| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 4.0, | |
| "eval_steps": 100, | |
| "global_step": 1540, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.282051282051282e-07, | |
| "logits/chosen": -1.7278180122375488, | |
| "logits/rejected": -1.7377450466156006, | |
| "logps/chosen": -29.553977966308594, | |
| "logps/rejected": -42.813133239746094, | |
| "loss": 0.5, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.282051282051282e-06, | |
| "logits/chosen": -1.866452693939209, | |
| "logits/rejected": -1.8707809448242188, | |
| "logps/chosen": -36.98366165161133, | |
| "logps/rejected": -33.666290283203125, | |
| "loss": 0.4914, | |
| "rewards/accuracies": 0.5555555820465088, | |
| "rewards/chosen": 0.01834939420223236, | |
| "rewards/margins": 0.04732292518019676, | |
| "rewards/rejected": -0.02897353284060955, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.564102564102564e-06, | |
| "logits/chosen": -1.9971332550048828, | |
| "logits/rejected": -1.999770164489746, | |
| "logps/chosen": -29.6362247467041, | |
| "logps/rejected": -29.044448852539062, | |
| "loss": 0.5009, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.004776014480739832, | |
| "rewards/margins": -0.01042011845856905, | |
| "rewards/rejected": 0.015196132473647594, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 3.846153846153847e-06, | |
| "logits/chosen": -1.9208379983901978, | |
| "logits/rejected": -1.9181442260742188, | |
| "logps/chosen": -31.425155639648438, | |
| "logps/rejected": -33.237945556640625, | |
| "loss": 0.4997, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.007249836809933186, | |
| "rewards/margins": 0.008341209962964058, | |
| "rewards/rejected": -0.015591045841574669, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.999896948438434e-06, | |
| "logits/chosen": -2.0180625915527344, | |
| "logits/rejected": -2.009308338165283, | |
| "logps/chosen": -32.559749603271484, | |
| "logps/rejected": -32.524681091308594, | |
| "loss": 0.4962, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.013646525330841541, | |
| "rewards/margins": 0.023250887170433998, | |
| "rewards/rejected": -0.009604359045624733, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.987541037542187e-06, | |
| "logits/chosen": -1.8639278411865234, | |
| "logits/rejected": -1.8531490564346313, | |
| "logps/chosen": -33.53788375854492, | |
| "logps/rejected": -35.42852020263672, | |
| "loss": 0.5015, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.015677569434046745, | |
| "rewards/margins": -0.0033827773295342922, | |
| "rewards/rejected": 0.019060343503952026, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.954691471941119e-06, | |
| "logits/chosen": -1.9435577392578125, | |
| "logits/rejected": -1.9454940557479858, | |
| "logps/chosen": -32.49355697631836, | |
| "logps/rejected": -33.1856575012207, | |
| "loss": 0.4817, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.08530057966709137, | |
| "rewards/margins": 0.09672373533248901, | |
| "rewards/rejected": -0.011423162184655666, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.901618883413549e-06, | |
| "logits/chosen": -2.075368881225586, | |
| "logits/rejected": -2.0803427696228027, | |
| "logps/chosen": -33.89010238647461, | |
| "logps/rejected": -36.549957275390625, | |
| "loss": 0.4868, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.07870620489120483, | |
| "rewards/margins": 0.06357409805059433, | |
| "rewards/rejected": 0.015132094733417034, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.828760511501322e-06, | |
| "logits/chosen": -1.9372116327285767, | |
| "logits/rejected": -1.9403302669525146, | |
| "logps/chosen": -34.205604553222656, | |
| "logps/rejected": -34.49314880371094, | |
| "loss": 0.4788, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.17776472866535187, | |
| "rewards/margins": 0.09835849702358246, | |
| "rewards/rejected": 0.0794062465429306, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.7367166013034295e-06, | |
| "logits/chosen": -1.945416808128357, | |
| "logits/rejected": -1.9499365091323853, | |
| "logps/chosen": -32.26136016845703, | |
| "logps/rejected": -32.26020050048828, | |
| "loss": 0.4788, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.16484788060188293, | |
| "rewards/margins": 0.08588258177042007, | |
| "rewards/rejected": 0.07896529138088226, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.626245458345211e-06, | |
| "logits/chosen": -2.043586254119873, | |
| "logits/rejected": -2.0416104793548584, | |
| "logps/chosen": -31.968154907226562, | |
| "logps/rejected": -31.14252281188965, | |
| "loss": 0.4749, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.21814338862895966, | |
| "rewards/margins": 0.11919467151165009, | |
| "rewards/rejected": 0.09894871711730957, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_logits/chosen": -2.2375268936157227, | |
| "eval_logits/rejected": -2.2326817512512207, | |
| "eval_logps/chosen": -33.851219177246094, | |
| "eval_logps/rejected": -37.3537483215332, | |
| "eval_loss": 0.49629145860671997, | |
| "eval_rewards/accuracies": 0.5336378812789917, | |
| "eval_rewards/chosen": 0.14666710793972015, | |
| "eval_rewards/margins": 0.016369162127375603, | |
| "eval_rewards/rejected": 0.1302979290485382, | |
| "eval_runtime": 145.7938, | |
| "eval_samples_per_second": 2.353, | |
| "eval_steps_per_second": 0.295, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.498257201263691e-06, | |
| "logits/chosen": -1.998853087425232, | |
| "logits/rejected": -1.9964749813079834, | |
| "logps/chosen": -32.937965393066406, | |
| "logps/rejected": -33.846004486083984, | |
| "loss": 0.474, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.2446821928024292, | |
| "rewards/margins": 0.09798828512430191, | |
| "rewards/rejected": 0.1466939002275467, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.353806263777678e-06, | |
| "logits/chosen": -2.0093610286712646, | |
| "logits/rejected": -2.0010437965393066, | |
| "logps/chosen": -32.1511116027832, | |
| "logps/rejected": -31.939533233642578, | |
| "loss": 0.4873, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.2350343018770218, | |
| "rewards/margins": 0.057456426322460175, | |
| "rewards/rejected": 0.17757786810398102, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.1940827077152755e-06, | |
| "logits/chosen": -2.0374419689178467, | |
| "logits/rejected": -2.029500961303711, | |
| "logps/chosen": -30.157424926757812, | |
| "logps/rejected": -31.90500831604004, | |
| "loss": 0.4716, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.2546332776546478, | |
| "rewards/margins": 0.1362351030111313, | |
| "rewards/rejected": 0.11839816719293594, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.0204024186666215e-06, | |
| "logits/chosen": -1.9675471782684326, | |
| "logits/rejected": -1.9777450561523438, | |
| "logps/chosen": -31.059162139892578, | |
| "logps/rejected": -32.384300231933594, | |
| "loss": 0.4616, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.28839823603630066, | |
| "rewards/margins": 0.16793587803840637, | |
| "rewards/rejected": 0.12046238034963608, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 3.834196265035119e-06, | |
| "logits/chosen": -1.8793624639511108, | |
| "logits/rejected": -1.8805418014526367, | |
| "logps/chosen": -33.646915435791016, | |
| "logps/rejected": -34.56272888183594, | |
| "loss": 0.4385, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.4366222321987152, | |
| "rewards/margins": 0.2809991240501404, | |
| "rewards/rejected": 0.15562310814857483, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.636998309800573e-06, | |
| "logits/chosen": -1.9307949542999268, | |
| "logits/rejected": -1.9274256229400635, | |
| "logps/chosen": -35.765926361083984, | |
| "logps/rejected": -32.48193359375, | |
| "loss": 0.4685, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.32460084557533264, | |
| "rewards/margins": 0.13194666802883148, | |
| "rewards/rejected": 0.19265416264533997, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.4304331721118078e-06, | |
| "logits/chosen": -2.0318870544433594, | |
| "logits/rejected": -2.0245723724365234, | |
| "logps/chosen": -33.25593566894531, | |
| "logps/rejected": -31.189266204833984, | |
| "loss": 0.4315, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.4335424304008484, | |
| "rewards/margins": 0.3034728467464447, | |
| "rewards/rejected": 0.1300695389509201, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.2162026428305436e-06, | |
| "logits/chosen": -2.0387840270996094, | |
| "logits/rejected": -2.0439517498016357, | |
| "logps/chosen": -31.977941513061523, | |
| "logps/rejected": -32.183441162109375, | |
| "loss": 0.4499, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.4534761905670166, | |
| "rewards/margins": 0.2121172845363617, | |
| "rewards/rejected": 0.2413589060306549, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 2.996071664294641e-06, | |
| "logits/chosen": -2.039411783218384, | |
| "logits/rejected": -2.036684513092041, | |
| "logps/chosen": -31.094594955444336, | |
| "logps/rejected": -31.10666275024414, | |
| "loss": 0.4653, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.3154846429824829, | |
| "rewards/margins": 0.1462438404560089, | |
| "rewards/rejected": 0.169240802526474, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.7718537898066833e-06, | |
| "logits/chosen": -1.909691572189331, | |
| "logits/rejected": -1.914358139038086, | |
| "logps/chosen": -31.08693504333496, | |
| "logps/rejected": -32.61079406738281, | |
| "loss": 0.4376, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.41547876596450806, | |
| "rewards/margins": 0.27136775851249695, | |
| "rewards/rejected": 0.14411096274852753, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "eval_logits/chosen": -2.233895778656006, | |
| "eval_logits/rejected": -2.2290823459625244, | |
| "eval_logps/chosen": -33.78962326049805, | |
| "eval_logps/rejected": -37.29551696777344, | |
| "eval_loss": 0.4955826997756958, | |
| "eval_rewards/accuracies": 0.5485880374908447, | |
| "eval_rewards/chosen": 0.19594170153141022, | |
| "eval_rewards/margins": 0.019058095291256905, | |
| "eval_rewards/rejected": 0.17688362300395966, | |
| "eval_runtime": 145.7262, | |
| "eval_samples_per_second": 2.354, | |
| "eval_steps_per_second": 0.295, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2.5453962426402006e-06, | |
| "logits/chosen": -2.0220179557800293, | |
| "logits/rejected": -2.0326101779937744, | |
| "logps/chosen": -31.51509666442871, | |
| "logps/rejected": -33.655662536621094, | |
| "loss": 0.4554, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.36496439576148987, | |
| "rewards/margins": 0.19933317601680756, | |
| "rewards/rejected": 0.16563120484352112, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.3185646976551794e-06, | |
| "logits/chosen": -1.9142780303955078, | |
| "logits/rejected": -1.9289791584014893, | |
| "logps/chosen": -29.601001739501953, | |
| "logps/rejected": -31.418704986572266, | |
| "loss": 0.4339, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.4084321856498718, | |
| "rewards/margins": 0.2968555688858032, | |
| "rewards/rejected": 0.11157669872045517, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.0932279108998323e-06, | |
| "logits/chosen": -1.9722967147827148, | |
| "logits/rejected": -1.9762967824935913, | |
| "logps/chosen": -32.825157165527344, | |
| "logps/rejected": -31.445331573486328, | |
| "loss": 0.4142, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.47385087609291077, | |
| "rewards/margins": 0.39067354798316956, | |
| "rewards/rejected": 0.0831773579120636, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.8712423238279358e-06, | |
| "logits/chosen": -1.9708006381988525, | |
| "logits/rejected": -1.9490963220596313, | |
| "logps/chosen": -33.599205017089844, | |
| "logps/rejected": -34.856910705566406, | |
| "loss": 0.4183, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.4436335563659668, | |
| "rewards/margins": 0.36852845549583435, | |
| "rewards/rejected": 0.07510510087013245, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.6544367689701824e-06, | |
| "logits/chosen": -2.011307954788208, | |
| "logits/rejected": -2.008012533187866, | |
| "logps/chosen": -32.494956970214844, | |
| "logps/rejected": -35.98812484741211, | |
| "loss": 0.4641, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.3440879285335541, | |
| "rewards/margins": 0.15464913845062256, | |
| "rewards/rejected": 0.18943879008293152, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.4445974030621963e-06, | |
| "logits/chosen": -1.8780018091201782, | |
| "logits/rejected": -1.8755817413330078, | |
| "logps/chosen": -33.72126388549805, | |
| "logps/rejected": -35.290069580078125, | |
| "loss": 0.4544, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.37370795011520386, | |
| "rewards/margins": 0.20206721127033234, | |
| "rewards/rejected": 0.17164072394371033, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.243452991757889e-06, | |
| "logits/chosen": -1.8628206253051758, | |
| "logits/rejected": -1.8603187799453735, | |
| "logps/chosen": -33.929344177246094, | |
| "logps/rejected": -31.61127281188965, | |
| "loss": 0.4451, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.3746057152748108, | |
| "rewards/margins": 0.246560737490654, | |
| "rewards/rejected": 0.128044992685318, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0526606671603523e-06, | |
| "logits/chosen": -1.9666433334350586, | |
| "logits/rejected": -1.9561984539031982, | |
| "logps/chosen": -34.74022674560547, | |
| "logps/rejected": -31.632495880126953, | |
| "loss": 0.4239, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.4931296706199646, | |
| "rewards/margins": 0.3263343572616577, | |
| "rewards/rejected": 0.16679534316062927, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.737922755071455e-07, | |
| "logits/chosen": -2.0619094371795654, | |
| "logits/rejected": -2.0470869541168213, | |
| "logps/chosen": -30.43317222595215, | |
| "logps/rejected": -32.3420524597168, | |
| "loss": 0.4679, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.39276057481765747, | |
| "rewards/margins": 0.1449340283870697, | |
| "rewards/rejected": 0.24782654643058777, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 7.08321427484816e-07, | |
| "logits/chosen": -1.9332126379013062, | |
| "logits/rejected": -1.9307750463485718, | |
| "logps/chosen": -32.100502014160156, | |
| "logps/rejected": -30.6780948638916, | |
| "loss": 0.3835, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.6612862348556519, | |
| "rewards/margins": 0.5503143072128296, | |
| "rewards/rejected": 0.1109718531370163, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "eval_logits/chosen": -2.231203556060791, | |
| "eval_logits/rejected": -2.226395606994629, | |
| "eval_logps/chosen": -33.778892517089844, | |
| "eval_logps/rejected": -37.28717803955078, | |
| "eval_loss": 0.4949820339679718, | |
| "eval_rewards/accuracies": 0.5245016813278198, | |
| "eval_rewards/chosen": 0.204526886343956, | |
| "eval_rewards/margins": 0.020971858873963356, | |
| "eval_rewards/rejected": 0.1835550218820572, | |
| "eval_runtime": 145.6189, | |
| "eval_samples_per_second": 2.355, | |
| "eval_steps_per_second": 0.295, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 4.84533120650964e-06, | |
| "logits/chosen": -1.9174331426620483, | |
| "logits/rejected": -1.9142847061157227, | |
| "logps/chosen": -31.049779891967773, | |
| "logps/rejected": -33.54362487792969, | |
| "loss": 0.4366, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.4428572654724121, | |
| "rewards/margins": 0.28403452038764954, | |
| "rewards/rejected": 0.15882274508476257, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.825108134172131e-06, | |
| "logits/chosen": -1.9670381546020508, | |
| "logits/rejected": -1.9549226760864258, | |
| "logps/chosen": -34.00459671020508, | |
| "logps/rejected": -33.429378509521484, | |
| "loss": 0.4145, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.4731821119785309, | |
| "rewards/margins": 0.38561564683914185, | |
| "rewards/rejected": 0.08756652474403381, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 4.80369052967602e-06, | |
| "logits/chosen": -2.001861095428467, | |
| "logits/rejected": -2.0005335807800293, | |
| "logps/chosen": -32.78104019165039, | |
| "logps/rejected": -32.248443603515625, | |
| "loss": 0.4183, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.5638058185577393, | |
| "rewards/margins": 0.37889450788497925, | |
| "rewards/rejected": 0.18491129577159882, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 4.781089396387968e-06, | |
| "logits/chosen": -2.0892767906188965, | |
| "logits/rejected": -2.073693037033081, | |
| "logps/chosen": -33.377323150634766, | |
| "logps/rejected": -32.779090881347656, | |
| "loss": 0.4167, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.641654908657074, | |
| "rewards/margins": 0.37049978971481323, | |
| "rewards/rejected": 0.2711551785469055, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 4.757316345716554e-06, | |
| "logits/chosen": -1.9618675708770752, | |
| "logits/rejected": -1.9610551595687866, | |
| "logps/chosen": -32.51572799682617, | |
| "logps/rejected": -32.19974899291992, | |
| "loss": 0.4137, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.6130937933921814, | |
| "rewards/margins": 0.4007722735404968, | |
| "rewards/rejected": 0.21232159435749054, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 4.73238359114687e-06, | |
| "logits/chosen": -1.916259765625, | |
| "logits/rejected": -1.9264857769012451, | |
| "logps/chosen": -31.4000244140625, | |
| "logps/rejected": -35.018882751464844, | |
| "loss": 0.4065, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.6437316536903381, | |
| "rewards/margins": 0.427828311920166, | |
| "rewards/rejected": 0.21590332686901093, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 4.706303941965804e-06, | |
| "logits/chosen": -2.050283908843994, | |
| "logits/rejected": -2.043849468231201, | |
| "logps/chosen": -32.88298797607422, | |
| "logps/rejected": -28.950063705444336, | |
| "loss": 0.4093, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.6151864528656006, | |
| "rewards/margins": 0.3940979838371277, | |
| "rewards/rejected": 0.2210884839296341, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 4.679090796681225e-06, | |
| "logits/chosen": -1.9082043170928955, | |
| "logits/rejected": -1.9103626012802124, | |
| "logps/chosen": -33.2851676940918, | |
| "logps/rejected": -30.665771484375, | |
| "loss": 0.3759, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.7631045579910278, | |
| "rewards/margins": 0.5976192355155945, | |
| "rewards/rejected": 0.1654852330684662, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "learning_rate": 4.650758136138454e-06, | |
| "logits/chosen": -1.9357115030288696, | |
| "logits/rejected": -1.9344127178192139, | |
| "logps/chosen": -33.36685562133789, | |
| "logps/rejected": -35.62567138671875, | |
| "loss": 0.3409, | |
| "rewards/accuracies": 0.7583333253860474, | |
| "rewards/chosen": 0.8143496513366699, | |
| "rewards/margins": 0.790077805519104, | |
| "rewards/rejected": 0.02427184209227562, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 4.621320516337559e-06, | |
| "logits/chosen": -1.8698232173919678, | |
| "logits/rejected": -1.8616269826889038, | |
| "logps/chosen": -30.623443603515625, | |
| "logps/rejected": -36.03126907348633, | |
| "loss": 0.3117, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.920426070690155, | |
| "rewards/margins": 0.9929585456848145, | |
| "rewards/rejected": -0.07253243774175644, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "eval_logits/chosen": -2.2160162925720215, | |
| "eval_logits/rejected": -2.2111997604370117, | |
| "eval_logps/chosen": -33.652793884277344, | |
| "eval_logps/rejected": -37.193359375, | |
| "eval_loss": 0.4891091287136078, | |
| "eval_rewards/accuracies": 0.5651993155479431, | |
| "eval_rewards/chosen": 0.30540817975997925, | |
| "eval_rewards/margins": 0.04680241644382477, | |
| "eval_rewards/rejected": 0.2586057484149933, | |
| "eval_runtime": 146.2599, | |
| "eval_samples_per_second": 2.345, | |
| "eval_steps_per_second": 0.294, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "learning_rate": 4.590793060955158e-06, | |
| "logits/chosen": -2.044127941131592, | |
| "logits/rejected": -2.046945571899414, | |
| "logps/chosen": -31.79901695251465, | |
| "logps/rejected": -34.87114334106445, | |
| "loss": 0.299, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.003937005996704, | |
| "rewards/margins": 1.0665733814239502, | |
| "rewards/rejected": -0.0626363754272461, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 4.559191453574582e-06, | |
| "logits/chosen": -1.8841445446014404, | |
| "logits/rejected": -1.8827756643295288, | |
| "logps/chosen": -27.981430053710938, | |
| "logps/rejected": -32.39961624145508, | |
| "loss": 0.3151, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 0.868902325630188, | |
| "rewards/margins": 0.9344655871391296, | |
| "rewards/rejected": -0.06556323915719986, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 4.52653192962838e-06, | |
| "logits/chosen": -1.8422701358795166, | |
| "logits/rejected": -1.835268259048462, | |
| "logps/chosen": -32.6194953918457, | |
| "logps/rejected": -34.0600700378418, | |
| "loss": 0.2904, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.1636759042739868, | |
| "rewards/margins": 1.0155143737792969, | |
| "rewards/rejected": 0.14816154539585114, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "learning_rate": 4.492831268057307e-06, | |
| "logits/chosen": -2.0123181343078613, | |
| "logits/rejected": -2.0072364807128906, | |
| "logps/chosen": -30.41460609436035, | |
| "logps/rejected": -32.033443450927734, | |
| "loss": 0.284, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.0949517488479614, | |
| "rewards/margins": 1.1476168632507324, | |
| "rewards/rejected": -0.05266512185335159, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 4.458106782690094e-06, | |
| "logits/chosen": -1.8972448110580444, | |
| "logits/rejected": -1.9013574123382568, | |
| "logps/chosen": -33.069793701171875, | |
| "logps/rejected": -32.7395133972168, | |
| "loss": 0.263, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.156842589378357, | |
| "rewards/margins": 1.2775455713272095, | |
| "rewards/rejected": -0.12070278078317642, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 4.422376313348405e-06, | |
| "logits/chosen": -1.9023644924163818, | |
| "logits/rejected": -1.8966976404190063, | |
| "logps/chosen": -33.858001708984375, | |
| "logps/rejected": -35.19942092895508, | |
| "loss": 0.2537, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.249218225479126, | |
| "rewards/margins": 1.4494789838790894, | |
| "rewards/rejected": -0.20026080310344696, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 4.3856582166815696e-06, | |
| "logits/chosen": -1.9303083419799805, | |
| "logits/rejected": -1.9299156665802002, | |
| "logps/chosen": -32.68412399291992, | |
| "logps/rejected": -34.166542053222656, | |
| "loss": 0.2809, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.1873619556427002, | |
| "rewards/margins": 1.1722075939178467, | |
| "rewards/rejected": 0.01515440084040165, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 4.347971356735789e-06, | |
| "logits/chosen": -1.97918701171875, | |
| "logits/rejected": -1.960694670677185, | |
| "logps/chosen": -32.545894622802734, | |
| "logps/rejected": -33.19091033935547, | |
| "loss": 0.2501, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.291896104812622, | |
| "rewards/margins": 1.3886420726776123, | |
| "rewards/rejected": -0.09674612432718277, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "learning_rate": 4.309335095262675e-06, | |
| "logits/chosen": -1.9456336498260498, | |
| "logits/rejected": -1.9451408386230469, | |
| "logps/chosen": -30.120983123779297, | |
| "logps/rejected": -31.22079849243164, | |
| "loss": 0.2773, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.1922577619552612, | |
| "rewards/margins": 1.1720596551895142, | |
| "rewards/rejected": 0.020198041573166847, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 4.269769281772082e-06, | |
| "logits/chosen": -1.9087337255477905, | |
| "logits/rejected": -1.9019416570663452, | |
| "logps/chosen": -30.988162994384766, | |
| "logps/rejected": -34.79794692993164, | |
| "loss": 0.2459, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.345572829246521, | |
| "rewards/margins": 1.4825364351272583, | |
| "rewards/rejected": -0.13696344196796417, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "eval_logits/chosen": -2.1905555725097656, | |
| "eval_logits/rejected": -2.1857681274414062, | |
| "eval_logps/chosen": -33.6363525390625, | |
| "eval_logps/rejected": -37.182701110839844, | |
| "eval_loss": 0.4884692430496216, | |
| "eval_rewards/accuracies": 0.5622923374176025, | |
| "eval_rewards/chosen": 0.3185593783855438, | |
| "eval_rewards/margins": 0.05142458155751228, | |
| "eval_rewards/rejected": 0.26713478565216064, | |
| "eval_runtime": 145.9464, | |
| "eval_samples_per_second": 2.35, | |
| "eval_steps_per_second": 0.295, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 4.22929424333435e-06, | |
| "logits/chosen": -1.903969168663025, | |
| "logits/rejected": -1.9078088998794556, | |
| "logps/chosen": -27.756145477294922, | |
| "logps/rejected": -33.003501892089844, | |
| "loss": 0.2721, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.1349422931671143, | |
| "rewards/margins": 1.2503130435943604, | |
| "rewards/rejected": -0.11537061631679535, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "learning_rate": 4.1879307741372085e-06, | |
| "logits/chosen": -1.9105024337768555, | |
| "logits/rejected": -1.9210312366485596, | |
| "logps/chosen": -31.711151123046875, | |
| "logps/rejected": -30.852153778076172, | |
| "loss": 0.2636, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.240944743156433, | |
| "rewards/margins": 1.4013227224349976, | |
| "rewards/rejected": -0.1603778600692749, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "learning_rate": 4.145700124802693e-06, | |
| "logits/chosen": -1.8562660217285156, | |
| "logits/rejected": -1.8539501428604126, | |
| "logps/chosen": -30.094646453857422, | |
| "logps/rejected": -30.31314468383789, | |
| "loss": 0.2542, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.2579472064971924, | |
| "rewards/margins": 1.3670498132705688, | |
| "rewards/rejected": -0.10910268872976303, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 4.102623991469562e-06, | |
| "logits/chosen": -1.9388301372528076, | |
| "logits/rejected": -1.931828260421753, | |
| "logps/chosen": -32.52373504638672, | |
| "logps/rejected": -33.19480514526367, | |
| "loss": 0.234, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.3672724962234497, | |
| "rewards/margins": 1.5091286897659302, | |
| "rewards/rejected": -0.14185625314712524, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "learning_rate": 4.058724504646834e-06, | |
| "logits/chosen": -1.9054105281829834, | |
| "logits/rejected": -1.9118045568466187, | |
| "logps/chosen": -30.23921775817871, | |
| "logps/rejected": -32.79788589477539, | |
| "loss": 0.2692, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.211173176765442, | |
| "rewards/margins": 1.2424962520599365, | |
| "rewards/rejected": -0.03132311627268791, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "learning_rate": 4.014024217844167e-06, | |
| "logits/chosen": -1.9779608249664307, | |
| "logits/rejected": -1.9552392959594727, | |
| "logps/chosen": -29.897476196289062, | |
| "logps/rejected": -33.047569274902344, | |
| "loss": 0.2672, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.2320712804794312, | |
| "rewards/margins": 1.273280143737793, | |
| "rewards/rejected": -0.041208960115909576, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "learning_rate": 3.968546095984911e-06, | |
| "logits/chosen": -1.9123926162719727, | |
| "logits/rejected": -1.9074643850326538, | |
| "logps/chosen": -30.69790267944336, | |
| "logps/rejected": -32.02743911743164, | |
| "loss": 0.2612, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.3750633001327515, | |
| "rewards/margins": 1.3002793788909912, | |
| "rewards/rejected": 0.07478378713130951, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 3.922313503607806e-06, | |
| "logits/chosen": -1.9546325206756592, | |
| "logits/rejected": -1.956498384475708, | |
| "logps/chosen": -32.70453643798828, | |
| "logps/rejected": -35.04417037963867, | |
| "loss": 0.2395, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.337005853652954, | |
| "rewards/margins": 1.5316107273101807, | |
| "rewards/rejected": -0.19460472464561462, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "learning_rate": 3.875350192863368e-06, | |
| "logits/chosen": -1.9385898113250732, | |
| "logits/rejected": -1.9379469156265259, | |
| "logps/chosen": -28.891321182250977, | |
| "logps/rejected": -31.504505157470703, | |
| "loss": 0.2493, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.3605870008468628, | |
| "rewards/margins": 1.3948794603347778, | |
| "rewards/rejected": -0.034292541444301605, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 3.8276802913111436e-06, | |
| "logits/chosen": -1.9496517181396484, | |
| "logits/rejected": -1.9475892782211304, | |
| "logps/chosen": -31.307537078857422, | |
| "logps/rejected": -32.269371032714844, | |
| "loss": 0.2639, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.3525952100753784, | |
| "rewards/margins": 1.3281266689300537, | |
| "rewards/rejected": 0.024468522518873215, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "eval_logits/chosen": -2.183182716369629, | |
| "eval_logits/rejected": -2.178420066833496, | |
| "eval_logps/chosen": -33.58167266845703, | |
| "eval_logps/rejected": -37.20375442504883, | |
| "eval_loss": 0.47496160864830017, | |
| "eval_rewards/accuracies": 0.5855481624603271, | |
| "eval_rewards/chosen": 0.3623029887676239, | |
| "eval_rewards/margins": 0.11201038211584091, | |
| "eval_rewards/rejected": 0.2502925992012024, | |
| "eval_runtime": 145.9651, | |
| "eval_samples_per_second": 2.35, | |
| "eval_steps_per_second": 0.295, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "learning_rate": 3.7793282895240927e-06, | |
| "logits/chosen": -1.9928886890411377, | |
| "logits/rejected": -1.9994275569915771, | |
| "logps/chosen": -30.57297706604004, | |
| "logps/rejected": -32.0395622253418, | |
| "loss": 0.2611, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.3315242528915405, | |
| "rewards/margins": 1.3126256465911865, | |
| "rewards/rejected": 0.01889852061867714, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 3.730319028506478e-06, | |
| "logits/chosen": -1.9484421014785767, | |
| "logits/rejected": -1.946126937866211, | |
| "logps/chosen": -32.853248596191406, | |
| "logps/rejected": -30.977664947509766, | |
| "loss": 0.2377, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.5221078395843506, | |
| "rewards/margins": 1.5158333778381348, | |
| "rewards/rejected": 0.0062743364833295345, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "learning_rate": 3.6806776869317074e-06, | |
| "logits/chosen": -1.8969027996063232, | |
| "logits/rejected": -1.8902422189712524, | |
| "logps/chosen": -33.44438171386719, | |
| "logps/rejected": -32.37852478027344, | |
| "loss": 0.2252, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.6158740520477295, | |
| "rewards/margins": 1.6329265832901, | |
| "rewards/rejected": -0.017052406445145607, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "learning_rate": 3.6304297682067146e-06, | |
| "logits/chosen": -1.9018688201904297, | |
| "logits/rejected": -1.9080841541290283, | |
| "logps/chosen": -32.16619110107422, | |
| "logps/rejected": -33.13113021850586, | |
| "loss": 0.2565, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.450488567352295, | |
| "rewards/margins": 1.3631832599639893, | |
| "rewards/rejected": 0.08730525523424149, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "learning_rate": 3.579601087369492e-06, | |
| "logits/chosen": -1.98688542842865, | |
| "logits/rejected": -2.0009474754333496, | |
| "logps/chosen": -30.202587127685547, | |
| "logps/rejected": -31.93996238708496, | |
| "loss": 0.2613, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 1.3199015855789185, | |
| "rewards/margins": 1.2737209796905518, | |
| "rewards/rejected": 0.046180594712495804, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 3.5282177578265295e-06, | |
| "logits/chosen": -1.861903429031372, | |
| "logits/rejected": -1.8589446544647217, | |
| "logps/chosen": -31.801767349243164, | |
| "logps/rejected": -34.892234802246094, | |
| "loss": 0.1937, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.6969530582427979, | |
| "rewards/margins": 1.8505403995513916, | |
| "rewards/rejected": -0.15358731150627136, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 3.476306177936961e-06, | |
| "logits/chosen": -1.9529634714126587, | |
| "logits/rejected": -1.9532558917999268, | |
| "logps/chosen": -29.594532012939453, | |
| "logps/rejected": -33.946571350097656, | |
| "loss": 0.2422, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.3528481721878052, | |
| "rewards/margins": 1.449000597000122, | |
| "rewards/rejected": -0.09615238755941391, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "learning_rate": 3.423893017450324e-06, | |
| "logits/chosen": -1.9031528234481812, | |
| "logits/rejected": -1.9001739025115967, | |
| "logps/chosen": -29.16634178161621, | |
| "logps/rejected": -32.86220169067383, | |
| "loss": 0.2577, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.3648045063018799, | |
| "rewards/margins": 1.40032160282135, | |
| "rewards/rejected": -0.03551710769534111, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "learning_rate": 3.3710052038048794e-06, | |
| "logits/chosen": -1.9385061264038086, | |
| "logits/rejected": -1.9385082721710205, | |
| "logps/chosen": -27.996692657470703, | |
| "logps/rejected": -30.74009132385254, | |
| "loss": 0.2085, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.7331215143203735, | |
| "rewards/margins": 1.7887967824935913, | |
| "rewards/rejected": -0.05567514896392822, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 3.3176699082935546e-06, | |
| "logits/chosen": -1.8473536968231201, | |
| "logits/rejected": -1.850669503211975, | |
| "logps/chosen": -32.4169921875, | |
| "logps/rejected": -31.4881649017334, | |
| "loss": 0.2437, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.6491864919662476, | |
| "rewards/margins": 1.6679694652557373, | |
| "rewards/rejected": -0.01878293789923191, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "eval_logits/chosen": -2.178642988204956, | |
| "eval_logits/rejected": -2.173909902572632, | |
| "eval_logps/chosen": -33.59923553466797, | |
| "eval_logps/rejected": -37.22935485839844, | |
| "eval_loss": 0.4741733968257904, | |
| "eval_rewards/accuracies": 0.5747508406639099, | |
| "eval_rewards/chosen": 0.34825241565704346, | |
| "eval_rewards/margins": 0.1184391975402832, | |
| "eval_rewards/rejected": 0.22981324791908264, | |
| "eval_runtime": 145.9455, | |
| "eval_samples_per_second": 2.35, | |
| "eval_steps_per_second": 0.295, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "learning_rate": 3.2639145321045933e-06, | |
| "logits/chosen": -1.9303743839263916, | |
| "logits/rejected": -1.9213898181915283, | |
| "logps/chosen": -34.50909423828125, | |
| "logps/rejected": -31.8797550201416, | |
| "loss": 0.2249, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.602341890335083, | |
| "rewards/margins": 1.5814943313598633, | |
| "rewards/rejected": 0.02084733545780182, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "learning_rate": 3.2097666922441107e-06, | |
| "logits/chosen": -1.9459857940673828, | |
| "logits/rejected": -1.9470113515853882, | |
| "logps/chosen": -34.359092712402344, | |
| "logps/rejected": -33.046634674072266, | |
| "loss": 0.2173, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.6610389947891235, | |
| "rewards/margins": 1.6375271081924438, | |
| "rewards/rejected": 0.02351185865700245, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "learning_rate": 3.1552542073477554e-06, | |
| "logits/chosen": -1.962262511253357, | |
| "logits/rejected": -1.9598932266235352, | |
| "logps/chosen": -30.564239501953125, | |
| "logps/rejected": -32.997047424316406, | |
| "loss": 0.23, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.5992707014083862, | |
| "rewards/margins": 1.5993889570236206, | |
| "rewards/rejected": -0.00011831596202682704, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 3.100405083388799e-06, | |
| "logits/chosen": -1.9474010467529297, | |
| "logits/rejected": -1.952619194984436, | |
| "logps/chosen": -29.694049835205078, | |
| "logps/rejected": -33.282554626464844, | |
| "loss": 0.2069, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.6733427047729492, | |
| "rewards/margins": 1.7755340337753296, | |
| "rewards/rejected": -0.10219136625528336, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "learning_rate": 3.0452474992899645e-06, | |
| "logits/chosen": -1.892179250717163, | |
| "logits/rejected": -1.8910681009292603, | |
| "logps/chosen": -31.141448974609375, | |
| "logps/rejected": -34.69000244140625, | |
| "loss": 0.24, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.5868273973464966, | |
| "rewards/margins": 1.5900630950927734, | |
| "rewards/rejected": -0.0032357454765588045, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 2.989809792446417e-06, | |
| "logits/chosen": -1.7668031454086304, | |
| "logits/rejected": -1.7617915868759155, | |
| "logps/chosen": -33.82783889770508, | |
| "logps/rejected": -35.33191680908203, | |
| "loss": 0.2193, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.7364822626113892, | |
| "rewards/margins": 1.8352861404418945, | |
| "rewards/rejected": -0.0988037958741188, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 2.9341204441673267e-06, | |
| "logits/chosen": -1.9073164463043213, | |
| "logits/rejected": -1.9110107421875, | |
| "logps/chosen": -33.157711029052734, | |
| "logps/rejected": -33.576969146728516, | |
| "loss": 0.2221, | |
| "rewards/accuracies": 0.8708333969116211, | |
| "rewards/chosen": 1.6447594165802002, | |
| "rewards/margins": 1.6304632425308228, | |
| "rewards/rejected": 0.014296052046120167, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "learning_rate": 2.878208065043501e-06, | |
| "logits/chosen": -1.8515024185180664, | |
| "logits/rejected": -1.8497060537338257, | |
| "logps/chosen": -31.297021865844727, | |
| "logps/rejected": -35.01959991455078, | |
| "loss": 0.1312, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.0836594104766846, | |
| "rewards/margins": 2.575618028640747, | |
| "rewards/rejected": -0.4919588565826416, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "learning_rate": 2.8221013802485974e-06, | |
| "logits/chosen": -1.899592638015747, | |
| "logits/rejected": -1.8983027935028076, | |
| "logps/chosen": -30.751636505126953, | |
| "logps/rejected": -33.40520477294922, | |
| "loss": 0.1431, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.0992822647094727, | |
| "rewards/margins": 2.354191541671753, | |
| "rewards/rejected": -0.25490933656692505, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 2.76582921478147e-06, | |
| "logits/chosen": -1.83371901512146, | |
| "logits/rejected": -1.8279550075531006, | |
| "logps/chosen": -31.871252059936523, | |
| "logps/rejected": -31.73343849182129, | |
| "loss": 0.1567, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 2.0094635486602783, | |
| "rewards/margins": 2.224806547164917, | |
| "rewards/rejected": -0.21534299850463867, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "eval_logits/chosen": -2.1802878379821777, | |
| "eval_logits/rejected": -2.1755456924438477, | |
| "eval_logps/chosen": -33.54963302612305, | |
| "eval_logps/rejected": -37.20660400390625, | |
| "eval_loss": 0.4695200026035309, | |
| "eval_rewards/accuracies": 0.5826411843299866, | |
| "eval_rewards/chosen": 0.3879339396953583, | |
| "eval_rewards/margins": 0.1399237960577011, | |
| "eval_rewards/rejected": 0.24801018834114075, | |
| "eval_runtime": 145.8879, | |
| "eval_samples_per_second": 2.351, | |
| "eval_steps_per_second": 0.295, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "learning_rate": 2.7094204786572254e-06, | |
| "logits/chosen": -1.9300823211669922, | |
| "logits/rejected": -1.9375168085098267, | |
| "logps/chosen": -29.62691307067871, | |
| "logps/rejected": -34.63849639892578, | |
| "loss": 0.1613, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.0010743141174316, | |
| "rewards/margins": 2.3072001934051514, | |
| "rewards/rejected": -0.3061259686946869, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 2.6529041520546072e-06, | |
| "logits/chosen": -1.9036821126937866, | |
| "logits/rejected": -1.9064573049545288, | |
| "logps/chosen": -30.34005355834961, | |
| "logps/rejected": -33.30119705200195, | |
| "loss": 0.1999, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.784820556640625, | |
| "rewards/margins": 1.8419349193572998, | |
| "rewards/rejected": -0.05711423233151436, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "learning_rate": 2.5963092704273302e-06, | |
| "logits/chosen": -1.805069923400879, | |
| "logits/rejected": -1.8091821670532227, | |
| "logps/chosen": -30.127094268798828, | |
| "logps/rejected": -35.30133056640625, | |
| "loss": 0.1557, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.9714915752410889, | |
| "rewards/margins": 2.2844126224517822, | |
| "rewards/rejected": -0.31292054057121277, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "learning_rate": 2.53966490958702e-06, | |
| "logits/chosen": -1.8699356317520142, | |
| "logits/rejected": -1.865962028503418, | |
| "logps/chosen": -30.581180572509766, | |
| "logps/rejected": -33.02351379394531, | |
| "loss": 0.1821, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.9099775552749634, | |
| "rewards/margins": 1.9330689907073975, | |
| "rewards/rejected": -0.02309143915772438, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "learning_rate": 2.4830001707654135e-06, | |
| "logits/chosen": -1.9513721466064453, | |
| "logits/rejected": -1.9533140659332275, | |
| "logps/chosen": -30.263479232788086, | |
| "logps/rejected": -36.080474853515625, | |
| "loss": 0.1402, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 2.0094857215881348, | |
| "rewards/margins": 2.344318389892578, | |
| "rewards/rejected": -0.3348326086997986, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "learning_rate": 2.4263441656635054e-06, | |
| "logits/chosen": -1.7577623128890991, | |
| "logits/rejected": -1.7520900964736938, | |
| "logps/chosen": -33.64806365966797, | |
| "logps/rejected": -33.127098083496094, | |
| "loss": 0.157, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 2.106886625289917, | |
| "rewards/margins": 2.2950429916381836, | |
| "rewards/rejected": -0.18815645575523376, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "learning_rate": 2.3697260014953107e-06, | |
| "logits/chosen": -1.8190982341766357, | |
| "logits/rejected": -1.8191255331039429, | |
| "logps/chosen": -33.16907501220703, | |
| "logps/rejected": -35.244049072265625, | |
| "loss": 0.1428, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 2.138986110687256, | |
| "rewards/margins": 2.457213878631592, | |
| "rewards/rejected": -0.31822749972343445, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "learning_rate": 2.3131747660339396e-06, | |
| "logits/chosen": -1.867693543434143, | |
| "logits/rejected": -1.8564653396606445, | |
| "logps/chosen": -31.62196922302246, | |
| "logps/rejected": -33.34123992919922, | |
| "loss": 0.1484, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.9876899719238281, | |
| "rewards/margins": 2.3230578899383545, | |
| "rewards/rejected": -0.3353681266307831, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.31, | |
| "learning_rate": 2.256719512667651e-06, | |
| "logits/chosen": -1.9699180126190186, | |
| "logits/rejected": -1.9743999242782593, | |
| "logps/chosen": -30.770462036132812, | |
| "logps/rejected": -32.943992614746094, | |
| "loss": 0.1449, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.0475802421569824, | |
| "rewards/margins": 2.453592300415039, | |
| "rewards/rejected": -0.40601205825805664, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 2.2003892454735786e-06, | |
| "logits/chosen": -1.8890403509140015, | |
| "logits/rejected": -1.881980538368225, | |
| "logps/chosen": -32.13774490356445, | |
| "logps/rejected": -32.89883041381836, | |
| "loss": 0.131, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.173067808151245, | |
| "rewards/margins": 2.5704257488250732, | |
| "rewards/rejected": -0.397358238697052, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "eval_logits/chosen": -2.1704888343811035, | |
| "eval_logits/rejected": -2.165771961212158, | |
| "eval_logps/chosen": -33.59296798706055, | |
| "eval_logps/rejected": -37.2408447265625, | |
| "eval_loss": 0.4716451168060303, | |
| "eval_rewards/accuracies": 0.5859634280204773, | |
| "eval_rewards/chosen": 0.3532681465148926, | |
| "eval_rewards/margins": 0.1326470971107483, | |
| "eval_rewards/rejected": 0.22062106430530548, | |
| "eval_runtime": 145.952, | |
| "eval_samples_per_second": 2.35, | |
| "eval_steps_per_second": 0.295, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "learning_rate": 2.1442129043167877e-06, | |
| "logits/chosen": -1.8803346157073975, | |
| "logits/rejected": -1.8807852268218994, | |
| "logps/chosen": -28.69057846069336, | |
| "logps/rejected": -35.19638442993164, | |
| "loss": 0.1531, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.9812660217285156, | |
| "rewards/margins": 2.3866193294525146, | |
| "rewards/rejected": -0.4053533673286438, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "learning_rate": 2.088219349982323e-06, | |
| "logits/chosen": -1.8385648727416992, | |
| "logits/rejected": -1.8306169509887695, | |
| "logps/chosen": -29.56528091430664, | |
| "logps/rejected": -34.04883575439453, | |
| "loss": 0.1693, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.8189325332641602, | |
| "rewards/margins": 2.2009072303771973, | |
| "rewards/rejected": -0.3819747567176819, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "learning_rate": 2.0324373493478803e-06, | |
| "logits/chosen": -2.004300117492676, | |
| "logits/rejected": -2.004765033721924, | |
| "logps/chosen": -27.645904541015625, | |
| "logps/rejected": -33.17176818847656, | |
| "loss": 0.1901, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.8809070587158203, | |
| "rewards/margins": 2.126433849334717, | |
| "rewards/rejected": -0.24552695453166962, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "learning_rate": 1.976895560604729e-06, | |
| "logits/chosen": -1.883927583694458, | |
| "logits/rejected": -1.893393874168396, | |
| "logps/chosen": -31.66170883178711, | |
| "logps/rejected": -33.47935104370117, | |
| "loss": 0.1308, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.2106852531433105, | |
| "rewards/margins": 2.602447509765625, | |
| "rewards/rejected": -0.3917620778083801, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 1.921622518534466e-06, | |
| "logits/chosen": -1.9267995357513428, | |
| "logits/rejected": -1.9306179285049438, | |
| "logps/chosen": -28.294620513916016, | |
| "logps/rejected": -32.00605010986328, | |
| "loss": 0.1628, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 1.882483720779419, | |
| "rewards/margins": 2.0984606742858887, | |
| "rewards/rejected": -0.21597695350646973, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.49, | |
| "learning_rate": 1.8666466198491794e-06, | |
| "logits/chosen": -1.9250777959823608, | |
| "logits/rejected": -1.921263337135315, | |
| "logps/chosen": -31.393346786499023, | |
| "logps/rejected": -34.266326904296875, | |
| "loss": 0.1531, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.145495891571045, | |
| "rewards/margins": 2.4334449768066406, | |
| "rewards/rejected": -0.28794899582862854, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 2.52, | |
| "learning_rate": 1.8119961086025376e-06, | |
| "logits/chosen": -1.8407471179962158, | |
| "logits/rejected": -1.84335458278656, | |
| "logps/chosen": -30.177082061767578, | |
| "logps/rejected": -35.5606803894043, | |
| "loss": 0.15, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.1085331439971924, | |
| "rewards/margins": 2.411801815032959, | |
| "rewards/rejected": -0.3032683730125427, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 1.7576990616793139e-06, | |
| "logits/chosen": -1.8759149312973022, | |
| "logits/rejected": -1.8695703744888306, | |
| "logps/chosen": -32.879905700683594, | |
| "logps/rejected": -36.688812255859375, | |
| "loss": 0.1591, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 2.020131826400757, | |
| "rewards/margins": 2.3616156578063965, | |
| "rewards/rejected": -0.34148353338241577, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 1.7037833743707892e-06, | |
| "logits/chosen": -1.8502181768417358, | |
| "logits/rejected": -1.845171332359314, | |
| "logps/chosen": -28.556514739990234, | |
| "logps/rejected": -36.38850784301758, | |
| "loss": 0.167, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.9571492671966553, | |
| "rewards/margins": 2.2760462760925293, | |
| "rewards/rejected": -0.31889697909355164, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 1.6502767460434588e-06, | |
| "logits/chosen": -1.8279926776885986, | |
| "logits/rejected": -1.8174269199371338, | |
| "logps/chosen": -29.500579833984375, | |
| "logps/rejected": -29.74460220336914, | |
| "loss": 0.1784, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.798485517501831, | |
| "rewards/margins": 1.9143791198730469, | |
| "rewards/rejected": -0.11589355766773224, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "eval_logits/chosen": -2.169346570968628, | |
| "eval_logits/rejected": -2.1646294593811035, | |
| "eval_logps/chosen": -33.57487106323242, | |
| "eval_logps/rejected": -37.250404357910156, | |
| "eval_loss": 0.4672938883304596, | |
| "eval_rewards/accuracies": 0.5859634280204773, | |
| "eval_rewards/chosen": 0.3677443265914917, | |
| "eval_rewards/margins": 0.15477292239665985, | |
| "eval_rewards/rejected": 0.21297141909599304, | |
| "eval_runtime": 145.9529, | |
| "eval_samples_per_second": 2.35, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "learning_rate": 1.5972066659083796e-06, | |
| "logits/chosen": -1.938071846961975, | |
| "logits/rejected": -1.9374967813491821, | |
| "logps/chosen": -29.38279151916504, | |
| "logps/rejected": -30.51906394958496, | |
| "loss": 0.1659, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.953122854232788, | |
| "rewards/margins": 2.1615920066833496, | |
| "rewards/rejected": -0.20846888422966003, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "learning_rate": 1.5446003988985041e-06, | |
| "logits/chosen": -1.9777361154556274, | |
| "logits/rejected": -1.9785385131835938, | |
| "logps/chosen": -29.455398559570312, | |
| "logps/rejected": -31.60211753845215, | |
| "loss": 0.1457, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 1.9905421733856201, | |
| "rewards/margins": 2.3395895957946777, | |
| "rewards/rejected": -0.3490474820137024, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 2.68, | |
| "learning_rate": 1.4924849716612211e-06, | |
| "logits/chosen": -1.948656439781189, | |
| "logits/rejected": -1.951843500137329, | |
| "logps/chosen": -29.8292236328125, | |
| "logps/rejected": -28.163440704345703, | |
| "loss": 0.1712, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.960677146911621, | |
| "rewards/margins": 2.1044540405273438, | |
| "rewards/rejected": -0.14377683401107788, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 2.7, | |
| "learning_rate": 1.440887158673332e-06, | |
| "logits/chosen": -1.9582103490829468, | |
| "logits/rejected": -1.9505630731582642, | |
| "logps/chosen": -28.690776824951172, | |
| "logps/rejected": -33.91931915283203, | |
| "loss": 0.1486, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.9896939992904663, | |
| "rewards/margins": 2.3249640464782715, | |
| "rewards/rejected": -0.33527034521102905, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "learning_rate": 1.3898334684855647e-06, | |
| "logits/chosen": -1.8943002223968506, | |
| "logits/rejected": -1.905225157737732, | |
| "logps/chosen": -30.873615264892578, | |
| "logps/rejected": -32.47500228881836, | |
| "loss": 0.1427, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.143594264984131, | |
| "rewards/margins": 2.3275885581970215, | |
| "rewards/rejected": -0.18399406969547272, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 1.3393501301037245e-06, | |
| "logits/chosen": -1.9732818603515625, | |
| "logits/rejected": -1.9640705585479736, | |
| "logps/chosen": -30.988611221313477, | |
| "logps/rejected": -36.728416442871094, | |
| "loss": 0.171, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 2.009733200073242, | |
| "rewards/margins": 2.327359199523926, | |
| "rewards/rejected": -0.31762614846229553, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 1.2894630795134454e-06, | |
| "logits/chosen": -1.8804638385772705, | |
| "logits/rejected": -1.8824098110198975, | |
| "logps/chosen": -33.24834442138672, | |
| "logps/rejected": -33.089820861816406, | |
| "loss": 0.1391, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.1927542686462402, | |
| "rewards/margins": 2.506317615509033, | |
| "rewards/rejected": -0.31356340646743774, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "learning_rate": 1.2401979463554984e-06, | |
| "logits/chosen": -2.0114994049072266, | |
| "logits/rejected": -2.012619972229004, | |
| "logps/chosen": -30.447467803955078, | |
| "logps/rejected": -34.64774703979492, | |
| "loss": 0.1376, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 2.088136911392212, | |
| "rewards/margins": 2.513786792755127, | |
| "rewards/rejected": -0.42564988136291504, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "learning_rate": 1.1915800407584705e-06, | |
| "logits/chosen": -1.9806941747665405, | |
| "logits/rejected": -1.9849655628204346, | |
| "logps/chosen": -28.488088607788086, | |
| "logps/rejected": -33.84292984008789, | |
| "loss": 0.1691, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.9241825342178345, | |
| "rewards/margins": 2.160259485244751, | |
| "rewards/rejected": -0.23607663810253143, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "learning_rate": 1.1436343403356019e-06, | |
| "logits/chosen": -1.9700733423233032, | |
| "logits/rejected": -1.9752953052520752, | |
| "logps/chosen": -31.235626220703125, | |
| "logps/rejected": -30.409204483032227, | |
| "loss": 0.1956, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.8667011260986328, | |
| "rewards/margins": 1.8727245330810547, | |
| "rewards/rejected": -0.0060233683325350285, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "eval_logits/chosen": -2.1669461727142334, | |
| "eval_logits/rejected": -2.162247896194458, | |
| "eval_logps/chosen": -33.58707809448242, | |
| "eval_logps/rejected": -37.24416732788086, | |
| "eval_loss": 0.47056397795677185, | |
| "eval_rewards/accuracies": 0.5859634280204773, | |
| "eval_rewards/chosen": 0.3579804301261902, | |
| "eval_rewards/margins": 0.14001740515232086, | |
| "eval_rewards/rejected": 0.21796302497386932, | |
| "eval_runtime": 145.9563, | |
| "eval_samples_per_second": 2.35, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "learning_rate": 1.0963854773524548e-06, | |
| "logits/chosen": -1.9610811471939087, | |
| "logits/rejected": -1.9610563516616821, | |
| "logps/chosen": -30.302623748779297, | |
| "logps/rejected": -31.24424171447754, | |
| "loss": 0.1495, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.1100215911865234, | |
| "rewards/margins": 2.2926793098449707, | |
| "rewards/rejected": -0.18265783786773682, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "learning_rate": 1.049857726072005e-06, | |
| "logits/chosen": -1.8055435419082642, | |
| "logits/rejected": -1.8071558475494385, | |
| "logps/chosen": -31.964313507080078, | |
| "logps/rejected": -33.263916015625, | |
| "loss": 0.1722, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 2.08052921295166, | |
| "rewards/margins": 2.258647918701172, | |
| "rewards/rejected": -0.17811879515647888, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "learning_rate": 1.0040749902836508e-06, | |
| "logits/chosen": -1.8441641330718994, | |
| "logits/rejected": -1.8414533138275146, | |
| "logps/chosen": -28.680904388427734, | |
| "logps/rejected": -31.291534423828125, | |
| "loss": 0.1972, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.8646504878997803, | |
| "rewards/margins": 1.9652849435806274, | |
| "rewards/rejected": -0.10063423216342926, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 9.59060791022566e-07, | |
| "logits/chosen": -1.979723334312439, | |
| "logits/rejected": -1.9741865396499634, | |
| "logps/chosen": -30.334341049194336, | |
| "logps/rejected": -33.13202667236328, | |
| "loss": 0.1522, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.1643624305725098, | |
| "rewards/margins": 2.3067753314971924, | |
| "rewards/rejected": -0.14241299033164978, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "learning_rate": 9.148382544856885e-07, | |
| "logits/chosen": -1.8336684703826904, | |
| "logits/rejected": -1.823862075805664, | |
| "logps/chosen": -30.90520668029785, | |
| "logps/rejected": -31.456531524658203, | |
| "loss": 0.1763, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.037224292755127, | |
| "rewards/margins": 2.1685280799865723, | |
| "rewards/rejected": -0.13130377233028412, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 3.01, | |
| "learning_rate": 8.714301001505568e-07, | |
| "logits/chosen": -1.9085521697998047, | |
| "logits/rejected": -1.908399224281311, | |
| "logps/chosen": -31.05959701538086, | |
| "logps/rejected": -31.56801414489746, | |
| "loss": 0.1469, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.09771990776062, | |
| "rewards/margins": 2.3909716606140137, | |
| "rewards/rejected": -0.2932516634464264, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 3.04, | |
| "learning_rate": 8.288586291031025e-07, | |
| "logits/chosen": -1.9817787408828735, | |
| "logits/rejected": -1.9764562845230103, | |
| "logps/chosen": -30.92386817932129, | |
| "logps/rejected": -33.19076919555664, | |
| "loss": 0.1496, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.0816802978515625, | |
| "rewards/margins": 2.2813220024108887, | |
| "rewards/rejected": -0.19964155554771423, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "learning_rate": 7.871457125803897e-07, | |
| "logits/chosen": -1.8389742374420166, | |
| "logits/rejected": -1.8466594219207764, | |
| "logps/chosen": -30.842212677001953, | |
| "logps/rejected": -32.71555709838867, | |
| "loss": 0.1498, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 1.9926265478134155, | |
| "rewards/margins": 2.2759203910827637, | |
| "rewards/rejected": -0.28329357504844666, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "learning_rate": 7.463127807341966e-07, | |
| "logits/chosen": -1.9125378131866455, | |
| "logits/rejected": -1.9067026376724243, | |
| "logps/chosen": -29.305688858032227, | |
| "logps/rejected": -33.628910064697266, | |
| "loss": 0.1274, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.294342279434204, | |
| "rewards/margins": 2.570711612701416, | |
| "rewards/rejected": -0.27636927366256714, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "learning_rate": 7.063808116212021e-07, | |
| "logits/chosen": -1.8545643091201782, | |
| "logits/rejected": -1.8566617965698242, | |
| "logps/chosen": -30.835662841796875, | |
| "logps/rejected": -33.26459503173828, | |
| "loss": 0.137, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.1238210201263428, | |
| "rewards/margins": 2.5274574756622314, | |
| "rewards/rejected": -0.403636634349823, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "eval_logits/chosen": -2.167179584503174, | |
| "eval_logits/rejected": -2.1624755859375, | |
| "eval_logps/chosen": -33.572837829589844, | |
| "eval_logps/rejected": -37.24382400512695, | |
| "eval_loss": 0.4679512083530426, | |
| "eval_rewards/accuracies": 0.6063122749328613, | |
| "eval_rewards/chosen": 0.36937177181243896, | |
| "eval_rewards/margins": 0.15113388001918793, | |
| "eval_rewards/rejected": 0.21823787689208984, | |
| "eval_runtime": 145.8851, | |
| "eval_samples_per_second": 2.351, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.14, | |
| "learning_rate": 6.673703204254348e-07, | |
| "logits/chosen": -1.7856515645980835, | |
| "logits/rejected": -1.784690260887146, | |
| "logps/chosen": -33.143531799316406, | |
| "logps/rejected": -33.10202407836914, | |
| "loss": 0.1273, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.3271608352661133, | |
| "rewards/margins": 2.641291618347168, | |
| "rewards/rejected": -0.3141304850578308, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "learning_rate": 6.293013489185315e-07, | |
| "logits/chosen": -1.9604772329330444, | |
| "logits/rejected": -1.9546642303466797, | |
| "logps/chosen": -28.979385375976562, | |
| "logps/rejected": -33.37938690185547, | |
| "loss": 0.1352, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.1794400215148926, | |
| "rewards/margins": 2.5801169872283936, | |
| "rewards/rejected": -0.40067729353904724, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "learning_rate": 5.921934551632086e-07, | |
| "logits/chosen": -1.8098831176757812, | |
| "logits/rejected": -1.798370361328125, | |
| "logps/chosen": -31.50105857849121, | |
| "logps/rejected": -33.25225067138672, | |
| "loss": 0.1311, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.3076980113983154, | |
| "rewards/margins": 2.5944409370422363, | |
| "rewards/rejected": -0.2867427468299866, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 3.22, | |
| "learning_rate": 5.560657034652405e-07, | |
| "logits/chosen": -1.9010969400405884, | |
| "logits/rejected": -1.8942989110946655, | |
| "logps/chosen": -28.258697509765625, | |
| "logps/rejected": -29.34195327758789, | |
| "loss": 0.1603, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.9532296657562256, | |
| "rewards/margins": 2.33274507522583, | |
| "rewards/rejected": -0.3795151114463806, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "learning_rate": 5.2093665457911e-07, | |
| "logits/chosen": -1.9308841228485107, | |
| "logits/rejected": -1.9373613595962524, | |
| "logps/chosen": -32.524845123291016, | |
| "logps/rejected": -31.774621963500977, | |
| "loss": 0.1315, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.2964370250701904, | |
| "rewards/margins": 2.5014843940734863, | |
| "rewards/rejected": -0.20504704117774963, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 3.27, | |
| "learning_rate": 4.868243561723535e-07, | |
| "logits/chosen": -1.9048112630844116, | |
| "logits/rejected": -1.9054410457611084, | |
| "logps/chosen": -30.393224716186523, | |
| "logps/rejected": -33.286048889160156, | |
| "loss": 0.1317, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.1082680225372314, | |
| "rewards/margins": 2.485210418701172, | |
| "rewards/rejected": -0.37694239616394043, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 3.3, | |
| "learning_rate": 4.537463335535161e-07, | |
| "logits/chosen": -1.846727728843689, | |
| "logits/rejected": -1.8463557958602905, | |
| "logps/chosen": -30.23922348022461, | |
| "logps/rejected": -33.92576217651367, | |
| "loss": 0.1224, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.2812910079956055, | |
| "rewards/margins": 2.6054136753082275, | |
| "rewards/rejected": -0.3241223692893982, | |
| "step": 1270 | |
| }, | |
| { | |
| "epoch": 3.32, | |
| "learning_rate": 4.217195806684629e-07, | |
| "logits/chosen": -1.730064034461975, | |
| "logits/rejected": -1.7253376245498657, | |
| "logps/chosen": -32.30054473876953, | |
| "logps/rejected": -31.246471405029297, | |
| "loss": 0.1154, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.349520206451416, | |
| "rewards/margins": 2.605508327484131, | |
| "rewards/rejected": -0.2559877932071686, | |
| "step": 1280 | |
| }, | |
| { | |
| "epoch": 3.35, | |
| "learning_rate": 3.907605513696808e-07, | |
| "logits/chosen": -1.9358218908309937, | |
| "logits/rejected": -1.9203758239746094, | |
| "logps/chosen": -31.7402400970459, | |
| "logps/rejected": -35.388179779052734, | |
| "loss": 0.1234, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 2.085869550704956, | |
| "rewards/margins": 2.5841598510742188, | |
| "rewards/rejected": -0.4982902407646179, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 3.38, | |
| "learning_rate": 3.6088515096305675e-07, | |
| "logits/chosen": -1.879227638244629, | |
| "logits/rejected": -1.8840656280517578, | |
| "logps/chosen": -30.69500732421875, | |
| "logps/rejected": -36.70278549194336, | |
| "loss": 0.1211, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 2.2580931186676025, | |
| "rewards/margins": 2.7377448081970215, | |
| "rewards/rejected": -0.4796522259712219, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.38, | |
| "eval_logits/chosen": -2.1668777465820312, | |
| "eval_logits/rejected": -2.1621785163879395, | |
| "eval_logps/chosen": -33.58045959472656, | |
| "eval_logps/rejected": -37.23927688598633, | |
| "eval_loss": 0.47049835324287415, | |
| "eval_rewards/accuracies": 0.5917773842811584, | |
| "eval_rewards/chosen": 0.3632733225822449, | |
| "eval_rewards/margins": 0.14139726758003235, | |
| "eval_rewards/rejected": 0.22187604010105133, | |
| "eval_runtime": 145.9507, | |
| "eval_samples_per_second": 2.35, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "learning_rate": 3.321087280364757e-07, | |
| "logits/chosen": -1.8520374298095703, | |
| "logits/rejected": -1.8522069454193115, | |
| "logps/chosen": -33.18568801879883, | |
| "logps/rejected": -37.46868896484375, | |
| "loss": 0.134, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.446664333343506, | |
| "rewards/margins": 2.7543227672576904, | |
| "rewards/rejected": -0.30765873193740845, | |
| "step": 1310 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "learning_rate": 3.044460665744284e-07, | |
| "logits/chosen": -1.9394071102142334, | |
| "logits/rejected": -1.9380232095718384, | |
| "logps/chosen": -29.4769344329834, | |
| "logps/rejected": -31.30813980102539, | |
| "loss": 0.159, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.093536853790283, | |
| "rewards/margins": 2.3773601055145264, | |
| "rewards/rejected": -0.2838229537010193, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 3.45, | |
| "learning_rate": 2.779113783626916e-07, | |
| "logits/chosen": -1.844186782836914, | |
| "logits/rejected": -1.8454091548919678, | |
| "logps/chosen": -31.52813720703125, | |
| "logps/rejected": -33.989559173583984, | |
| "loss": 0.1366, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.276893138885498, | |
| "rewards/margins": 2.549605131149292, | |
| "rewards/rejected": -0.27271172404289246, | |
| "step": 1330 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 2.5251829568697204e-07, | |
| "logits/chosen": -1.9051072597503662, | |
| "logits/rejected": -1.9038896560668945, | |
| "logps/chosen": -28.674198150634766, | |
| "logps/rejected": -32.49821090698242, | |
| "loss": 0.1373, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.01957631111145, | |
| "rewards/margins": 2.424564838409424, | |
| "rewards/rejected": -0.4049888551235199, | |
| "step": 1340 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "learning_rate": 2.2827986432927774e-07, | |
| "logits/chosen": -1.9206342697143555, | |
| "logits/rejected": -1.9062645435333252, | |
| "logps/chosen": -31.68826675415039, | |
| "logps/rejected": -36.69408416748047, | |
| "loss": 0.1396, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.102721691131592, | |
| "rewards/margins": 2.4464478492736816, | |
| "rewards/rejected": -0.34372615814208984, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 3.53, | |
| "learning_rate": 2.0520853686560177e-07, | |
| "logits/chosen": -1.921002984046936, | |
| "logits/rejected": -1.9334627389907837, | |
| "logps/chosen": -29.323871612548828, | |
| "logps/rejected": -32.485260009765625, | |
| "loss": 0.1368, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 2.195680856704712, | |
| "rewards/margins": 2.3920950889587402, | |
| "rewards/rejected": -0.1964142769575119, | |
| "step": 1360 | |
| }, | |
| { | |
| "epoch": 3.56, | |
| "learning_rate": 1.833161662683672e-07, | |
| "logits/chosen": -2.0142505168914795, | |
| "logits/rejected": -2.0135538578033447, | |
| "logps/chosen": -29.253097534179688, | |
| "logps/rejected": -36.559532165527344, | |
| "loss": 0.1185, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 2.2218658924102783, | |
| "rewards/margins": 2.8047871589660645, | |
| "rewards/rejected": -0.5829211473464966, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 3.58, | |
| "learning_rate": 1.626139998169246e-07, | |
| "logits/chosen": -1.8797496557235718, | |
| "logits/rejected": -1.8874984979629517, | |
| "logps/chosen": -31.32064437866211, | |
| "logps/rejected": -37.89208221435547, | |
| "loss": 0.144, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 2.287868022918701, | |
| "rewards/margins": 2.641819477081299, | |
| "rewards/rejected": -0.3539513051509857, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 3.61, | |
| "learning_rate": 1.4311267331922535e-07, | |
| "logits/chosen": -1.8338594436645508, | |
| "logits/rejected": -1.8298187255859375, | |
| "logps/chosen": -31.842514038085938, | |
| "logps/rejected": -31.77047348022461, | |
| "loss": 0.1419, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 2.3397603034973145, | |
| "rewards/margins": 2.4782257080078125, | |
| "rewards/rejected": -0.13846543431282043, | |
| "step": 1390 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "learning_rate": 1.2482220564763669e-07, | |
| "logits/chosen": -1.9958465099334717, | |
| "logits/rejected": -1.9933058023452759, | |
| "logps/chosen": -28.800048828125, | |
| "logps/rejected": -32.510841369628906, | |
| "loss": 0.1553, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.9972022771835327, | |
| "rewards/margins": 2.2903003692626953, | |
| "rewards/rejected": -0.29309794306755066, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "eval_logits/chosen": -2.1668412685394287, | |
| "eval_logits/rejected": -2.162137985229492, | |
| "eval_logps/chosen": -33.57229232788086, | |
| "eval_logps/rejected": -37.258155822753906, | |
| "eval_loss": 0.46543803811073303, | |
| "eval_rewards/accuracies": 0.6034052968025208, | |
| "eval_rewards/chosen": 0.36980658769607544, | |
| "eval_rewards/margins": 0.16303294897079468, | |
| "eval_rewards/rejected": 0.20677369832992554, | |
| "eval_runtime": 145.9296, | |
| "eval_samples_per_second": 2.35, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.66, | |
| "learning_rate": 1.0775199359171346e-07, | |
| "logits/chosen": -1.9372856616973877, | |
| "logits/rejected": -1.9331865310668945, | |
| "logps/chosen": -30.8883113861084, | |
| "logps/rejected": -29.84084129333496, | |
| "loss": 0.1363, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.2033774852752686, | |
| "rewards/margins": 2.4324822425842285, | |
| "rewards/rejected": -0.22910502552986145, | |
| "step": 1410 | |
| }, | |
| { | |
| "epoch": 3.69, | |
| "learning_rate": 9.191080703056604e-08, | |
| "logits/chosen": -1.8948686122894287, | |
| "logits/rejected": -1.8960151672363281, | |
| "logps/chosen": -30.679889678955078, | |
| "logps/rejected": -35.14348602294922, | |
| "loss": 0.152, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.1479902267456055, | |
| "rewards/margins": 2.323575735092163, | |
| "rewards/rejected": -0.17558541893959045, | |
| "step": 1420 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "learning_rate": 7.730678442730539e-08, | |
| "logits/chosen": -1.8465667963027954, | |
| "logits/rejected": -1.8398901224136353, | |
| "logps/chosen": -31.448232650756836, | |
| "logps/rejected": -37.158573150634766, | |
| "loss": 0.1367, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.1548283100128174, | |
| "rewards/margins": 2.530158519744873, | |
| "rewards/rejected": -0.3753301799297333, | |
| "step": 1430 | |
| }, | |
| { | |
| "epoch": 3.74, | |
| "learning_rate": 6.394742864787806e-08, | |
| "logits/chosen": -1.859572172164917, | |
| "logits/rejected": -1.8541902303695679, | |
| "logps/chosen": -26.713485717773438, | |
| "logps/rejected": -32.027252197265625, | |
| "loss": 0.1488, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 2.0798051357269287, | |
| "rewards/margins": 2.317080020904541, | |
| "rewards/rejected": -0.2372746467590332, | |
| "step": 1440 | |
| }, | |
| { | |
| "epoch": 3.77, | |
| "learning_rate": 5.183960310644748e-08, | |
| "logits/chosen": -1.8826467990875244, | |
| "logits/rejected": -1.8725192546844482, | |
| "logps/chosen": -30.08587074279785, | |
| "logps/rejected": -36.018455505371094, | |
| "loss": 0.1431, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.9165277481079102, | |
| "rewards/margins": 2.503002405166626, | |
| "rewards/rejected": -0.5864745378494263, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "learning_rate": 4.098952823928693e-08, | |
| "logits/chosen": -1.8568382263183594, | |
| "logits/rejected": -1.8540499210357666, | |
| "logps/chosen": -30.78998374938965, | |
| "logps/rejected": -31.02651596069336, | |
| "loss": 0.1648, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.0158348083496094, | |
| "rewards/margins": 2.124389886856079, | |
| "rewards/rejected": -0.10855510085821152, | |
| "step": 1460 | |
| }, | |
| { | |
| "epoch": 3.82, | |
| "learning_rate": 3.1402778309014284e-08, | |
| "logits/chosen": -1.9178192615509033, | |
| "logits/rejected": -1.924325704574585, | |
| "logps/chosen": -29.135822296142578, | |
| "logps/rejected": -33.322288513183594, | |
| "loss": 0.1271, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.3466856479644775, | |
| "rewards/margins": 2.634577989578247, | |
| "rewards/rejected": -0.28789228200912476, | |
| "step": 1470 | |
| }, | |
| { | |
| "epoch": 3.84, | |
| "learning_rate": 2.3084278540791427e-08, | |
| "logits/chosen": -1.9184119701385498, | |
| "logits/rejected": -1.9287493228912354, | |
| "logps/chosen": -29.0635986328125, | |
| "logps/rejected": -29.924457550048828, | |
| "loss": 0.1386, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.1053242683410645, | |
| "rewards/margins": 2.342036008834839, | |
| "rewards/rejected": -0.2367115318775177, | |
| "step": 1480 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "learning_rate": 1.6038302591975807e-08, | |
| "logits/chosen": -1.8501653671264648, | |
| "logits/rejected": -1.8432512283325195, | |
| "logps/chosen": -31.104766845703125, | |
| "logps/rejected": -32.60163879394531, | |
| "loss": 0.144, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.0413756370544434, | |
| "rewards/margins": 2.3543033599853516, | |
| "rewards/rejected": -0.31292763352394104, | |
| "step": 1490 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "learning_rate": 1.0268470356514237e-08, | |
| "logits/chosen": -1.9025371074676514, | |
| "logits/rejected": -1.8992702960968018, | |
| "logps/chosen": -31.016992568969727, | |
| "logps/rejected": -33.745033264160156, | |
| "loss": 0.1447, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.086864709854126, | |
| "rewards/margins": 2.474714517593384, | |
| "rewards/rejected": -0.3878494203090668, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "eval_logits/chosen": -2.1669082641601562, | |
| "eval_logits/rejected": -2.1621992588043213, | |
| "eval_logps/chosen": -33.575931549072266, | |
| "eval_logps/rejected": -37.24650192260742, | |
| "eval_loss": 0.46842658519744873, | |
| "eval_rewards/accuracies": 0.574335515499115, | |
| "eval_rewards/chosen": 0.3668961524963379, | |
| "eval_rewards/margins": 0.15079911053180695, | |
| "eval_rewards/rejected": 0.21609705686569214, | |
| "eval_runtime": 145.9178, | |
| "eval_samples_per_second": 2.351, | |
| "eval_steps_per_second": 0.295, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.92, | |
| "learning_rate": 5.777746105209147e-09, | |
| "logits/chosen": -1.9817174673080444, | |
| "logits/rejected": -1.9823877811431885, | |
| "logps/chosen": -27.261791229248047, | |
| "logps/rejected": -33.12643814086914, | |
| "loss": 0.1654, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.0435855388641357, | |
| "rewards/margins": 2.35237717628479, | |
| "rewards/rejected": -0.30879148840904236, | |
| "step": 1510 | |
| }, | |
| { | |
| "epoch": 3.95, | |
| "learning_rate": 2.5684369628148352e-09, | |
| "logits/chosen": -1.834775686264038, | |
| "logits/rejected": -1.8347667455673218, | |
| "logps/chosen": -30.5068416595459, | |
| "logps/rejected": -34.2733154296875, | |
| "loss": 0.1565, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.0280237197875977, | |
| "rewards/margins": 2.347689151763916, | |
| "rewards/rejected": -0.31966525316238403, | |
| "step": 1520 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "learning_rate": 6.421917227455999e-10, | |
| "logits/chosen": -1.986581563949585, | |
| "logits/rejected": -1.9839556217193604, | |
| "logps/chosen": -29.06634521484375, | |
| "logps/rejected": -31.9202938079834, | |
| "loss": 0.1594, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.926171898841858, | |
| "rewards/margins": 2.2754859924316406, | |
| "rewards/rejected": -0.34931421279907227, | |
| "step": 1530 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -1.9706960916519165, | |
| "logits/rejected": -1.9718106985092163, | |
| "logps/chosen": -28.249774932861328, | |
| "logps/rejected": -30.143047332763672, | |
| "loss": 0.1594, | |
| "rewards/accuracies": 0.9583333730697632, | |
| "rewards/chosen": 1.8961349725723267, | |
| "rewards/margins": 2.178429126739502, | |
| "rewards/rejected": -0.2822941541671753, | |
| "step": 1540 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 1540, | |
| "total_flos": 0.0, | |
| "train_loss": 0.16076272354497537, | |
| "train_runtime": 10802.3033, | |
| "train_samples_per_second": 1.14, | |
| "train_steps_per_second": 0.143 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1540, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |