| { | |
| "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": 2500.0, | |
| "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": 0.8362942934036255, | |
| "logits/rejected": 0.8542055487632751, | |
| "logps/chosen": -236.253662109375, | |
| "logps/rejected": -221.88853454589844, | |
| "loss": 2503.2357, | |
| "rewards/accuracies": 0.3958333432674408, | |
| "rewards/chosen": 8.874866762198508e-05, | |
| "rewards/margins": -0.00026307348161935806, | |
| "rewards/rejected": 0.0003518221783451736, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 1.0416666666666667e-06, | |
| "logits/chosen": 0.8335070610046387, | |
| "logits/rejected": 0.9283286929130554, | |
| "logps/chosen": -254.7803192138672, | |
| "logps/rejected": -247.91357421875, | |
| "loss": 2498.9305, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.0006878537242300808, | |
| "rewards/margins": 0.00017734414723236114, | |
| "rewards/rejected": -0.0008651980315335095, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 1.5625e-06, | |
| "logits/chosen": 0.8601231575012207, | |
| "logits/rejected": 0.9173057675361633, | |
| "logps/chosen": -260.49664306640625, | |
| "logps/rejected": -232.03378295898438, | |
| "loss": 2500.6068, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": 0.0006941998144611716, | |
| "rewards/margins": 0.0007489144918508828, | |
| "rewards/rejected": -5.471452095662244e-05, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 2.0833333333333334e-06, | |
| "logits/chosen": 0.811628520488739, | |
| "logits/rejected": 0.9033697843551636, | |
| "logps/chosen": -280.28118896484375, | |
| "logps/rejected": -228.5680694580078, | |
| "loss": 2502.0561, | |
| "rewards/accuracies": 0.4312500059604645, | |
| "rewards/chosen": -0.000692486937623471, | |
| "rewards/margins": -0.0004873524303548038, | |
| "rewards/rejected": -0.00020513453637249768, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 2.604166666666667e-06, | |
| "logits/chosen": 0.8561393022537231, | |
| "logits/rejected": 0.9358364343643188, | |
| "logps/chosen": -257.61163330078125, | |
| "logps/rejected": -219.1778106689453, | |
| "loss": 2492.4258, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.0008641455206088722, | |
| "rewards/margins": 0.0018391588237136602, | |
| "rewards/rejected": -0.000975013361312449, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 3.125e-06, | |
| "logits/chosen": 0.8744305372238159, | |
| "logits/rejected": 0.9009464383125305, | |
| "logps/chosen": -237.00228881835938, | |
| "logps/rejected": -237.4504852294922, | |
| "loss": 2497.3361, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -9.825383131101262e-06, | |
| "rewards/margins": 0.00019281035929452628, | |
| "rewards/rejected": -0.0002026357251452282, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 3.6458333333333333e-06, | |
| "logits/chosen": 0.8502357602119446, | |
| "logits/rejected": 0.8783925771713257, | |
| "logps/chosen": -260.8014221191406, | |
| "logps/rejected": -227.61328125, | |
| "loss": 2500.9129, | |
| "rewards/accuracies": 0.4437499940395355, | |
| "rewards/chosen": -0.0007933862507343292, | |
| "rewards/margins": -0.00043974880827590823, | |
| "rewards/rejected": -0.000353637442458421, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.166666666666667e-06, | |
| "logits/chosen": 0.8865741491317749, | |
| "logits/rejected": 0.9179280996322632, | |
| "logps/chosen": -251.12197875976562, | |
| "logps/rejected": -231.300048828125, | |
| "loss": 2494.3678, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -0.00037043695920147, | |
| "rewards/margins": 0.0005658747395500541, | |
| "rewards/rejected": -0.0009363117860630155, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.6875000000000004e-06, | |
| "logits/chosen": 0.8638327717781067, | |
| "logits/rejected": 0.9173502922058105, | |
| "logps/chosen": -225.3396453857422, | |
| "logps/rejected": -241.39352416992188, | |
| "loss": 2497.2648, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": -0.0007204846478998661, | |
| "rewards/margins": 0.0005254354909993708, | |
| "rewards/rejected": -0.001245920080691576, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.999731868769027e-06, | |
| "logits/chosen": 0.9247162938117981, | |
| "logits/rejected": 0.9241034388542175, | |
| "logps/chosen": -242.1609344482422, | |
| "logps/rejected": -221.4512939453125, | |
| "loss": 2494.2439, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.00025123285013251007, | |
| "rewards/margins": 0.0006884234608151019, | |
| "rewards/rejected": -0.0009396563400514424, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": 0.8141916990280151, | |
| "eval_logits/rejected": 0.9164313673973083, | |
| "eval_logps/chosen": -260.2576599121094, | |
| "eval_logps/rejected": -231.54052734375, | |
| "eval_loss": 2494.119384765625, | |
| "eval_rewards/accuracies": 0.5479999780654907, | |
| "eval_rewards/chosen": -0.00013494741870090365, | |
| "eval_rewards/margins": 0.0009091641986742616, | |
| "eval_rewards/rejected": -0.0010441114427521825, | |
| "eval_runtime": 325.9254, | |
| "eval_samples_per_second": 6.136, | |
| "eval_steps_per_second": 0.384, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.996716052911017e-06, | |
| "logits/chosen": 0.8349224328994751, | |
| "logits/rejected": 0.8758266568183899, | |
| "logps/chosen": -264.0060119628906, | |
| "logps/rejected": -219.6302490234375, | |
| "loss": 2491.0754, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": 0.00016143513494171202, | |
| "rewards/margins": 0.0016861247131600976, | |
| "rewards/rejected": -0.001524689607322216, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.9903533134293035e-06, | |
| "logits/chosen": 0.8607719540596008, | |
| "logits/rejected": 0.9709636569023132, | |
| "logps/chosen": -254.9365234375, | |
| "logps/rejected": -219.2154541015625, | |
| "loss": 2483.0162, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -4.470603380468674e-05, | |
| "rewards/margins": 0.0026740250177681446, | |
| "rewards/rejected": -0.002718730829656124, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.9806521797692184e-06, | |
| "logits/chosen": 0.8791080713272095, | |
| "logits/rejected": 0.8794806599617004, | |
| "logps/chosen": -264.69219970703125, | |
| "logps/rejected": -247.05224609375, | |
| "loss": 2478.3258, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.000567199953366071, | |
| "rewards/margins": 0.0015755310887470841, | |
| "rewards/rejected": -0.002142731100320816, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.967625656594782e-06, | |
| "logits/chosen": 0.8544095754623413, | |
| "logits/rejected": 0.9317782521247864, | |
| "logps/chosen": -222.06851196289062, | |
| "logps/rejected": -232.3370819091797, | |
| "loss": 2474.4896, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.0013934863964095712, | |
| "rewards/margins": 0.0016968228155747056, | |
| "rewards/rejected": -0.0030903094448149204, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.95129120635556e-06, | |
| "logits/chosen": 0.8754276037216187, | |
| "logits/rejected": 0.905910849571228, | |
| "logps/chosen": -258.3515625, | |
| "logps/rejected": -215.86328125, | |
| "loss": 2470.2957, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": -0.0012219983618706465, | |
| "rewards/margins": 0.0023179189302027225, | |
| "rewards/rejected": -0.003539917292073369, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.93167072587771e-06, | |
| "logits/chosen": 0.7798808813095093, | |
| "logits/rejected": 0.8481999635696411, | |
| "logps/chosen": -257.81036376953125, | |
| "logps/rejected": -250.8020477294922, | |
| "loss": 2466.5924, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.0015566629590466619, | |
| "rewards/margins": 0.002945856424048543, | |
| "rewards/rejected": -0.004502518568187952, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.908790517010637e-06, | |
| "logits/chosen": 0.9313274621963501, | |
| "logits/rejected": 0.9725171327590942, | |
| "logps/chosen": -239.09017944335938, | |
| "logps/rejected": -253.09255981445312, | |
| "loss": 2458.4684, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.0023754839785397053, | |
| "rewards/margins": 0.003910133149474859, | |
| "rewards/rejected": -0.006285616662353277, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 4.882681251368549e-06, | |
| "logits/chosen": 0.8546341061592102, | |
| "logits/rejected": 0.8491582870483398, | |
| "logps/chosen": -270.9264831542969, | |
| "logps/rejected": -257.13836669921875, | |
| "loss": 2442.4984, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.0015862795989960432, | |
| "rewards/margins": 0.005088582634925842, | |
| "rewards/rejected": -0.006674862466752529, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 4.853377929214243e-06, | |
| "logits/chosen": 0.7889066934585571, | |
| "logits/rejected": 0.8611849546432495, | |
| "logps/chosen": -252.2519989013672, | |
| "logps/rejected": -245.2388458251953, | |
| "loss": 2436.8207, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.0018671129364520311, | |
| "rewards/margins": 0.0070212846621870995, | |
| "rewards/rejected": -0.008888397365808487, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 4.8209198325401815e-06, | |
| "logits/chosen": 0.8786072731018066, | |
| "logits/rejected": 0.8954092264175415, | |
| "logps/chosen": -237.77017211914062, | |
| "logps/rejected": -233.86325073242188, | |
| "loss": 2425.7957, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.0038148313760757446, | |
| "rewards/margins": 0.00766246672719717, | |
| "rewards/rejected": -0.01147729717195034, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": 0.8182709813117981, | |
| "eval_logits/rejected": 0.9217536449432373, | |
| "eval_logps/chosen": -260.7673034667969, | |
| "eval_logps/rejected": -232.97280883789062, | |
| "eval_loss": 2420.32958984375, | |
| "eval_rewards/accuracies": 0.656000018119812, | |
| "eval_rewards/chosen": -0.00523131899535656, | |
| "eval_rewards/margins": 0.010135524906218052, | |
| "eval_rewards/rejected": -0.015366843901574612, | |
| "eval_runtime": 326.0274, | |
| "eval_samples_per_second": 6.134, | |
| "eval_steps_per_second": 0.383, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 4.785350472409792e-06, | |
| "logits/chosen": 0.9068363904953003, | |
| "logits/rejected": 0.911398708820343, | |
| "logps/chosen": -233.5305633544922, | |
| "logps/rejected": -229.12158203125, | |
| "loss": 2425.1516, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.009300420060753822, | |
| "rewards/margins": 0.007756868842989206, | |
| "rewards/rejected": -0.01705729030072689, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 4.746717530629565e-06, | |
| "logits/chosen": 0.8564063906669617, | |
| "logits/rejected": 0.8920175433158875, | |
| "logps/chosen": -260.2020263671875, | |
| "logps/rejected": -240.0993194580078, | |
| "loss": 2418.3254, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.008474646136164665, | |
| "rewards/margins": 0.012274968437850475, | |
| "rewards/rejected": -0.020749617367982864, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 4.7050727958301505e-06, | |
| "logits/chosen": 0.8939176797866821, | |
| "logits/rejected": 0.8774939775466919, | |
| "logps/chosen": -245.0527801513672, | |
| "logps/rejected": -230.6669921875, | |
| "loss": 2386.8547, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.011173086240887642, | |
| "rewards/margins": 0.013905773870646954, | |
| "rewards/rejected": -0.02507885918021202, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 4.660472094042121e-06, | |
| "logits/chosen": 0.8556815385818481, | |
| "logits/rejected": 0.863630473613739, | |
| "logps/chosen": -282.67669677734375, | |
| "logps/rejected": -236.6685028076172, | |
| "loss": 2355.6057, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.013796107843518257, | |
| "rewards/margins": 0.017949409782886505, | |
| "rewards/rejected": -0.03174551948904991, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 4.612975213859487e-06, | |
| "logits/chosen": 0.8427609205245972, | |
| "logits/rejected": 0.915958046913147, | |
| "logps/chosen": -269.03228759765625, | |
| "logps/rejected": -244.0561981201172, | |
| "loss": 2353.825, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.014950485900044441, | |
| "rewards/margins": 0.016987096518278122, | |
| "rewards/rejected": -0.031937580555677414, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 4.5626458262912745e-06, | |
| "logits/chosen": 0.8542389869689941, | |
| "logits/rejected": 0.8951870203018188, | |
| "logps/chosen": -275.2603454589844, | |
| "logps/rejected": -262.59820556640625, | |
| "loss": 2344.6988, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.010937584564089775, | |
| "rewards/margins": 0.02299124002456665, | |
| "rewards/rejected": -0.033928822726011276, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 4.509551399408598e-06, | |
| "logits/chosen": 0.9320127367973328, | |
| "logits/rejected": 0.9399789571762085, | |
| "logps/chosen": -254.00363159179688, | |
| "logps/rejected": -211.70822143554688, | |
| "loss": 2326.323, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.017307719215750694, | |
| "rewards/margins": 0.020856201648712158, | |
| "rewards/rejected": -0.038163922727108, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 4.453763107901676e-06, | |
| "logits/chosen": 0.9306808710098267, | |
| "logits/rejected": 0.8955329060554504, | |
| "logps/chosen": -248.00350952148438, | |
| "logps/rejected": -256.06072998046875, | |
| "loss": 2353.0586, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.0263187438249588, | |
| "rewards/margins": 0.0119154192507267, | |
| "rewards/rejected": -0.0382341630756855, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 4.3953557376679856e-06, | |
| "logits/chosen": 0.8539811372756958, | |
| "logits/rejected": 0.860480785369873, | |
| "logps/chosen": -262.2434387207031, | |
| "logps/rejected": -258.9552917480469, | |
| "loss": 2328.1365, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.026760926470160484, | |
| "rewards/margins": 0.01638859696686268, | |
| "rewards/rejected": -0.04314952343702316, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 4.33440758555951e-06, | |
| "logits/chosen": 0.8304749727249146, | |
| "logits/rejected": 0.9085506200790405, | |
| "logps/chosen": -251.0150604248047, | |
| "logps/rejected": -247.31289672851562, | |
| "loss": 2310.102, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.019448721781373024, | |
| "rewards/margins": 0.029306888580322266, | |
| "rewards/rejected": -0.04875560849905014, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": 0.8040502071380615, | |
| "eval_logits/rejected": 0.9088209271430969, | |
| "eval_logps/chosen": -263.2439880371094, | |
| "eval_logps/rejected": -237.19593811035156, | |
| "eval_loss": 2309.945068359375, | |
| "eval_rewards/accuracies": 0.6679999828338623, | |
| "eval_rewards/chosen": -0.02999839559197426, | |
| "eval_rewards/margins": 0.02759976126253605, | |
| "eval_rewards/rejected": -0.05759815126657486, | |
| "eval_runtime": 325.793, | |
| "eval_samples_per_second": 6.139, | |
| "eval_steps_per_second": 0.384, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 4.2710003544234255e-06, | |
| "logits/chosen": 0.8506165742874146, | |
| "logits/rejected": 0.8639786839485168, | |
| "logps/chosen": -238.32284545898438, | |
| "logps/rejected": -230.3314971923828, | |
| "loss": 2280.0254, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.034138236194849014, | |
| "rewards/margins": 0.023389272391796112, | |
| "rewards/rejected": -0.05752750486135483, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 4.205219043576955e-06, | |
| "logits/chosen": 0.8560878038406372, | |
| "logits/rejected": 0.878685474395752, | |
| "logps/chosen": -226.87521362304688, | |
| "logps/rejected": -220.47201538085938, | |
| "loss": 2272.9057, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.03677995875477791, | |
| "rewards/margins": 0.026556292548775673, | |
| "rewards/rejected": -0.06333625316619873, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 4.137151834863213e-06, | |
| "logits/chosen": 0.8369635343551636, | |
| "logits/rejected": 0.8179060220718384, | |
| "logps/chosen": -255.9115753173828, | |
| "logps/rejected": -263.99334716796875, | |
| "loss": 2288.366, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.03619007021188736, | |
| "rewards/margins": 0.036657560616731644, | |
| "rewards/rejected": -0.07284761965274811, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 4.066889974440757e-06, | |
| "logits/chosen": 0.8635396957397461, | |
| "logits/rejected": 0.8681753873825073, | |
| "logps/chosen": -268.9547119140625, | |
| "logps/rejected": -254.25814819335938, | |
| "loss": 2217.676, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.03367992490530014, | |
| "rewards/margins": 0.03633836284279823, | |
| "rewards/rejected": -0.07001828402280807, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 3.994527650465352e-06, | |
| "logits/chosen": 0.8572274446487427, | |
| "logits/rejected": 0.8776391744613647, | |
| "logps/chosen": -226.3207550048828, | |
| "logps/rejected": -211.4627685546875, | |
| "loss": 2251.0934, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.046974435448646545, | |
| "rewards/margins": 0.03458600863814354, | |
| "rewards/rejected": -0.08156044781208038, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 3.92016186682789e-06, | |
| "logits/chosen": 0.7831335067749023, | |
| "logits/rejected": 0.7933910489082336, | |
| "logps/chosen": -213.00491333007812, | |
| "logps/rejected": -242.6522979736328, | |
| "loss": 2245.7613, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.05574915558099747, | |
| "rewards/margins": 0.030930276960134506, | |
| "rewards/rejected": -0.08667943626642227, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 3.843892313117724e-06, | |
| "logits/chosen": 0.8648300170898438, | |
| "logits/rejected": 0.895352840423584, | |
| "logps/chosen": -272.79400634765625, | |
| "logps/rejected": -249.0972900390625, | |
| "loss": 2275.2465, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.05149676650762558, | |
| "rewards/margins": 0.0362628772854805, | |
| "rewards/rejected": -0.08775965869426727, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 3.7658212309857576e-06, | |
| "logits/chosen": 0.8409647941589355, | |
| "logits/rejected": 0.8629050254821777, | |
| "logps/chosen": -243.03182983398438, | |
| "logps/rejected": -222.12338256835938, | |
| "loss": 2199.2453, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.053347665816545486, | |
| "rewards/margins": 0.0404227040708065, | |
| "rewards/rejected": -0.09377036988735199, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 3.686053277086401e-06, | |
| "logits/chosen": 0.7878540754318237, | |
| "logits/rejected": 0.875022292137146, | |
| "logps/chosen": -266.7389831542969, | |
| "logps/rejected": -244.7388916015625, | |
| "loss": 2180.4859, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.05637942627072334, | |
| "rewards/margins": 0.04192977398633957, | |
| "rewards/rejected": -0.09830919653177261, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.604695382782159e-06, | |
| "logits/chosen": 0.7989987134933472, | |
| "logits/rejected": 0.8174804449081421, | |
| "logps/chosen": -282.3369445800781, | |
| "logps/rejected": -260.7422790527344, | |
| "loss": 2159.0707, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.05871356651186943, | |
| "rewards/margins": 0.03764244168996811, | |
| "rewards/rejected": -0.09635601192712784, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": 0.7577926516532898, | |
| "eval_logits/rejected": 0.8636941313743591, | |
| "eval_logps/chosen": -266.5838623046875, | |
| "eval_logps/rejected": -242.28573608398438, | |
| "eval_loss": 2236.27587890625, | |
| "eval_rewards/accuracies": 0.6840000152587891, | |
| "eval_rewards/chosen": -0.06339714676141739, | |
| "eval_rewards/margins": 0.04509904235601425, | |
| "eval_rewards/rejected": -0.10849618166685104, | |
| "eval_runtime": 325.6827, | |
| "eval_samples_per_second": 6.141, | |
| "eval_steps_per_second": 0.384, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.5218566107988872e-06, | |
| "logits/chosen": 0.8149998784065247, | |
| "logits/rejected": 0.8398739695549011, | |
| "logps/chosen": -274.44647216796875, | |
| "logps/rejected": -244.11441040039062, | |
| "loss": 2235.4445, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.060248058289289474, | |
| "rewards/margins": 0.0490226149559021, | |
| "rewards/rejected": -0.10927066951990128, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 3.437648009023905e-06, | |
| "logits/chosen": 0.7635517716407776, | |
| "logits/rejected": 0.8307437896728516, | |
| "logps/chosen": -219.5519256591797, | |
| "logps/rejected": -215.0430145263672, | |
| "loss": 2210.1293, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.07285571843385696, | |
| "rewards/margins": 0.03309701010584831, | |
| "rewards/rejected": -0.10595273971557617, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 3.352182461642929e-06, | |
| "logits/chosen": 0.7868200540542603, | |
| "logits/rejected": 0.8455543518066406, | |
| "logps/chosen": -240.498779296875, | |
| "logps/rejected": -228.08535766601562, | |
| "loss": 2132.6711, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.06197371333837509, | |
| "rewards/margins": 0.048869095742702484, | |
| "rewards/rejected": -0.11084280163049698, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 3.265574537815398e-06, | |
| "logits/chosen": 0.7955938577651978, | |
| "logits/rejected": 0.8357489705085754, | |
| "logps/chosen": -285.87493896484375, | |
| "logps/rejected": -252.0729522705078, | |
| "loss": 2188.6035, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.05718296021223068, | |
| "rewards/margins": 0.05849381536245346, | |
| "rewards/rejected": -0.11567678302526474, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 3.177940338091043e-06, | |
| "logits/chosen": 0.8076552152633667, | |
| "logits/rejected": 0.8619295954704285, | |
| "logps/chosen": -259.4091796875, | |
| "logps/rejected": -231.5826873779297, | |
| "loss": 2190.8027, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.07417537271976471, | |
| "rewards/margins": 0.042368099093437195, | |
| "rewards/rejected": -0.1165434867143631, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 3.089397338773569e-06, | |
| "logits/chosen": 0.787534236907959, | |
| "logits/rejected": 0.8202370405197144, | |
| "logps/chosen": -265.3134460449219, | |
| "logps/rejected": -237.35116577148438, | |
| "loss": 2146.2295, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.0778832882642746, | |
| "rewards/margins": 0.05344442278146744, | |
| "rewards/rejected": -0.13132771849632263, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 3.0000642344401115e-06, | |
| "logits/chosen": 0.7577365040779114, | |
| "logits/rejected": 0.803280234336853, | |
| "logps/chosen": -246.31301879882812, | |
| "logps/rejected": -235.45675659179688, | |
| "loss": 2136.8623, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.0812118723988533, | |
| "rewards/margins": 0.04804684966802597, | |
| "rewards/rejected": -0.12925872206687927, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 2.9100607788275547e-06, | |
| "logits/chosen": 0.7916151881217957, | |
| "logits/rejected": 0.8478500247001648, | |
| "logps/chosen": -251.5865020751953, | |
| "logps/rejected": -246.8653564453125, | |
| "loss": 2219.0154, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.08595123142004013, | |
| "rewards/margins": 0.04087045416235924, | |
| "rewards/rejected": -0.12682169675827026, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "learning_rate": 2.8195076242990124e-06, | |
| "logits/chosen": 0.7650834321975708, | |
| "logits/rejected": 0.8132265210151672, | |
| "logps/chosen": -246.82644653320312, | |
| "logps/rejected": -227.9211883544922, | |
| "loss": 2162.0559, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.0883156806230545, | |
| "rewards/margins": 0.050779860466718674, | |
| "rewards/rejected": -0.13909552991390228, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "learning_rate": 2.72852616010567e-06, | |
| "logits/chosen": 0.7439101338386536, | |
| "logits/rejected": 0.7856588363647461, | |
| "logps/chosen": -261.5977478027344, | |
| "logps/rejected": -247.032470703125, | |
| "loss": 2176.8641, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.08485061675310135, | |
| "rewards/margins": 0.05461747199296951, | |
| "rewards/rejected": -0.13946808874607086, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "eval_logits/chosen": 0.7125015258789062, | |
| "eval_logits/rejected": 0.8179839849472046, | |
| "eval_logps/chosen": -269.2715759277344, | |
| "eval_logps/rejected": -246.0634002685547, | |
| "eval_loss": 2197.5419921875, | |
| "eval_rewards/accuracies": 0.6980000138282776, | |
| "eval_rewards/chosen": -0.09027400612831116, | |
| "eval_rewards/margins": 0.05599898844957352, | |
| "eval_rewards/rejected": -0.14627300202846527, | |
| "eval_runtime": 325.7893, | |
| "eval_samples_per_second": 6.139, | |
| "eval_steps_per_second": 0.384, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 2.637238349660819e-06, | |
| "logits/chosen": 0.7544692754745483, | |
| "logits/rejected": 0.8448120355606079, | |
| "logps/chosen": -245.3660430908203, | |
| "logps/rejected": -210.2096405029297, | |
| "loss": 2188.5398, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.0969894677400589, | |
| "rewards/margins": 0.04615020379424095, | |
| "rewards/rejected": -0.14313964545726776, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 2.5457665670441937e-06, | |
| "logits/chosen": 0.8052291870117188, | |
| "logits/rejected": 0.8244439959526062, | |
| "logps/chosen": -257.635009765625, | |
| "logps/rejected": -238.87258911132812, | |
| "loss": 2140.0687, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.082728311419487, | |
| "rewards/margins": 0.061602912843227386, | |
| "rewards/rejected": -0.14433124661445618, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 2.4542334329558075e-06, | |
| "logits/chosen": 0.7256805896759033, | |
| "logits/rejected": 0.7552824020385742, | |
| "logps/chosen": -250.42422485351562, | |
| "logps/rejected": -241.7630615234375, | |
| "loss": 2133.6596, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.0894266813993454, | |
| "rewards/margins": 0.05599946528673172, | |
| "rewards/rejected": -0.14542615413665771, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "learning_rate": 2.3627616503391813e-06, | |
| "logits/chosen": 0.7206599116325378, | |
| "logits/rejected": 0.7505000233650208, | |
| "logps/chosen": -267.65875244140625, | |
| "logps/rejected": -228.6749725341797, | |
| "loss": 2182.2404, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.09456731379032135, | |
| "rewards/margins": 0.04886298626661301, | |
| "rewards/rejected": -0.14343029260635376, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "learning_rate": 2.271473839894331e-06, | |
| "logits/chosen": 0.7229181528091431, | |
| "logits/rejected": 0.749284029006958, | |
| "logps/chosen": -276.0394287109375, | |
| "logps/rejected": -262.62982177734375, | |
| "loss": 2171.0582, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.09430189430713654, | |
| "rewards/margins": 0.05570146441459656, | |
| "rewards/rejected": -0.1500033438205719, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 2.1804923757009885e-06, | |
| "logits/chosen": 0.694362998008728, | |
| "logits/rejected": 0.7259857654571533, | |
| "logps/chosen": -261.21038818359375, | |
| "logps/rejected": -239.57852172851562, | |
| "loss": 2163.352, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.100721076130867, | |
| "rewards/margins": 0.05086208134889603, | |
| "rewards/rejected": -0.15158315002918243, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 2.089939221172446e-06, | |
| "logits/chosen": 0.7141777276992798, | |
| "logits/rejected": 0.7269617915153503, | |
| "logps/chosen": -279.9233093261719, | |
| "logps/rejected": -245.4532928466797, | |
| "loss": 2219.2641, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.10629091411828995, | |
| "rewards/margins": 0.06119798496365547, | |
| "rewards/rejected": -0.16748890280723572, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 1.9999357655598894e-06, | |
| "logits/chosen": 0.7328698635101318, | |
| "logits/rejected": 0.7043576240539551, | |
| "logps/chosen": -250.1948699951172, | |
| "logps/rejected": -239.837890625, | |
| "loss": 2180.0646, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.10338902473449707, | |
| "rewards/margins": 0.0403132289648056, | |
| "rewards/rejected": -0.14370223879814148, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "learning_rate": 1.9106026612264316e-06, | |
| "logits/chosen": 0.7637673616409302, | |
| "logits/rejected": 0.8286052942276001, | |
| "logps/chosen": -227.8063507080078, | |
| "logps/rejected": -227.64193725585938, | |
| "loss": 2125.3691, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.10240204632282257, | |
| "rewards/margins": 0.04882645606994629, | |
| "rewards/rejected": -0.15122851729393005, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "learning_rate": 1.8220596619089576e-06, | |
| "logits/chosen": 0.7344295978546143, | |
| "logits/rejected": 0.7085897922515869, | |
| "logps/chosen": -271.92706298828125, | |
| "logps/rejected": -245.1657257080078, | |
| "loss": 2066.3285, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.10316000878810883, | |
| "rewards/margins": 0.061250198632478714, | |
| "rewards/rejected": -0.16441020369529724, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "eval_logits/chosen": 0.6878580451011658, | |
| "eval_logits/rejected": 0.7927125096321106, | |
| "eval_logps/chosen": -270.385498046875, | |
| "eval_logps/rejected": -247.7128448486328, | |
| "eval_loss": 2177.3388671875, | |
| "eval_rewards/accuracies": 0.6959999799728394, | |
| "eval_rewards/chosen": -0.10141333192586899, | |
| "eval_rewards/margins": 0.0613539032638073, | |
| "eval_rewards/rejected": -0.16276724636554718, | |
| "eval_runtime": 325.4497, | |
| "eval_samples_per_second": 6.145, | |
| "eval_steps_per_second": 0.384, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 1.7344254621846018e-06, | |
| "logits/chosen": 0.7047083377838135, | |
| "logits/rejected": 0.7504035830497742, | |
| "logps/chosen": -273.6443176269531, | |
| "logps/rejected": -265.7873840332031, | |
| "loss": 2070.9469, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.09533126652240753, | |
| "rewards/margins": 0.06477675586938858, | |
| "rewards/rejected": -0.16010800004005432, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 1.647817538357072e-06, | |
| "logits/chosen": 0.7320979237556458, | |
| "logits/rejected": 0.8024471998214722, | |
| "logps/chosen": -257.2734680175781, | |
| "logps/rejected": -243.01651000976562, | |
| "loss": 2189.2789, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.10891245305538177, | |
| "rewards/margins": 0.04389963299036026, | |
| "rewards/rejected": -0.15281209349632263, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 1.5623519909760953e-06, | |
| "logits/chosen": 0.7088596820831299, | |
| "logits/rejected": 0.7686936259269714, | |
| "logps/chosen": -253.4821014404297, | |
| "logps/rejected": -253.3447723388672, | |
| "loss": 2140.6246, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.11572308838367462, | |
| "rewards/margins": 0.046286530792713165, | |
| "rewards/rejected": -0.16200962662696838, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "learning_rate": 1.4781433892011132e-06, | |
| "logits/chosen": 0.7846838235855103, | |
| "logits/rejected": 0.8031150698661804, | |
| "logps/chosen": -251.593505859375, | |
| "logps/rejected": -272.357177734375, | |
| "loss": 2168.2102, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.10245855897665024, | |
| "rewards/margins": 0.04714034125208855, | |
| "rewards/rejected": -0.1495988965034485, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "learning_rate": 1.3953046172178413e-06, | |
| "logits/chosen": 0.7221434116363525, | |
| "logits/rejected": 0.7379263043403625, | |
| "logps/chosen": -264.2921142578125, | |
| "logps/rejected": -261.3705749511719, | |
| "loss": 2137.5834, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.11888917535543442, | |
| "rewards/margins": 0.05505634471774101, | |
| "rewards/rejected": -0.17394550144672394, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "learning_rate": 1.3139467229135999e-06, | |
| "logits/chosen": 0.6965005397796631, | |
| "logits/rejected": 0.7608405351638794, | |
| "logps/chosen": -230.50830078125, | |
| "logps/rejected": -261.0777282714844, | |
| "loss": 2178.6738, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.10801998525857925, | |
| "rewards/margins": 0.049364469945430756, | |
| "rewards/rejected": -0.15738445520401, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 1.2341787690142436e-06, | |
| "logits/chosen": 0.6697909235954285, | |
| "logits/rejected": 0.7529794573783875, | |
| "logps/chosen": -309.4987487792969, | |
| "logps/rejected": -255.5410919189453, | |
| "loss": 2163.5801, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.09704665839672089, | |
| "rewards/margins": 0.06983451545238495, | |
| "rewards/rejected": -0.16688117384910583, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "learning_rate": 1.1561076868822756e-06, | |
| "logits/chosen": 0.6984275579452515, | |
| "logits/rejected": 0.7366929650306702, | |
| "logps/chosen": -253.47988891601562, | |
| "logps/rejected": -230.54940795898438, | |
| "loss": 2114.4482, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.10325287282466888, | |
| "rewards/margins": 0.05643494054675102, | |
| "rewards/rejected": -0.1596878170967102, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "learning_rate": 1.079838133172111e-06, | |
| "logits/chosen": 0.760990560054779, | |
| "logits/rejected": 0.7594455480575562, | |
| "logps/chosen": -267.423583984375, | |
| "logps/rejected": -251.7685089111328, | |
| "loss": 2198.5418, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.0984167829155922, | |
| "rewards/margins": 0.05027080327272415, | |
| "rewards/rejected": -0.14868757128715515, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "learning_rate": 1.0054723495346484e-06, | |
| "logits/chosen": 0.6380269527435303, | |
| "logits/rejected": 0.7157927751541138, | |
| "logps/chosen": -269.24908447265625, | |
| "logps/rejected": -241.45919799804688, | |
| "loss": 2119.5369, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.10670281946659088, | |
| "rewards/margins": 0.05142299085855484, | |
| "rewards/rejected": -0.15812578797340393, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "eval_logits/chosen": 0.6726287007331848, | |
| "eval_logits/rejected": 0.7770729660987854, | |
| "eval_logps/chosen": -270.7824401855469, | |
| "eval_logps/rejected": -248.45333862304688, | |
| "eval_loss": 2166.385498046875, | |
| "eval_rewards/accuracies": 0.6959999799728394, | |
| "eval_rewards/chosen": -0.10538262128829956, | |
| "eval_rewards/margins": 0.06478944420814514, | |
| "eval_rewards/rejected": -0.17017203569412231, | |
| "eval_runtime": 325.443, | |
| "eval_samples_per_second": 6.145, | |
| "eval_steps_per_second": 0.384, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "learning_rate": 9.331100255592437e-07, | |
| "logits/chosen": 0.7291372418403625, | |
| "logits/rejected": 0.7911130785942078, | |
| "logps/chosen": -288.43621826171875, | |
| "logps/rejected": -244.11181640625, | |
| "loss": 2094.7482, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.09699388593435287, | |
| "rewards/margins": 0.0718715712428093, | |
| "rewards/rejected": -0.16886545717716217, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 8.628481651367876e-07, | |
| "logits/chosen": 0.7129195928573608, | |
| "logits/rejected": 0.7519146203994751, | |
| "logps/chosen": -273.88372802734375, | |
| "logps/rejected": -259.33465576171875, | |
| "loss": 2053.8975, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.10259418189525604, | |
| "rewards/margins": 0.06398696452379227, | |
| "rewards/rejected": -0.1665811538696289, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "learning_rate": 7.947809564230446e-07, | |
| "logits/chosen": 0.7380334138870239, | |
| "logits/rejected": 0.7657966017723083, | |
| "logps/chosen": -275.5352478027344, | |
| "logps/rejected": -267.64544677734375, | |
| "loss": 2090.3867, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.10346021503210068, | |
| "rewards/margins": 0.06666620075702667, | |
| "rewards/rejected": -0.17012640833854675, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "learning_rate": 7.289996455765749e-07, | |
| "logits/chosen": 0.694438099861145, | |
| "logits/rejected": 0.7096751928329468, | |
| "logps/chosen": -285.7628479003906, | |
| "logps/rejected": -266.2068176269531, | |
| "loss": 2179.6143, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.1050759106874466, | |
| "rewards/margins": 0.04846997186541557, | |
| "rewards/rejected": -0.15354588627815247, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "learning_rate": 6.655924144404907e-07, | |
| "logits/chosen": 0.6481191515922546, | |
| "logits/rejected": 0.7124420404434204, | |
| "logps/chosen": -269.7061767578125, | |
| "logps/rejected": -250.74905395507812, | |
| "loss": 2143.8697, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.09301155805587769, | |
| "rewards/margins": 0.06966546177864075, | |
| "rewards/rejected": -0.16267701983451843, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "learning_rate": 6.046442623320145e-07, | |
| "logits/chosen": 0.7415434122085571, | |
| "logits/rejected": 0.77794349193573, | |
| "logps/chosen": -267.1199035644531, | |
| "logps/rejected": -233.93765258789062, | |
| "loss": 2100.39, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.11892955005168915, | |
| "rewards/margins": 0.048730865120887756, | |
| "rewards/rejected": -0.1676604300737381, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 5.462368920983249e-07, | |
| "logits/chosen": 0.7082683444023132, | |
| "logits/rejected": 0.7616415023803711, | |
| "logps/chosen": -252.79940795898438, | |
| "logps/rejected": -231.410400390625, | |
| "loss": 2186.8764, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.11736402660608292, | |
| "rewards/margins": 0.03412212058901787, | |
| "rewards/rejected": -0.1514861285686493, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "learning_rate": 4.904486005914027e-07, | |
| "logits/chosen": 0.7068939805030823, | |
| "logits/rejected": 0.7558518052101135, | |
| "logps/chosen": -293.97174072265625, | |
| "logps/rejected": -274.1463928222656, | |
| "loss": 2085.8768, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.0985812246799469, | |
| "rewards/margins": 0.07491330802440643, | |
| "rewards/rejected": -0.17349454760551453, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 4.373541737087264e-07, | |
| "logits/chosen": 0.6898752450942993, | |
| "logits/rejected": 0.7903083562850952, | |
| "logps/chosen": -275.10650634765625, | |
| "logps/rejected": -239.94729614257812, | |
| "loss": 2145.707, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.09976668655872345, | |
| "rewards/margins": 0.06850672513246536, | |
| "rewards/rejected": -0.16827340424060822, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "learning_rate": 3.8702478614051353e-07, | |
| "logits/chosen": 0.668254554271698, | |
| "logits/rejected": 0.6736531257629395, | |
| "logps/chosen": -248.01626586914062, | |
| "logps/rejected": -245.8029327392578, | |
| "loss": 2096.7854, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.10279978811740875, | |
| "rewards/margins": 0.07554516196250916, | |
| "rewards/rejected": -0.1783449351787567, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "eval_logits/chosen": 0.6641319990158081, | |
| "eval_logits/rejected": 0.7684468030929565, | |
| "eval_logps/chosen": -271.1501159667969, | |
| "eval_logps/rejected": -248.99647521972656, | |
| "eval_loss": 2159.71044921875, | |
| "eval_rewards/accuracies": 0.6959999799728394, | |
| "eval_rewards/chosen": -0.10905998200178146, | |
| "eval_rewards/margins": 0.06654350459575653, | |
| "eval_rewards/rejected": -0.1756034791469574, | |
| "eval_runtime": 325.5727, | |
| "eval_samples_per_second": 6.143, | |
| "eval_steps_per_second": 0.384, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "learning_rate": 3.3952790595787986e-07, | |
| "logits/chosen": 0.7833539247512817, | |
| "logits/rejected": 0.7525036931037903, | |
| "logps/chosen": -260.330810546875, | |
| "logps/rejected": -247.55337524414062, | |
| "loss": 2140.7756, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.11206915229558945, | |
| "rewards/margins": 0.04849858209490776, | |
| "rewards/rejected": -0.1605677306652069, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "learning_rate": 2.9492720416985004e-07, | |
| "logits/chosen": 0.6317057609558105, | |
| "logits/rejected": 0.6787868738174438, | |
| "logps/chosen": -283.4757080078125, | |
| "logps/rejected": -266.5099182128906, | |
| "loss": 2079.7037, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.09024739265441895, | |
| "rewards/margins": 0.06939631700515747, | |
| "rewards/rejected": -0.15964370965957642, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 2.5328246937043526e-07, | |
| "logits/chosen": 0.7278770208358765, | |
| "logits/rejected": 0.7481415867805481, | |
| "logps/chosen": -248.8776397705078, | |
| "logps/rejected": -233.3114776611328, | |
| "loss": 2126.3594, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.12706544995307922, | |
| "rewards/margins": 0.0513269305229187, | |
| "rewards/rejected": -0.17839238047599792, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "learning_rate": 2.1464952759020857e-07, | |
| "logits/chosen": 0.711814820766449, | |
| "logits/rejected": 0.7446814775466919, | |
| "logps/chosen": -260.2667541503906, | |
| "logps/rejected": -262.43939208984375, | |
| "loss": 2154.4643, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.10629498958587646, | |
| "rewards/margins": 0.06066171079874039, | |
| "rewards/rejected": -0.16695669293403625, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "learning_rate": 1.790801674598186e-07, | |
| "logits/chosen": 0.6856343746185303, | |
| "logits/rejected": 0.7043182849884033, | |
| "logps/chosen": -278.94085693359375, | |
| "logps/rejected": -251.6495819091797, | |
| "loss": 2141.4652, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.10617993772029877, | |
| "rewards/margins": 0.0626487284898758, | |
| "rewards/rejected": -0.16882868111133575, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 1.4662207078575685e-07, | |
| "logits/chosen": 0.6835039258003235, | |
| "logits/rejected": 0.7542312145233154, | |
| "logps/chosen": -242.60501098632812, | |
| "logps/rejected": -255.89266967773438, | |
| "loss": 2120.8357, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.10873384773731232, | |
| "rewards/margins": 0.06333796679973602, | |
| "rewards/rejected": -0.17207179963588715, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 1.1731874863145143e-07, | |
| "logits/chosen": 0.7217626571655273, | |
| "logits/rejected": 0.7742848992347717, | |
| "logps/chosen": -271.91668701171875, | |
| "logps/rejected": -239.55130004882812, | |
| "loss": 2097.9199, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.10101070255041122, | |
| "rewards/margins": 0.06306958198547363, | |
| "rewards/rejected": -0.16408027708530426, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "learning_rate": 9.120948298936422e-08, | |
| "logits/chosen": 0.7404865026473999, | |
| "logits/rejected": 0.7562910914421082, | |
| "logps/chosen": -260.2977294921875, | |
| "logps/rejected": -240.30502319335938, | |
| "loss": 2009.9664, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.10928317159414291, | |
| "rewards/margins": 0.05941414088010788, | |
| "rewards/rejected": -0.1686973124742508, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 6.832927412229017e-08, | |
| "logits/chosen": 0.6707200407981873, | |
| "logits/rejected": 0.7818647623062134, | |
| "logps/chosen": -271.86065673828125, | |
| "logps/rejected": -243.53182983398438, | |
| "loss": 2081.6182, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.10177429020404816, | |
| "rewards/margins": 0.06641246378421783, | |
| "rewards/rejected": -0.16818676888942719, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "learning_rate": 4.870879364444109e-08, | |
| "logits/chosen": 0.6861158609390259, | |
| "logits/rejected": 0.7574108839035034, | |
| "logps/chosen": -261.988525390625, | |
| "logps/rejected": -223.02371215820312, | |
| "loss": 2094.5041, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.10657148063182831, | |
| "rewards/margins": 0.07416997849941254, | |
| "rewards/rejected": -0.18074145913124084, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "eval_logits/chosen": 0.6645969748497009, | |
| "eval_logits/rejected": 0.7690178155899048, | |
| "eval_logps/chosen": -271.27447509765625, | |
| "eval_logps/rejected": -249.114013671875, | |
| "eval_loss": 2158.6298828125, | |
| "eval_rewards/accuracies": 0.6980000138282776, | |
| "eval_rewards/chosen": -0.11030303686857224, | |
| "eval_rewards/margins": 0.06647594273090363, | |
| "eval_rewards/rejected": -0.17677900195121765, | |
| "eval_runtime": 325.3231, | |
| "eval_samples_per_second": 6.148, | |
| "eval_steps_per_second": 0.384, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "learning_rate": 3.237434340521789e-08, | |
| "logits/chosen": 0.6756221055984497, | |
| "logits/rejected": 0.7528306245803833, | |
| "logps/chosen": -272.76104736328125, | |
| "logps/rejected": -262.8546447753906, | |
| "loss": 2083.5836, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.10101411491632462, | |
| "rewards/margins": 0.06638970226049423, | |
| "rewards/rejected": -0.16740381717681885, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "learning_rate": 1.93478202307823e-08, | |
| "logits/chosen": 0.6799421310424805, | |
| "logits/rejected": 0.7030155658721924, | |
| "logps/chosen": -252.42404174804688, | |
| "logps/rejected": -260.2074890136719, | |
| "loss": 2119.1219, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.11061377823352814, | |
| "rewards/margins": 0.04686294496059418, | |
| "rewards/rejected": -0.15747670829296112, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "learning_rate": 9.646686570697062e-09, | |
| "logits/chosen": 0.7364012002944946, | |
| "logits/rejected": 0.7563687562942505, | |
| "logps/chosen": -267.6513366699219, | |
| "logps/rejected": -265.22686767578125, | |
| "loss": 2177.6246, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.11080427467823029, | |
| "rewards/margins": 0.05903978273272514, | |
| "rewards/rejected": -0.16984406113624573, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 3.283947088983663e-09, | |
| "logits/chosen": 0.7186964750289917, | |
| "logits/rejected": 0.7220372557640076, | |
| "logps/chosen": -249.5286407470703, | |
| "logps/rejected": -258.6728515625, | |
| "loss": 2177.0199, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.12288296222686768, | |
| "rewards/margins": 0.04801628738641739, | |
| "rewards/rejected": -0.17089924216270447, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 1.99, | |
| "learning_rate": 2.681312309735229e-10, | |
| "logits/chosen": 0.6878038644790649, | |
| "logits/rejected": 0.7842324376106262, | |
| "logps/chosen": -242.4557342529297, | |
| "logps/rejected": -244.35446166992188, | |
| "loss": 2182.6682, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.11521060764789581, | |
| "rewards/margins": 0.04983743280172348, | |
| "rewards/rejected": -0.1650480479001999, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 954, | |
| "total_flos": 0.0, | |
| "train_loss": 2246.599348344143, | |
| "train_runtime": 18130.6033, | |
| "train_samples_per_second": 3.372, | |
| "train_steps_per_second": 0.053 | |
| } | |
| ], | |
| "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 | |
| } | |