| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.9968602825745683, | |
| "eval_steps": 100, | |
| "global_step": 954, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": 0.952304482460022, | |
| "logits/rejected": 0.5888463854789734, | |
| "logps/chosen": -223.79486083984375, | |
| "logps/rejected": -209.482666015625, | |
| "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": 5.208333333333334e-07, | |
| "logits/chosen": 0.836148738861084, | |
| "logits/rejected": 0.8545415997505188, | |
| "logps/chosen": -236.217529296875, | |
| "logps/rejected": -221.92974853515625, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.4444444477558136, | |
| "rewards/chosen": 0.00045002170372754335, | |
| "rewards/margins": 0.0005102179129607975, | |
| "rewards/rejected": -6.019628563080914e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 1.0416666666666667e-06, | |
| "logits/chosen": 0.8339964151382446, | |
| "logits/rejected": 0.9286314249038696, | |
| "logps/chosen": -254.7573699951172, | |
| "logps/rejected": -247.9052734375, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": -0.0004581348621286452, | |
| "rewards/margins": 0.00032393785659223795, | |
| "rewards/rejected": -0.0007820727187208831, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 1.5625e-06, | |
| "logits/chosen": 0.8594380617141724, | |
| "logits/rejected": 0.916830837726593, | |
| "logps/chosen": -260.4742736816406, | |
| "logps/rejected": -232.0736846923828, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.0009176501189358532, | |
| "rewards/margins": 0.0013712994987145066, | |
| "rewards/rejected": -0.00045364940888248384, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 2.0833333333333334e-06, | |
| "logits/chosen": 0.8113815188407898, | |
| "logits/rejected": 0.9032772183418274, | |
| "logps/chosen": -280.20867919921875, | |
| "logps/rejected": -228.60598754882812, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": 3.258117794757709e-05, | |
| "rewards/margins": 0.0006167444516904652, | |
| "rewards/rejected": -0.0005841633537784219, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 2.604166666666667e-06, | |
| "logits/chosen": 0.8574679493904114, | |
| "logits/rejected": 0.9363411068916321, | |
| "logps/chosen": -257.6195373535156, | |
| "logps/rejected": -219.19448852539062, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.0007853487622924149, | |
| "rewards/margins": 0.001927538076415658, | |
| "rewards/rejected": -0.0011421891395002604, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 3.125e-06, | |
| "logits/chosen": 0.8745417594909668, | |
| "logits/rejected": 0.901624858379364, | |
| "logps/chosen": -237.000244140625, | |
| "logps/rejected": -237.52395629882812, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 1.0722555089159869e-05, | |
| "rewards/margins": 0.0009478582069277763, | |
| "rewards/rejected": -0.0009371357737109065, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 3.6458333333333333e-06, | |
| "logits/chosen": 0.8507300615310669, | |
| "logits/rejected": 0.8796240091323853, | |
| "logps/chosen": -260.81085205078125, | |
| "logps/rejected": -227.607421875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.0008873953483998775, | |
| "rewards/margins": -0.0005922773270867765, | |
| "rewards/rejected": -0.00029511802131310105, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.166666666666667e-06, | |
| "logits/chosen": 0.8868101239204407, | |
| "logits/rejected": 0.9193674921989441, | |
| "logps/chosen": -251.10498046875, | |
| "logps/rejected": -231.2660675048828, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -0.00020074064377695322, | |
| "rewards/margins": 0.0003954143321607262, | |
| "rewards/rejected": -0.0005961551214568317, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.6875000000000004e-06, | |
| "logits/chosen": 0.8633348345756531, | |
| "logits/rejected": 0.9175950884819031, | |
| "logps/chosen": -225.3305206298828, | |
| "logps/rejected": -241.3758087158203, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.0006292141624726355, | |
| "rewards/margins": 0.0004391434194985777, | |
| "rewards/rejected": -0.0010683576110750437, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.999731868769027e-06, | |
| "logits/chosen": 0.9241905212402344, | |
| "logits/rejected": 0.9234986305236816, | |
| "logps/chosen": -242.17343139648438, | |
| "logps/rejected": -221.4640350341797, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.00037607658305205405, | |
| "rewards/margins": 0.0006910150987096131, | |
| "rewards/rejected": -0.0010670917108654976, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": 0.8145024180412292, | |
| "eval_logits/rejected": 0.9167963862419128, | |
| "eval_logps/chosen": -260.2240295410156, | |
| "eval_logps/rejected": -231.53602600097656, | |
| "eval_loss": 0.6927996873855591, | |
| "eval_rewards/accuracies": 0.5320000052452087, | |
| "eval_rewards/chosen": 0.00020143474102951586, | |
| "eval_rewards/margins": 0.0012005382450297475, | |
| "eval_rewards/rejected": -0.000999103649519384, | |
| "eval_runtime": 326.2051, | |
| "eval_samples_per_second": 6.131, | |
| "eval_steps_per_second": 0.383, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.996716052911017e-06, | |
| "logits/chosen": 0.8340839147567749, | |
| "logits/rejected": 0.8756756782531738, | |
| "logps/chosen": -264.01495361328125, | |
| "logps/rejected": -219.5539093017578, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 7.205537986010313e-05, | |
| "rewards/margins": 0.0008332778816111386, | |
| "rewards/rejected": -0.0007612224435433745, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.9903533134293035e-06, | |
| "logits/chosen": 0.8605300784111023, | |
| "logits/rejected": 0.9697392582893372, | |
| "logps/chosen": -254.9059295654297, | |
| "logps/rejected": -219.1450958251953, | |
| "loss": 0.6924, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.0002610751544125378, | |
| "rewards/margins": 0.002276130486279726, | |
| "rewards/rejected": -0.002015055390074849, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.9806521797692184e-06, | |
| "logits/chosen": 0.877278208732605, | |
| "logits/rejected": 0.8787814974784851, | |
| "logps/chosen": -264.6486511230469, | |
| "logps/rejected": -247.06497192382812, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -0.00013167767610866576, | |
| "rewards/margins": 0.002138237003237009, | |
| "rewards/rejected": -0.0022699148394167423, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.967625656594782e-06, | |
| "logits/chosen": 0.8538626432418823, | |
| "logits/rejected": 0.9312636256217957, | |
| "logps/chosen": -222.04147338867188, | |
| "logps/rejected": -232.23892211914062, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.0011232274118810892, | |
| "rewards/margins": 0.0009854320669546723, | |
| "rewards/rejected": -0.0021086593624204397, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.95129120635556e-06, | |
| "logits/chosen": 0.8747261166572571, | |
| "logits/rejected": 0.9058718681335449, | |
| "logps/chosen": -258.2791442871094, | |
| "logps/rejected": -215.7833251953125, | |
| "loss": 0.6919, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.0004979403456673026, | |
| "rewards/margins": 0.002242505783215165, | |
| "rewards/rejected": -0.0027404462452977896, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.93167072587771e-06, | |
| "logits/chosen": 0.7788752317428589, | |
| "logits/rejected": 0.8470233082771301, | |
| "logps/chosen": -257.824951171875, | |
| "logps/rejected": -250.79281616210938, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.0017024253029376268, | |
| "rewards/margins": 0.0027075002435594797, | |
| "rewards/rejected": -0.0044099255464971066, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.908790517010637e-06, | |
| "logits/chosen": 0.9306305050849915, | |
| "logits/rejected": 0.9719738960266113, | |
| "logps/chosen": -239.07101440429688, | |
| "logps/rejected": -252.9778289794922, | |
| "loss": 0.6911, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.002183904405683279, | |
| "rewards/margins": 0.002954904455691576, | |
| "rewards/rejected": -0.005138809327036142, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 4.882681251368549e-06, | |
| "logits/chosen": 0.853600800037384, | |
| "logits/rejected": 0.8488330841064453, | |
| "logps/chosen": -270.89105224609375, | |
| "logps/rejected": -257.0163269042969, | |
| "loss": 0.6906, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.0012318979715928435, | |
| "rewards/margins": 0.004222923889756203, | |
| "rewards/rejected": -0.0054548210464417934, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 4.853377929214243e-06, | |
| "logits/chosen": 0.7871135473251343, | |
| "logits/rejected": 0.8603521585464478, | |
| "logps/chosen": -252.1376190185547, | |
| "logps/rejected": -245.2152862548828, | |
| "loss": 0.6898, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.0007233443320728838, | |
| "rewards/margins": 0.007929561659693718, | |
| "rewards/rejected": -0.008652905933558941, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 4.8209198325401815e-06, | |
| "logits/chosen": 0.8782480955123901, | |
| "logits/rejected": 0.8951404690742493, | |
| "logps/chosen": -237.61087036132812, | |
| "logps/rejected": -233.7824249267578, | |
| "loss": 0.6893, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.002221657894551754, | |
| "rewards/margins": 0.00844704918563366, | |
| "rewards/rejected": -0.010668707080185413, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": 0.8204554915428162, | |
| "eval_logits/rejected": 0.9234155416488647, | |
| "eval_logps/chosen": -260.62249755859375, | |
| "eval_logps/rejected": -232.77418518066406, | |
| "eval_loss": 0.6890601515769958, | |
| "eval_rewards/accuracies": 0.6499999761581421, | |
| "eval_rewards/chosen": -0.0037833875976502895, | |
| "eval_rewards/margins": 0.009597329422831535, | |
| "eval_rewards/rejected": -0.013380718417465687, | |
| "eval_runtime": 326.0716, | |
| "eval_samples_per_second": 6.134, | |
| "eval_steps_per_second": 0.383, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 4.785350472409792e-06, | |
| "logits/chosen": 0.906767725944519, | |
| "logits/rejected": 0.9123601913452148, | |
| "logps/chosen": -233.3804168701172, | |
| "logps/rejected": -228.94284057617188, | |
| "loss": 0.689, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.00779892411082983, | |
| "rewards/margins": 0.007471003569662571, | |
| "rewards/rejected": -0.015269925817847252, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 4.746717530629565e-06, | |
| "logits/chosen": 0.8570274114608765, | |
| "logits/rejected": 0.8925272226333618, | |
| "logps/chosen": -259.993896484375, | |
| "logps/rejected": -239.8516082763672, | |
| "loss": 0.6887, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.006393183022737503, | |
| "rewards/margins": 0.01187940128147602, | |
| "rewards/rejected": -0.018272582441568375, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 4.7050727958301505e-06, | |
| "logits/chosen": 0.8952150344848633, | |
| "logits/rejected": 0.8783475756645203, | |
| "logps/chosen": -244.82919311523438, | |
| "logps/rejected": -230.5172882080078, | |
| "loss": 0.687, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.008937228471040726, | |
| "rewards/margins": 0.014644329436123371, | |
| "rewards/rejected": -0.023581556975841522, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 4.660472094042121e-06, | |
| "logits/chosen": 0.857439398765564, | |
| "logits/rejected": 0.8656377792358398, | |
| "logps/chosen": -282.5680847167969, | |
| "logps/rejected": -236.55111694335938, | |
| "loss": 0.6853, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.012710051611065865, | |
| "rewards/margins": 0.017861289903521538, | |
| "rewards/rejected": -0.030571341514587402, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 4.612975213859487e-06, | |
| "logits/chosen": 0.8455727696418762, | |
| "logits/rejected": 0.9185819625854492, | |
| "logps/chosen": -268.9493713378906, | |
| "logps/rejected": -243.8383331298828, | |
| "loss": 0.6852, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.014121539890766144, | |
| "rewards/margins": 0.015637289732694626, | |
| "rewards/rejected": -0.02975882962346077, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 4.5626458262912745e-06, | |
| "logits/chosen": 0.858254075050354, | |
| "logits/rejected": 0.8988651037216187, | |
| "logps/chosen": -274.9979553222656, | |
| "logps/rejected": -262.4846496582031, | |
| "loss": 0.6842, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.008313321508467197, | |
| "rewards/margins": 0.024480264633893967, | |
| "rewards/rejected": -0.03279358521103859, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 4.509551399408598e-06, | |
| "logits/chosen": 0.9360445737838745, | |
| "logits/rejected": 0.9437822103500366, | |
| "logps/chosen": -253.78689575195312, | |
| "logps/rejected": -211.69509887695312, | |
| "loss": 0.683, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.015140311792492867, | |
| "rewards/margins": 0.022892246022820473, | |
| "rewards/rejected": -0.03803255409002304, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 4.453763107901676e-06, | |
| "logits/chosen": 0.9350343942642212, | |
| "logits/rejected": 0.8994883298873901, | |
| "logps/chosen": -247.9213104248047, | |
| "logps/rejected": -256.08056640625, | |
| "loss": 0.6842, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.025496864691376686, | |
| "rewards/margins": 0.012935856357216835, | |
| "rewards/rejected": -0.03843272104859352, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 4.3953557376679856e-06, | |
| "logits/chosen": 0.8583111763000488, | |
| "logits/rejected": 0.8643442988395691, | |
| "logps/chosen": -262.2920837402344, | |
| "logps/rejected": -258.9486083984375, | |
| "loss": 0.6827, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.027247220277786255, | |
| "rewards/margins": 0.015835126861929893, | |
| "rewards/rejected": -0.0430823490023613, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 4.33440758555951e-06, | |
| "logits/chosen": 0.8355463743209839, | |
| "logits/rejected": 0.9161020517349243, | |
| "logps/chosen": -250.9317626953125, | |
| "logps/rejected": -247.6198272705078, | |
| "loss": 0.6809, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.018615344539284706, | |
| "rewards/margins": 0.033209316432476044, | |
| "rewards/rejected": -0.0518246591091156, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": 0.8092363476753235, | |
| "eval_logits/rejected": 0.9150914549827576, | |
| "eval_logps/chosen": -263.3646545410156, | |
| "eval_logps/rejected": -237.54312133789062, | |
| "eval_loss": 0.6809768080711365, | |
| "eval_rewards/accuracies": 0.6679999828338623, | |
| "eval_rewards/chosen": -0.03120502457022667, | |
| "eval_rewards/margins": 0.029865048825740814, | |
| "eval_rewards/rejected": -0.061070073395967484, | |
| "eval_runtime": 325.9788, | |
| "eval_samples_per_second": 6.135, | |
| "eval_steps_per_second": 0.383, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 4.2710003544234255e-06, | |
| "logits/chosen": 0.855351448059082, | |
| "logits/rejected": 0.8703336715698242, | |
| "logps/chosen": -238.54061889648438, | |
| "logps/rejected": -230.7489776611328, | |
| "loss": 0.6787, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.036315977573394775, | |
| "rewards/margins": 0.025386247783899307, | |
| "rewards/rejected": -0.061702221632003784, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 4.205219043576955e-06, | |
| "logits/chosen": 0.8615506887435913, | |
| "logits/rejected": 0.8854458928108215, | |
| "logps/chosen": -227.1722869873047, | |
| "logps/rejected": -220.97412109375, | |
| "loss": 0.6785, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.03975047916173935, | |
| "rewards/margins": 0.028606727719306946, | |
| "rewards/rejected": -0.06835721433162689, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 4.137151834863213e-06, | |
| "logits/chosen": 0.8440972566604614, | |
| "logits/rejected": 0.8271803855895996, | |
| "logps/chosen": -256.4277038574219, | |
| "logps/rejected": -264.7651062011719, | |
| "loss": 0.6787, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.04135146737098694, | |
| "rewards/margins": 0.039213813841342926, | |
| "rewards/rejected": -0.08056528866291046, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 4.066889974440757e-06, | |
| "logits/chosen": 0.8700094223022461, | |
| "logits/rejected": 0.8766329884529114, | |
| "logps/chosen": -269.4642333984375, | |
| "logps/rejected": -255.2469024658203, | |
| "loss": 0.6737, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.03877495974302292, | |
| "rewards/margins": 0.04113053157925606, | |
| "rewards/rejected": -0.07990548759698868, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 3.994527650465352e-06, | |
| "logits/chosen": 0.8648789525032043, | |
| "logits/rejected": 0.8878329396247864, | |
| "logps/chosen": -227.1282958984375, | |
| "logps/rejected": -212.8804931640625, | |
| "loss": 0.6755, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.05504993349313736, | |
| "rewards/margins": 0.040687672793865204, | |
| "rewards/rejected": -0.09573759883642197, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 3.92016186682789e-06, | |
| "logits/chosen": 0.7903240919113159, | |
| "logits/rejected": 0.8026289939880371, | |
| "logps/chosen": -214.0880889892578, | |
| "logps/rejected": -244.24618530273438, | |
| "loss": 0.6752, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.06658105552196503, | |
| "rewards/margins": 0.03603743016719818, | |
| "rewards/rejected": -0.1026184931397438, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 3.843892313117724e-06, | |
| "logits/chosen": 0.8764356374740601, | |
| "logits/rejected": 0.9107359051704407, | |
| "logps/chosen": -273.88873291015625, | |
| "logps/rejected": -251.10812377929688, | |
| "loss": 0.6758, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.06244384124875069, | |
| "rewards/margins": 0.04542430490255356, | |
| "rewards/rejected": -0.10786814987659454, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 3.7658212309857576e-06, | |
| "logits/chosen": 0.8514490127563477, | |
| "logits/rejected": 0.8745658993721008, | |
| "logps/chosen": -244.4405059814453, | |
| "logps/rejected": -224.3061981201172, | |
| "loss": 0.6708, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.06743437796831131, | |
| "rewards/margins": 0.048164453357458115, | |
| "rewards/rejected": -0.11559884250164032, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 3.686053277086401e-06, | |
| "logits/chosen": 0.8000567555427551, | |
| "logits/rejected": 0.8884351849555969, | |
| "logps/chosen": -268.4157409667969, | |
| "logps/rejected": -247.18115234375, | |
| "loss": 0.6696, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.07314668595790863, | |
| "rewards/margins": 0.049585141241550446, | |
| "rewards/rejected": -0.12273182719945908, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.604695382782159e-06, | |
| "logits/chosen": 0.8107595443725586, | |
| "logits/rejected": 0.8327080607414246, | |
| "logps/chosen": -284.17218017578125, | |
| "logps/rejected": -263.6735534667969, | |
| "loss": 0.6671, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.0770660936832428, | |
| "rewards/margins": 0.0486028790473938, | |
| "rewards/rejected": -0.1256689727306366, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": 0.7712985873222351, | |
| "eval_logits/rejected": 0.8790242671966553, | |
| "eval_logps/chosen": -268.7867431640625, | |
| "eval_logps/rejected": -245.51242065429688, | |
| "eval_loss": 0.6722846627235413, | |
| "eval_rewards/accuracies": 0.6639999747276306, | |
| "eval_rewards/chosen": -0.0854262262582779, | |
| "eval_rewards/margins": 0.055336710065603256, | |
| "eval_rewards/rejected": -0.14076292514801025, | |
| "eval_runtime": 325.7205, | |
| "eval_samples_per_second": 6.14, | |
| "eval_steps_per_second": 0.384, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.5218566107988872e-06, | |
| "logits/chosen": 0.8288625478744507, | |
| "logits/rejected": 0.8551315069198608, | |
| "logps/chosen": -276.74462890625, | |
| "logps/rejected": -247.576416015625, | |
| "loss": 0.6718, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.08322989195585251, | |
| "rewards/margins": 0.06066075712442398, | |
| "rewards/rejected": -0.1438906490802765, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 3.437648009023905e-06, | |
| "logits/chosen": 0.7773796916007996, | |
| "logits/rejected": 0.8443312644958496, | |
| "logps/chosen": -222.3345184326172, | |
| "logps/rejected": -218.86825561523438, | |
| "loss": 0.6696, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.10068164020776749, | |
| "rewards/margins": 0.04352349415421486, | |
| "rewards/rejected": -0.14420512318611145, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 3.352182461642929e-06, | |
| "logits/chosen": 0.8025181889533997, | |
| "logits/rejected": 0.8646955490112305, | |
| "logps/chosen": -243.1133270263672, | |
| "logps/rejected": -232.02047729492188, | |
| "loss": 0.6638, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.08811922371387482, | |
| "rewards/margins": 0.06207479164004326, | |
| "rewards/rejected": -0.15019401907920837, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 3.265574537815398e-06, | |
| "logits/chosen": 0.8092749714851379, | |
| "logits/rejected": 0.8513230085372925, | |
| "logps/chosen": -288.71636962890625, | |
| "logps/rejected": -256.467529296875, | |
| "loss": 0.6676, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.0855972021818161, | |
| "rewards/margins": 0.07402561604976654, | |
| "rewards/rejected": -0.15962281823158264, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 3.177940338091043e-06, | |
| "logits/chosen": 0.8230969309806824, | |
| "logits/rejected": 0.8771616816520691, | |
| "logps/chosen": -262.885009765625, | |
| "logps/rejected": -236.41366577148438, | |
| "loss": 0.6661, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.10893376171588898, | |
| "rewards/margins": 0.05591960996389389, | |
| "rewards/rejected": -0.16485336422920227, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 3.089397338773569e-06, | |
| "logits/chosen": 0.8003429174423218, | |
| "logits/rejected": 0.8373273015022278, | |
| "logps/chosen": -269.59130859375, | |
| "logps/rejected": -243.3413543701172, | |
| "loss": 0.6635, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.12066192924976349, | |
| "rewards/margins": 0.07056786119937897, | |
| "rewards/rejected": -0.19122979044914246, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 3.0000642344401115e-06, | |
| "logits/chosen": 0.7690973877906799, | |
| "logits/rejected": 0.8160909414291382, | |
| "logps/chosen": -250.65017700195312, | |
| "logps/rejected": -241.54763793945312, | |
| "loss": 0.6624, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.12458320707082748, | |
| "rewards/margins": 0.06558409333229065, | |
| "rewards/rejected": -0.19016727805137634, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 2.9100607788275547e-06, | |
| "logits/chosen": 0.8020121455192566, | |
| "logits/rejected": 0.8617356419563293, | |
| "logps/chosen": -256.21771240234375, | |
| "logps/rejected": -253.4419708251953, | |
| "loss": 0.6658, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.1322631537914276, | |
| "rewards/margins": 0.06032481789588928, | |
| "rewards/rejected": -0.1925879716873169, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "learning_rate": 2.8195076242990124e-06, | |
| "logits/chosen": 0.7743502259254456, | |
| "logits/rejected": 0.8271788358688354, | |
| "logps/chosen": -252.33804321289062, | |
| "logps/rejected": -235.423828125, | |
| "loss": 0.6624, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.14343151450157166, | |
| "rewards/margins": 0.07069016247987747, | |
| "rewards/rejected": -0.21412166953086853, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "learning_rate": 2.72852616010567e-06, | |
| "logits/chosen": 0.7551236748695374, | |
| "logits/rejected": 0.7969120144844055, | |
| "logps/chosen": -266.84375, | |
| "logps/rejected": -254.37460327148438, | |
| "loss": 0.6627, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.1373104304075241, | |
| "rewards/margins": 0.07557874917984009, | |
| "rewards/rejected": -0.212889164686203, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "eval_logits/chosen": 0.7216536998748779, | |
| "eval_logits/rejected": 0.8293780088424683, | |
| "eval_logps/chosen": -275.1849365234375, | |
| "eval_logps/rejected": -254.370361328125, | |
| "eval_loss": 0.664509654045105, | |
| "eval_rewards/accuracies": 0.6679999828338623, | |
| "eval_rewards/chosen": -0.14940780401229858, | |
| "eval_rewards/margins": 0.07993472367525101, | |
| "eval_rewards/rejected": -0.2293425351381302, | |
| "eval_runtime": 325.7214, | |
| "eval_samples_per_second": 6.14, | |
| "eval_steps_per_second": 0.384, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 2.637238349660819e-06, | |
| "logits/chosen": 0.7647446393966675, | |
| "logits/rejected": 0.8582034111022949, | |
| "logps/chosen": -251.79904174804688, | |
| "logps/rejected": -218.86587524414062, | |
| "loss": 0.665, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.16131970286369324, | |
| "rewards/margins": 0.06838177144527435, | |
| "rewards/rejected": -0.2297014743089676, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 2.5457665670441937e-06, | |
| "logits/chosen": 0.8116461038589478, | |
| "logits/rejected": 0.8353972434997559, | |
| "logps/chosen": -263.5601501464844, | |
| "logps/rejected": -247.66909790039062, | |
| "loss": 0.6591, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.14197978377342224, | |
| "rewards/margins": 0.09031648933887482, | |
| "rewards/rejected": -0.23229627311229706, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 2.4542334329558075e-06, | |
| "logits/chosen": 0.7308142781257629, | |
| "logits/rejected": 0.7632499933242798, | |
| "logps/chosen": -257.2229919433594, | |
| "logps/rejected": -250.84469604492188, | |
| "loss": 0.6607, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.15741421282291412, | |
| "rewards/margins": 0.07882841676473618, | |
| "rewards/rejected": -0.2362426221370697, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "learning_rate": 2.3627616503391813e-06, | |
| "logits/chosen": 0.7235521674156189, | |
| "logits/rejected": 0.7589094042778015, | |
| "logps/chosen": -274.9698791503906, | |
| "logps/rejected": -238.1727752685547, | |
| "loss": 0.6614, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.16767871379852295, | |
| "rewards/margins": 0.07072968035936356, | |
| "rewards/rejected": -0.2384084016084671, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "learning_rate": 2.271473839894331e-06, | |
| "logits/chosen": 0.7224727869033813, | |
| "logits/rejected": 0.7499375939369202, | |
| "logps/chosen": -283.6424560546875, | |
| "logps/rejected": -272.5013122558594, | |
| "loss": 0.6593, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.17033258080482483, | |
| "rewards/margins": 0.07838577777147293, | |
| "rewards/rejected": -0.24871835112571716, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 2.1804923757009885e-06, | |
| "logits/chosen": 0.6960411667823792, | |
| "logits/rejected": 0.7274879813194275, | |
| "logps/chosen": -269.6304016113281, | |
| "logps/rejected": -250.3975830078125, | |
| "loss": 0.6614, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.18492145836353302, | |
| "rewards/margins": 0.07485240697860718, | |
| "rewards/rejected": -0.259773850440979, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 2.089939221172446e-06, | |
| "logits/chosen": 0.7146167755126953, | |
| "logits/rejected": 0.7337481379508972, | |
| "logps/chosen": -288.8470153808594, | |
| "logps/rejected": -258.131591796875, | |
| "loss": 0.664, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.19552810490131378, | |
| "rewards/margins": 0.09874384850263596, | |
| "rewards/rejected": -0.29427194595336914, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 1.9999357655598894e-06, | |
| "logits/chosen": 0.7304507493972778, | |
| "logits/rejected": 0.7039176225662231, | |
| "logps/chosen": -259.52008056640625, | |
| "logps/rejected": -250.97146606445312, | |
| "loss": 0.6591, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.19664093852043152, | |
| "rewards/margins": 0.058397077023983, | |
| "rewards/rejected": -0.2550380527973175, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "learning_rate": 1.9106026612264316e-06, | |
| "logits/chosen": 0.7533366084098816, | |
| "logits/rejected": 0.8193215131759644, | |
| "logps/chosen": -236.71243286132812, | |
| "logps/rejected": -238.93679809570312, | |
| "loss": 0.653, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.1914629340171814, | |
| "rewards/margins": 0.07271396368741989, | |
| "rewards/rejected": -0.2641769051551819, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "learning_rate": 1.8220596619089576e-06, | |
| "logits/chosen": 0.7307643890380859, | |
| "logits/rejected": 0.7107076048851013, | |
| "logps/chosen": -281.09710693359375, | |
| "logps/rejected": -258.6228942871094, | |
| "loss": 0.6476, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.19486048817634583, | |
| "rewards/margins": 0.10412123054265976, | |
| "rewards/rejected": -0.2989817261695862, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "eval_logits/chosen": 0.6828243136405945, | |
| "eval_logits/rejected": 0.7883425354957581, | |
| "eval_logps/chosen": -280.03369140625, | |
| "eval_logps/rejected": -261.11236572265625, | |
| "eval_loss": 0.659087061882019, | |
| "eval_rewards/accuracies": 0.6639999747276306, | |
| "eval_rewards/chosen": -0.1978950798511505, | |
| "eval_rewards/margins": 0.09886746108531952, | |
| "eval_rewards/rejected": -0.2967625558376312, | |
| "eval_runtime": 325.9412, | |
| "eval_samples_per_second": 6.136, | |
| "eval_steps_per_second": 0.384, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 1.7344254621846018e-06, | |
| "logits/chosen": 0.700169026851654, | |
| "logits/rejected": 0.7478510141372681, | |
| "logps/chosen": -283.26641845703125, | |
| "logps/rejected": -278.6813049316406, | |
| "loss": 0.6514, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.191552072763443, | |
| "rewards/margins": 0.09749529510736465, | |
| "rewards/rejected": -0.28904733061790466, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 1.647817538357072e-06, | |
| "logits/chosen": 0.7241575121879578, | |
| "logits/rejected": 0.7943159341812134, | |
| "logps/chosen": -267.4305725097656, | |
| "logps/rejected": -256.3984069824219, | |
| "loss": 0.6591, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.2104833871126175, | |
| "rewards/margins": 0.07614756375551224, | |
| "rewards/rejected": -0.2866309583187103, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 1.5623519909760953e-06, | |
| "logits/chosen": 0.6998089551925659, | |
| "logits/rejected": 0.7587723731994629, | |
| "logps/chosen": -264.82110595703125, | |
| "logps/rejected": -267.1747741699219, | |
| "loss": 0.6551, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.22911302745342255, | |
| "rewards/margins": 0.07119645923376083, | |
| "rewards/rejected": -0.30030950903892517, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "learning_rate": 1.4781433892011132e-06, | |
| "logits/chosen": 0.7701424360275269, | |
| "logits/rejected": 0.7888699173927307, | |
| "logps/chosen": -262.2229309082031, | |
| "logps/rejected": -286.0958557128906, | |
| "loss": 0.658, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.2087526023387909, | |
| "rewards/margins": 0.07823307812213898, | |
| "rewards/rejected": -0.28698569536209106, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "learning_rate": 1.3953046172178413e-06, | |
| "logits/chosen": 0.7079032063484192, | |
| "logits/rejected": 0.7274273037910461, | |
| "logps/chosen": -276.0498352050781, | |
| "logps/rejected": -277.2353820800781, | |
| "loss": 0.6536, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.23646607995033264, | |
| "rewards/margins": 0.09612749516963959, | |
| "rewards/rejected": -0.33259356021881104, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "learning_rate": 1.3139467229135999e-06, | |
| "logits/chosen": 0.6826360821723938, | |
| "logits/rejected": 0.751067578792572, | |
| "logps/chosen": -241.80044555664062, | |
| "logps/rejected": -275.70989990234375, | |
| "loss": 0.6556, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.22094163298606873, | |
| "rewards/margins": 0.08276458084583282, | |
| "rewards/rejected": -0.30370622873306274, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 1.2341787690142436e-06, | |
| "logits/chosen": 0.6527940630912781, | |
| "logits/rejected": 0.7365330457687378, | |
| "logps/chosen": -320.76385498046875, | |
| "logps/rejected": -271.78790283203125, | |
| "loss": 0.6552, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.20969745516777039, | |
| "rewards/margins": 0.11965203285217285, | |
| "rewards/rejected": -0.32934948801994324, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "learning_rate": 1.1561076868822756e-06, | |
| "logits/chosen": 0.6766701936721802, | |
| "logits/rejected": 0.7141178250312805, | |
| "logps/chosen": -265.7345886230469, | |
| "logps/rejected": -245.7202606201172, | |
| "loss": 0.6518, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.22580020129680634, | |
| "rewards/margins": 0.0855962485074997, | |
| "rewards/rejected": -0.31139642000198364, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "learning_rate": 1.079838133172111e-06, | |
| "logits/chosen": 0.7425110936164856, | |
| "logits/rejected": 0.7384303212165833, | |
| "logps/chosen": -278.8888854980469, | |
| "logps/rejected": -266.8258972167969, | |
| "loss": 0.6598, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.21307018399238586, | |
| "rewards/margins": 0.08619161695241928, | |
| "rewards/rejected": -0.29926180839538574, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "learning_rate": 1.0054723495346484e-06, | |
| "logits/chosen": 0.6221153736114502, | |
| "logits/rejected": 0.6998614072799683, | |
| "logps/chosen": -280.8089599609375, | |
| "logps/rejected": -257.3545837402344, | |
| "loss": 0.6488, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.22230124473571777, | |
| "rewards/margins": 0.09477819502353668, | |
| "rewards/rejected": -0.31707945466041565, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "eval_logits/chosen": 0.651095986366272, | |
| "eval_logits/rejected": 0.7549068927764893, | |
| "eval_logps/chosen": -283.3440246582031, | |
| "eval_logps/rejected": -265.57830810546875, | |
| "eval_loss": 0.6559058427810669, | |
| "eval_rewards/accuracies": 0.6620000004768372, | |
| "eval_rewards/chosen": -0.23099879920482635, | |
| "eval_rewards/margins": 0.11042327433824539, | |
| "eval_rewards/rejected": -0.34142205119132996, | |
| "eval_runtime": 325.7526, | |
| "eval_samples_per_second": 6.14, | |
| "eval_steps_per_second": 0.384, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "learning_rate": 9.331100255592437e-07, | |
| "logits/chosen": 0.7064228653907776, | |
| "logits/rejected": 0.7716813683509827, | |
| "logps/chosen": -300.17694091796875, | |
| "logps/rejected": -261.4855651855469, | |
| "loss": 0.6475, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.2144012451171875, | |
| "rewards/margins": 0.12820136547088623, | |
| "rewards/rejected": -0.34260261058807373, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 8.628481651367876e-07, | |
| "logits/chosen": 0.6913928985595703, | |
| "logits/rejected": 0.7301486730575562, | |
| "logps/chosen": -286.8462829589844, | |
| "logps/rejected": -276.6960754394531, | |
| "loss": 0.6457, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.23222024738788605, | |
| "rewards/margins": 0.10797496885061264, | |
| "rewards/rejected": -0.3401952087879181, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "learning_rate": 7.947809564230446e-07, | |
| "logits/chosen": 0.7156798839569092, | |
| "logits/rejected": 0.7468422055244446, | |
| "logps/chosen": -288.5824890136719, | |
| "logps/rejected": -286.32293701171875, | |
| "loss": 0.6439, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.23393268883228302, | |
| "rewards/margins": 0.12296883016824722, | |
| "rewards/rejected": -0.35690149664878845, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "learning_rate": 7.289996455765749e-07, | |
| "logits/chosen": 0.668700098991394, | |
| "logits/rejected": 0.6845619678497314, | |
| "logps/chosen": -299.6327819824219, | |
| "logps/rejected": -283.09295654296875, | |
| "loss": 0.6584, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.2437756061553955, | |
| "rewards/margins": 0.07863186299800873, | |
| "rewards/rejected": -0.32240745425224304, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "learning_rate": 6.655924144404907e-07, | |
| "logits/chosen": 0.6234613656997681, | |
| "logits/rejected": 0.6867505311965942, | |
| "logps/chosen": -281.64007568359375, | |
| "logps/rejected": -268.5241394042969, | |
| "loss": 0.6515, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.21235065162181854, | |
| "rewards/margins": 0.12807723879814148, | |
| "rewards/rejected": -0.34042784571647644, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "learning_rate": 6.046442623320145e-07, | |
| "logits/chosen": 0.7112501859664917, | |
| "logits/rejected": 0.7498377561569214, | |
| "logps/chosen": -281.40875244140625, | |
| "logps/rejected": -251.820068359375, | |
| "loss": 0.6496, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.26181793212890625, | |
| "rewards/margins": 0.08466657996177673, | |
| "rewards/rejected": -0.3464844822883606, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 5.462368920983249e-07, | |
| "logits/chosen": 0.6798941493034363, | |
| "logits/rejected": 0.7346147298812866, | |
| "logps/chosen": -266.92315673828125, | |
| "logps/rejected": -247.5948028564453, | |
| "loss": 0.6566, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.25860121846199036, | |
| "rewards/margins": 0.05472899600863457, | |
| "rewards/rejected": -0.31333020329475403, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "learning_rate": 4.904486005914027e-07, | |
| "logits/chosen": 0.6729373931884766, | |
| "logits/rejected": 0.7290818095207214, | |
| "logps/chosen": -307.30926513671875, | |
| "logps/rejected": -292.6742248535156, | |
| "loss": 0.6463, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.2319561243057251, | |
| "rewards/margins": 0.1268167346715927, | |
| "rewards/rejected": -0.3587728440761566, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 4.373541737087264e-07, | |
| "logits/chosen": 0.6597692966461182, | |
| "logits/rejected": 0.7631260752677917, | |
| "logps/chosen": -288.7844543457031, | |
| "logps/rejected": -258.93450927734375, | |
| "loss": 0.6505, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.2365461140871048, | |
| "rewards/margins": 0.12159979343414307, | |
| "rewards/rejected": -0.35814589262008667, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "learning_rate": 3.8702478614051353e-07, | |
| "logits/chosen": 0.6355623006820679, | |
| "logits/rejected": 0.64263516664505, | |
| "logps/chosen": -261.3099670410156, | |
| "logps/rejected": -265.4583435058594, | |
| "loss": 0.6449, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.23573680222034454, | |
| "rewards/margins": 0.13916215300559998, | |
| "rewards/rejected": -0.3748989403247833, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "eval_logits/chosen": 0.6347466111183167, | |
| "eval_logits/rejected": 0.7372007966041565, | |
| "eval_logps/chosen": -285.4225769042969, | |
| "eval_logps/rejected": -268.3813781738281, | |
| "eval_loss": 0.6541773676872253, | |
| "eval_rewards/accuracies": 0.656000018119812, | |
| "eval_rewards/chosen": -0.2517840266227722, | |
| "eval_rewards/margins": 0.11766859143972397, | |
| "eval_rewards/rejected": -0.3694525957107544, | |
| "eval_runtime": 325.7549, | |
| "eval_samples_per_second": 6.14, | |
| "eval_steps_per_second": 0.384, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "learning_rate": 3.3952790595787986e-07, | |
| "logits/chosen": 0.7447446584701538, | |
| "logits/rejected": 0.7175893783569336, | |
| "logps/chosen": -275.09906005859375, | |
| "logps/rejected": -265.5423889160156, | |
| "loss": 0.6508, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.25975173711776733, | |
| "rewards/margins": 0.08070627599954605, | |
| "rewards/rejected": -0.3404580354690552, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "learning_rate": 2.9492720416985004e-07, | |
| "logits/chosen": 0.5959675908088684, | |
| "logits/rejected": 0.6474281549453735, | |
| "logps/chosen": -294.9203186035156, | |
| "logps/rejected": -284.7016906738281, | |
| "loss": 0.6455, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.2046935260295868, | |
| "rewards/margins": 0.13686785101890564, | |
| "rewards/rejected": -0.34156137704849243, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 2.5328246937043526e-07, | |
| "logits/chosen": 0.6916497349739075, | |
| "logits/rejected": 0.7139278650283813, | |
| "logps/chosen": -264.2368469238281, | |
| "logps/rejected": -253.6294403076172, | |
| "loss": 0.6506, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.28065750002861023, | |
| "rewards/margins": 0.10091479122638702, | |
| "rewards/rejected": -0.38157230615615845, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "learning_rate": 2.1464952759020857e-07, | |
| "logits/chosen": 0.6801538467407227, | |
| "logits/rejected": 0.7134484648704529, | |
| "logps/chosen": -274.87481689453125, | |
| "logps/rejected": -282.79766845703125, | |
| "loss": 0.6522, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.2523757517337799, | |
| "rewards/margins": 0.11816386878490448, | |
| "rewards/rejected": -0.3705395758152008, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "learning_rate": 1.790801674598186e-07, | |
| "logits/chosen": 0.6462175846099854, | |
| "logits/rejected": 0.6658031940460205, | |
| "logps/chosen": -292.3714599609375, | |
| "logps/rejected": -269.7221984863281, | |
| "loss": 0.6522, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.24048607051372528, | |
| "rewards/margins": 0.10906896740198135, | |
| "rewards/rejected": -0.34955504536628723, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 1.4662207078575685e-07, | |
| "logits/chosen": 0.6442946791648865, | |
| "logits/rejected": 0.7183449864387512, | |
| "logps/chosen": -256.558837890625, | |
| "logps/rejected": -274.9573669433594, | |
| "loss": 0.6476, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.2482720911502838, | |
| "rewards/margins": 0.1144469827413559, | |
| "rewards/rejected": -0.3627190887928009, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 1.1731874863145143e-07, | |
| "logits/chosen": 0.68730229139328, | |
| "logits/rejected": 0.7417377829551697, | |
| "logps/chosen": -285.8447570800781, | |
| "logps/rejected": -258.12408447265625, | |
| "loss": 0.6483, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.24029162526130676, | |
| "rewards/margins": 0.10951662063598633, | |
| "rewards/rejected": -0.3498082160949707, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "learning_rate": 9.120948298936422e-08, | |
| "logits/chosen": 0.7039914131164551, | |
| "logits/rejected": 0.720660388469696, | |
| "logps/chosen": -275.22564697265625, | |
| "logps/rejected": -259.50921630859375, | |
| "loss": 0.6391, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.2585623860359192, | |
| "rewards/margins": 0.10217638313770294, | |
| "rewards/rejected": -0.36073875427246094, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 6.832927412229017e-08, | |
| "logits/chosen": 0.634792685508728, | |
| "logits/rejected": 0.7409111857414246, | |
| "logps/chosen": -285.78076171875, | |
| "logps/rejected": -263.0660095214844, | |
| "loss": 0.6441, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.24097540974617004, | |
| "rewards/margins": 0.12255308777093887, | |
| "rewards/rejected": -0.3635285198688507, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "learning_rate": 4.870879364444109e-08, | |
| "logits/chosen": 0.6492586731910706, | |
| "logits/rejected": 0.7190669775009155, | |
| "logps/chosen": -275.6029357910156, | |
| "logps/rejected": -243.29281616210938, | |
| "loss": 0.6487, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.24271516501903534, | |
| "rewards/margins": 0.14071747660636902, | |
| "rewards/rejected": -0.38343262672424316, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "eval_logits/chosen": 0.6298695802688599, | |
| "eval_logits/rejected": 0.7320161461830139, | |
| "eval_logps/chosen": -285.9532165527344, | |
| "eval_logps/rejected": -269.07244873046875, | |
| "eval_loss": 0.6538792848587036, | |
| "eval_rewards/accuracies": 0.656000018119812, | |
| "eval_rewards/chosen": -0.25709065794944763, | |
| "eval_rewards/margins": 0.11927253752946854, | |
| "eval_rewards/rejected": -0.37636318802833557, | |
| "eval_runtime": 326.0646, | |
| "eval_samples_per_second": 6.134, | |
| "eval_steps_per_second": 0.383, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "learning_rate": 3.237434340521789e-08, | |
| "logits/chosen": 0.6427167654037476, | |
| "logits/rejected": 0.7207670211791992, | |
| "logps/chosen": -286.7334289550781, | |
| "logps/rejected": -282.4278259277344, | |
| "loss": 0.6455, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.24073800444602966, | |
| "rewards/margins": 0.1223975196480751, | |
| "rewards/rejected": -0.36313554644584656, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "learning_rate": 1.93478202307823e-08, | |
| "logits/chosen": 0.6422888040542603, | |
| "logits/rejected": 0.6643663644790649, | |
| "logps/chosen": -267.14398193359375, | |
| "logps/rejected": -278.8194274902344, | |
| "loss": 0.6473, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.257813036441803, | |
| "rewards/margins": 0.0857832059264183, | |
| "rewards/rejected": -0.34359627962112427, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "learning_rate": 9.646686570697062e-09, | |
| "logits/chosen": 0.6971439123153687, | |
| "logits/rejected": 0.722652792930603, | |
| "logps/chosen": -283.13067626953125, | |
| "logps/rejected": -285.1507568359375, | |
| "loss": 0.6572, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.2655975818634033, | |
| "rewards/margins": 0.10348521173000336, | |
| "rewards/rejected": -0.36908283829689026, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 3.283947088983663e-09, | |
| "logits/chosen": 0.6804937720298767, | |
| "logits/rejected": 0.6877144575119019, | |
| "logps/chosen": -265.1544494628906, | |
| "logps/rejected": -278.4422302246094, | |
| "loss": 0.6544, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.27914106845855713, | |
| "rewards/margins": 0.08945213258266449, | |
| "rewards/rejected": -0.3685932457447052, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 1.99, | |
| "learning_rate": 2.681312309735229e-10, | |
| "logits/chosen": 0.6534776091575623, | |
| "logits/rejected": 0.7444522976875305, | |
| "logps/chosen": -257.7799377441406, | |
| "logps/rejected": -263.00030517578125, | |
| "loss": 0.654, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.2684522867202759, | |
| "rewards/margins": 0.08305440098047256, | |
| "rewards/rejected": -0.35150665044784546, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 954, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6680307045922589, | |
| "train_runtime": 18174.3674, | |
| "train_samples_per_second": 3.364, | |
| "train_steps_per_second": 0.052 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 954, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |