -0.345708429813385, |
| "logits/rejected": -0.447677880525589, |
| "logps/chosen": -8.686080932617188, |
| "logps/rejected": -23.45962142944336, |
| "loss": 0.9978072047233582, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5644747018814087, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.18989036977291107, |
| "rewards/margins": 1.1108869314193726, |
| "rewards/rejected": -0.9209966659545898, |
| "step": 54, |
| "train_speed(iter/s)": 0.005613 |
| }, |
| { |
| "epoch": 0.24608501118568232, |
| "grad_norm": 2.1220943927764893, |
| "learning_rate": 0.0001971891102300738, |
| "logits/chosen": -0.3758760690689087, |
| "logits/rejected": -0.4768344461917877, |
| "logps/chosen": -7.041341304779053, |
| "logps/rejected": -22.334056854248047, |
| "loss": 1.0144665241241455, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5401796698570251, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.1189827248454094, |
| "rewards/margins": 0.939737856388092, |
| "rewards/rejected": -0.820755124092102, |
| "step": 55, |
| "train_speed(iter/s)": 0.005613 |
| }, |
| { |
| "epoch": 0.2505592841163311, |
| "grad_norm": 1.843038558959961, |
| "learning_rate": 0.00019701157849175228, |
| "logits/chosen": -0.37703099846839905, |
| "logits/rejected": -0.46533748507499695, |
| "logps/chosen": -8.012426376342773, |
| "logps/rejected": -22.621742248535156, |
| "loss": 1.0123478174209595, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5707098245620728, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.15585632622241974, |
| "rewards/margins": 0.9487887620925903, |
| "rewards/rejected": -0.792932391166687, |
| "step": 56, |
| "train_speed(iter/s)": 0.005614 |
| }, |
| { |
| "epoch": 0.2550335570469799, |
| "grad_norm": 2.230550527572632, |
| "learning_rate": 0.00019682869568400684, |
| "logits/chosen": -0.27077603340148926, |
| "logits/rejected": -0.43189266324043274, |
| "logps/chosen": -6.5583930015563965, |
| "logps/rejected": -29.555484771728516, |
| "loss": 0.9619537591934204, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5382556915283203, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.012836463749408722, |
| "rewards/margins": 1.0998661518096924, |
| "rewards/rejected": -1.0870296955108643, |
| "step": 57, |
| "train_speed(iter/s)": 0.005614 |
| }, |
| { |
| "epoch": 0.2595078299776286, |
| "grad_norm": 2.252734661102295, |
| "learning_rate": 0.00019664047189448493, |
| "logits/chosen": -0.30061304569244385, |
| "logits/rejected": -0.4390646517276764, |
| "logps/chosen": -5.5507659912109375, |
| "logps/rejected": -26.51093292236328, |
| "loss": 0.9082356095314026, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.41620010137557983, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": -0.027434132993221283, |
| "rewards/margins": 0.9383218288421631, |
| "rewards/rejected": -0.9657559394836426, |
| "step": 58, |
| "train_speed(iter/s)": 0.005615 |
| }, |
| { |
| "epoch": 0.2639821029082774, |
| "grad_norm": 2.943512201309204, |
| "learning_rate": 0.00019644691750543767, |
| "logits/chosen": -0.2617231011390686, |
| "logits/rejected": -0.38011983036994934, |
| "logps/chosen": -10.003762245178223, |
| "logps/rejected": -27.785303115844727, |
| "loss": 1.1763017177581787, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6572035551071167, |
| "rewards/accuracies": 0.671875, |
| "rewards/chosen": -0.07653994858264923, |
| "rewards/margins": 0.9367327690124512, |
| "rewards/rejected": -1.0132726430892944, |
| "step": 59, |
| "train_speed(iter/s)": 0.005615 |
| }, |
| { |
| "epoch": 0.2684563758389262, |
| "grad_norm": 1.736883521080017, |
| "learning_rate": 0.00019624804319314705, |
| "logits/chosen": -0.3555205464363098, |
| "logits/rejected": -0.4549139738082886, |
| "logps/chosen": -7.022237300872803, |
| "logps/rejected": -28.04250717163086, |
| "loss": 0.9159735441207886, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.47989311814308167, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.060450125485658646, |
| "rewards/margins": 1.1664522886276245, |
| "rewards/rejected": -1.1060020923614502, |
| "step": 60, |
| "train_speed(iter/s)": 0.005616 |
| }, |
| { |
| "epoch": 0.27293064876957496, |
| "grad_norm": 1.8152859210968018, |
| "learning_rate": 0.00019604385992733715, |
| "logits/chosen": -0.2845993638038635, |
| "logits/rejected": -0.3513396680355072, |
| "logps/chosen": -9.077548027038574, |
| "logps/rejected": -22.780668258666992, |
| "loss": 0.9680572748184204, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4960174560546875, |
| "rewards/accuracies": 0.734375, |
| "rewards/chosen": 0.06926427036523819, |
| "rewards/margins": 0.9312706589698792, |
| "rewards/rejected": -0.862006425857544, |
| "step": 61, |
| "train_speed(iter/s)": 0.005616 |
| }, |
| { |
| "epoch": 0.27740492170022374, |
| "grad_norm": 2.46378755569458, |
| "learning_rate": 0.00019583437897056915, |
| "logits/chosen": -0.3284386098384857, |
| "logits/rejected": -0.4009109139442444, |
| "logps/chosen": -8.893238067626953, |
| "logps/rejected": -21.27986717224121, |
| "loss": 1.128482699394226, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5518131256103516, |
| "rewards/accuracies": 0.609375, |
| "rewards/chosen": 0.010286218486726284, |
| "rewards/margins": 0.5274713635444641, |
| "rewards/rejected": -0.5171850919723511, |
| "step": 62, |
| "train_speed(iter/s)": 0.005617 |
| }, |
| { |
| "epoch": 0.28187919463087246, |
| "grad_norm": 2.1732046604156494, |
| "learning_rate": 0.00019561961187761985, |
| "logits/chosen": -0.24422433972358704, |
| "logits/rejected": -0.375420480966568, |
| "logps/chosen": -6.340362071990967, |
| "logps/rejected": -20.184795379638672, |
| "loss": 0.987078070640564, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4728015065193176, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.168831467628479, |
| "rewards/margins": 0.6444362998008728, |
| "rewards/rejected": -0.4756048321723938, |
| "step": 63, |
| "train_speed(iter/s)": 0.005617 |
| }, |
| { |
| "epoch": 0.28635346756152125, |
| "grad_norm": 2.231933355331421, |
| "learning_rate": 0.00019539957049484458, |
| "logits/chosen": -0.22632239758968353, |
| "logits/rejected": -0.3298073410987854, |
| "logps/chosen": -5.790369987487793, |
| "logps/rejected": -21.647865295410156, |
| "loss": 0.9348481893539429, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4110260307788849, |
| "rewards/accuracies": 0.765625, |
| "rewards/chosen": 0.11635202169418335, |
| "rewards/margins": 0.6466613411903381, |
| "rewards/rejected": -0.5303093194961548, |
| "step": 64, |
| "train_speed(iter/s)": 0.005617 |
| }, |
| { |
| "epoch": 0.29082774049217003, |
| "grad_norm": 2.592883825302124, |
| "learning_rate": 0.00019517426695952358, |
| "logits/chosen": -0.2631714344024658, |
| "logits/rejected": -0.3835735321044922, |
| "logps/chosen": -8.772464752197266, |
| "logps/rejected": -23.13977813720703, |
| "loss": 1.0820575952529907, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6241056323051453, |
| "rewards/accuracies": 0.828125, |
| "rewards/chosen": 0.1670496016740799, |
| "rewards/margins": 0.9779586791992188, |
| "rewards/rejected": -0.8109091520309448, |
| "step": 65, |
| "train_speed(iter/s)": 0.005617 |
| }, |
| { |
| "epoch": 0.2953020134228188, |
| "grad_norm": 2.1508402824401855, |
| "learning_rate": 0.0001949437136991925, |
| "logits/chosen": -0.28555068373680115, |
| "logits/rejected": -0.42386704683303833, |
| "logps/chosen": -4.788064002990723, |
| "logps/rejected": -26.989431381225586, |
| "loss": 0.7586299777030945, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.3784462809562683, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.09788702428340912, |
| "rewards/margins": 1.3373408317565918, |
| "rewards/rejected": -1.239453673362732, |
| "step": 66, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.29977628635346754, |
| "grad_norm": 4.065814971923828, |
| "learning_rate": 0.00019470792343095718, |
| "logits/chosen": -0.3336545526981354, |
| "logits/rejected": -0.3990672826766968, |
| "logps/chosen": -8.33244514465332, |
| "logps/rejected": -29.013874053955078, |
| "loss": 1.0058256387710571, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6062498092651367, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": -0.020362887531518936, |
| "rewards/margins": 1.5735464096069336, |
| "rewards/rejected": -1.5939093828201294, |
| "step": 67, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.3042505592841163, |
| "grad_norm": 2.998826265335083, |
| "learning_rate": 0.0001944669091607919, |
| "logits/chosen": -0.17095105350017548, |
| "logits/rejected": -0.33402499556541443, |
| "logps/chosen": -7.172041893005371, |
| "logps/rejected": -36.20872497558594, |
| "loss": 0.8435176610946655, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5263906121253967, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.01939973793923855, |
| "rewards/margins": 2.0239157676696777, |
| "rewards/rejected": -2.004516124725342, |
| "step": 68, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.3087248322147651, |
| "grad_norm": 2.8316242694854736, |
| "learning_rate": 0.00019422068418282202, |
| "logits/chosen": -0.23822994530200958, |
| "logits/rejected": -0.3510916531085968, |
| "logps/chosen": -8.204334259033203, |
| "logps/rejected": -26.101930618286133, |
| "loss": 0.9929779171943665, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.592566967010498, |
| "rewards/accuracies": 0.765625, |
| "rewards/chosen": 0.07919630408287048, |
| "rewards/margins": 1.4216609001159668, |
| "rewards/rejected": -1.3424646854400635, |
| "step": 69, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.3131991051454139, |
| "grad_norm": 2.1727654933929443, |
| "learning_rate": 0.00019396926207859084, |
| "logits/chosen": -0.19429409503936768, |
| "logits/rejected": -0.33817699551582336, |
| "logps/chosen": -7.17151403427124, |
| "logps/rejected": -29.4168701171875, |
| "loss": 0.8446124196052551, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4697123169898987, |
| "rewards/accuracies": 0.859375, |
| "rewards/chosen": 0.09260577708482742, |
| "rewards/margins": 1.373153567314148, |
| "rewards/rejected": -1.280547857284546, |
| "step": 70, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.31767337807606266, |
| "grad_norm": 2.131971597671509, |
| "learning_rate": 0.00019371265671631037, |
| "logits/chosen": -0.2500344514846802, |
| "logits/rejected": -0.391252338886261, |
| "logps/chosen": -6.454396724700928, |
| "logps/rejected": -31.325042724609375, |
| "loss": 0.857275128364563, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.45489931106567383, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.12299442291259766, |
| "rewards/margins": 1.4080641269683838, |
| "rewards/rejected": -1.2850697040557861, |
| "step": 71, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.3221476510067114, |
| "grad_norm": 2.6743502616882324, |
| "learning_rate": 0.00019345088225009626, |
| "logits/chosen": -0.20935188233852386, |
| "logits/rejected": -0.3504689633846283, |
| "logps/chosen": -8.028803825378418, |
| "logps/rejected": -33.473976135253906, |
| "loss": 0.951294481754303, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5516812205314636, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.002318259794265032, |
| "rewards/margins": 1.4767383337020874, |
| "rewards/rejected": -1.4744200706481934, |
| "step": 72, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.32662192393736017, |
| "grad_norm": 2.1057374477386475, |
| "learning_rate": 0.0001931839531191873, |
| "logits/chosen": -0.21616998314857483, |
| "logits/rejected": -0.3215841054916382, |
| "logps/chosen": -6.386109352111816, |
| "logps/rejected": -26.058643341064453, |
| "loss": 0.9168461561203003, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5115207433700562, |
| "rewards/accuracies": 0.859375, |
| "rewards/chosen": 0.16049177944660187, |
| "rewards/margins": 1.1623506546020508, |
| "rewards/rejected": -1.0018587112426758, |
| "step": 73, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.33109619686800895, |
| "grad_norm": 1.8313034772872925, |
| "learning_rate": 0.00019291188404714878, |
| "logits/chosen": -0.2762296795845032, |
| "logits/rejected": -0.3759133815765381, |
| "logps/chosen": -7.998234748840332, |
| "logps/rejected": -29.753068923950195, |
| "loss": 0.9865212440490723, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5828378796577454, |
| "rewards/accuracies": 0.828125, |
| "rewards/chosen": 0.04827752709388733, |
| "rewards/margins": 1.679243564605713, |
| "rewards/rejected": -1.6309659481048584, |
| "step": 74, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.33557046979865773, |
| "grad_norm": 2.7458698749542236, |
| "learning_rate": 0.0001926346900410604, |
| "logits/chosen": -0.26945143938064575, |
| "logits/rejected": -0.319683313369751, |
| "logps/chosen": -9.152531623840332, |
| "logps/rejected": -22.447328567504883, |
| "loss": 1.118042230606079, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6387289762496948, |
| "rewards/accuracies": 0.765625, |
| "rewards/chosen": 0.012468192726373672, |
| "rewards/margins": 0.904109537601471, |
| "rewards/rejected": -0.8916412591934204, |
| "step": 75, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.3400447427293065, |
| "grad_norm": 2.5607635974884033, |
| "learning_rate": 0.00019235238639068856, |
| "logits/chosen": -0.19932252168655396, |
| "logits/rejected": -0.37197503447532654, |
| "logps/chosen": -7.51588773727417, |
| "logps/rejected": -33.714637756347656, |
| "loss": 0.9608560800552368, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5613290667533875, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.0078031644225120544, |
| "rewards/margins": 1.3550933599472046, |
| "rewards/rejected": -1.347290277481079, |
| "step": 76, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.34451901565995524, |
| "grad_norm": 2.152766227722168, |
| "learning_rate": 0.00019206498866764288, |
| "logits/chosen": -0.18406139314174652, |
| "logits/rejected": -0.36746495962142944, |
| "logps/chosen": -4.768266677856445, |
| "logps/rejected": -35.82173156738281, |
| "loss": 0.8752678632736206, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5213672518730164, |
| "rewards/accuracies": 0.859375, |
| "rewards/chosen": -0.024980902671813965, |
| "rewards/margins": 1.6603260040283203, |
| "rewards/rejected": -1.6853069067001343, |
| "step": 77, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.348993288590604, |
| "grad_norm": 1.957088589668274, |
| "learning_rate": 0.0001917725127245174, |
| "logits/chosen": -0.2368219494819641, |
| "logits/rejected": -0.31324779987335205, |
| "logps/chosen": -7.644411087036133, |
| "logps/rejected": -25.68874740600586, |
| "loss": 0.9143100380897522, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.47288480401039124, |
| "rewards/accuracies": 0.796875, |
| "rewards/chosen": 0.04239855334162712, |
| "rewards/margins": 0.925534188747406, |
| "rewards/rejected": -0.8831356167793274, |
| "step": 78, |
| "train_speed(iter/s)": 0.005618 |
| }, |
| { |
| "epoch": 0.3534675615212528, |
| "grad_norm": 2.0980629920959473, |
| "learning_rate": 0.0001914749746940161, |
| "logits/chosen": -0.3016536831855774, |
| "logits/rejected": -0.3982566297054291, |
| "logps/chosen": -7.566556930541992, |
| "logps/rejected": -26.53971290588379, |
| "loss": 0.9413108825683594, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5161536931991577, |
| "rewards/accuracies": 0.765625, |
| "rewards/chosen": 0.1749323159456253, |
| "rewards/margins": 1.1360664367675781, |
| "rewards/rejected": -0.9611341953277588, |
| "step": 79, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.3579418344519016, |
| "grad_norm": 2.1753523349761963, |
| "learning_rate": 0.00019117239098806295, |
| "logits/chosen": -0.25202423334121704, |
| "logits/rejected": -0.3296201527118683, |
| "logps/chosen": -8.623554229736328, |
| "logps/rejected": -24.96216583251953, |
| "loss": 0.9156360626220703, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.521922767162323, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.17782656848430634, |
| "rewards/margins": 1.1749255657196045, |
| "rewards/rejected": -0.997098982334137, |
| "step": 80, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.3624161073825503, |
| "grad_norm": 3.2816970348358154, |
| "learning_rate": 0.00019086477829689685, |
| "logits/chosen": -0.22371181845664978, |
| "logits/rejected": -0.3009035885334015, |
| "logps/chosen": -10.19958782196045, |
| "logps/rejected": -26.129743576049805, |
| "loss": 1.161747932434082, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.7456332445144653, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.06418374180793762, |
| "rewards/margins": 1.1640924215316772, |
| "rewards/rejected": -1.099908709526062, |
| "step": 81, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.3668903803131991, |
| "grad_norm": 4.27427339553833, |
| "learning_rate": 0.0001905521535881509, |
| "logits/chosen": -0.15121029317378998, |
| "logits/rejected": -0.3036370873451233, |
| "logps/chosen": -7.66187858581543, |
| "logps/rejected": -30.927831649780273, |
| "loss": 1.1210309267044067, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6641408205032349, |
| "rewards/accuracies": 0.765625, |
| "rewards/chosen": -0.00867693591862917, |
| "rewards/margins": 1.272807002067566, |
| "rewards/rejected": -1.2814840078353882, |
| "step": 82, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.3713646532438479, |
| "grad_norm": 2.29978084564209, |
| "learning_rate": 0.00019023453410591635, |
| "logits/chosen": -0.21486631035804749, |
| "logits/rejected": -0.3746005892753601, |
| "logps/chosen": -9.569436073303223, |
| "logps/rejected": -32.21381378173828, |
| "loss": 1.1229475736618042, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6794871687889099, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.0131741464138031, |
| "rewards/margins": 1.4518442153930664, |
| "rewards/rejected": -1.4386701583862305, |
| "step": 83, |
| "train_speed(iter/s)": 0.005619 |
| }, |
| { |
| "epoch": 0.37583892617449666, |
| "grad_norm": 2.6593234539031982, |
| "learning_rate": 0.00018991193736979175, |
| "logits/chosen": -0.12227392941713333, |
| "logits/rejected": -0.2726154327392578, |
| "logps/chosen": -7.078165054321289, |
| "logps/rejected": -30.666397094726562, |
| "loss": 0.7911984920501709, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.4572717845439911, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.06052548438310623, |
| "rewards/margins": 1.428847074508667, |
| "rewards/rejected": -1.3683216571807861, |
| "step": 84, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.38031319910514544, |
| "grad_norm": 2.8958120346069336, |
| "learning_rate": 0.00018958438117391618, |
| "logits/chosen": -0.0972181037068367, |
| "logits/rejected": -0.2590752840042114, |
| "logps/chosen": -8.193891525268555, |
| "logps/rejected": -35.56235885620117, |
| "loss": 1.0837842226028442, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6906030774116516, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": -0.06835082173347473, |
| "rewards/margins": 1.79606294631958, |
| "rewards/rejected": -1.8644137382507324, |
| "step": 85, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.38478747203579416, |
| "grad_norm": 3.5083513259887695, |
| "learning_rate": 0.00018925188358598813, |
| "logits/chosen": -0.17531399428844452, |
| "logits/rejected": -0.25995591282844543, |
| "logps/chosen": -8.639050483703613, |
| "logps/rejected": -29.226892471313477, |
| "loss": 1.0494576692581177, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6720495820045471, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": -0.03724604845046997, |
| "rewards/margins": 1.4595208168029785, |
| "rewards/rejected": -1.4967668056488037, |
| "step": 86, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.38926174496644295, |
| "grad_norm": 2.0132222175598145, |
| "learning_rate": 0.00018891446294626866, |
| "logits/chosen": -0.05051232874393463, |
| "logits/rejected": -0.25516456365585327, |
| "logps/chosen": -4.587294578552246, |
| "logps/rejected": -36.072322845458984, |
| "loss": 0.602871298789978, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.32869917154312134, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.17169487476348877, |
| "rewards/margins": 2.032069444656372, |
| "rewards/rejected": -1.8603744506835938, |
| "step": 87, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.39373601789709173, |
| "grad_norm": 3.1251611709594727, |
| "learning_rate": 0.00018857213786656985, |
| "logits/chosen": -0.16258063912391663, |
| "logits/rejected": -0.2599830627441406, |
| "logps/chosen": -9.219761848449707, |
| "logps/rejected": -27.451026916503906, |
| "loss": 1.0880504846572876, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6885117292404175, |
| "rewards/accuracies": 0.859375, |
| "rewards/chosen": 0.04242437332868576, |
| "rewards/margins": 1.207942247390747, |
| "rewards/rejected": -1.1655179262161255, |
| "step": 88, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.3982102908277405, |
| "grad_norm": 2.0855207443237305, |
| "learning_rate": 0.0001882249272292282, |
| "logits/chosen": -0.1683093011379242, |
| "logits/rejected": -0.3001120686531067, |
| "logps/chosen": -7.336337089538574, |
| "logps/rejected": -24.294296264648438, |
| "loss": 0.9775981903076172, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.515565812587738, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.09023457765579224, |
| "rewards/margins": 0.9091135859489441, |
| "rewards/rejected": -0.8188790082931519, |
| "step": 89, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.40268456375838924, |
| "grad_norm": 1.8228232860565186, |
| "learning_rate": 0.00018787285018606297, |
| "logits/chosen": -0.1059221625328064, |
| "logits/rejected": -0.30855828523635864, |
| "logps/chosen": -3.3506364822387695, |
| "logps/rejected": -32.792762756347656, |
| "loss": 0.6121629476547241, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.2924441993236542, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.11779851466417313, |
| "rewards/margins": 1.558281660079956, |
| "rewards/rejected": -1.4404832124710083, |
| "step": 90, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.407158836689038, |
| "grad_norm": 2.038564443588257, |
| "learning_rate": 0.00018751592615732005, |
| "logits/chosen": -0.23549917340278625, |
| "logits/rejected": -0.28893160820007324, |
| "logps/chosen": -7.049911975860596, |
| "logps/rejected": -23.307397842407227, |
| "loss": 0.8128443956375122, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.42805472016334534, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.235755056142807, |
| "rewards/margins": 1.322766900062561, |
| "rewards/rejected": -1.0870119333267212, |
| "step": 91, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.4116331096196868, |
| "grad_norm": 3.0157337188720703, |
| "learning_rate": 0.0001871541748306005, |
| "logits/chosen": -0.16576483845710754, |
| "logits/rejected": -0.3216527998447418, |
| "logps/chosen": -3.5259528160095215, |
| "logps/rejected": -34.49477005004883, |
| "loss": 0.6575559377670288, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.324476957321167, |
| "rewards/accuracies": 0.796875, |
| "rewards/chosen": 0.0956890657544136, |
| "rewards/margins": 2.1255316734313965, |
| "rewards/rejected": -2.0298423767089844, |
| "step": 92, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.4161073825503356, |
| "grad_norm": 4.184682846069336, |
| "learning_rate": 0.00018678761615977468, |
| "logits/chosen": -0.08753462135791779, |
| "logits/rejected": -0.2798512876033783, |
| "logps/chosen": -9.157657623291016, |
| "logps/rejected": -36.74687576293945, |
| "loss": 1.0873180627822876, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.7294520139694214, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.07273296266794205, |
| "rewards/margins": 1.809075117111206, |
| "rewards/rejected": -1.736342430114746, |
| "step": 93, |
| "train_speed(iter/s)": 0.00562 |
| }, |
| { |
| "epoch": 0.42058165548098436, |
| "grad_norm": 3.6453614234924316, |
| "learning_rate": 0.00018641627036388169, |
| "logits/chosen": -0.26587241888046265, |
| "logits/rejected": -0.35532236099243164, |
| "logps/chosen": -6.705877780914307, |
| "logps/rejected": -34.05841827392578, |
| "loss": 0.8682178258895874, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.5556175112724304, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.0699356198310852, |
| "rewards/margins": 2.082659959793091, |
| "rewards/rejected": -2.0127243995666504, |
| "step": 94, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.4250559284116331, |
| "grad_norm": 3.926342725753784, |
| "learning_rate": 0.00018604015792601396, |
| "logits/chosen": -0.18198826909065247, |
| "logits/rejected": -0.28424450755119324, |
| "logps/chosen": -9.228765487670898, |
| "logps/rejected": -28.494712829589844, |
| "loss": 1.100342869758606, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6449709534645081, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.0022647292353212833, |
| "rewards/margins": 1.3801387548446655, |
| "rewards/rejected": -1.3778741359710693, |
| "step": 95, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.42953020134228187, |
| "grad_norm": 6.2491960525512695, |
| "learning_rate": 0.00018565929959218758, |
| "logits/chosen": -0.162868469953537, |
| "logits/rejected": -0.35664084553718567, |
| "logps/chosen": -5.505702495574951, |
| "logps/rejected": -40.82157516479492, |
| "loss": 0.8373050689697266, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.48504531383514404, |
| "rewards/accuracies": 0.828125, |
| "rewards/chosen": 0.03489302098751068, |
| "rewards/margins": 1.9277093410491943, |
| "rewards/rejected": -1.892816424369812, |
| "step": 96, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.43400447427293065, |
| "grad_norm": 3.155294895172119, |
| "learning_rate": 0.0001852737163701979, |
| "logits/chosen": -0.14200377464294434, |
| "logits/rejected": -0.3238810896873474, |
| "logps/chosen": -6.1366353034973145, |
| "logps/rejected": -36.68225860595703, |
| "loss": 0.7804915904998779, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.47483596205711365, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.05882677063345909, |
| "rewards/margins": 1.9737701416015625, |
| "rewards/rejected": -1.9149433374404907, |
| "step": 97, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.43847874720357943, |
| "grad_norm": 2.6832268238067627, |
| "learning_rate": 0.00018488342952846073, |
| "logits/chosen": -0.12941551208496094, |
| "logits/rejected": -0.2959509789943695, |
| "logps/chosen": -12.032288551330566, |
| "logps/rejected": -35.381710052490234, |
| "loss": 0.838473916053772, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.49494707584381104, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.032307837158441544, |
| "rewards/margins": 1.5512968301773071, |
| "rewards/rejected": -1.5189889669418335, |
| "step": 98, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.4429530201342282, |
| "grad_norm": 2.0462474822998047, |
| "learning_rate": 0.0001844884605948392, |
| "logits/chosen": -0.23558180034160614, |
| "logits/rejected": -0.3692886531352997, |
| "logps/chosen": -9.6895112991333, |
| "logps/rejected": -26.33759117126465, |
| "loss": 1.0117727518081665, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.6607507467269897, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.148264542222023, |
| "rewards/margins": 1.4499516487121582, |
| "rewards/rejected": -1.3016871213912964, |
| "step": 99, |
| "train_speed(iter/s)": 0.005621 |
| }, |
| { |
| "epoch": 0.44742729306487694, |
| "grad_norm": 2.056586503982544, |
| "learning_rate": 0.00018408883135545632, |
| "logits/chosen": -0.20895381271839142, |
| "logits/rejected": -0.350538432598114, |
| "logps/chosen": -7.372556209564209, |
| "logps/rejected": -30.752912521362305, |
| "loss": 0.8857133388519287, |
| "memory(GiB)": 46.57, |
| "nll_loss": 0.49547722935676575, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.0997147411108017, |
| "rewards/margins": 1.3622983694076538, |
| "rewards/rejected": -1.262583613395691, |
| "step": 100, |
| "train_speed(iter/s)": 0.005621 |
| } |
| ], |
| "logging_steps": 1, |
| "max_steps": 446, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 2, |
| "save_steps": 50, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 1.1766177885008691e+17, |
| "train_batch_size": 4, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
