| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9996020692399522, | |
| "eval_steps": 100, | |
| "global_step": 1256, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 3.968253968253968e-09, | |
| "logits/chosen": -2.7193620204925537, | |
| "logits/rejected": -2.698728084564209, | |
| "logps/chosen": -182.0961456298828, | |
| "logps/rejected": -172.47128295898438, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 3.968253968253968e-08, | |
| "logits/chosen": -2.7041964530944824, | |
| "logits/rejected": -2.6794540882110596, | |
| "logps/chosen": -162.45831298828125, | |
| "logps/rejected": -140.5693359375, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.5486111044883728, | |
| "rewards/chosen": 0.00032037965138442814, | |
| "rewards/margins": 0.0004935775068588555, | |
| "rewards/rejected": -0.00017319784092251211, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 7.936507936507936e-08, | |
| "logits/chosen": -2.7177577018737793, | |
| "logits/rejected": -2.7136425971984863, | |
| "logps/chosen": -134.47242736816406, | |
| "logps/rejected": -143.55604553222656, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 8.780837379163131e-05, | |
| "rewards/margins": 0.00010721785656642169, | |
| "rewards/rejected": -1.940951551659964e-05, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.1904761904761903e-07, | |
| "logits/chosen": -2.6898293495178223, | |
| "logits/rejected": -2.676154613494873, | |
| "logps/chosen": -140.94692993164062, | |
| "logps/rejected": -136.50369262695312, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": 0.0005466601578518748, | |
| "rewards/margins": -0.00021456097601912916, | |
| "rewards/rejected": 0.0007612211629748344, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.5873015873015872e-07, | |
| "logits/chosen": -2.6958394050598145, | |
| "logits/rejected": -2.686532974243164, | |
| "logps/chosen": -134.98963928222656, | |
| "logps/rejected": -144.46652221679688, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": 0.0015748919686302543, | |
| "rewards/margins": 0.0009769219905138016, | |
| "rewards/rejected": 0.0005979698617011309, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 1.984126984126984e-07, | |
| "logits/chosen": -2.7042899131774902, | |
| "logits/rejected": -2.6861345767974854, | |
| "logps/chosen": -149.71768188476562, | |
| "logps/rejected": -145.0757293701172, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.005199921317398548, | |
| "rewards/margins": 0.0022330707870423794, | |
| "rewards/rejected": 0.0029668500646948814, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.3809523809523806e-07, | |
| "logits/chosen": -2.705153703689575, | |
| "logits/rejected": -2.685439348220825, | |
| "logps/chosen": -154.3783416748047, | |
| "logps/rejected": -151.54519653320312, | |
| "loss": 0.6912, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 0.00569504126906395, | |
| "rewards/margins": 0.0022000311873853207, | |
| "rewards/rejected": 0.003495010081678629, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 2.7777777777777776e-07, | |
| "logits/chosen": -2.7017154693603516, | |
| "logits/rejected": -2.6924962997436523, | |
| "logps/chosen": -146.3284149169922, | |
| "logps/rejected": -138.79405212402344, | |
| "loss": 0.6885, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.010588793084025383, | |
| "rewards/margins": 0.010192448273301125, | |
| "rewards/rejected": 0.00039634370477870107, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.1746031746031743e-07, | |
| "logits/chosen": -2.7155232429504395, | |
| "logits/rejected": -2.696071147918701, | |
| "logps/chosen": -141.80067443847656, | |
| "logps/rejected": -147.0068817138672, | |
| "loss": 0.6867, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": 0.0049073463305830956, | |
| "rewards/margins": 0.013599385507404804, | |
| "rewards/rejected": -0.008692039176821709, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 3.5714285714285716e-07, | |
| "logits/chosen": -2.7175304889678955, | |
| "logits/rejected": -2.7080624103546143, | |
| "logps/chosen": -153.12509155273438, | |
| "logps/rejected": -146.53590393066406, | |
| "loss": 0.6847, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.028871387243270874, | |
| "rewards/margins": 0.017175236716866493, | |
| "rewards/rejected": -0.046046625822782516, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 3.968253968253968e-07, | |
| "logits/chosen": -2.7524733543395996, | |
| "logits/rejected": -2.7452526092529297, | |
| "logps/chosen": -163.88070678710938, | |
| "logps/rejected": -163.61032104492188, | |
| "loss": 0.6789, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.0662173330783844, | |
| "rewards/margins": 0.02977912127971649, | |
| "rewards/rejected": -0.09599645435810089, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "eval_logits/chosen": -2.7336502075195312, | |
| "eval_logits/rejected": -2.7255024909973145, | |
| "eval_logps/chosen": -155.19271850585938, | |
| "eval_logps/rejected": -165.35523986816406, | |
| "eval_loss": 0.6769910454750061, | |
| "eval_rewards/accuracies": 0.5914179086685181, | |
| "eval_rewards/chosen": -0.10619194805622101, | |
| "eval_rewards/margins": 0.03601696714758873, | |
| "eval_rewards/rejected": -0.14220890402793884, | |
| "eval_runtime": 184.3291, | |
| "eval_samples_per_second": 46.395, | |
| "eval_steps_per_second": 0.727, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 4.365079365079365e-07, | |
| "logits/chosen": -2.738532543182373, | |
| "logits/rejected": -2.7273170948028564, | |
| "logps/chosen": -164.2928009033203, | |
| "logps/rejected": -160.19398498535156, | |
| "loss": 0.6738, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.16211798787117004, | |
| "rewards/margins": 0.03163355216383934, | |
| "rewards/rejected": -0.19375154376029968, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.761904761904761e-07, | |
| "logits/chosen": -2.7289297580718994, | |
| "logits/rejected": -2.705962657928467, | |
| "logps/chosen": -196.69662475585938, | |
| "logps/rejected": -197.2833251953125, | |
| "loss": 0.661, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.2917623221874237, | |
| "rewards/margins": 0.08966299891471863, | |
| "rewards/rejected": -0.38142532110214233, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.999845414634076e-07, | |
| "logits/chosen": -2.658005475997925, | |
| "logits/rejected": -2.6317684650421143, | |
| "logps/chosen": -187.4532928466797, | |
| "logps/rejected": -188.37689208984375, | |
| "loss": 0.6542, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.3956056833267212, | |
| "rewards/margins": 0.12092368304729462, | |
| "rewards/rejected": -0.5165294408798218, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 4.998106548810311e-07, | |
| "logits/chosen": -2.6906683444976807, | |
| "logits/rejected": -2.6913747787475586, | |
| "logps/chosen": -199.67568969726562, | |
| "logps/rejected": -253.02487182617188, | |
| "loss": 0.6171, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.4227059781551361, | |
| "rewards/margins": 0.27536457777023315, | |
| "rewards/rejected": -0.6980706453323364, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 4.994436933879359e-07, | |
| "logits/chosen": -2.6662166118621826, | |
| "logits/rejected": -2.644784927368164, | |
| "logps/chosen": -197.07180786132812, | |
| "logps/rejected": -198.4012908935547, | |
| "loss": 0.6395, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.3597154915332794, | |
| "rewards/margins": 0.13716872036457062, | |
| "rewards/rejected": -0.49688419699668884, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.988839406031596e-07, | |
| "logits/chosen": -2.647681474685669, | |
| "logits/rejected": -2.6395888328552246, | |
| "logps/chosen": -182.04420471191406, | |
| "logps/rejected": -206.59780883789062, | |
| "loss": 0.629, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.3573477864265442, | |
| "rewards/margins": 0.2222837507724762, | |
| "rewards/rejected": -0.579631507396698, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 4.981318291512395e-07, | |
| "logits/chosen": -2.619232654571533, | |
| "logits/rejected": -2.598362684249878, | |
| "logps/chosen": -227.0933380126953, | |
| "logps/rejected": -230.9747772216797, | |
| "loss": 0.6242, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.7434185743331909, | |
| "rewards/margins": 0.21749505400657654, | |
| "rewards/rejected": -0.9609137773513794, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 4.971879403278432e-07, | |
| "logits/chosen": -2.5654754638671875, | |
| "logits/rejected": -2.5364232063293457, | |
| "logps/chosen": -241.6617431640625, | |
| "logps/rejected": -245.66268920898438, | |
| "loss": 0.6151, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.7342535257339478, | |
| "rewards/margins": 0.23685339093208313, | |
| "rewards/rejected": -0.9711068868637085, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.960530036504941e-07, | |
| "logits/chosen": -2.5271048545837402, | |
| "logits/rejected": -2.486818790435791, | |
| "logps/chosen": -235.6089630126953, | |
| "logps/rejected": -251.17758178710938, | |
| "loss": 0.6215, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.807177722454071, | |
| "rewards/margins": 0.28561535477638245, | |
| "rewards/rejected": -1.0927931070327759, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.947278962947386e-07, | |
| "logits/chosen": -2.4217896461486816, | |
| "logits/rejected": -2.413295269012451, | |
| "logps/chosen": -251.0736083984375, | |
| "logps/rejected": -268.6098937988281, | |
| "loss": 0.6062, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.086307406425476, | |
| "rewards/margins": 0.24874301254749298, | |
| "rewards/rejected": -1.3350504636764526, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "eval_logits/chosen": -2.3855514526367188, | |
| "eval_logits/rejected": -2.369593858718872, | |
| "eval_logps/chosen": -246.6970672607422, | |
| "eval_logps/rejected": -289.8621826171875, | |
| "eval_loss": 0.6079375743865967, | |
| "eval_rewards/accuracies": 0.66697758436203, | |
| "eval_rewards/chosen": -1.021235704421997, | |
| "eval_rewards/margins": 0.3660426437854767, | |
| "eval_rewards/rejected": -1.3872781991958618, | |
| "eval_runtime": 184.172, | |
| "eval_samples_per_second": 46.435, | |
| "eval_steps_per_second": 0.728, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.932136424161899e-07, | |
| "logits/chosen": -2.3366785049438477, | |
| "logits/rejected": -2.3228511810302734, | |
| "logps/chosen": -266.292236328125, | |
| "logps/rejected": -300.22894287109375, | |
| "loss": 0.5893, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -1.2130026817321777, | |
| "rewards/margins": 0.3487839996814728, | |
| "rewards/rejected": -1.5617868900299072, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.915114123589732e-07, | |
| "logits/chosen": -2.321228504180908, | |
| "logits/rejected": -2.3033699989318848, | |
| "logps/chosen": -336.34161376953125, | |
| "logps/rejected": -373.39935302734375, | |
| "loss": 0.612, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.9529145956039429, | |
| "rewards/margins": 0.2863468527793884, | |
| "rewards/rejected": -2.2392613887786865, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.896225217511849e-07, | |
| "logits/chosen": -2.4310107231140137, | |
| "logits/rejected": -2.422048568725586, | |
| "logps/chosen": -291.1025695800781, | |
| "logps/rejected": -328.18963623046875, | |
| "loss": 0.6079, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.4314143657684326, | |
| "rewards/margins": 0.3364001214504242, | |
| "rewards/rejected": -1.7678143978118896, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.875484304880629e-07, | |
| "logits/chosen": -2.3412394523620605, | |
| "logits/rejected": -2.309183120727539, | |
| "logps/chosen": -280.8785705566406, | |
| "logps/rejected": -308.54132080078125, | |
| "loss": 0.613, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -1.309309482574463, | |
| "rewards/margins": 0.3731766939163208, | |
| "rewards/rejected": -1.6824861764907837, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.852907416036558e-07, | |
| "logits/chosen": -2.415271282196045, | |
| "logits/rejected": -2.4072234630584717, | |
| "logps/chosen": -243.56332397460938, | |
| "logps/rejected": -298.7532043457031, | |
| "loss": 0.591, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.967076301574707, | |
| "rewards/margins": 0.4581146240234375, | |
| "rewards/rejected": -1.4251911640167236, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.828512000318616e-07, | |
| "logits/chosen": -2.3924427032470703, | |
| "logits/rejected": -2.3613152503967285, | |
| "logps/chosen": -266.86572265625, | |
| "logps/rejected": -304.2983093261719, | |
| "loss": 0.5986, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -1.2214807271957397, | |
| "rewards/margins": 0.4553411602973938, | |
| "rewards/rejected": -1.6768219470977783, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.802316912577946e-07, | |
| "logits/chosen": -2.4108529090881348, | |
| "logits/rejected": -2.3902478218078613, | |
| "logps/chosen": -252.7959442138672, | |
| "logps/rejected": -295.266357421875, | |
| "loss": 0.5917, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.0824626684188843, | |
| "rewards/margins": 0.39643940329551697, | |
| "rewards/rejected": -1.4789022207260132, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.774342398605221e-07, | |
| "logits/chosen": -2.3505263328552246, | |
| "logits/rejected": -2.2942967414855957, | |
| "logps/chosen": -279.871337890625, | |
| "logps/rejected": -300.4220886230469, | |
| "loss": 0.5979, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.1787078380584717, | |
| "rewards/margins": 0.430286169052124, | |
| "rewards/rejected": -1.6089938879013062, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.744610079482978e-07, | |
| "logits/chosen": -2.3269264698028564, | |
| "logits/rejected": -2.2910802364349365, | |
| "logps/chosen": -255.27706909179688, | |
| "logps/rejected": -281.60137939453125, | |
| "loss": 0.5853, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.1753785610198975, | |
| "rewards/margins": 0.3495523929595947, | |
| "rewards/rejected": -1.5249310731887817, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.713142934875005e-07, | |
| "logits/chosen": -2.2868428230285645, | |
| "logits/rejected": -2.2631592750549316, | |
| "logps/chosen": -284.2200012207031, | |
| "logps/rejected": -322.45269775390625, | |
| "loss": 0.5965, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.336501955986023, | |
| "rewards/margins": 0.3968800902366638, | |
| "rewards/rejected": -1.733382225036621, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "eval_logits/chosen": -2.265592098236084, | |
| "eval_logits/rejected": -2.244987964630127, | |
| "eval_logps/chosen": -282.3620910644531, | |
| "eval_logps/rejected": -331.2099609375, | |
| "eval_loss": 0.5907339453697205, | |
| "eval_rewards/accuracies": 0.6623134613037109, | |
| "eval_rewards/chosen": -1.3778856992721558, | |
| "eval_rewards/margins": 0.42287060618400574, | |
| "eval_rewards/rejected": -1.8007562160491943, | |
| "eval_runtime": 184.1739, | |
| "eval_samples_per_second": 46.434, | |
| "eval_steps_per_second": 0.728, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.679965285265706e-07, | |
| "logits/chosen": -2.2354235649108887, | |
| "logits/rejected": -2.23685884475708, | |
| "logps/chosen": -277.09283447265625, | |
| "logps/rejected": -347.7145080566406, | |
| "loss": 0.5612, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.3514426946640015, | |
| "rewards/margins": 0.4907970428466797, | |
| "rewards/rejected": -1.8422397375106812, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.64510277316316e-07, | |
| "logits/chosen": -2.2262344360351562, | |
| "logits/rejected": -2.226029634475708, | |
| "logps/chosen": -271.74212646484375, | |
| "logps/rejected": -332.5010986328125, | |
| "loss": 0.5903, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -1.3337775468826294, | |
| "rewards/margins": 0.39512914419174194, | |
| "rewards/rejected": -1.7289068698883057, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.6085823432804137e-07, | |
| "logits/chosen": -2.2451891899108887, | |
| "logits/rejected": -2.2502384185791016, | |
| "logps/chosen": -250.6347198486328, | |
| "logps/rejected": -333.8939208984375, | |
| "loss": 0.5722, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.1326004266738892, | |
| "rewards/margins": 0.5066065192222595, | |
| "rewards/rejected": -1.639206886291504, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.570432221710314e-07, | |
| "logits/chosen": -2.0656931400299072, | |
| "logits/rejected": -2.0213730335235596, | |
| "logps/chosen": -318.232177734375, | |
| "logps/rejected": -369.13311767578125, | |
| "loss": 0.5766, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.584176778793335, | |
| "rewards/margins": 0.5901076197624207, | |
| "rewards/rejected": -2.1742844581604004, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 4.5306818941099866e-07, | |
| "logits/chosen": -1.9084612131118774, | |
| "logits/rejected": -1.8514792919158936, | |
| "logps/chosen": -316.9821472167969, | |
| "logps/rejected": -352.9412841796875, | |
| "loss": 0.5825, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -1.5627154111862183, | |
| "rewards/margins": 0.5152220726013184, | |
| "rewards/rejected": -2.077937364578247, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.4893620829118124e-07, | |
| "logits/chosen": -1.8860156536102295, | |
| "logits/rejected": -1.8301204442977905, | |
| "logps/chosen": -309.8200378417969, | |
| "logps/rejected": -362.0408935546875, | |
| "loss": 0.5755, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.5869390964508057, | |
| "rewards/margins": 0.49348369240760803, | |
| "rewards/rejected": -2.080422878265381, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.4465047235785185e-07, | |
| "logits/chosen": -1.6610889434814453, | |
| "logits/rejected": -1.585129737854004, | |
| "logps/chosen": -321.8608703613281, | |
| "logps/rejected": -380.31036376953125, | |
| "loss": 0.5697, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.774713158607483, | |
| "rewards/margins": 0.6593302488327026, | |
| "rewards/rejected": -2.4340434074401855, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.40214293992074e-07, | |
| "logits/chosen": -1.385825753211975, | |
| "logits/rejected": -1.31913161277771, | |
| "logps/chosen": -377.07269287109375, | |
| "logps/rejected": -459.5557556152344, | |
| "loss": 0.5818, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -2.1890993118286133, | |
| "rewards/margins": 0.7521292567253113, | |
| "rewards/rejected": -2.9412286281585693, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.3563110184961234e-07, | |
| "logits/chosen": -1.5089499950408936, | |
| "logits/rejected": -1.4075387716293335, | |
| "logps/chosen": -338.3626708984375, | |
| "logps/rejected": -396.67578125, | |
| "loss": 0.5584, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.9042552709579468, | |
| "rewards/margins": 0.5932050347328186, | |
| "rewards/rejected": -2.49746036529541, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.3090443821097566e-07, | |
| "logits/chosen": -1.2587625980377197, | |
| "logits/rejected": -1.2017955780029297, | |
| "logps/chosen": -309.43377685546875, | |
| "logps/rejected": -372.00531005859375, | |
| "loss": 0.5729, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.9031288623809814, | |
| "rewards/margins": 0.5602144598960876, | |
| "rewards/rejected": -2.4633431434631348, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "eval_logits/chosen": -1.3760210275650024, | |
| "eval_logits/rejected": -1.2920024394989014, | |
| "eval_logps/chosen": -312.20635986328125, | |
| "eval_logps/rejected": -375.1720275878906, | |
| "eval_loss": 0.5711147785186768, | |
| "eval_rewards/accuracies": 0.6828358173370361, | |
| "eval_rewards/chosen": -1.676328182220459, | |
| "eval_rewards/margins": 0.5640482306480408, | |
| "eval_rewards/rejected": -2.2403764724731445, | |
| "eval_runtime": 184.465, | |
| "eval_samples_per_second": 46.361, | |
| "eval_steps_per_second": 0.726, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.2603795624364195e-07, | |
| "logits/chosen": -1.2894772291183472, | |
| "logits/rejected": -1.23129141330719, | |
| "logps/chosen": -299.457275390625, | |
| "logps/rejected": -370.8555908203125, | |
| "loss": 0.5666, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.6054102182388306, | |
| "rewards/margins": 0.5984233021736145, | |
| "rewards/rejected": -2.203833818435669, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.210354171785795e-07, | |
| "logits/chosen": -1.022984266281128, | |
| "logits/rejected": -0.9285897016525269, | |
| "logps/chosen": -324.4284973144531, | |
| "logps/rejected": -385.0074157714844, | |
| "loss": 0.5596, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.921677589416504, | |
| "rewards/margins": 0.5404387712478638, | |
| "rewards/rejected": -2.4621164798736572, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.15900687403248e-07, | |
| "logits/chosen": -0.8059805631637573, | |
| "logits/rejected": -0.7196700572967529, | |
| "logps/chosen": -353.788330078125, | |
| "logps/rejected": -411.4853515625, | |
| "loss": 0.5865, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -2.1321234703063965, | |
| "rewards/margins": 0.463266521692276, | |
| "rewards/rejected": -2.5953898429870605, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.1063773547332584e-07, | |
| "logits/chosen": -0.9645301699638367, | |
| "logits/rejected": -0.7601315975189209, | |
| "logps/chosen": -346.8272705078125, | |
| "logps/rejected": -392.2935791015625, | |
| "loss": 0.5591, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.9153356552124023, | |
| "rewards/margins": 0.5891679525375366, | |
| "rewards/rejected": -2.5045037269592285, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.0525062904547276e-07, | |
| "logits/chosen": -0.608537495136261, | |
| "logits/rejected": -0.47767123579978943, | |
| "logps/chosen": -341.55364990234375, | |
| "logps/rejected": -434.1073303222656, | |
| "loss": 0.5687, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.105318069458008, | |
| "rewards/margins": 0.6994394659996033, | |
| "rewards/rejected": -2.8047571182250977, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 3.997435317334988e-07, | |
| "logits/chosen": -0.6356207132339478, | |
| "logits/rejected": -0.25634175539016724, | |
| "logps/chosen": -384.43780517578125, | |
| "logps/rejected": -419.24176025390625, | |
| "loss": 0.5608, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.2970900535583496, | |
| "rewards/margins": 0.6535100340843201, | |
| "rewards/rejected": -2.9506001472473145, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 3.941206998903701e-07, | |
| "logits/chosen": -1.0318920612335205, | |
| "logits/rejected": -0.7451022267341614, | |
| "logps/chosen": -338.9430236816406, | |
| "logps/rejected": -384.64111328125, | |
| "loss": 0.5678, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.9643396139144897, | |
| "rewards/margins": 0.5402536392211914, | |
| "rewards/rejected": -2.5045928955078125, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.8838647931853684e-07, | |
| "logits/chosen": -0.6847028732299805, | |
| "logits/rejected": -0.5548251867294312, | |
| "logps/chosen": -339.61456298828125, | |
| "logps/rejected": -435.32061767578125, | |
| "loss": 0.5814, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.1349122524261475, | |
| "rewards/margins": 0.7573872804641724, | |
| "rewards/rejected": -2.8923001289367676, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 3.825453019111281e-07, | |
| "logits/chosen": -0.5378957986831665, | |
| "logits/rejected": -0.28533270955085754, | |
| "logps/chosen": -363.78570556640625, | |
| "logps/rejected": -430.11749267578125, | |
| "loss": 0.5327, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.134934425354004, | |
| "rewards/margins": 0.6089809536933899, | |
| "rewards/rejected": -2.743915319442749, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.7660168222660824e-07, | |
| "logits/chosen": -0.6318235397338867, | |
| "logits/rejected": -0.5071814656257629, | |
| "logps/chosen": -350.5252380371094, | |
| "logps/rejected": -421.93353271484375, | |
| "loss": 0.5645, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -2.062009572982788, | |
| "rewards/margins": 0.5333147048950195, | |
| "rewards/rejected": -2.5953242778778076, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_logits/chosen": -0.7860146760940552, | |
| "eval_logits/rejected": -0.6090859770774841, | |
| "eval_logps/chosen": -351.7882995605469, | |
| "eval_logps/rejected": -419.81939697265625, | |
| "eval_loss": 0.5639454126358032, | |
| "eval_rewards/accuracies": 0.6986940503120422, | |
| "eval_rewards/chosen": -2.0721471309661865, | |
| "eval_rewards/margins": 0.6147031188011169, | |
| "eval_rewards/rejected": -2.6868505477905273, | |
| "eval_runtime": 184.5772, | |
| "eval_samples_per_second": 46.333, | |
| "eval_steps_per_second": 0.726, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 3.705602139995416e-07, | |
| "logits/chosen": -0.7258490920066833, | |
| "logits/rejected": -0.4828409254550934, | |
| "logps/chosen": -388.1371154785156, | |
| "logps/rejected": -422.11181640625, | |
| "loss": 0.574, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -2.290266513824463, | |
| "rewards/margins": 0.4104091227054596, | |
| "rewards/rejected": -2.7006754875183105, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 3.6442556659016475e-07, | |
| "logits/chosen": -0.5335447192192078, | |
| "logits/rejected": -0.33706527948379517, | |
| "logps/chosen": -378.86492919921875, | |
| "logps/rejected": -429.67724609375, | |
| "loss": 0.5608, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.236337423324585, | |
| "rewards/margins": 0.556148886680603, | |
| "rewards/rejected": -2.7924864292144775, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.582024813755076e-07, | |
| "logits/chosen": -0.39548322558403015, | |
| "logits/rejected": -0.10662730038166046, | |
| "logps/chosen": -368.8847961425781, | |
| "logps/rejected": -473.3500061035156, | |
| "loss": 0.5485, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.3263449668884277, | |
| "rewards/margins": 0.8236624598503113, | |
| "rewards/rejected": -3.150007724761963, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.5189576808485404e-07, | |
| "logits/chosen": 0.15742243826389313, | |
| "logits/rejected": 0.31491726636886597, | |
| "logps/chosen": -394.34930419921875, | |
| "logps/rejected": -492.82232666015625, | |
| "loss": 0.5478, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.6109700202941895, | |
| "rewards/margins": 0.8250136375427246, | |
| "rewards/rejected": -3.435983657836914, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.4551030108237433e-07, | |
| "logits/chosen": -0.2550584375858307, | |
| "logits/rejected": -0.06936412304639816, | |
| "logps/chosen": -406.5508728027344, | |
| "logps/rejected": -448.47576904296875, | |
| "loss": 0.5562, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -2.5152666568756104, | |
| "rewards/margins": 0.4819938540458679, | |
| "rewards/rejected": -2.997260332107544, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.390510155998023e-07, | |
| "logits/chosen": -0.5292027592658997, | |
| "logits/rejected": -0.2619571387767792, | |
| "logps/chosen": -371.6798095703125, | |
| "logps/rejected": -420.7915954589844, | |
| "loss": 0.5492, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.1147050857543945, | |
| "rewards/margins": 0.6524336338043213, | |
| "rewards/rejected": -2.7671384811401367, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.325229039220684e-07, | |
| "logits/chosen": -0.5881962776184082, | |
| "logits/rejected": -0.4658876061439514, | |
| "logps/chosen": -343.7039794921875, | |
| "logps/rejected": -406.14178466796875, | |
| "loss": 0.57, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -2.0860273838043213, | |
| "rewards/margins": 0.498068630695343, | |
| "rewards/rejected": -2.5840957164764404, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.2593101152883795e-07, | |
| "logits/chosen": -0.6565806269645691, | |
| "logits/rejected": -0.2549567222595215, | |
| "logps/chosen": -374.8047180175781, | |
| "logps/rejected": -430.33221435546875, | |
| "loss": 0.5512, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.2211391925811768, | |
| "rewards/margins": 0.6813799142837524, | |
| "rewards/rejected": -2.9025187492370605, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.192804331949349e-07, | |
| "logits/chosen": -0.07184700667858124, | |
| "logits/rejected": 0.1699156016111374, | |
| "logps/chosen": -422.27081298828125, | |
| "logps/rejected": -490.69134521484375, | |
| "loss": 0.535, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.730973720550537, | |
| "rewards/margins": 0.7726518511772156, | |
| "rewards/rejected": -3.5036251544952393, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.125763090526674e-07, | |
| "logits/chosen": -0.029465889558196068, | |
| "logits/rejected": 0.15842057764530182, | |
| "logps/chosen": -417.373046875, | |
| "logps/rejected": -478.73291015625, | |
| "loss": 0.5513, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -2.8006317615509033, | |
| "rewards/margins": 0.6451797485351562, | |
| "rewards/rejected": -3.4458115100860596, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "eval_logits/chosen": -0.10542195290327072, | |
| "eval_logits/rejected": 0.12242482602596283, | |
| "eval_logps/chosen": -436.9386291503906, | |
| "eval_logps/rejected": -505.02227783203125, | |
| "eval_loss": 0.5582411885261536, | |
| "eval_rewards/accuracies": 0.7108209133148193, | |
| "eval_rewards/chosen": -2.9236514568328857, | |
| "eval_rewards/margins": 0.6152271032333374, | |
| "eval_rewards/rejected": -3.5388784408569336, | |
| "eval_runtime": 184.6013, | |
| "eval_samples_per_second": 46.327, | |
| "eval_steps_per_second": 0.726, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 3.0582382061909623e-07, | |
| "logits/chosen": -0.2445104569196701, | |
| "logits/rejected": -0.018268002197146416, | |
| "logps/chosen": -441.7857971191406, | |
| "logps/rejected": -502.60791015625, | |
| "loss": 0.5594, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -2.858261823654175, | |
| "rewards/margins": 0.5510683655738831, | |
| "rewards/rejected": -3.409330368041992, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 2.9902818679131775e-07, | |
| "logits/chosen": -0.4190225601196289, | |
| "logits/rejected": -0.22823679447174072, | |
| "logps/chosen": -399.03924560546875, | |
| "logps/rejected": -498.6724548339844, | |
| "loss": 0.5499, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.650449275970459, | |
| "rewards/margins": 0.7673205137252808, | |
| "rewards/rejected": -3.4177703857421875, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.921946598128571e-07, | |
| "logits/chosen": -0.43653860688209534, | |
| "logits/rejected": -0.20837187767028809, | |
| "logps/chosen": -402.82781982421875, | |
| "logps/rejected": -485.4117736816406, | |
| "loss": 0.5739, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.573000192642212, | |
| "rewards/margins": 0.7478531002998352, | |
| "rewards/rejected": -3.3208529949188232, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 2.8532852121428733e-07, | |
| "logits/chosen": -0.43430274724960327, | |
| "logits/rejected": -0.13240045309066772, | |
| "logps/chosen": -397.2491149902344, | |
| "logps/rejected": -442.12384033203125, | |
| "loss": 0.5462, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.4125733375549316, | |
| "rewards/margins": 0.5821372866630554, | |
| "rewards/rejected": -2.9947104454040527, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.7843507773121414e-07, | |
| "logits/chosen": -0.4247920513153076, | |
| "logits/rejected": -0.21372787654399872, | |
| "logps/chosen": -389.4237976074219, | |
| "logps/rejected": -458.3169860839844, | |
| "loss": 0.5373, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -2.444688320159912, | |
| "rewards/margins": 0.7236617207527161, | |
| "rewards/rejected": -3.1683506965637207, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 2.715196572027789e-07, | |
| "logits/chosen": -0.6697942614555359, | |
| "logits/rejected": -0.4933086931705475, | |
| "logps/chosen": -387.529296875, | |
| "logps/rejected": -472.73944091796875, | |
| "loss": 0.5685, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.3496451377868652, | |
| "rewards/margins": 0.7728831171989441, | |
| "rewards/rejected": -3.122528314590454, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 2.645876044538521e-07, | |
| "logits/chosen": -1.0338900089263916, | |
| "logits/rejected": -0.8813627362251282, | |
| "logps/chosen": -372.53118896484375, | |
| "logps/rejected": -426.54241943359375, | |
| "loss": 0.5725, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -2.201908588409424, | |
| "rewards/margins": 0.5010865926742554, | |
| "rewards/rejected": -2.7029950618743896, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.5764427716409815e-07, | |
| "logits/chosen": -0.9278701543807983, | |
| "logits/rejected": -0.7282145023345947, | |
| "logps/chosen": -347.2828674316406, | |
| "logps/rejected": -416.9349060058594, | |
| "loss": 0.5479, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.0276436805725098, | |
| "rewards/margins": 0.743033230304718, | |
| "rewards/rejected": -2.770677089691162, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2.5069504172710494e-07, | |
| "logits/chosen": -0.5008482336997986, | |
| "logits/rejected": -0.34875133633613586, | |
| "logps/chosen": -373.7621154785156, | |
| "logps/rejected": -485.12884521484375, | |
| "loss": 0.5217, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.347053050994873, | |
| "rewards/margins": 0.9024646878242493, | |
| "rewards/rejected": -3.2495174407958984, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.4374526910277886e-07, | |
| "logits/chosen": 0.06850005686283112, | |
| "logits/rejected": 0.41385045647621155, | |
| "logps/chosen": -411.46246337890625, | |
| "logps/rejected": -476.6162109375, | |
| "loss": 0.5571, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -2.6757898330688477, | |
| "rewards/margins": 0.8085702657699585, | |
| "rewards/rejected": -3.4843602180480957, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "eval_logits/chosen": 0.035554468631744385, | |
| "eval_logits/rejected": 0.2980235815048218, | |
| "eval_logps/chosen": -424.2823486328125, | |
| "eval_logps/rejected": -505.6960754394531, | |
| "eval_loss": 0.5558871626853943, | |
| "eval_rewards/accuracies": 0.704291045665741, | |
| "eval_rewards/chosen": -2.797088146209717, | |
| "eval_rewards/margins": 0.748529314994812, | |
| "eval_rewards/rejected": -3.5456173419952393, | |
| "eval_runtime": 184.3728, | |
| "eval_samples_per_second": 46.384, | |
| "eval_steps_per_second": 0.727, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.368003306662104e-07, | |
| "logits/chosen": 0.07857178151607513, | |
| "logits/rejected": 0.3302653729915619, | |
| "logps/chosen": -413.8836975097656, | |
| "logps/rejected": -535.0875244140625, | |
| "loss": 0.5287, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.7193782329559326, | |
| "rewards/margins": 1.0089346170425415, | |
| "rewards/rejected": -3.7283127307891846, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.2986559405621886e-07, | |
| "logits/chosen": 0.2789291739463806, | |
| "logits/rejected": 0.4242584705352783, | |
| "logps/chosen": -422.7801818847656, | |
| "logps/rejected": -522.7840576171875, | |
| "loss": 0.5551, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.925621509552002, | |
| "rewards/margins": 0.8043605089187622, | |
| "rewards/rejected": -3.729982376098633, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.2294641902678443e-07, | |
| "logits/chosen": -0.19327735900878906, | |
| "logits/rejected": 0.043265581130981445, | |
| "logps/chosen": -363.1488342285156, | |
| "logps/rejected": -470.94970703125, | |
| "loss": 0.5284, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.39530873298645, | |
| "rewards/margins": 0.908363938331604, | |
| "rewards/rejected": -3.3036727905273438, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.160481533045751e-07, | |
| "logits/chosen": -0.37412697076797485, | |
| "logits/rejected": -0.17320053279399872, | |
| "logps/chosen": -390.2896423339844, | |
| "logps/rejected": -428.08099365234375, | |
| "loss": 0.5572, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -2.3005330562591553, | |
| "rewards/margins": 0.48462891578674316, | |
| "rewards/rejected": -2.7851624488830566, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.0917612845576882e-07, | |
| "logits/chosen": -0.26352375745773315, | |
| "logits/rejected": -0.0010178961092606187, | |
| "logps/chosen": -373.3875427246094, | |
| "logps/rejected": -440.09442138671875, | |
| "loss": 0.5534, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -2.3315823078155518, | |
| "rewards/margins": 0.6843063235282898, | |
| "rewards/rejected": -3.0158886909484863, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.0233565576536564e-07, | |
| "logits/chosen": -0.3354080021381378, | |
| "logits/rejected": -0.006600166670978069, | |
| "logps/chosen": -360.56463623046875, | |
| "logps/rejected": -440.66961669921875, | |
| "loss": 0.5328, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.1626803874969482, | |
| "rewards/margins": 0.8473829030990601, | |
| "rewards/rejected": -3.010063409805298, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 1.9553202213217537e-07, | |
| "logits/chosen": -0.021420275792479515, | |
| "logits/rejected": 0.19946305453777313, | |
| "logps/chosen": -389.1043395996094, | |
| "logps/rejected": -448.04998779296875, | |
| "loss": 0.5523, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.391838788986206, | |
| "rewards/margins": 0.6678962707519531, | |
| "rewards/rejected": -3.059735059738159, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.887704859826528e-07, | |
| "logits/chosen": -0.15253478288650513, | |
| "logits/rejected": -0.00011998042464256287, | |
| "logps/chosen": -394.9501953125, | |
| "logps/rejected": -462.32843017578125, | |
| "loss": 0.5443, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.476644992828369, | |
| "rewards/margins": 0.566824734210968, | |
| "rewards/rejected": -3.0434699058532715, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.8205627320673836e-07, | |
| "logits/chosen": -0.17955633997917175, | |
| "logits/rejected": 0.18167546391487122, | |
| "logps/chosen": -390.32244873046875, | |
| "logps/rejected": -444.895263671875, | |
| "loss": 0.5566, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.4376220703125, | |
| "rewards/margins": 0.7008293271064758, | |
| "rewards/rejected": -3.138451099395752, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.7539457311884675e-07, | |
| "logits/chosen": -0.09838727861642838, | |
| "logits/rejected": 0.11829495429992676, | |
| "logps/chosen": -402.4017333984375, | |
| "logps/rejected": -451.49346923828125, | |
| "loss": 0.5609, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.4441986083984375, | |
| "rewards/margins": 0.5067789554595947, | |
| "rewards/rejected": -2.9509775638580322, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_logits/chosen": -0.03116540051996708, | |
| "eval_logits/rejected": 0.1922437697649002, | |
| "eval_logps/chosen": -387.7091979980469, | |
| "eval_logps/rejected": -459.44390869140625, | |
| "eval_loss": 0.5468714833259583, | |
| "eval_rewards/accuracies": 0.7108209133148193, | |
| "eval_rewards/chosen": -2.431356430053711, | |
| "eval_rewards/margins": 0.6517390012741089, | |
| "eval_rewards/rejected": -3.0830955505371094, | |
| "eval_runtime": 184.5893, | |
| "eval_samples_per_second": 46.33, | |
| "eval_steps_per_second": 0.726, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.687905344471226e-07, | |
| "logits/chosen": 0.07735608518123627, | |
| "logits/rejected": 0.3973601460456848, | |
| "logps/chosen": -408.05999755859375, | |
| "logps/rejected": -459.011474609375, | |
| "loss": 0.5384, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -2.5008435249328613, | |
| "rewards/margins": 0.6535352468490601, | |
| "rewards/rejected": -3.154379367828369, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.6224926135406693e-07, | |
| "logits/chosen": 0.1125444769859314, | |
| "logits/rejected": 0.3865428566932678, | |
| "logps/chosen": -404.16058349609375, | |
| "logps/rejected": -484.68621826171875, | |
| "loss": 0.5448, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -2.4405789375305176, | |
| "rewards/margins": 0.718208909034729, | |
| "rewards/rejected": -3.158787727355957, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1.557758094916053e-07, | |
| "logits/chosen": 0.11989516019821167, | |
| "logits/rejected": 0.30926594138145447, | |
| "logps/chosen": -370.29876708984375, | |
| "logps/rejected": -452.27911376953125, | |
| "loss": 0.5418, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.3860089778900146, | |
| "rewards/margins": 0.7260924577713013, | |
| "rewards/rejected": -3.1121015548706055, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.4937518209365108e-07, | |
| "logits/chosen": -0.14239154756069183, | |
| "logits/rejected": 0.14250756800174713, | |
| "logps/chosen": -395.55755615234375, | |
| "logps/rejected": -447.6368713378906, | |
| "loss": 0.5573, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.378154754638672, | |
| "rewards/margins": 0.6160937547683716, | |
| "rewards/rejected": -2.994248390197754, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.4305232610918045e-07, | |
| "logits/chosen": -0.16526366770267487, | |
| "logits/rejected": 0.16432161629199982, | |
| "logps/chosen": -373.45330810546875, | |
| "logps/rejected": -436.6773376464844, | |
| "loss": 0.5415, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.3134028911590576, | |
| "rewards/margins": 0.774810791015625, | |
| "rewards/rejected": -3.0882136821746826, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.3681212837880977e-07, | |
| "logits/chosen": -0.1321481615304947, | |
| "logits/rejected": 0.23287932574748993, | |
| "logps/chosen": -364.96990966796875, | |
| "logps/rejected": -447.7923278808594, | |
| "loss": 0.5396, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.176964282989502, | |
| "rewards/margins": 0.8955341577529907, | |
| "rewards/rejected": -3.0724985599517822, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3065941185782977e-07, | |
| "logits/chosen": 0.05437428876757622, | |
| "logits/rejected": 0.2819867432117462, | |
| "logps/chosen": -383.08599853515625, | |
| "logps/rejected": -439.3629455566406, | |
| "loss": 0.5505, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -2.405247211456299, | |
| "rewards/margins": 0.5403125882148743, | |
| "rewards/rejected": -2.9455599784851074, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.2459893188861613e-07, | |
| "logits/chosen": -0.12052659690380096, | |
| "logits/rejected": 0.12284734100103378, | |
| "logps/chosen": -367.1181640625, | |
| "logps/rejected": -468.1044921875, | |
| "loss": 0.5185, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.39152455329895, | |
| "rewards/margins": 0.9137696027755737, | |
| "rewards/rejected": -3.3052947521209717, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.1863537252529548e-07, | |
| "logits/chosen": 0.14598000049591064, | |
| "logits/rejected": 0.38815659284591675, | |
| "logps/chosen": -397.891357421875, | |
| "logps/rejected": -472.38677978515625, | |
| "loss": 0.5323, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.512676239013672, | |
| "rewards/margins": 0.7713057994842529, | |
| "rewards/rejected": -3.2839818000793457, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1.1277334291351145e-07, | |
| "logits/chosen": 0.15319526195526123, | |
| "logits/rejected": 0.35974830389022827, | |
| "logps/chosen": -380.77783203125, | |
| "logps/rejected": -449.54315185546875, | |
| "loss": 0.5514, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -2.3706305027008057, | |
| "rewards/margins": 0.6724039912223816, | |
| "rewards/rejected": -3.043034076690674, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "eval_logits/chosen": 0.28598034381866455, | |
| "eval_logits/rejected": 0.5382024645805359, | |
| "eval_logps/chosen": -392.3096008300781, | |
| "eval_logps/rejected": -471.95330810546875, | |
| "eval_loss": 0.5473664402961731, | |
| "eval_rewards/accuracies": 0.6996268630027771, | |
| "eval_rewards/chosen": -2.4773612022399902, | |
| "eval_rewards/margins": 0.7308279275894165, | |
| "eval_rewards/rejected": -3.2081892490386963, | |
| "eval_runtime": 184.5275, | |
| "eval_samples_per_second": 46.345, | |
| "eval_steps_per_second": 0.726, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1.0701737372808431e-07, | |
| "logits/chosen": 0.15951867401599884, | |
| "logits/rejected": 0.46630391478538513, | |
| "logps/chosen": -383.52850341796875, | |
| "logps/rejected": -467.2303771972656, | |
| "loss": 0.5362, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.2782187461853027, | |
| "rewards/margins": 0.8473943471908569, | |
| "rewards/rejected": -3.125612735748291, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0137191367132078e-07, | |
| "logits/chosen": 0.2791319191455841, | |
| "logits/rejected": 0.45174160599708557, | |
| "logps/chosen": -372.1945495605469, | |
| "logps/rejected": -446.6507263183594, | |
| "loss": 0.5458, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -2.3160648345947266, | |
| "rewards/margins": 0.68004310131073, | |
| "rewards/rejected": -2.996107816696167, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 9.584132603467827e-08, | |
| "logits/chosen": -0.12192128598690033, | |
| "logits/rejected": 0.1477951854467392, | |
| "logps/chosen": -366.48321533203125, | |
| "logps/rejected": -453.130126953125, | |
| "loss": 0.5467, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.200005531311035, | |
| "rewards/margins": 0.7978888750076294, | |
| "rewards/rejected": -2.997894287109375, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 9.042988532644249e-08, | |
| "logits/chosen": -0.03106372058391571, | |
| "logits/rejected": 0.07721444219350815, | |
| "logps/chosen": -344.21270751953125, | |
| "logps/rejected": -438.11077880859375, | |
| "loss": 0.5161, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.149094581604004, | |
| "rewards/margins": 0.7353444695472717, | |
| "rewards/rejected": -2.884438991546631, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 8.514177396802428e-08, | |
| "logits/chosen": 0.006801058538258076, | |
| "logits/rejected": 0.20282092690467834, | |
| "logps/chosen": -358.15167236328125, | |
| "logps/rejected": -436.4964294433594, | |
| "loss": 0.5385, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.2220425605773926, | |
| "rewards/margins": 0.7004804611206055, | |
| "rewards/rejected": -2.922523260116577, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 7.998107906142839e-08, | |
| "logits/chosen": 0.41448846459388733, | |
| "logits/rejected": 0.705254852771759, | |
| "logps/chosen": -371.27801513671875, | |
| "logps/rejected": -434.56866455078125, | |
| "loss": 0.5236, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.2166616916656494, | |
| "rewards/margins": 0.6714047193527222, | |
| "rewards/rejected": -2.888066530227661, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.495178923039396e-08, | |
| "logits/chosen": 0.23847150802612305, | |
| "logits/rejected": 0.48661884665489197, | |
| "logps/chosen": -366.28179931640625, | |
| "logps/rejected": -462.679443359375, | |
| "loss": 0.5459, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.1916985511779785, | |
| "rewards/margins": 0.8472123146057129, | |
| "rewards/rejected": -3.038910388946533, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 7.005779153764682e-08, | |
| "logits/chosen": 0.41438961029052734, | |
| "logits/rejected": 0.6912784576416016, | |
| "logps/chosen": -382.70123291015625, | |
| "logps/rejected": -461.8614807128906, | |
| "loss": 0.5453, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.4304287433624268, | |
| "rewards/margins": 0.7116767764091492, | |
| "rewards/rejected": -3.1421055793762207, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 6.530286848064698e-08, | |
| "logits/chosen": 0.36573725938796997, | |
| "logits/rejected": 0.5834362506866455, | |
| "logps/chosen": -384.49749755859375, | |
| "logps/rejected": -466.30096435546875, | |
| "loss": 0.5528, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.5111565589904785, | |
| "rewards/margins": 0.7234699130058289, | |
| "rewards/rejected": -3.234626054763794, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 6.069069506815325e-08, | |
| "logits/chosen": 0.45530566573143005, | |
| "logits/rejected": 0.5909157991409302, | |
| "logps/chosen": -379.1433410644531, | |
| "logps/rejected": -468.88458251953125, | |
| "loss": 0.527, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.5361268520355225, | |
| "rewards/margins": 0.7407721281051636, | |
| "rewards/rejected": -3.2768986225128174, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_logits/chosen": 0.3871051073074341, | |
| "eval_logits/rejected": 0.6372014284133911, | |
| "eval_logps/chosen": -394.97113037109375, | |
| "eval_logps/rejected": -471.8453674316406, | |
| "eval_loss": 0.5453863739967346, | |
| "eval_rewards/accuracies": 0.70802241563797, | |
| "eval_rewards/chosen": -2.503976345062256, | |
| "eval_rewards/margins": 0.7031334638595581, | |
| "eval_rewards/rejected": -3.2071101665496826, | |
| "eval_runtime": 184.4226, | |
| "eval_samples_per_second": 46.372, | |
| "eval_steps_per_second": 0.727, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 5.6224835979863714e-08, | |
| "logits/chosen": 0.31174296140670776, | |
| "logits/rejected": 0.6193565130233765, | |
| "logps/chosen": -390.387451171875, | |
| "logps/rejected": -468.4959411621094, | |
| "loss": 0.5568, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.32766056060791, | |
| "rewards/margins": 0.68747878074646, | |
| "rewards/rejected": -3.015139102935791, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 5.190874281132851e-08, | |
| "logits/chosen": 0.22277125716209412, | |
| "logits/rejected": 0.6487134099006653, | |
| "logps/chosen": -402.0958557128906, | |
| "logps/rejected": -448.5992736816406, | |
| "loss": 0.5408, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.359062671661377, | |
| "rewards/margins": 0.6533006429672241, | |
| "rewards/rejected": -3.0123631954193115, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.774575140626316e-08, | |
| "logits/chosen": 0.23170511424541473, | |
| "logits/rejected": 0.47184085845947266, | |
| "logps/chosen": -363.46917724609375, | |
| "logps/rejected": -442.47918701171875, | |
| "loss": 0.5309, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.142770290374756, | |
| "rewards/margins": 0.7513145208358765, | |
| "rewards/rejected": -2.894084930419922, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.373907927832513e-08, | |
| "logits/chosen": 0.07573021948337555, | |
| "logits/rejected": 0.32997313141822815, | |
| "logps/chosen": -381.45599365234375, | |
| "logps/rejected": -443.0684509277344, | |
| "loss": 0.5407, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.2033116817474365, | |
| "rewards/margins": 0.710732638835907, | |
| "rewards/rejected": -2.914044141769409, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.9891823124345665e-08, | |
| "logits/chosen": 0.23884686827659607, | |
| "logits/rejected": 0.6128005385398865, | |
| "logps/chosen": -364.00567626953125, | |
| "logps/rejected": -433.3273010253906, | |
| "loss": 0.5471, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.2081527709960938, | |
| "rewards/margins": 0.7639263868331909, | |
| "rewards/rejected": -2.972079038619995, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.620695643093924e-08, | |
| "logits/chosen": 0.21963253617286682, | |
| "logits/rejected": 0.6894062757492065, | |
| "logps/chosen": -399.5767517089844, | |
| "logps/rejected": -452.88909912109375, | |
| "loss": 0.5154, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.3384335041046143, | |
| "rewards/margins": 0.7010769844055176, | |
| "rewards/rejected": -3.0395102500915527, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 3.268732717634032e-08, | |
| "logits/chosen": 0.3474286198616028, | |
| "logits/rejected": 0.695271372795105, | |
| "logps/chosen": -368.0654602050781, | |
| "logps/rejected": -431.47222900390625, | |
| "loss": 0.5499, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.1910276412963867, | |
| "rewards/margins": 0.7267633080482483, | |
| "rewards/rejected": -2.9177908897399902, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 2.9335655629243645e-08, | |
| "logits/chosen": 0.2347393035888672, | |
| "logits/rejected": 0.5894696712493896, | |
| "logps/chosen": -388.94757080078125, | |
| "logps/rejected": -447.3855895996094, | |
| "loss": 0.525, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.254683017730713, | |
| "rewards/margins": 0.7334609031677246, | |
| "rewards/rejected": -2.9881439208984375, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 2.6154532246349476e-08, | |
| "logits/chosen": 0.25378522276878357, | |
| "logits/rejected": 0.5771256685256958, | |
| "logps/chosen": -358.50640869140625, | |
| "logps/rejected": -431.145751953125, | |
| "loss": 0.5462, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.1666626930236816, | |
| "rewards/margins": 0.756801426410675, | |
| "rewards/rejected": -2.923464059829712, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.31464156702382e-08, | |
| "logits/chosen": 0.35370689630508423, | |
| "logits/rejected": 0.5671936273574829, | |
| "logps/chosen": -363.0, | |
| "logps/rejected": -438.209228515625, | |
| "loss": 0.5487, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.2130119800567627, | |
| "rewards/margins": 0.7499077916145325, | |
| "rewards/rejected": -2.9629194736480713, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "eval_logits/chosen": 0.1857856959104538, | |
| "eval_logits/rejected": 0.43363669514656067, | |
| "eval_logps/chosen": -373.08306884765625, | |
| "eval_logps/rejected": -450.7598876953125, | |
| "eval_loss": 0.5444055199623108, | |
| "eval_rewards/accuracies": 0.7089552283287048, | |
| "eval_rewards/chosen": -2.285095453262329, | |
| "eval_rewards/margins": 0.711159884929657, | |
| "eval_rewards/rejected": -2.996255397796631, | |
| "eval_runtime": 184.462, | |
| "eval_samples_per_second": 46.362, | |
| "eval_steps_per_second": 0.726, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.031363082912252e-08, | |
| "logits/chosen": 0.070524200797081, | |
| "logits/rejected": 0.4635602533817291, | |
| "logps/chosen": -373.29327392578125, | |
| "logps/rejected": -426.85552978515625, | |
| "loss": 0.5513, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -2.2541089057922363, | |
| "rewards/margins": 0.6198171973228455, | |
| "rewards/rejected": -2.8739261627197266, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 1.7658367139945228e-08, | |
| "logits/chosen": 0.2600646913051605, | |
| "logits/rejected": 0.5517584681510925, | |
| "logps/chosen": -390.8568115234375, | |
| "logps/rejected": -462.80828857421875, | |
| "loss": 0.5471, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -2.377396821975708, | |
| "rewards/margins": 0.6719989776611328, | |
| "rewards/rejected": -3.049395799636841, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.5182676816211632e-08, | |
| "logits/chosen": 0.04413030296564102, | |
| "logits/rejected": 0.30151715874671936, | |
| "logps/chosen": -382.0662536621094, | |
| "logps/rejected": -447.08673095703125, | |
| "loss": 0.5431, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -2.228654384613037, | |
| "rewards/margins": 0.6926024556159973, | |
| "rewards/rejected": -2.9212570190429688, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.2888473281864597e-08, | |
| "logits/chosen": 0.14212054014205933, | |
| "logits/rejected": 0.47429710626602173, | |
| "logps/chosen": -367.8409729003906, | |
| "logps/rejected": -435.02764892578125, | |
| "loss": 0.5369, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.2534115314483643, | |
| "rewards/margins": 0.7273036241531372, | |
| "rewards/rejected": -2.980715274810791, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 1.0777529692427679e-08, | |
| "logits/chosen": 0.04115242511034012, | |
| "logits/rejected": 0.28970104455947876, | |
| "logps/chosen": -372.7949523925781, | |
| "logps/rejected": -456.10675048828125, | |
| "loss": 0.5265, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.300356388092041, | |
| "rewards/margins": 0.8059718012809753, | |
| "rewards/rejected": -3.106328248977661, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 8.851477564560061e-09, | |
| "logits/chosen": 0.0867738351225853, | |
| "logits/rejected": 0.4068300127983093, | |
| "logps/chosen": -372.08636474609375, | |
| "logps/rejected": -426.42388916015625, | |
| "loss": 0.5342, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -2.331385850906372, | |
| "rewards/margins": 0.6490964293479919, | |
| "rewards/rejected": -2.9804821014404297, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 7.111805515081531e-09, | |
| "logits/chosen": 0.02022993005812168, | |
| "logits/rejected": 0.41968393325805664, | |
| "logps/chosen": -363.818603515625, | |
| "logps/rejected": -447.7919006347656, | |
| "loss": 0.5312, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.2372307777404785, | |
| "rewards/margins": 0.8540315628051758, | |
| "rewards/rejected": -3.0912623405456543, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 5.559858110443016e-09, | |
| "logits/chosen": 0.29695388674736023, | |
| "logits/rejected": 0.714096188545227, | |
| "logps/chosen": -372.5519714355469, | |
| "logps/rejected": -442.5354919433594, | |
| "loss": 0.5383, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -2.3107995986938477, | |
| "rewards/margins": 0.8070123791694641, | |
| "rewards/rejected": -3.117811918258667, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 4.196834827531276e-09, | |
| "logits/chosen": 0.140055850148201, | |
| "logits/rejected": 0.3409932255744934, | |
| "logps/chosen": -355.64324951171875, | |
| "logps/rejected": -447.585693359375, | |
| "loss": 0.5152, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.149151563644409, | |
| "rewards/margins": 0.7904965877532959, | |
| "rewards/rejected": -2.939648151397705, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 3.023789126611137e-09, | |
| "logits/chosen": 0.03294936567544937, | |
| "logits/rejected": 0.2933207154273987, | |
| "logps/chosen": -363.29290771484375, | |
| "logps/rejected": -435.640380859375, | |
| "loss": 0.5483, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.281057357788086, | |
| "rewards/margins": 0.7091296911239624, | |
| "rewards/rejected": -2.990186929702759, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "eval_logits/chosen": 0.07418080419301987, | |
| "eval_logits/rejected": 0.32435521483421326, | |
| "eval_logps/chosen": -373.978515625, | |
| "eval_logps/rejected": -451.6764831542969, | |
| "eval_loss": 0.5440130829811096, | |
| "eval_rewards/accuracies": 0.7089552283287048, | |
| "eval_rewards/chosen": -2.2940499782562256, | |
| "eval_rewards/margins": 0.7113713622093201, | |
| "eval_rewards/rejected": -3.0054211616516113, | |
| "eval_runtime": 184.3589, | |
| "eval_samples_per_second": 46.388, | |
| "eval_steps_per_second": 0.727, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.041627637121929e-09, | |
| "logits/chosen": 0.10010697692632675, | |
| "logits/rejected": 0.3795483410358429, | |
| "logps/chosen": -348.8675231933594, | |
| "logps/rejected": -437.20361328125, | |
| "loss": 0.5398, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.087364673614502, | |
| "rewards/margins": 0.828387439250946, | |
| "rewards/rejected": -2.9157521724700928, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 1.2511094569571668e-09, | |
| "logits/chosen": 0.09991980344057083, | |
| "logits/rejected": 0.4467397630214691, | |
| "logps/chosen": -380.14520263671875, | |
| "logps/rejected": -440.24658203125, | |
| "loss": 0.5345, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.253425121307373, | |
| "rewards/margins": 0.702509880065918, | |
| "rewards/rejected": -2.955935001373291, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 6.528455657691112e-10, | |
| "logits/chosen": 0.11626466363668442, | |
| "logits/rejected": 0.41348797082901, | |
| "logps/chosen": -372.7298889160156, | |
| "logps/rejected": -427.22576904296875, | |
| "loss": 0.549, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.2800345420837402, | |
| "rewards/margins": 0.6291176080703735, | |
| "rewards/rejected": -2.909151792526245, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 2.4729835275189016e-10, | |
| "logits/chosen": 0.06715863198041916, | |
| "logits/rejected": 0.29241910576820374, | |
| "logps/chosen": -393.8903503417969, | |
| "logps/rejected": -477.9420471191406, | |
| "loss": 0.5462, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.323488473892212, | |
| "rewards/margins": 0.8067766427993774, | |
| "rewards/rejected": -3.1302647590637207, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 3.478125926756337e-11, | |
| "logits/chosen": 0.25983649492263794, | |
| "logits/rejected": 0.4905417561531067, | |
| "logps/chosen": -364.73431396484375, | |
| "logps/rejected": -443.79296875, | |
| "loss": 0.5474, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.2353272438049316, | |
| "rewards/margins": 0.771331787109375, | |
| "rewards/rejected": -3.0066590309143066, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 1256, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5712926928784438, | |
| "train_runtime": 11573.5451, | |
| "train_samples_per_second": 13.894, | |
| "train_steps_per_second": 0.109 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1256, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |