diff --git "a/trainer_state.json" "b/trainer_state.json" new file mode 100644--- /dev/null +++ "b/trainer_state.json" @@ -0,0 +1,7648 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 3.0, + "eval_steps": 500, + "global_step": 846, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0035555555555555557, + "grad_norm": 17.495248794555664, + "learning_rate": 2.5000000000000004e-07, + "loss": 0.9483866095542908, + "loss_sft": 0.9483866095542908, + "step": 1, + "token_acc": 0.7468001312766656 + }, + { + "epoch": 0.0071111111111111115, + "grad_norm": 16.598894119262695, + "learning_rate": 5.000000000000001e-07, + "loss": 0.9053852558135986, + "loss_sft": 0.9053852558135986, + "step": 2, + "token_acc": 0.7559464426358624 + }, + { + "epoch": 0.010666666666666666, + "grad_norm": 17.457666397094727, + "learning_rate": 7.5e-07, + "loss": 0.9427951574325562, + "loss_sft": 0.9427951574325562, + "step": 3, + "token_acc": 0.7498739919354839 + }, + { + "epoch": 0.014222222222222223, + "grad_norm": 17.3460693359375, + "learning_rate": 1.0000000000000002e-06, + "loss": 0.8845763802528381, + "loss_sft": 0.8845763802528381, + "step": 4, + "token_acc": 0.7514659685863875 + }, + { + "epoch": 0.017777777777777778, + "grad_norm": 13.943445205688477, + "learning_rate": 1.25e-06, + "loss": 0.8871760964393616, + "loss_sft": 0.8871760964393616, + "step": 5, + "token_acc": 0.7511533242876527 + }, + { + "epoch": 0.021333333333333333, + "grad_norm": 10.018475532531738, + "learning_rate": 1.5e-06, + "loss": 0.8330506682395935, + "loss_sft": 0.8330506682395935, + "step": 6, + "token_acc": 0.7620556939098936 + }, + { + "epoch": 0.024888888888888887, + "grad_norm": 8.78338623046875, + "learning_rate": 1.75e-06, + "loss": 0.8371806144714355, + "loss_sft": 0.8371806144714355, + "step": 7, + "token_acc": 0.7672026819705665 + }, + { + "epoch": 0.028444444444444446, + "grad_norm": 5.931888103485107, + "learning_rate": 2.0000000000000003e-06, + "loss": 0.769527792930603, + "loss_sft": 0.769527792930603, + "step": 8, + "token_acc": 0.7749389024661186 + }, + { + "epoch": 0.032, + "grad_norm": 5.365172386169434, + "learning_rate": 2.25e-06, + "loss": 0.7864335179328918, + "loss_sft": 0.7864335179328918, + "step": 9, + "token_acc": 0.7820643705335554 + }, + { + "epoch": 0.035555555555555556, + "grad_norm": 6.1973466873168945, + "learning_rate": 2.5e-06, + "loss": 0.7547352910041809, + "loss_sft": 0.7547352910041809, + "step": 10, + "token_acc": 0.782741879551497 + }, + { + "epoch": 0.03911111111111111, + "grad_norm": 3.893557071685791, + "learning_rate": 2.7500000000000004e-06, + "loss": 0.6970406770706177, + "loss_sft": 0.6970406770706177, + "step": 11, + "token_acc": 0.7988962244687435 + }, + { + "epoch": 0.042666666666666665, + "grad_norm": 3.581576108932495, + "learning_rate": 3e-06, + "loss": 0.7086485028266907, + "loss_sft": 0.7086485028266907, + "step": 12, + "token_acc": 0.8016231323229066 + }, + { + "epoch": 0.04622222222222222, + "grad_norm": 3.867741823196411, + "learning_rate": 3.2500000000000002e-06, + "loss": 0.7034077644348145, + "loss_sft": 0.7034077644348145, + "step": 13, + "token_acc": 0.799099606077659 + }, + { + "epoch": 0.049777777777777775, + "grad_norm": 2.989391565322876, + "learning_rate": 3.5e-06, + "loss": 0.6772717237472534, + "loss_sft": 0.6772717237472534, + "step": 14, + "token_acc": 0.8074077905954167 + }, + { + "epoch": 0.05333333333333334, + "grad_norm": 3.1689505577087402, + "learning_rate": 3.7500000000000005e-06, + "loss": 0.6785804033279419, + "loss_sft": 0.6785804033279419, + "step": 15, + "token_acc": 0.8026548672566372 + }, + { + "epoch": 0.05688888888888889, + "grad_norm": 2.881948947906494, + "learning_rate": 4.000000000000001e-06, + "loss": 0.6801018714904785, + "loss_sft": 0.6801018714904785, + "step": 16, + "token_acc": 0.8072206364309124 + }, + { + "epoch": 0.060444444444444446, + "grad_norm": 3.312012195587158, + "learning_rate": 4.25e-06, + "loss": 0.6272972822189331, + "loss_sft": 0.6272972822189331, + "step": 17, + "token_acc": 0.8205128205128205 + }, + { + "epoch": 0.064, + "grad_norm": 2.6100964546203613, + "learning_rate": 4.5e-06, + "loss": 0.6777547597885132, + "loss_sft": 0.6777547597885132, + "step": 18, + "token_acc": 0.8047497224501617 + }, + { + "epoch": 0.06755555555555555, + "grad_norm": 2.7741847038269043, + "learning_rate": 4.75e-06, + "loss": 0.6672918200492859, + "loss_sft": 0.6672918200492859, + "step": 19, + "token_acc": 0.8097198399085191 + }, + { + "epoch": 0.07111111111111111, + "grad_norm": 2.7010674476623535, + "learning_rate": 5e-06, + "loss": 0.6928048729896545, + "loss_sft": 0.6928048729896545, + "step": 20, + "token_acc": 0.8019967650175693 + }, + { + "epoch": 0.07466666666666667, + "grad_norm": 2.527926206588745, + "learning_rate": 5.2500000000000006e-06, + "loss": 0.6631382703781128, + "loss_sft": 0.6631382703781128, + "step": 21, + "token_acc": 0.8087185851690971 + }, + { + "epoch": 0.07822222222222222, + "grad_norm": 2.437958002090454, + "learning_rate": 5.500000000000001e-06, + "loss": 0.6603161096572876, + "loss_sft": 0.6603161096572876, + "step": 22, + "token_acc": 0.8108370359021501 + }, + { + "epoch": 0.08177777777777778, + "grad_norm": 2.62800669670105, + "learning_rate": 5.75e-06, + "loss": 0.670966386795044, + "loss_sft": 0.670966386795044, + "step": 23, + "token_acc": 0.803944530046225 + }, + { + "epoch": 0.08533333333333333, + "grad_norm": 2.581692695617676, + "learning_rate": 6e-06, + "loss": 0.6421648263931274, + "loss_sft": 0.6421648263931274, + "step": 24, + "token_acc": 0.8138567293777135 + }, + { + "epoch": 0.08888888888888889, + "grad_norm": 2.5290439128875732, + "learning_rate": 6.25e-06, + "loss": 0.6531534194946289, + "loss_sft": 0.6531534194946289, + "step": 25, + "token_acc": 0.8090807174887893 + }, + { + "epoch": 0.09244444444444444, + "grad_norm": 2.3785524368286133, + "learning_rate": 6.5000000000000004e-06, + "loss": 0.6678711175918579, + "loss_sft": 0.6678711175918579, + "step": 26, + "token_acc": 0.8087300315221456 + }, + { + "epoch": 0.096, + "grad_norm": 2.4084808826446533, + "learning_rate": 6.750000000000001e-06, + "loss": 0.6548430919647217, + "loss_sft": 0.6548430919647217, + "step": 27, + "token_acc": 0.808735583084152 + }, + { + "epoch": 0.09955555555555555, + "grad_norm": 2.3656678199768066, + "learning_rate": 7e-06, + "loss": 0.663451075553894, + "loss_sft": 0.663451075553894, + "step": 28, + "token_acc": 0.8122356644044457 + }, + { + "epoch": 0.10311111111111111, + "grad_norm": 2.322953701019287, + "learning_rate": 7.25e-06, + "loss": 0.658333420753479, + "loss_sft": 0.658333420753479, + "step": 29, + "token_acc": 0.8060753719350993 + }, + { + "epoch": 0.10666666666666667, + "grad_norm": 2.4030601978302, + "learning_rate": 7.500000000000001e-06, + "loss": 0.633331298828125, + "loss_sft": 0.633331298828125, + "step": 30, + "token_acc": 0.8110839272417828 + }, + { + "epoch": 0.11022222222222222, + "grad_norm": 2.5537326335906982, + "learning_rate": 7.75e-06, + "loss": 0.6702368259429932, + "loss_sft": 0.6702368259429932, + "step": 31, + "token_acc": 0.8042502568539447 + }, + { + "epoch": 0.11377777777777778, + "grad_norm": 2.438399076461792, + "learning_rate": 8.000000000000001e-06, + "loss": 0.6565305590629578, + "loss_sft": 0.6565305590629578, + "step": 32, + "token_acc": 0.8108615265743142 + }, + { + "epoch": 0.11733333333333333, + "grad_norm": 2.4852664470672607, + "learning_rate": 8.25e-06, + "loss": 0.6853891611099243, + "loss_sft": 0.6853891611099243, + "step": 33, + "token_acc": 0.8058776259901275 + }, + { + "epoch": 0.12088888888888889, + "grad_norm": 2.532184362411499, + "learning_rate": 8.5e-06, + "loss": 0.6334942579269409, + "loss_sft": 0.6334942579269409, + "step": 34, + "token_acc": 0.8146627936347562 + }, + { + "epoch": 0.12444444444444444, + "grad_norm": 2.295111656188965, + "learning_rate": 8.750000000000001e-06, + "loss": 0.6364755630493164, + "loss_sft": 0.6364755630493164, + "step": 35, + "token_acc": 0.8140566703640214 + }, + { + "epoch": 0.128, + "grad_norm": 2.223219871520996, + "learning_rate": 9e-06, + "loss": 0.6513509750366211, + "loss_sft": 0.6513509750366211, + "step": 36, + "token_acc": 0.8078332215284586 + }, + { + "epoch": 0.13155555555555556, + "grad_norm": 2.330840826034546, + "learning_rate": 9.250000000000001e-06, + "loss": 0.6392641067504883, + "loss_sft": 0.6392641067504883, + "step": 37, + "token_acc": 0.8155442727222005 + }, + { + "epoch": 0.1351111111111111, + "grad_norm": 2.365277051925659, + "learning_rate": 9.5e-06, + "loss": 0.6332631707191467, + "loss_sft": 0.6332631707191467, + "step": 38, + "token_acc": 0.8176216216216217 + }, + { + "epoch": 0.13866666666666666, + "grad_norm": 2.391047477722168, + "learning_rate": 9.75e-06, + "loss": 0.6312863230705261, + "loss_sft": 0.6312863230705261, + "step": 39, + "token_acc": 0.8118370810352142 + }, + { + "epoch": 0.14222222222222222, + "grad_norm": 2.532961130142212, + "learning_rate": 1e-05, + "loss": 0.6740774512290955, + "loss_sft": 0.6740774512290955, + "step": 40, + "token_acc": 0.8025179098187948 + }, + { + "epoch": 0.14577777777777778, + "grad_norm": 2.4365673065185547, + "learning_rate": 9.99996201876164e-06, + "loss": 0.632836103439331, + "loss_sft": 0.632836103439331, + "step": 41, + "token_acc": 0.8131967120819728 + }, + { + "epoch": 0.14933333333333335, + "grad_norm": 2.3039770126342773, + "learning_rate": 9.999848075623584e-06, + "loss": 0.662885308265686, + "loss_sft": 0.662885308265686, + "step": 42, + "token_acc": 0.8080118240660517 + }, + { + "epoch": 0.15288888888888888, + "grad_norm": 2.457951784133911, + "learning_rate": 9.999658172316915e-06, + "loss": 0.6900539398193359, + "loss_sft": 0.6900539398193359, + "step": 43, + "token_acc": 0.8015775728828013 + }, + { + "epoch": 0.15644444444444444, + "grad_norm": 2.2119102478027344, + "learning_rate": 9.999392311726738e-06, + "loss": 0.6972055435180664, + "loss_sft": 0.6972055435180664, + "step": 44, + "token_acc": 0.7983933787731257 + }, + { + "epoch": 0.16, + "grad_norm": 2.4633448123931885, + "learning_rate": 9.99905049789214e-06, + "loss": 0.6264879703521729, + "loss_sft": 0.6264879703521729, + "step": 45, + "token_acc": 0.8173712198528227 + }, + { + "epoch": 0.16355555555555557, + "grad_norm": 2.430586099624634, + "learning_rate": 9.998632736006124e-06, + "loss": 0.6588121652603149, + "loss_sft": 0.6588121652603149, + "step": 46, + "token_acc": 0.8072451066316098 + }, + { + "epoch": 0.1671111111111111, + "grad_norm": 2.3461453914642334, + "learning_rate": 9.998139032415534e-06, + "loss": 0.6481072902679443, + "loss_sft": 0.6481072902679443, + "step": 47, + "token_acc": 0.8127968866250787 + }, + { + "epoch": 0.17066666666666666, + "grad_norm": 3.693047523498535, + "learning_rate": 9.997569394620965e-06, + "loss": 0.6446748971939087, + "loss_sft": 0.6446748971939087, + "step": 48, + "token_acc": 0.8112321375037473 + }, + { + "epoch": 0.17422222222222222, + "grad_norm": 2.0652964115142822, + "learning_rate": 9.996923831276632e-06, + "loss": 0.6227473616600037, + "loss_sft": 0.6227473616600037, + "step": 49, + "token_acc": 0.817675647619497 + }, + { + "epoch": 0.17777777777777778, + "grad_norm": 2.402451276779175, + "learning_rate": 9.996202352190256e-06, + "loss": 0.6587773561477661, + "loss_sft": 0.6587773561477661, + "step": 50, + "token_acc": 0.8099698265023887 + }, + { + "epoch": 0.18133333333333335, + "grad_norm": 2.2670416831970215, + "learning_rate": 9.995404968322902e-06, + "loss": 0.678132951259613, + "loss_sft": 0.678132951259613, + "step": 51, + "token_acc": 0.7998999749937484 + }, + { + "epoch": 0.18488888888888888, + "grad_norm": 2.270219087600708, + "learning_rate": 9.994531691788822e-06, + "loss": 0.6402097940444946, + "loss_sft": 0.6402097940444946, + "step": 52, + "token_acc": 0.8093608566066894 + }, + { + "epoch": 0.18844444444444444, + "grad_norm": 2.130141258239746, + "learning_rate": 9.993582535855265e-06, + "loss": 0.6793354153633118, + "loss_sft": 0.6793354153633118, + "step": 53, + "token_acc": 0.8070858079049764 + }, + { + "epoch": 0.192, + "grad_norm": 2.125152349472046, + "learning_rate": 9.992557514942278e-06, + "loss": 0.661788821220398, + "loss_sft": 0.661788821220398, + "step": 54, + "token_acc": 0.8044157124856188 + }, + { + "epoch": 0.19555555555555557, + "grad_norm": 2.3009588718414307, + "learning_rate": 9.991456644622489e-06, + "loss": 0.6258248090744019, + "loss_sft": 0.6258248090744019, + "step": 55, + "token_acc": 0.8175399479747306 + }, + { + "epoch": 0.1991111111111111, + "grad_norm": 2.1973323822021484, + "learning_rate": 9.990279941620861e-06, + "loss": 0.6419394016265869, + "loss_sft": 0.6419394016265869, + "step": 56, + "token_acc": 0.8090412206786609 + }, + { + "epoch": 0.20266666666666666, + "grad_norm": 2.158705711364746, + "learning_rate": 9.989027423814454e-06, + "loss": 0.6332802772521973, + "loss_sft": 0.6332802772521973, + "step": 57, + "token_acc": 0.8108580508474577 + }, + { + "epoch": 0.20622222222222222, + "grad_norm": 2.1737403869628906, + "learning_rate": 9.987699110232134e-06, + "loss": 0.6522477865219116, + "loss_sft": 0.6522477865219116, + "step": 58, + "token_acc": 0.8111087887971575 + }, + { + "epoch": 0.20977777777777779, + "grad_norm": 2.262225389480591, + "learning_rate": 9.986295021054302e-06, + "loss": 0.6412181854248047, + "loss_sft": 0.6412181854248047, + "step": 59, + "token_acc": 0.813780621808545 + }, + { + "epoch": 0.21333333333333335, + "grad_norm": 2.3398172855377197, + "learning_rate": 9.984815177612574e-06, + "loss": 0.6547109484672546, + "loss_sft": 0.6547109484672546, + "step": 60, + "token_acc": 0.8076034819683781 + }, + { + "epoch": 0.21688888888888888, + "grad_norm": 2.313699960708618, + "learning_rate": 9.983259602389469e-06, + "loss": 0.6451797485351562, + "loss_sft": 0.6451797485351562, + "step": 61, + "token_acc": 0.814795259885017 + }, + { + "epoch": 0.22044444444444444, + "grad_norm": 2.1824562549591064, + "learning_rate": 9.98162831901805e-06, + "loss": 0.6661728620529175, + "loss_sft": 0.6661728620529175, + "step": 62, + "token_acc": 0.8045642146151979 + }, + { + "epoch": 0.224, + "grad_norm": 2.1974189281463623, + "learning_rate": 9.979921352281585e-06, + "loss": 0.7027256488800049, + "loss_sft": 0.7027256488800049, + "step": 63, + "token_acc": 0.7961222091656874 + }, + { + "epoch": 0.22755555555555557, + "grad_norm": 2.083812952041626, + "learning_rate": 9.97813872811316e-06, + "loss": 0.6293184757232666, + "loss_sft": 0.6293184757232666, + "step": 64, + "token_acc": 0.8159653531519008 + }, + { + "epoch": 0.2311111111111111, + "grad_norm": 2.170288324356079, + "learning_rate": 9.976280473595284e-06, + "loss": 0.6319900751113892, + "loss_sft": 0.6319900751113892, + "step": 65, + "token_acc": 0.8104868913857678 + }, + { + "epoch": 0.23466666666666666, + "grad_norm": 2.3147213459014893, + "learning_rate": 9.974346616959476e-06, + "loss": 0.6747528314590454, + "loss_sft": 0.6747528314590454, + "step": 66, + "token_acc": 0.807041863044081 + }, + { + "epoch": 0.23822222222222222, + "grad_norm": 2.2782130241394043, + "learning_rate": 9.972337187585848e-06, + "loss": 0.66471266746521, + "loss_sft": 0.66471266746521, + "step": 67, + "token_acc": 0.8058387395736794 + }, + { + "epoch": 0.24177777777777779, + "grad_norm": 2.2280936241149902, + "learning_rate": 9.970252216002647e-06, + "loss": 0.6100447177886963, + "loss_sft": 0.6100447177886963, + "step": 68, + "token_acc": 0.8156276931916117 + }, + { + "epoch": 0.24533333333333332, + "grad_norm": 2.1991140842437744, + "learning_rate": 9.96809173388579e-06, + "loss": 0.6443502902984619, + "loss_sft": 0.6443502902984619, + "step": 69, + "token_acc": 0.8088529014844804 + }, + { + "epoch": 0.24888888888888888, + "grad_norm": 2.138368844985962, + "learning_rate": 9.965855774058395e-06, + "loss": 0.6707169413566589, + "loss_sft": 0.6707169413566589, + "step": 70, + "token_acc": 0.8019056170807144 + }, + { + "epoch": 0.25244444444444447, + "grad_norm": 2.185886859893799, + "learning_rate": 9.96354437049027e-06, + "loss": 0.7145326137542725, + "loss_sft": 0.7145326137542725, + "step": 71, + "token_acc": 0.7981840490797546 + }, + { + "epoch": 0.256, + "grad_norm": 1.9851183891296387, + "learning_rate": 9.961157558297404e-06, + "loss": 0.6444203853607178, + "loss_sft": 0.6444203853607178, + "step": 72, + "token_acc": 0.8122939693620322 + }, + { + "epoch": 0.25955555555555554, + "grad_norm": 2.0093414783477783, + "learning_rate": 9.958695373741428e-06, + "loss": 0.621960461139679, + "loss_sft": 0.621960461139679, + "step": 73, + "token_acc": 0.814341565748069 + }, + { + "epoch": 0.26311111111111113, + "grad_norm": 2.067918300628662, + "learning_rate": 9.956157854229072e-06, + "loss": 0.644780158996582, + "loss_sft": 0.644780158996582, + "step": 74, + "token_acc": 0.812317016795932 + }, + { + "epoch": 0.26666666666666666, + "grad_norm": 2.119696855545044, + "learning_rate": 9.953545038311587e-06, + "loss": 0.6622623205184937, + "loss_sft": 0.6622623205184937, + "step": 75, + "token_acc": 0.8045563549160671 + }, + { + "epoch": 0.2702222222222222, + "grad_norm": 2.231426954269409, + "learning_rate": 9.950856965684167e-06, + "loss": 0.6669363379478455, + "loss_sft": 0.6669363379478455, + "step": 76, + "token_acc": 0.804698820120058 + }, + { + "epoch": 0.2737777777777778, + "grad_norm": 2.0908682346343994, + "learning_rate": 9.948093677185345e-06, + "loss": 0.5849854946136475, + "loss_sft": 0.5849854946136475, + "step": 77, + "token_acc": 0.8281780902167546 + }, + { + "epoch": 0.2773333333333333, + "grad_norm": 2.106154680252075, + "learning_rate": 9.945255214796366e-06, + "loss": 0.6207313537597656, + "loss_sft": 0.6207313537597656, + "step": 78, + "token_acc": 0.8166472319395858 + }, + { + "epoch": 0.2808888888888889, + "grad_norm": 2.150275230407715, + "learning_rate": 9.942341621640558e-06, + "loss": 0.6709491014480591, + "loss_sft": 0.6709491014480591, + "step": 79, + "token_acc": 0.8041940990002439 + }, + { + "epoch": 0.28444444444444444, + "grad_norm": 2.0800719261169434, + "learning_rate": 9.939352941982671e-06, + "loss": 0.6028379201889038, + "loss_sft": 0.6028379201889038, + "step": 80, + "token_acc": 0.8237361901573486 + }, + { + "epoch": 0.288, + "grad_norm": 2.2775790691375732, + "learning_rate": 9.936289221228207e-06, + "loss": 0.6904752254486084, + "loss_sft": 0.6904752254486084, + "step": 81, + "token_acc": 0.7974901960784314 + }, + { + "epoch": 0.29155555555555557, + "grad_norm": 2.1881532669067383, + "learning_rate": 9.93315050592273e-06, + "loss": 0.6303409337997437, + "loss_sft": 0.6303409337997437, + "step": 82, + "token_acc": 0.8094717099093467 + }, + { + "epoch": 0.2951111111111111, + "grad_norm": 2.125056266784668, + "learning_rate": 9.929936843751158e-06, + "loss": 0.7039297819137573, + "loss_sft": 0.7039297819137573, + "step": 83, + "token_acc": 0.801138281965948 + }, + { + "epoch": 0.2986666666666667, + "grad_norm": 2.0034618377685547, + "learning_rate": 9.926648283537037e-06, + "loss": 0.598913311958313, + "loss_sft": 0.598913311958313, + "step": 84, + "token_acc": 0.8245642006802721 + }, + { + "epoch": 0.3022222222222222, + "grad_norm": 2.112959861755371, + "learning_rate": 9.923284875241802e-06, + "loss": 0.6272166967391968, + "loss_sft": 0.6272166967391968, + "step": 85, + "token_acc": 0.8157683884407885 + }, + { + "epoch": 0.30577777777777776, + "grad_norm": 2.084275960922241, + "learning_rate": 9.91984666996402e-06, + "loss": 0.6814351081848145, + "loss_sft": 0.6814351081848145, + "step": 86, + "token_acc": 0.7992861207387099 + }, + { + "epoch": 0.30933333333333335, + "grad_norm": 2.0872268676757812, + "learning_rate": 9.916333719938608e-06, + "loss": 0.6420176029205322, + "loss_sft": 0.6420176029205322, + "step": 87, + "token_acc": 0.8083574304019958 + }, + { + "epoch": 0.3128888888888889, + "grad_norm": 1.9719644784927368, + "learning_rate": 9.912746078536044e-06, + "loss": 0.6596208214759827, + "loss_sft": 0.6596208214759827, + "step": 88, + "token_acc": 0.8058035714285714 + }, + { + "epoch": 0.3164444444444444, + "grad_norm": 2.207044839859009, + "learning_rate": 9.90908380026155e-06, + "loss": 0.6528189778327942, + "loss_sft": 0.6528189778327942, + "step": 89, + "token_acc": 0.8052219633148116 + }, + { + "epoch": 0.32, + "grad_norm": 1.9884454011917114, + "learning_rate": 9.905346940754274e-06, + "loss": 0.6233652830123901, + "loss_sft": 0.6233652830123901, + "step": 90, + "token_acc": 0.8167207541338278 + }, + { + "epoch": 0.32355555555555554, + "grad_norm": 2.0874171257019043, + "learning_rate": 9.90153555678644e-06, + "loss": 0.6191186904907227, + "loss_sft": 0.6191186904907227, + "step": 91, + "token_acc": 0.8176688515246509 + }, + { + "epoch": 0.32711111111111113, + "grad_norm": 1.953534483909607, + "learning_rate": 9.897649706262474e-06, + "loss": 0.6781367063522339, + "loss_sft": 0.6781367063522339, + "step": 92, + "token_acc": 0.8023043686989918 + }, + { + "epoch": 0.33066666666666666, + "grad_norm": 2.1874265670776367, + "learning_rate": 9.893689448218146e-06, + "loss": 0.670505166053772, + "loss_sft": 0.670505166053772, + "step": 93, + "token_acc": 0.8013241439192489 + }, + { + "epoch": 0.3342222222222222, + "grad_norm": 2.011523962020874, + "learning_rate": 9.889654842819658e-06, + "loss": 0.6426149606704712, + "loss_sft": 0.6426149606704712, + "step": 94, + "token_acc": 0.8115926615261898 + }, + { + "epoch": 0.3377777777777778, + "grad_norm": 2.045391321182251, + "learning_rate": 9.885545951362733e-06, + "loss": 0.6748344898223877, + "loss_sft": 0.6748344898223877, + "step": 95, + "token_acc": 0.802775300936246 + }, + { + "epoch": 0.3413333333333333, + "grad_norm": 1.9540326595306396, + "learning_rate": 9.881362836271686e-06, + "loss": 0.6798080205917358, + "loss_sft": 0.6798080205917358, + "step": 96, + "token_acc": 0.8037626185377791 + }, + { + "epoch": 0.3448888888888889, + "grad_norm": 1.875409483909607, + "learning_rate": 9.877105561098473e-06, + "loss": 0.6294456124305725, + "loss_sft": 0.6294456124305725, + "step": 97, + "token_acc": 0.8121143562320033 + }, + { + "epoch": 0.34844444444444445, + "grad_norm": 2.1824817657470703, + "learning_rate": 9.872774190521727e-06, + "loss": 0.5989289283752441, + "loss_sft": 0.5989289283752441, + "step": 98, + "token_acc": 0.8252853380158033 + }, + { + "epoch": 0.352, + "grad_norm": 2.101071834564209, + "learning_rate": 9.868368790345777e-06, + "loss": 0.6433327198028564, + "loss_sft": 0.6433327198028564, + "step": 99, + "token_acc": 0.8120848385969002 + }, + { + "epoch": 0.35555555555555557, + "grad_norm": 2.03118634223938, + "learning_rate": 9.863889427499641e-06, + "loss": 0.616054117679596, + "loss_sft": 0.616054117679596, + "step": 100, + "token_acc": 0.8193526710717886 + }, + { + "epoch": 0.3591111111111111, + "grad_norm": 2.0384483337402344, + "learning_rate": 9.859336170036022e-06, + "loss": 0.611056387424469, + "loss_sft": 0.611056387424469, + "step": 101, + "token_acc": 0.8185125971663267 + }, + { + "epoch": 0.3626666666666667, + "grad_norm": 2.136881113052368, + "learning_rate": 9.854709087130261e-06, + "loss": 0.5821218490600586, + "loss_sft": 0.5821218490600586, + "step": 102, + "token_acc": 0.8229521887396478 + }, + { + "epoch": 0.3662222222222222, + "grad_norm": 2.26222825050354, + "learning_rate": 9.850008249079295e-06, + "loss": 0.6048051118850708, + "loss_sft": 0.6048051118850708, + "step": 103, + "token_acc": 0.8232665133092343 + }, + { + "epoch": 0.36977777777777776, + "grad_norm": 2.1815664768218994, + "learning_rate": 9.84523372730058e-06, + "loss": 0.6702889800071716, + "loss_sft": 0.6702889800071716, + "step": 104, + "token_acc": 0.8005819823029872 + }, + { + "epoch": 0.37333333333333335, + "grad_norm": 2.0779106616973877, + "learning_rate": 9.840385594331022e-06, + "loss": 0.7162292003631592, + "loss_sft": 0.7162292003631592, + "step": 105, + "token_acc": 0.7944361860267573 + }, + { + "epoch": 0.3768888888888889, + "grad_norm": 2.0308585166931152, + "learning_rate": 9.835463923825854e-06, + "loss": 0.6526224613189697, + "loss_sft": 0.6526224613189697, + "step": 106, + "token_acc": 0.8078534593897723 + }, + { + "epoch": 0.3804444444444444, + "grad_norm": 1.9520409107208252, + "learning_rate": 9.830468790557536e-06, + "loss": 0.6701065301895142, + "loss_sft": 0.6701065301895142, + "step": 107, + "token_acc": 0.8039615051682876 + }, + { + "epoch": 0.384, + "grad_norm": 2.019804000854492, + "learning_rate": 9.825400270414602e-06, + "loss": 0.6107809543609619, + "loss_sft": 0.6107809543609619, + "step": 108, + "token_acc": 0.8163130768308762 + }, + { + "epoch": 0.38755555555555554, + "grad_norm": 1.9157466888427734, + "learning_rate": 9.820258440400525e-06, + "loss": 0.6843827962875366, + "loss_sft": 0.6843827962875366, + "step": 109, + "token_acc": 0.804978649068323 + }, + { + "epoch": 0.39111111111111113, + "grad_norm": 1.8669224977493286, + "learning_rate": 9.81504337863253e-06, + "loss": 0.6194301247596741, + "loss_sft": 0.6194301247596741, + "step": 110, + "token_acc": 0.8224382553862323 + }, + { + "epoch": 0.39466666666666667, + "grad_norm": 2.007814645767212, + "learning_rate": 9.809755164340423e-06, + "loss": 0.6097871661186218, + "loss_sft": 0.6097871661186218, + "step": 111, + "token_acc": 0.8162617577471291 + }, + { + "epoch": 0.3982222222222222, + "grad_norm": 1.8429173231124878, + "learning_rate": 9.804393877865373e-06, + "loss": 0.6324461102485657, + "loss_sft": 0.6324461102485657, + "step": 112, + "token_acc": 0.8144855144855145 + }, + { + "epoch": 0.4017777777777778, + "grad_norm": 2.086236000061035, + "learning_rate": 9.798959600658697e-06, + "loss": 0.623206615447998, + "loss_sft": 0.623206615447998, + "step": 113, + "token_acc": 0.8170318205529473 + }, + { + "epoch": 0.4053333333333333, + "grad_norm": 2.0972321033477783, + "learning_rate": 9.79345241528063e-06, + "loss": 0.5557901263237, + "loss_sft": 0.5557901263237, + "step": 114, + "token_acc": 0.8360274308106784 + }, + { + "epoch": 0.4088888888888889, + "grad_norm": 1.9284859895706177, + "learning_rate": 9.787872405399059e-06, + "loss": 0.6329312920570374, + "loss_sft": 0.6329312920570374, + "step": 115, + "token_acc": 0.8117099135063207 + }, + { + "epoch": 0.41244444444444445, + "grad_norm": 1.9095735549926758, + "learning_rate": 9.782219655788257e-06, + "loss": 0.6140768527984619, + "loss_sft": 0.6140768527984619, + "step": 116, + "token_acc": 0.8201055322852635 + }, + { + "epoch": 0.416, + "grad_norm": 1.8941130638122559, + "learning_rate": 9.776494252327597e-06, + "loss": 0.6196410655975342, + "loss_sft": 0.6196410655975342, + "step": 117, + "token_acc": 0.8153930841215032 + }, + { + "epoch": 0.41955555555555557, + "grad_norm": 1.9722214937210083, + "learning_rate": 9.770696282000245e-06, + "loss": 0.6696373224258423, + "loss_sft": 0.6696373224258423, + "step": 118, + "token_acc": 0.8047451945739276 + }, + { + "epoch": 0.4231111111111111, + "grad_norm": 2.2128162384033203, + "learning_rate": 9.764825832891837e-06, + "loss": 0.7269062399864197, + "loss_sft": 0.7269062399864197, + "step": 119, + "token_acc": 0.7861466759522151 + }, + { + "epoch": 0.4266666666666667, + "grad_norm": 2.00886607170105, + "learning_rate": 9.758882994189145e-06, + "loss": 0.6605268120765686, + "loss_sft": 0.6605268120765686, + "step": 120, + "token_acc": 0.8074494326591021 + }, + { + "epoch": 0.43022222222222223, + "grad_norm": 1.989458441734314, + "learning_rate": 9.752867856178719e-06, + "loss": 0.692130446434021, + "loss_sft": 0.692130446434021, + "step": 121, + "token_acc": 0.7941129758060247 + }, + { + "epoch": 0.43377777777777776, + "grad_norm": 1.9358607530593872, + "learning_rate": 9.746780510245512e-06, + "loss": 0.6602381467819214, + "loss_sft": 0.6602381467819214, + "step": 122, + "token_acc": 0.8074204946996466 + }, + { + "epoch": 0.43733333333333335, + "grad_norm": 2.0435028076171875, + "learning_rate": 9.740621048871501e-06, + "loss": 0.6339927911758423, + "loss_sft": 0.6339927911758423, + "step": 123, + "token_acc": 0.8094589996597482 + }, + { + "epoch": 0.4408888888888889, + "grad_norm": 1.8236545324325562, + "learning_rate": 9.734389565634277e-06, + "loss": 0.641081690788269, + "loss_sft": 0.641081690788269, + "step": 124, + "token_acc": 0.8104552449982624 + }, + { + "epoch": 0.4444444444444444, + "grad_norm": 1.9895737171173096, + "learning_rate": 9.728086155205614e-06, + "loss": 0.6381692886352539, + "loss_sft": 0.6381692886352539, + "step": 125, + "token_acc": 0.8126674912389198 + }, + { + "epoch": 0.448, + "grad_norm": 2.0070884227752686, + "learning_rate": 9.721710913350048e-06, + "loss": 0.6206820011138916, + "loss_sft": 0.6206820011138916, + "step": 126, + "token_acc": 0.8162042691793172 + }, + { + "epoch": 0.45155555555555554, + "grad_norm": 2.2156944274902344, + "learning_rate": 9.715263936923413e-06, + "loss": 0.6623772382736206, + "loss_sft": 0.6623772382736206, + "step": 127, + "token_acc": 0.8058734316834414 + }, + { + "epoch": 0.45511111111111113, + "grad_norm": 1.8213300704956055, + "learning_rate": 9.708745323871369e-06, + "loss": 0.6533066034317017, + "loss_sft": 0.6533066034317017, + "step": 128, + "token_acc": 0.8077866476385854 + }, + { + "epoch": 0.45866666666666667, + "grad_norm": 2.1231634616851807, + "learning_rate": 9.702155173227911e-06, + "loss": 0.6018190979957581, + "loss_sft": 0.6018190979957581, + "step": 129, + "token_acc": 0.8208051981144094 + }, + { + "epoch": 0.4622222222222222, + "grad_norm": 1.9561680555343628, + "learning_rate": 9.695493585113873e-06, + "loss": 0.6718469262123108, + "loss_sft": 0.6718469262123108, + "step": 130, + "token_acc": 0.8073103100579784 + }, + { + "epoch": 0.4657777777777778, + "grad_norm": 1.9308838844299316, + "learning_rate": 9.688760660735403e-06, + "loss": 0.657689094543457, + "loss_sft": 0.657689094543457, + "step": 131, + "token_acc": 0.8096878695942819 + }, + { + "epoch": 0.4693333333333333, + "grad_norm": 2.006679058074951, + "learning_rate": 9.681956502382423e-06, + "loss": 0.6275889873504639, + "loss_sft": 0.6275889873504639, + "step": 132, + "token_acc": 0.8074872645421158 + }, + { + "epoch": 0.4728888888888889, + "grad_norm": 2.0132620334625244, + "learning_rate": 9.675081213427076e-06, + "loss": 0.6728153228759766, + "loss_sft": 0.6728153228759766, + "step": 133, + "token_acc": 0.8038592259620656 + }, + { + "epoch": 0.47644444444444445, + "grad_norm": 1.9020133018493652, + "learning_rate": 9.668134898322157e-06, + "loss": 0.6304796934127808, + "loss_sft": 0.6304796934127808, + "step": 134, + "token_acc": 0.8128842700908646 + }, + { + "epoch": 0.48, + "grad_norm": 1.8926805257797241, + "learning_rate": 9.661117662599527e-06, + "loss": 0.6505844593048096, + "loss_sft": 0.6505844593048096, + "step": 135, + "token_acc": 0.808865969626721 + }, + { + "epoch": 0.48355555555555557, + "grad_norm": 2.0156068801879883, + "learning_rate": 9.654029612868507e-06, + "loss": 0.6537829637527466, + "loss_sft": 0.6537829637527466, + "step": 136, + "token_acc": 0.8107277075503929 + }, + { + "epoch": 0.4871111111111111, + "grad_norm": 1.8975149393081665, + "learning_rate": 9.646870856814259e-06, + "loss": 0.6245245933532715, + "loss_sft": 0.6245245933532715, + "step": 137, + "token_acc": 0.820132198073045 + }, + { + "epoch": 0.49066666666666664, + "grad_norm": 1.8280093669891357, + "learning_rate": 9.639641503196152e-06, + "loss": 0.6352229118347168, + "loss_sft": 0.6352229118347168, + "step": 138, + "token_acc": 0.8123537669630323 + }, + { + "epoch": 0.49422222222222223, + "grad_norm": 1.9423892498016357, + "learning_rate": 9.632341661846107e-06, + "loss": 0.6176074743270874, + "loss_sft": 0.6176074743270874, + "step": 139, + "token_acc": 0.8150616636053529 + }, + { + "epoch": 0.49777777777777776, + "grad_norm": 2.017029285430908, + "learning_rate": 9.62497144366693e-06, + "loss": 0.6960232257843018, + "loss_sft": 0.6960232257843018, + "step": 140, + "token_acc": 0.8044230479915752 + }, + { + "epoch": 0.5013333333333333, + "grad_norm": 2.022019624710083, + "learning_rate": 9.617530960630624e-06, + "loss": 0.6346350908279419, + "loss_sft": 0.6346350908279419, + "step": 141, + "token_acc": 0.8122124863088719 + }, + { + "epoch": 0.5048888888888889, + "grad_norm": 1.8554091453552246, + "learning_rate": 9.610020325776694e-06, + "loss": 0.6463602185249329, + "loss_sft": 0.6463602185249329, + "step": 142, + "token_acc": 0.8112031173891866 + }, + { + "epoch": 0.5084444444444445, + "grad_norm": 1.9547128677368164, + "learning_rate": 9.602439653210426e-06, + "loss": 0.5980263948440552, + "loss_sft": 0.5980263948440552, + "step": 143, + "token_acc": 0.8247139855241653 + }, + { + "epoch": 0.512, + "grad_norm": 1.7775309085845947, + "learning_rate": 9.594789058101154e-06, + "loss": 0.6348342895507812, + "loss_sft": 0.6348342895507812, + "step": 144, + "token_acc": 0.8106483469300129 + }, + { + "epoch": 0.5155555555555555, + "grad_norm": 1.913460373878479, + "learning_rate": 9.587068656680506e-06, + "loss": 0.6441361904144287, + "loss_sft": 0.6441361904144287, + "step": 145, + "token_acc": 0.8114478114478114 + }, + { + "epoch": 0.5191111111111111, + "grad_norm": 1.8312528133392334, + "learning_rate": 9.579278566240646e-06, + "loss": 0.6320350170135498, + "loss_sft": 0.6320350170135498, + "step": 146, + "token_acc": 0.8135441245870694 + }, + { + "epoch": 0.5226666666666666, + "grad_norm": 1.8235299587249756, + "learning_rate": 9.571418905132486e-06, + "loss": 0.6090279817581177, + "loss_sft": 0.6090279817581177, + "step": 147, + "token_acc": 0.8201888162672476 + }, + { + "epoch": 0.5262222222222223, + "grad_norm": 2.0719492435455322, + "learning_rate": 9.56348979276389e-06, + "loss": 0.628081202507019, + "loss_sft": 0.628081202507019, + "step": 148, + "token_acc": 0.8147393333771613 + }, + { + "epoch": 0.5297777777777778, + "grad_norm": 1.8468698263168335, + "learning_rate": 9.555491349597862e-06, + "loss": 0.6069492697715759, + "loss_sft": 0.6069492697715759, + "step": 149, + "token_acc": 0.8200987736529605 + }, + { + "epoch": 0.5333333333333333, + "grad_norm": 1.9947658777236938, + "learning_rate": 9.547423697150714e-06, + "loss": 0.6571695804595947, + "loss_sft": 0.6571695804595947, + "step": 150, + "token_acc": 0.8059766843631985 + }, + { + "epoch": 0.5368888888888889, + "grad_norm": 1.9690507650375366, + "learning_rate": 9.539286957990215e-06, + "loss": 0.6367045640945435, + "loss_sft": 0.6367045640945435, + "step": 151, + "token_acc": 0.8123455337177827 + }, + { + "epoch": 0.5404444444444444, + "grad_norm": 1.8724870681762695, + "learning_rate": 9.53108125573374e-06, + "loss": 0.6026374101638794, + "loss_sft": 0.6026374101638794, + "step": 152, + "token_acc": 0.8243138983798082 + }, + { + "epoch": 0.544, + "grad_norm": 2.0106894969940186, + "learning_rate": 9.52280671504638e-06, + "loss": 0.6913030743598938, + "loss_sft": 0.6913030743598938, + "step": 153, + "token_acc": 0.796599214438423 + }, + { + "epoch": 0.5475555555555556, + "grad_norm": 1.9416288137435913, + "learning_rate": 9.514463461639055e-06, + "loss": 0.672147810459137, + "loss_sft": 0.672147810459137, + "step": 154, + "token_acc": 0.8018356067136555 + }, + { + "epoch": 0.5511111111111111, + "grad_norm": 1.9028187990188599, + "learning_rate": 9.506051622266608e-06, + "loss": 0.6289101839065552, + "loss_sft": 0.6289101839065552, + "step": 155, + "token_acc": 0.8168169819621646 + }, + { + "epoch": 0.5546666666666666, + "grad_norm": 1.9217112064361572, + "learning_rate": 9.497571324725865e-06, + "loss": 0.6344020962715149, + "loss_sft": 0.6344020962715149, + "step": 156, + "token_acc": 0.8172468789528681 + }, + { + "epoch": 0.5582222222222222, + "grad_norm": 1.8635435104370117, + "learning_rate": 9.48902269785371e-06, + "loss": 0.6761628985404968, + "loss_sft": 0.6761628985404968, + "step": 157, + "token_acc": 0.8048990333518903 + }, + { + "epoch": 0.5617777777777778, + "grad_norm": 1.7905099391937256, + "learning_rate": 9.480405871525114e-06, + "loss": 0.6140603423118591, + "loss_sft": 0.6140603423118591, + "step": 158, + "token_acc": 0.8178053830227743 + }, + { + "epoch": 0.5653333333333334, + "grad_norm": 1.9200325012207031, + "learning_rate": 9.471720976651173e-06, + "loss": 0.6581777334213257, + "loss_sft": 0.6581777334213257, + "step": 159, + "token_acc": 0.8099182561307902 + }, + { + "epoch": 0.5688888888888889, + "grad_norm": 1.7868720293045044, + "learning_rate": 9.462968145177112e-06, + "loss": 0.62703937292099, + "loss_sft": 0.62703937292099, + "step": 160, + "token_acc": 0.8138444839155319 + }, + { + "epoch": 0.5724444444444444, + "grad_norm": 1.9432843923568726, + "learning_rate": 9.45414751008028e-06, + "loss": 0.5996054410934448, + "loss_sft": 0.5996054410934448, + "step": 161, + "token_acc": 0.8247170388498012 + }, + { + "epoch": 0.576, + "grad_norm": 2.04793643951416, + "learning_rate": 9.445259205368138e-06, + "loss": 0.6464821696281433, + "loss_sft": 0.6464821696281433, + "step": 162, + "token_acc": 0.812093023255814 + }, + { + "epoch": 0.5795555555555556, + "grad_norm": 2.112752676010132, + "learning_rate": 9.436303366076213e-06, + "loss": 0.6265243291854858, + "loss_sft": 0.6265243291854858, + "step": 163, + "token_acc": 0.8191344246031746 + }, + { + "epoch": 0.5831111111111111, + "grad_norm": 1.9839770793914795, + "learning_rate": 9.427280128266049e-06, + "loss": 0.6514317989349365, + "loss_sft": 0.6514317989349365, + "step": 164, + "token_acc": 0.8091738049184858 + }, + { + "epoch": 0.5866666666666667, + "grad_norm": 2.0614960193634033, + "learning_rate": 9.418189629023149e-06, + "loss": 0.6396028995513916, + "loss_sft": 0.6396028995513916, + "step": 165, + "token_acc": 0.8086371080245042 + }, + { + "epoch": 0.5902222222222222, + "grad_norm": 2.1397998332977295, + "learning_rate": 9.409032006454877e-06, + "loss": 0.6246759295463562, + "loss_sft": 0.6246759295463562, + "step": 166, + "token_acc": 0.8148710166919575 + }, + { + "epoch": 0.5937777777777777, + "grad_norm": 1.7479585409164429, + "learning_rate": 9.399807399688371e-06, + "loss": 0.5669630169868469, + "loss_sft": 0.5669630169868469, + "step": 167, + "token_acc": 0.832521857494372 + }, + { + "epoch": 0.5973333333333334, + "grad_norm": 2.1200382709503174, + "learning_rate": 9.39051594886843e-06, + "loss": 0.6390013098716736, + "loss_sft": 0.6390013098716736, + "step": 168, + "token_acc": 0.8099659378090319 + }, + { + "epoch": 0.6008888888888889, + "grad_norm": 2.092478036880493, + "learning_rate": 9.381157795155374e-06, + "loss": 0.6896307468414307, + "loss_sft": 0.6896307468414307, + "step": 169, + "token_acc": 0.8005585955312358 + }, + { + "epoch": 0.6044444444444445, + "grad_norm": 2.019629955291748, + "learning_rate": 9.371733080722911e-06, + "loss": 0.6213470697402954, + "loss_sft": 0.6213470697402954, + "step": 170, + "token_acc": 0.8172556693910441 + }, + { + "epoch": 0.608, + "grad_norm": 1.8019428253173828, + "learning_rate": 9.36224194875597e-06, + "loss": 0.639850378036499, + "loss_sft": 0.639850378036499, + "step": 171, + "token_acc": 0.8119359534206696 + }, + { + "epoch": 0.6115555555555555, + "grad_norm": 1.9416800737380981, + "learning_rate": 9.352684543448532e-06, + "loss": 0.6339808702468872, + "loss_sft": 0.6339808702468872, + "step": 172, + "token_acc": 0.8085903083700441 + }, + { + "epoch": 0.6151111111111112, + "grad_norm": 1.9458688497543335, + "learning_rate": 9.343061010001428e-06, + "loss": 0.6805344223976135, + "loss_sft": 0.6805344223976135, + "step": 173, + "token_acc": 0.7990944559510621 + }, + { + "epoch": 0.6186666666666667, + "grad_norm": 2.05065655708313, + "learning_rate": 9.33337149462015e-06, + "loss": 0.5782161951065063, + "loss_sft": 0.5782161951065063, + "step": 174, + "token_acc": 0.8284091725972252 + }, + { + "epoch": 0.6222222222222222, + "grad_norm": 1.895094394683838, + "learning_rate": 9.323616144512612e-06, + "loss": 0.6329742074012756, + "loss_sft": 0.6329742074012756, + "step": 175, + "token_acc": 0.8152491023339318 + }, + { + "epoch": 0.6257777777777778, + "grad_norm": 1.899973750114441, + "learning_rate": 9.313795107886925e-06, + "loss": 0.6038188934326172, + "loss_sft": 0.6038188934326172, + "step": 176, + "token_acc": 0.81999531999532 + }, + { + "epoch": 0.6293333333333333, + "grad_norm": 1.8051780462265015, + "learning_rate": 9.303908533949146e-06, + "loss": 0.5864337086677551, + "loss_sft": 0.5864337086677551, + "step": 177, + "token_acc": 0.8251345733624686 + }, + { + "epoch": 0.6328888888888888, + "grad_norm": 1.805015206336975, + "learning_rate": 9.293956572900999e-06, + "loss": 0.6283878087997437, + "loss_sft": 0.6283878087997437, + "step": 178, + "token_acc": 0.8134474943339209 + }, + { + "epoch": 0.6364444444444445, + "grad_norm": 1.8181071281433105, + "learning_rate": 9.283939375937609e-06, + "loss": 0.6102030277252197, + "loss_sft": 0.6102030277252197, + "step": 179, + "token_acc": 0.8173055859802848 + }, + { + "epoch": 0.64, + "grad_norm": 1.7828445434570312, + "learning_rate": 9.273857095245192e-06, + "loss": 0.606691837310791, + "loss_sft": 0.606691837310791, + "step": 180, + "token_acc": 0.8201410848941864 + }, + { + "epoch": 0.6435555555555555, + "grad_norm": 1.8373500108718872, + "learning_rate": 9.263709883998753e-06, + "loss": 0.5953468680381775, + "loss_sft": 0.5953468680381775, + "step": 181, + "token_acc": 0.8222874237717178 + }, + { + "epoch": 0.6471111111111111, + "grad_norm": 1.910115361213684, + "learning_rate": 9.253497896359749e-06, + "loss": 0.6695414781570435, + "loss_sft": 0.6695414781570435, + "step": 182, + "token_acc": 0.8046351451973194 + }, + { + "epoch": 0.6506666666666666, + "grad_norm": 1.9019252061843872, + "learning_rate": 9.243221287473755e-06, + "loss": 0.6111459732055664, + "loss_sft": 0.6111459732055664, + "step": 183, + "token_acc": 0.8168404835269021 + }, + { + "epoch": 0.6542222222222223, + "grad_norm": 1.7855130434036255, + "learning_rate": 9.232880213468106e-06, + "loss": 0.6460024118423462, + "loss_sft": 0.6460024118423462, + "step": 184, + "token_acc": 0.8109094401229233 + }, + { + "epoch": 0.6577777777777778, + "grad_norm": 1.9407767057418823, + "learning_rate": 9.222474831449519e-06, + "loss": 0.6573270559310913, + "loss_sft": 0.6573270559310913, + "step": 185, + "token_acc": 0.8080587640389632 + }, + { + "epoch": 0.6613333333333333, + "grad_norm": 2.007049798965454, + "learning_rate": 9.212005299501712e-06, + "loss": 0.6129324436187744, + "loss_sft": 0.6129324436187744, + "step": 186, + "token_acc": 0.8197684565710629 + }, + { + "epoch": 0.6648888888888889, + "grad_norm": 1.9091486930847168, + "learning_rate": 9.201471776682999e-06, + "loss": 0.5998786687850952, + "loss_sft": 0.5998786687850952, + "step": 187, + "token_acc": 0.8241437940247753 + }, + { + "epoch": 0.6684444444444444, + "grad_norm": 1.9870213270187378, + "learning_rate": 9.19087442302388e-06, + "loss": 0.6057083606719971, + "loss_sft": 0.6057083606719971, + "step": 188, + "token_acc": 0.8188895834517247 + }, + { + "epoch": 0.672, + "grad_norm": 1.8174588680267334, + "learning_rate": 9.180213399524599e-06, + "loss": 0.6237980723381042, + "loss_sft": 0.6237980723381042, + "step": 189, + "token_acc": 0.8141720510492871 + }, + { + "epoch": 0.6755555555555556, + "grad_norm": 1.8761476278305054, + "learning_rate": 9.169488868152704e-06, + "loss": 0.6264516115188599, + "loss_sft": 0.6264516115188599, + "step": 190, + "token_acc": 0.8093073593073593 + }, + { + "epoch": 0.6791111111111111, + "grad_norm": 1.8991554975509644, + "learning_rate": 9.15870099184059e-06, + "loss": 0.6026749610900879, + "loss_sft": 0.6026749610900879, + "step": 191, + "token_acc": 0.8177268683274022 + }, + { + "epoch": 0.6826666666666666, + "grad_norm": 1.851915955543518, + "learning_rate": 9.147849934483019e-06, + "loss": 0.6289232969284058, + "loss_sft": 0.6289232969284058, + "step": 192, + "token_acc": 0.8149988955157941 + }, + { + "epoch": 0.6862222222222222, + "grad_norm": 1.8582254648208618, + "learning_rate": 9.136935860934628e-06, + "loss": 0.6025111675262451, + "loss_sft": 0.6025111675262451, + "step": 193, + "token_acc": 0.8198704566136716 + }, + { + "epoch": 0.6897777777777778, + "grad_norm": 1.9214824438095093, + "learning_rate": 9.125958937007427e-06, + "loss": 0.6356956958770752, + "loss_sft": 0.6356956958770752, + "step": 194, + "token_acc": 0.8126490409538621 + }, + { + "epoch": 0.6933333333333334, + "grad_norm": 1.8750139474868774, + "learning_rate": 9.114919329468283e-06, + "loss": 0.593024730682373, + "loss_sft": 0.593024730682373, + "step": 195, + "token_acc": 0.8246136555212138 + }, + { + "epoch": 0.6968888888888889, + "grad_norm": 1.9743341207504272, + "learning_rate": 9.103817206036383e-06, + "loss": 0.6290537118911743, + "loss_sft": 0.6290537118911743, + "step": 196, + "token_acc": 0.8184636582845333 + }, + { + "epoch": 0.7004444444444444, + "grad_norm": 2.0942587852478027, + "learning_rate": 9.092652735380683e-06, + "loss": 0.608534574508667, + "loss_sft": 0.608534574508667, + "step": 197, + "token_acc": 0.8183300126788625 + }, + { + "epoch": 0.704, + "grad_norm": 2.145768642425537, + "learning_rate": 9.081426087117356e-06, + "loss": 0.6475411653518677, + "loss_sft": 0.6475411653518677, + "step": 198, + "token_acc": 0.8122179602888087 + }, + { + "epoch": 0.7075555555555556, + "grad_norm": 1.9645922183990479, + "learning_rate": 9.0701374318072e-06, + "loss": 0.6695002913475037, + "loss_sft": 0.6695002913475037, + "step": 199, + "token_acc": 0.8060507283284098 + }, + { + "epoch": 0.7111111111111111, + "grad_norm": 2.0839080810546875, + "learning_rate": 9.058786940953057e-06, + "loss": 0.6406234502792358, + "loss_sft": 0.6406234502792358, + "step": 200, + "token_acc": 0.8087468087468087 + }, + { + "epoch": 0.7146666666666667, + "grad_norm": 1.968871831893921, + "learning_rate": 9.04737478699721e-06, + "loss": 0.6253594160079956, + "loss_sft": 0.6253594160079956, + "step": 201, + "token_acc": 0.8123441396508728 + }, + { + "epoch": 0.7182222222222222, + "grad_norm": 1.8631529808044434, + "learning_rate": 9.035901143318753e-06, + "loss": 0.594918966293335, + "loss_sft": 0.594918966293335, + "step": 202, + "token_acc": 0.8236934178876011 + }, + { + "epoch": 0.7217777777777777, + "grad_norm": 1.9275434017181396, + "learning_rate": 9.024366184230964e-06, + "loss": 0.6467505693435669, + "loss_sft": 0.6467505693435669, + "step": 203, + "token_acc": 0.8100064006827395 + }, + { + "epoch": 0.7253333333333334, + "grad_norm": 1.8231463432312012, + "learning_rate": 9.012770084978657e-06, + "loss": 0.6207350492477417, + "loss_sft": 0.6207350492477417, + "step": 204, + "token_acc": 0.8175530279741777 + }, + { + "epoch": 0.7288888888888889, + "grad_norm": 1.8449485301971436, + "learning_rate": 9.001113021735512e-06, + "loss": 0.5855098366737366, + "loss_sft": 0.5855098366737366, + "step": 205, + "token_acc": 0.8254432966911561 + }, + { + "epoch": 0.7324444444444445, + "grad_norm": 1.9755560159683228, + "learning_rate": 8.989395171601413e-06, + "loss": 0.6692385673522949, + "loss_sft": 0.6692385673522949, + "step": 206, + "token_acc": 0.8056127846762069 + }, + { + "epoch": 0.736, + "grad_norm": 1.8819565773010254, + "learning_rate": 8.97761671259974e-06, + "loss": 0.6083887815475464, + "loss_sft": 0.6083887815475464, + "step": 207, + "token_acc": 0.8174640864300131 + }, + { + "epoch": 0.7395555555555555, + "grad_norm": 1.925837755203247, + "learning_rate": 8.965777823674679e-06, + "loss": 0.6189272403717041, + "loss_sft": 0.6189272403717041, + "step": 208, + "token_acc": 0.814784139144839 + }, + { + "epoch": 0.7431111111111111, + "grad_norm": 1.808821678161621, + "learning_rate": 8.953878684688492e-06, + "loss": 0.6046684980392456, + "loss_sft": 0.6046684980392456, + "step": 209, + "token_acc": 0.8202887740423038 + }, + { + "epoch": 0.7466666666666667, + "grad_norm": 1.9002490043640137, + "learning_rate": 8.941919476418797e-06, + "loss": 0.6286606192588806, + "loss_sft": 0.6286606192588806, + "step": 210, + "token_acc": 0.8194865173561107 + }, + { + "epoch": 0.7502222222222222, + "grad_norm": 1.9260332584381104, + "learning_rate": 8.929900380555805e-06, + "loss": 0.6806676387786865, + "loss_sft": 0.6806676387786865, + "step": 211, + "token_acc": 0.8044462127617248 + }, + { + "epoch": 0.7537777777777778, + "grad_norm": 2.277512788772583, + "learning_rate": 8.917821579699578e-06, + "loss": 0.6578326225280762, + "loss_sft": 0.6578326225280762, + "step": 212, + "token_acc": 0.8070809248554913 + }, + { + "epoch": 0.7573333333333333, + "grad_norm": 1.7587056159973145, + "learning_rate": 8.905683257357238e-06, + "loss": 0.6144957542419434, + "loss_sft": 0.6144957542419434, + "step": 213, + "token_acc": 0.816299494627349 + }, + { + "epoch": 0.7608888888888888, + "grad_norm": 1.9263427257537842, + "learning_rate": 8.893485597940195e-06, + "loss": 0.6236262321472168, + "loss_sft": 0.6236262321472168, + "step": 214, + "token_acc": 0.8131650869337675 + }, + { + "epoch": 0.7644444444444445, + "grad_norm": 1.9076762199401855, + "learning_rate": 8.881228786761329e-06, + "loss": 0.6595988273620605, + "loss_sft": 0.6595988273620605, + "step": 215, + "token_acc": 0.8055438670313486 + }, + { + "epoch": 0.768, + "grad_norm": 1.8322217464447021, + "learning_rate": 8.868913010032188e-06, + "loss": 0.5863650441169739, + "loss_sft": 0.5863650441169739, + "step": 216, + "token_acc": 0.8222607140861596 + }, + { + "epoch": 0.7715555555555556, + "grad_norm": 1.9024826288223267, + "learning_rate": 8.856538454860155e-06, + "loss": 0.6526978015899658, + "loss_sft": 0.6526978015899658, + "step": 217, + "token_acc": 0.8096035400346828 + }, + { + "epoch": 0.7751111111111111, + "grad_norm": 1.9115735292434692, + "learning_rate": 8.8441053092456e-06, + "loss": 0.6429371237754822, + "loss_sft": 0.6429371237754822, + "step": 218, + "token_acc": 0.8083616066542629 + }, + { + "epoch": 0.7786666666666666, + "grad_norm": 1.9409433603286743, + "learning_rate": 8.831613762079026e-06, + "loss": 0.6536085605621338, + "loss_sft": 0.6536085605621338, + "step": 219, + "token_acc": 0.8115160848733745 + }, + { + "epoch": 0.7822222222222223, + "grad_norm": 1.8302897214889526, + "learning_rate": 8.819064003138211e-06, + "loss": 0.6140881776809692, + "loss_sft": 0.6140881776809692, + "step": 220, + "token_acc": 0.8164424514200299 + }, + { + "epoch": 0.7857777777777778, + "grad_norm": 1.8768426179885864, + "learning_rate": 8.806456223085308e-06, + "loss": 0.5843682885169983, + "loss_sft": 0.5843682885169983, + "step": 221, + "token_acc": 0.8225374406428833 + }, + { + "epoch": 0.7893333333333333, + "grad_norm": 1.8818626403808594, + "learning_rate": 8.793790613463956e-06, + "loss": 0.6139037609100342, + "loss_sft": 0.6139037609100342, + "step": 222, + "token_acc": 0.8142438808773965 + }, + { + "epoch": 0.7928888888888889, + "grad_norm": 1.9720075130462646, + "learning_rate": 8.781067366696368e-06, + "loss": 0.6383850574493408, + "loss_sft": 0.6383850574493408, + "step": 223, + "token_acc": 0.8121648890170078 + }, + { + "epoch": 0.7964444444444444, + "grad_norm": 1.9403375387191772, + "learning_rate": 8.768286676080415e-06, + "loss": 0.5940805077552795, + "loss_sft": 0.5940805077552795, + "step": 224, + "token_acc": 0.82028833245783 + }, + { + "epoch": 0.8, + "grad_norm": 1.9513380527496338, + "learning_rate": 8.75544873578668e-06, + "loss": 0.6004047393798828, + "loss_sft": 0.6004047393798828, + "step": 225, + "token_acc": 0.8171078538612995 + }, + { + "epoch": 0.8035555555555556, + "grad_norm": 1.8336174488067627, + "learning_rate": 8.742553740855507e-06, + "loss": 0.5744870901107788, + "loss_sft": 0.5744870901107788, + "step": 226, + "token_acc": 0.8251822770611329 + }, + { + "epoch": 0.8071111111111111, + "grad_norm": 1.9177244901657104, + "learning_rate": 8.729601887194048e-06, + "loss": 0.6349228620529175, + "loss_sft": 0.6349228620529175, + "step": 227, + "token_acc": 0.8074149199909849 + }, + { + "epoch": 0.8106666666666666, + "grad_norm": 1.9450557231903076, + "learning_rate": 8.71659337157328e-06, + "loss": 0.609279215335846, + "loss_sft": 0.609279215335846, + "step": 228, + "token_acc": 0.8174922422710034 + }, + { + "epoch": 0.8142222222222222, + "grad_norm": 1.934356689453125, + "learning_rate": 8.70352839162502e-06, + "loss": 0.6055933237075806, + "loss_sft": 0.6055933237075806, + "step": 229, + "token_acc": 0.822868810617662 + }, + { + "epoch": 0.8177777777777778, + "grad_norm": 1.9567376375198364, + "learning_rate": 8.690407145838909e-06, + "loss": 0.6571247577667236, + "loss_sft": 0.6571247577667236, + "step": 230, + "token_acc": 0.8082120280693614 + }, + { + "epoch": 0.8213333333333334, + "grad_norm": 1.8213931322097778, + "learning_rate": 8.677229833559413e-06, + "loss": 0.6412234306335449, + "loss_sft": 0.6412234306335449, + "step": 231, + "token_acc": 0.8101429781298262 + }, + { + "epoch": 0.8248888888888889, + "grad_norm": 1.7699004411697388, + "learning_rate": 8.663996654982791e-06, + "loss": 0.646782398223877, + "loss_sft": 0.646782398223877, + "step": 232, + "token_acc": 0.811387397569935 + }, + { + "epoch": 0.8284444444444444, + "grad_norm": 1.8586689233779907, + "learning_rate": 8.650707811154045e-06, + "loss": 0.584469735622406, + "loss_sft": 0.584469735622406, + "step": 233, + "token_acc": 0.8251347517730496 + }, + { + "epoch": 0.832, + "grad_norm": 1.735683798789978, + "learning_rate": 8.637363503963873e-06, + "loss": 0.6591368913650513, + "loss_sft": 0.6591368913650513, + "step": 234, + "token_acc": 0.8071606166086525 + }, + { + "epoch": 0.8355555555555556, + "grad_norm": 1.8056318759918213, + "learning_rate": 8.6239639361456e-06, + "loss": 0.6457397937774658, + "loss_sft": 0.6457397937774658, + "step": 235, + "token_acc": 0.8091306698117844 + }, + { + "epoch": 0.8391111111111111, + "grad_norm": 1.9188008308410645, + "learning_rate": 8.610509311272099e-06, + "loss": 0.6361337304115295, + "loss_sft": 0.6361337304115295, + "step": 236, + "token_acc": 0.8134793928378444 + }, + { + "epoch": 0.8426666666666667, + "grad_norm": 1.865373134613037, + "learning_rate": 8.596999833752694e-06, + "loss": 0.5957101583480835, + "loss_sft": 0.5957101583480835, + "step": 237, + "token_acc": 0.8237246680642907 + }, + { + "epoch": 0.8462222222222222, + "grad_norm": 1.82596755027771, + "learning_rate": 8.583435708830058e-06, + "loss": 0.6771214604377747, + "loss_sft": 0.6771214604377747, + "step": 238, + "token_acc": 0.8015522610029279 + }, + { + "epoch": 0.8497777777777777, + "grad_norm": 1.8224360942840576, + "learning_rate": 8.569817142577099e-06, + "loss": 0.6583431959152222, + "loss_sft": 0.6583431959152222, + "step": 239, + "token_acc": 0.8075036658745007 + }, + { + "epoch": 0.8533333333333334, + "grad_norm": 1.759124517440796, + "learning_rate": 8.556144341893819e-06, + "loss": 0.5864856243133545, + "loss_sft": 0.5864856243133545, + "step": 240, + "token_acc": 0.8225327146655544 + }, + { + "epoch": 0.8568888888888889, + "grad_norm": 1.8329733610153198, + "learning_rate": 8.54241751450418e-06, + "loss": 0.6750659942626953, + "loss_sft": 0.6750659942626953, + "step": 241, + "token_acc": 0.8039346448816272 + }, + { + "epoch": 0.8604444444444445, + "grad_norm": 1.9416502714157104, + "learning_rate": 8.528636868952944e-06, + "loss": 0.643253743648529, + "loss_sft": 0.643253743648529, + "step": 242, + "token_acc": 0.8090988626421697 + }, + { + "epoch": 0.864, + "grad_norm": 1.8971896171569824, + "learning_rate": 8.514802614602503e-06, + "loss": 0.6241411566734314, + "loss_sft": 0.6241411566734314, + "step": 243, + "token_acc": 0.8132671576475611 + }, + { + "epoch": 0.8675555555555555, + "grad_norm": 1.9098659753799438, + "learning_rate": 8.5009149616297e-06, + "loss": 0.6896156072616577, + "loss_sft": 0.6896156072616577, + "step": 244, + "token_acc": 0.796642561983471 + }, + { + "epoch": 0.8711111111111111, + "grad_norm": 1.7779595851898193, + "learning_rate": 8.486974121022642e-06, + "loss": 0.6009849309921265, + "loss_sft": 0.6009849309921265, + "step": 245, + "token_acc": 0.8210103713616594 + }, + { + "epoch": 0.8746666666666667, + "grad_norm": 1.898873209953308, + "learning_rate": 8.472980304577483e-06, + "loss": 0.6380763053894043, + "loss_sft": 0.6380763053894043, + "step": 246, + "token_acc": 0.8108532542313732 + }, + { + "epoch": 0.8782222222222222, + "grad_norm": 1.8348407745361328, + "learning_rate": 8.458933724895216e-06, + "loss": 0.633634090423584, + "loss_sft": 0.633634090423584, + "step": 247, + "token_acc": 0.8149761242588025 + }, + { + "epoch": 0.8817777777777778, + "grad_norm": 1.8900325298309326, + "learning_rate": 8.444834595378434e-06, + "loss": 0.6224663257598877, + "loss_sft": 0.6224663257598877, + "step": 248, + "token_acc": 0.8163869693978283 + }, + { + "epoch": 0.8853333333333333, + "grad_norm": 1.826843500137329, + "learning_rate": 8.4306831302281e-06, + "loss": 0.6119165420532227, + "loss_sft": 0.6119165420532227, + "step": 249, + "token_acc": 0.8187933910984339 + }, + { + "epoch": 0.8888888888888888, + "grad_norm": 1.884302020072937, + "learning_rate": 8.416479544440279e-06, + "loss": 0.596698522567749, + "loss_sft": 0.596698522567749, + "step": 250, + "token_acc": 0.8179408884023737 + }, + { + "epoch": 0.8924444444444445, + "grad_norm": 1.9094375371932983, + "learning_rate": 8.402224053802884e-06, + "loss": 0.5831952095031738, + "loss_sft": 0.5831952095031738, + "step": 251, + "token_acc": 0.8270022406336304 + }, + { + "epoch": 0.896, + "grad_norm": 1.8364545106887817, + "learning_rate": 8.38791687489239e-06, + "loss": 0.5844613313674927, + "loss_sft": 0.5844613313674927, + "step": 252, + "token_acc": 0.8234582164273533 + }, + { + "epoch": 0.8995555555555556, + "grad_norm": 1.8941620588302612, + "learning_rate": 8.373558225070546e-06, + "loss": 0.6793165802955627, + "loss_sft": 0.6793165802955627, + "step": 253, + "token_acc": 0.8001790777495896 + }, + { + "epoch": 0.9031111111111111, + "grad_norm": 1.8025329113006592, + "learning_rate": 8.359148322481073e-06, + "loss": 0.6179349422454834, + "loss_sft": 0.6179349422454834, + "step": 254, + "token_acc": 0.8164189823766643 + }, + { + "epoch": 0.9066666666666666, + "grad_norm": 1.8716998100280762, + "learning_rate": 8.344687386046348e-06, + "loss": 0.6098579168319702, + "loss_sft": 0.6098579168319702, + "step": 255, + "token_acc": 0.8178296498634219 + }, + { + "epoch": 0.9102222222222223, + "grad_norm": 2.0504415035247803, + "learning_rate": 8.330175635464082e-06, + "loss": 0.5922345519065857, + "loss_sft": 0.5922345519065857, + "step": 256, + "token_acc": 0.8224268759322914 + }, + { + "epoch": 0.9137777777777778, + "grad_norm": 1.8238425254821777, + "learning_rate": 8.315613291203977e-06, + "loss": 0.6129311919212341, + "loss_sft": 0.6129311919212341, + "step": 257, + "token_acc": 0.8215840763010575 + }, + { + "epoch": 0.9173333333333333, + "grad_norm": 1.9828053712844849, + "learning_rate": 8.30100057450438e-06, + "loss": 0.6212902665138245, + "loss_sft": 0.6212902665138245, + "step": 258, + "token_acc": 0.8143126177024482 + }, + { + "epoch": 0.9208888888888889, + "grad_norm": 1.9384937286376953, + "learning_rate": 8.286337707368922e-06, + "loss": 0.6663275361061096, + "loss_sft": 0.6663275361061096, + "step": 259, + "token_acc": 0.8046350986084693 + }, + { + "epoch": 0.9244444444444444, + "grad_norm": 1.7904130220413208, + "learning_rate": 8.271624912563143e-06, + "loss": 0.5716975927352905, + "loss_sft": 0.5716975927352905, + "step": 260, + "token_acc": 0.8271734077173408 + }, + { + "epoch": 0.928, + "grad_norm": 1.8622715473175049, + "learning_rate": 8.256862413611113e-06, + "loss": 0.608354926109314, + "loss_sft": 0.608354926109314, + "step": 261, + "token_acc": 0.8201721755765756 + }, + { + "epoch": 0.9315555555555556, + "grad_norm": 1.8488192558288574, + "learning_rate": 8.242050434792022e-06, + "loss": 0.6060342788696289, + "loss_sft": 0.6060342788696289, + "step": 262, + "token_acc": 0.8161861466682617 + }, + { + "epoch": 0.9351111111111111, + "grad_norm": 1.9396274089813232, + "learning_rate": 8.227189201136796e-06, + "loss": 0.6129617691040039, + "loss_sft": 0.6129617691040039, + "step": 263, + "token_acc": 0.818019471167694 + }, + { + "epoch": 0.9386666666666666, + "grad_norm": 1.911159873008728, + "learning_rate": 8.212278938424654e-06, + "loss": 0.6359697580337524, + "loss_sft": 0.6359697580337524, + "step": 264, + "token_acc": 0.8096197842902277 + }, + { + "epoch": 0.9422222222222222, + "grad_norm": 1.6449636220932007, + "learning_rate": 8.197319873179694e-06, + "loss": 0.5893205404281616, + "loss_sft": 0.5893205404281616, + "step": 265, + "token_acc": 0.8212374744624964 + }, + { + "epoch": 0.9457777777777778, + "grad_norm": 1.7324434518814087, + "learning_rate": 8.182312232667446e-06, + "loss": 0.6007264256477356, + "loss_sft": 0.6007264256477356, + "step": 266, + "token_acc": 0.8256462512066931 + }, + { + "epoch": 0.9493333333333334, + "grad_norm": 1.9077322483062744, + "learning_rate": 8.167256244891416e-06, + "loss": 0.61815345287323, + "loss_sft": 0.61815345287323, + "step": 267, + "token_acc": 0.814988422657706 + }, + { + "epoch": 0.9528888888888889, + "grad_norm": 1.8453760147094727, + "learning_rate": 8.152152138589633e-06, + "loss": 0.5867645740509033, + "loss_sft": 0.5867645740509033, + "step": 268, + "token_acc": 0.8184016646982735 + }, + { + "epoch": 0.9564444444444444, + "grad_norm": 1.7990697622299194, + "learning_rate": 8.137000143231156e-06, + "loss": 0.6084282398223877, + "loss_sft": 0.6084282398223877, + "step": 269, + "token_acc": 0.8211048769961157 + }, + { + "epoch": 0.96, + "grad_norm": 1.7465487718582153, + "learning_rate": 8.121800489012608e-06, + "loss": 0.5811343193054199, + "loss_sft": 0.5811343193054199, + "step": 270, + "token_acc": 0.8251317175974711 + }, + { + "epoch": 0.9635555555555556, + "grad_norm": 1.8082103729248047, + "learning_rate": 8.106553406854664e-06, + "loss": 0.6276707053184509, + "loss_sft": 0.6276707053184509, + "step": 271, + "token_acc": 0.8127702362996481 + }, + { + "epoch": 0.9671111111111111, + "grad_norm": 2.0267155170440674, + "learning_rate": 8.091259128398548e-06, + "loss": 0.6073474884033203, + "loss_sft": 0.6073474884033203, + "step": 272, + "token_acc": 0.8221875885519977 + }, + { + "epoch": 0.9706666666666667, + "grad_norm": 1.9235782623291016, + "learning_rate": 8.075917886002514e-06, + "loss": 0.6014127135276794, + "loss_sft": 0.6014127135276794, + "step": 273, + "token_acc": 0.8197324958033249 + }, + { + "epoch": 0.9742222222222222, + "grad_norm": 1.88161301612854, + "learning_rate": 8.060529912738316e-06, + "loss": 0.5992648601531982, + "loss_sft": 0.5992648601531982, + "step": 274, + "token_acc": 0.8226065472513897 + }, + { + "epoch": 0.9777777777777777, + "grad_norm": 1.8518449068069458, + "learning_rate": 8.045095442387666e-06, + "loss": 0.6497490406036377, + "loss_sft": 0.6497490406036377, + "step": 275, + "token_acc": 0.8103678929765886 + }, + { + "epoch": 0.9813333333333333, + "grad_norm": 1.7776848077774048, + "learning_rate": 8.029614709438684e-06, + "loss": 0.6458376049995422, + "loss_sft": 0.6458376049995422, + "step": 276, + "token_acc": 0.8135398451161628 + }, + { + "epoch": 0.9848888888888889, + "grad_norm": 1.9263386726379395, + "learning_rate": 8.014087949082333e-06, + "loss": 0.5843530893325806, + "loss_sft": 0.5843530893325806, + "step": 277, + "token_acc": 0.8229918229918229 + }, + { + "epoch": 0.9884444444444445, + "grad_norm": 1.8184524774551392, + "learning_rate": 7.998515397208846e-06, + "loss": 0.6026554703712463, + "loss_sft": 0.6026554703712463, + "step": 278, + "token_acc": 0.8184083828943642 + }, + { + "epoch": 0.992, + "grad_norm": 1.9941420555114746, + "learning_rate": 7.982897290404146e-06, + "loss": 0.5620933771133423, + "loss_sft": 0.5620933771133423, + "step": 279, + "token_acc": 0.8317579054743285 + }, + { + "epoch": 0.9955555555555555, + "grad_norm": 1.9176608324050903, + "learning_rate": 7.967233865946249e-06, + "loss": 0.6180765628814697, + "loss_sft": 0.6180765628814697, + "step": 280, + "token_acc": 0.8099966714745368 + }, + { + "epoch": 0.9991111111111111, + "grad_norm": 1.812450885772705, + "learning_rate": 7.951525361801655e-06, + "loss": 0.6000022888183594, + "loss_sft": 0.6000022888183594, + "step": 281, + "token_acc": 0.8206473707467963 + }, + { + "epoch": 1.0, + "grad_norm": 3.4522907733917236, + "learning_rate": 7.935772016621744e-06, + "loss": 0.6548411846160889, + "loss_sft": 0.6548411846160889, + "step": 282, + "token_acc": 0.8114386101893422 + }, + { + "epoch": 1.0035555555555555, + "grad_norm": 1.7518024444580078, + "learning_rate": 7.919974069739136e-06, + "loss": 0.4313552975654602, + "loss_sft": 0.4313552975654602, + "step": 283, + "token_acc": 0.8667924528301887 + }, + { + "epoch": 1.007111111111111, + "grad_norm": 1.7726091146469116, + "learning_rate": 7.904131761164068e-06, + "loss": 0.4319254159927368, + "loss_sft": 0.4319254159927368, + "step": 284, + "token_acc": 0.8656733202136632 + }, + { + "epoch": 1.0106666666666666, + "grad_norm": 1.7077429294586182, + "learning_rate": 7.888245331580737e-06, + "loss": 0.4252268671989441, + "loss_sft": 0.4252268671989441, + "step": 285, + "token_acc": 0.861813110181311 + }, + { + "epoch": 1.0142222222222221, + "grad_norm": 1.7379792928695679, + "learning_rate": 7.872315022343654e-06, + "loss": 0.44653642177581787, + "loss_sft": 0.44653642177581787, + "step": 286, + "token_acc": 0.8608947848971481 + }, + { + "epoch": 1.0177777777777777, + "grad_norm": 1.8086405992507935, + "learning_rate": 7.856341075473963e-06, + "loss": 0.43224799633026123, + "loss_sft": 0.43224799633026123, + "step": 287, + "token_acc": 0.8643016040276315 + }, + { + "epoch": 1.0213333333333334, + "grad_norm": 1.8583446741104126, + "learning_rate": 7.84032373365578e-06, + "loss": 0.427482008934021, + "loss_sft": 0.427482008934021, + "step": 288, + "token_acc": 0.8636705622711541 + }, + { + "epoch": 1.024888888888889, + "grad_norm": 1.958483099937439, + "learning_rate": 7.824263240232497e-06, + "loss": 0.40691885352134705, + "loss_sft": 0.40691885352134705, + "step": 289, + "token_acc": 0.8714859437751004 + }, + { + "epoch": 1.0284444444444445, + "grad_norm": 2.176067352294922, + "learning_rate": 7.808159839203085e-06, + "loss": 0.42749154567718506, + "loss_sft": 0.42749154567718506, + "step": 290, + "token_acc": 0.8663954434499593 + }, + { + "epoch": 1.032, + "grad_norm": 2.434170961380005, + "learning_rate": 7.792013775218385e-06, + "loss": 0.43468087911605835, + "loss_sft": 0.43468087911605835, + "step": 291, + "token_acc": 0.8648891551827441 + }, + { + "epoch": 1.0355555555555556, + "grad_norm": 2.090839385986328, + "learning_rate": 7.775825293577407e-06, + "loss": 0.393714040517807, + "loss_sft": 0.393714040517807, + "step": 292, + "token_acc": 0.8750257997936016 + }, + { + "epoch": 1.039111111111111, + "grad_norm": 2.1758270263671875, + "learning_rate": 7.759594640223576e-06, + "loss": 0.4109739363193512, + "loss_sft": 0.4109739363193512, + "step": 293, + "token_acc": 0.8666896314157768 + }, + { + "epoch": 1.0426666666666666, + "grad_norm": 2.043649196624756, + "learning_rate": 7.743322061741024e-06, + "loss": 0.4008193612098694, + "loss_sft": 0.4008193612098694, + "step": 294, + "token_acc": 0.8699869868685673 + }, + { + "epoch": 1.0462222222222222, + "grad_norm": 1.9236005544662476, + "learning_rate": 7.727007805350815e-06, + "loss": 0.4134569764137268, + "loss_sft": 0.4134569764137268, + "step": 295, + "token_acc": 0.8673284757200841 + }, + { + "epoch": 1.0497777777777777, + "grad_norm": 2.062549352645874, + "learning_rate": 7.71065211890722e-06, + "loss": 0.4501626491546631, + "loss_sft": 0.4501626491546631, + "step": 296, + "token_acc": 0.859695839990532 + }, + { + "epoch": 1.0533333333333332, + "grad_norm": 1.843603253364563, + "learning_rate": 7.694255250893927e-06, + "loss": 0.3989909291267395, + "loss_sft": 0.3989909291267395, + "step": 297, + "token_acc": 0.8731678717998828 + }, + { + "epoch": 1.056888888888889, + "grad_norm": 1.8667809963226318, + "learning_rate": 7.677817450420279e-06, + "loss": 0.3915114402770996, + "loss_sft": 0.3915114402770996, + "step": 298, + "token_acc": 0.8766070565490575 + }, + { + "epoch": 1.0604444444444445, + "grad_norm": 1.9161550998687744, + "learning_rate": 7.661338967217483e-06, + "loss": 0.42347243428230286, + "loss_sft": 0.42347243428230286, + "step": 299, + "token_acc": 0.8663842268209223 + }, + { + "epoch": 1.064, + "grad_norm": 1.9624372720718384, + "learning_rate": 7.644820051634813e-06, + "loss": 0.378623902797699, + "loss_sft": 0.378623902797699, + "step": 300, + "token_acc": 0.8832217172925138 + }, + { + "epoch": 1.0675555555555556, + "grad_norm": 1.8872796297073364, + "learning_rate": 7.628260954635822e-06, + "loss": 0.39598548412323, + "loss_sft": 0.39598548412323, + "step": 301, + "token_acc": 0.8715932349455995 + }, + { + "epoch": 1.0711111111111111, + "grad_norm": 1.8708773851394653, + "learning_rate": 7.6116619277945135e-06, + "loss": 0.44171616435050964, + "loss_sft": 0.44171616435050964, + "step": 302, + "token_acc": 0.8634813925570228 + }, + { + "epoch": 1.0746666666666667, + "grad_norm": 2.0179035663604736, + "learning_rate": 7.595023223291525e-06, + "loss": 0.42614901065826416, + "loss_sft": 0.42614901065826416, + "step": 303, + "token_acc": 0.8678965949208518 + }, + { + "epoch": 1.0782222222222222, + "grad_norm": 1.9818003177642822, + "learning_rate": 7.578345093910298e-06, + "loss": 0.39987361431121826, + "loss_sft": 0.39987361431121826, + "step": 304, + "token_acc": 0.8700400043247919 + }, + { + "epoch": 1.0817777777777777, + "grad_norm": 2.2445433139801025, + "learning_rate": 7.561627793033233e-06, + "loss": 0.43376174569129944, + "loss_sft": 0.43376174569129944, + "step": 305, + "token_acc": 0.8650298862394755 + }, + { + "epoch": 1.0853333333333333, + "grad_norm": 1.9904202222824097, + "learning_rate": 7.54487157463785e-06, + "loss": 0.4143393635749817, + "loss_sft": 0.4143393635749817, + "step": 306, + "token_acc": 0.87069449119264 + }, + { + "epoch": 1.0888888888888888, + "grad_norm": 1.9291605949401855, + "learning_rate": 7.528076693292914e-06, + "loss": 0.4538661241531372, + "loss_sft": 0.4538661241531372, + "step": 307, + "token_acc": 0.8590569799248922 + }, + { + "epoch": 1.0924444444444443, + "grad_norm": 1.9268896579742432, + "learning_rate": 7.511243404154586e-06, + "loss": 0.4308316707611084, + "loss_sft": 0.4308316707611084, + "step": 308, + "token_acc": 0.8642501411921754 + }, + { + "epoch": 1.096, + "grad_norm": 1.9481511116027832, + "learning_rate": 7.49437196296253e-06, + "loss": 0.42588937282562256, + "loss_sft": 0.42588937282562256, + "step": 309, + "token_acc": 0.8608482120018417 + }, + { + "epoch": 1.0995555555555556, + "grad_norm": 1.8635936975479126, + "learning_rate": 7.4774626260360384e-06, + "loss": 0.39201420545578003, + "loss_sft": 0.39201420545578003, + "step": 310, + "token_acc": 0.8740149094781683 + }, + { + "epoch": 1.1031111111111112, + "grad_norm": 1.9378689527511597, + "learning_rate": 7.460515650270134e-06, + "loss": 0.38611549139022827, + "loss_sft": 0.38611549139022827, + "step": 311, + "token_acc": 0.8752483346967395 + }, + { + "epoch": 1.1066666666666667, + "grad_norm": 1.9042547941207886, + "learning_rate": 7.443531293131667e-06, + "loss": 0.4318225085735321, + "loss_sft": 0.4318225085735321, + "step": 312, + "token_acc": 0.8644313244959341 + }, + { + "epoch": 1.1102222222222222, + "grad_norm": 1.8838920593261719, + "learning_rate": 7.4265098126554065e-06, + "loss": 0.39200425148010254, + "loss_sft": 0.39200425148010254, + "step": 313, + "token_acc": 0.8776760160274757 + }, + { + "epoch": 1.1137777777777778, + "grad_norm": 2.061145544052124, + "learning_rate": 7.409451467440111e-06, + "loss": 0.3577609658241272, + "loss_sft": 0.3577609658241272, + "step": 314, + "token_acc": 0.8866403906112125 + }, + { + "epoch": 1.1173333333333333, + "grad_norm": 1.9201234579086304, + "learning_rate": 7.392356516644614e-06, + "loss": 0.4082680940628052, + "loss_sft": 0.4082680940628052, + "step": 315, + "token_acc": 0.8727016356457981 + }, + { + "epoch": 1.1208888888888888, + "grad_norm": 2.0884475708007812, + "learning_rate": 7.375225219983876e-06, + "loss": 0.4152630567550659, + "loss_sft": 0.4152630567550659, + "step": 316, + "token_acc": 0.8664407988587732 + }, + { + "epoch": 1.1244444444444444, + "grad_norm": 1.9312373399734497, + "learning_rate": 7.358057837725039e-06, + "loss": 0.41544800996780396, + "loss_sft": 0.41544800996780396, + "step": 317, + "token_acc": 0.8682948102921937 + }, + { + "epoch": 1.1280000000000001, + "grad_norm": 1.9792176485061646, + "learning_rate": 7.340854630683481e-06, + "loss": 0.40717118978500366, + "loss_sft": 0.40717118978500366, + "step": 318, + "token_acc": 0.8735221480741947 + }, + { + "epoch": 1.1315555555555556, + "grad_norm": 2.522533893585205, + "learning_rate": 7.323615860218844e-06, + "loss": 0.4196258783340454, + "loss_sft": 0.4196258783340454, + "step": 319, + "token_acc": 0.8673665343745999 + }, + { + "epoch": 1.1351111111111112, + "grad_norm": 2.01004695892334, + "learning_rate": 7.306341788231067e-06, + "loss": 0.389733225107193, + "loss_sft": 0.389733225107193, + "step": 320, + "token_acc": 0.8771724598930482 + }, + { + "epoch": 1.1386666666666667, + "grad_norm": 1.9227044582366943, + "learning_rate": 7.289032677156408e-06, + "loss": 0.42470553517341614, + "loss_sft": 0.42470553517341614, + "step": 321, + "token_acc": 0.8640462909868113 + }, + { + "epoch": 1.1422222222222222, + "grad_norm": 1.9080742597579956, + "learning_rate": 7.271688789963458e-06, + "loss": 0.3837636709213257, + "loss_sft": 0.3837636709213257, + "step": 322, + "token_acc": 0.8761488207286015 + }, + { + "epoch": 1.1457777777777778, + "grad_norm": 1.8660749197006226, + "learning_rate": 7.254310390149144e-06, + "loss": 0.40272265672683716, + "loss_sft": 0.40272265672683716, + "step": 323, + "token_acc": 0.8716034797551283 + }, + { + "epoch": 1.1493333333333333, + "grad_norm": 1.8732655048370361, + "learning_rate": 7.236897741734721e-06, + "loss": 0.3963853120803833, + "loss_sft": 0.3963853120803833, + "step": 324, + "token_acc": 0.8723799017149734 + }, + { + "epoch": 1.1528888888888889, + "grad_norm": 1.9640755653381348, + "learning_rate": 7.219451109261768e-06, + "loss": 0.39499202370643616, + "loss_sft": 0.39499202370643616, + "step": 325, + "token_acc": 0.8743733240060627 + }, + { + "epoch": 1.1564444444444444, + "grad_norm": 1.9264079332351685, + "learning_rate": 7.201970757788172e-06, + "loss": 0.40080487728118896, + "loss_sft": 0.40080487728118896, + "step": 326, + "token_acc": 0.8741408450704226 + }, + { + "epoch": 1.16, + "grad_norm": 1.9216946363449097, + "learning_rate": 7.184456952884089e-06, + "loss": 0.4075692296028137, + "loss_sft": 0.4075692296028137, + "step": 327, + "token_acc": 0.8681020733652313 + }, + { + "epoch": 1.1635555555555555, + "grad_norm": 1.903046727180481, + "learning_rate": 7.166909960627918e-06, + "loss": 0.42462170124053955, + "loss_sft": 0.42462170124053955, + "step": 328, + "token_acc": 0.8687612208258528 + }, + { + "epoch": 1.1671111111111112, + "grad_norm": 2.097158432006836, + "learning_rate": 7.149330047602258e-06, + "loss": 0.38915741443634033, + "loss_sft": 0.38915741443634033, + "step": 329, + "token_acc": 0.8772619047619048 + }, + { + "epoch": 1.1706666666666667, + "grad_norm": 1.876055359840393, + "learning_rate": 7.131717480889854e-06, + "loss": 0.38593608140945435, + "loss_sft": 0.38593608140945435, + "step": 330, + "token_acc": 0.8746732500726111 + }, + { + "epoch": 1.1742222222222223, + "grad_norm": 2.09281063079834, + "learning_rate": 7.114072528069549e-06, + "loss": 0.4185669422149658, + "loss_sft": 0.4185669422149658, + "step": 331, + "token_acc": 0.8686658697572891 + }, + { + "epoch": 1.1777777777777778, + "grad_norm": 2.2704269886016846, + "learning_rate": 7.0963954572122e-06, + "loss": 0.3780200481414795, + "loss_sft": 0.3780200481414795, + "step": 332, + "token_acc": 0.8780487804878049 + }, + { + "epoch": 1.1813333333333333, + "grad_norm": 1.9671108722686768, + "learning_rate": 7.078686536876627e-06, + "loss": 0.41342270374298096, + "loss_sft": 0.41342270374298096, + "step": 333, + "token_acc": 0.8691951360741169 + }, + { + "epoch": 1.1848888888888889, + "grad_norm": 1.8989192247390747, + "learning_rate": 7.06094603610552e-06, + "loss": 0.3884793817996979, + "loss_sft": 0.3884793817996979, + "step": 334, + "token_acc": 0.8763179020604763 + }, + { + "epoch": 1.1884444444444444, + "grad_norm": 1.9761707782745361, + "learning_rate": 7.043174224421353e-06, + "loss": 0.40490394830703735, + "loss_sft": 0.40490394830703735, + "step": 335, + "token_acc": 0.8705671028996664 + }, + { + "epoch": 1.192, + "grad_norm": 1.8867125511169434, + "learning_rate": 7.025371371822294e-06, + "loss": 0.38762935996055603, + "loss_sft": 0.38762935996055603, + "step": 336, + "token_acc": 0.8763580475912873 + }, + { + "epoch": 1.1955555555555555, + "grad_norm": 1.9159667491912842, + "learning_rate": 7.007537748778097e-06, + "loss": 0.37436068058013916, + "loss_sft": 0.37436068058013916, + "step": 337, + "token_acc": 0.8801699886275214 + }, + { + "epoch": 1.199111111111111, + "grad_norm": 1.9830975532531738, + "learning_rate": 6.989673626225997e-06, + "loss": 0.44501978158950806, + "loss_sft": 0.44501978158950806, + "step": 338, + "token_acc": 0.8615510649918078 + }, + { + "epoch": 1.2026666666666666, + "grad_norm": 1.8126604557037354, + "learning_rate": 6.971779275566593e-06, + "loss": 0.4045955538749695, + "loss_sft": 0.4045955538749695, + "step": 339, + "token_acc": 0.8720025241901557 + }, + { + "epoch": 1.2062222222222223, + "grad_norm": 1.9008846282958984, + "learning_rate": 6.953854968659726e-06, + "loss": 0.44734084606170654, + "loss_sft": 0.44734084606170654, + "step": 340, + "token_acc": 0.8613337082337504 + }, + { + "epoch": 1.2097777777777778, + "grad_norm": 1.8717738389968872, + "learning_rate": 6.935900977820341e-06, + "loss": 0.40877336263656616, + "loss_sft": 0.40877336263656616, + "step": 341, + "token_acc": 0.8696106741013642 + }, + { + "epoch": 1.2133333333333334, + "grad_norm": 1.9246532917022705, + "learning_rate": 6.917917575814364e-06, + "loss": 0.4054875373840332, + "loss_sft": 0.4054875373840332, + "step": 342, + "token_acc": 0.8714587986364171 + }, + { + "epoch": 1.216888888888889, + "grad_norm": 1.897621989250183, + "learning_rate": 6.899905035854544e-06, + "loss": 0.3885173797607422, + "loss_sft": 0.3885173797607422, + "step": 343, + "token_acc": 0.8749484536082475 + }, + { + "epoch": 1.2204444444444444, + "grad_norm": 1.917402744293213, + "learning_rate": 6.881863631596313e-06, + "loss": 0.4145117402076721, + "loss_sft": 0.4145117402076721, + "step": 344, + "token_acc": 0.8677187791111842 + }, + { + "epoch": 1.224, + "grad_norm": 1.9695416688919067, + "learning_rate": 6.863793637133618e-06, + "loss": 0.39646804332733154, + "loss_sft": 0.39646804332733154, + "step": 345, + "token_acc": 0.8743857324278063 + }, + { + "epoch": 1.2275555555555555, + "grad_norm": 1.9845435619354248, + "learning_rate": 6.845695326994768e-06, + "loss": 0.3647025227546692, + "loss_sft": 0.3647025227546692, + "step": 346, + "token_acc": 0.8817244312141187 + }, + { + "epoch": 1.231111111111111, + "grad_norm": 2.109989881515503, + "learning_rate": 6.827568976138255e-06, + "loss": 0.44945627450942993, + "loss_sft": 0.44945627450942993, + "step": 347, + "token_acc": 0.8615204286671132 + }, + { + "epoch": 1.2346666666666666, + "grad_norm": 2.0244266986846924, + "learning_rate": 6.809414859948579e-06, + "loss": 0.4137527048587799, + "loss_sft": 0.4137527048587799, + "step": 348, + "token_acc": 0.8664576469337868 + }, + { + "epoch": 1.2382222222222223, + "grad_norm": 2.043164014816284, + "learning_rate": 6.791233254232066e-06, + "loss": 0.3913249969482422, + "loss_sft": 0.3913249969482422, + "step": 349, + "token_acc": 0.8742678188250305 + }, + { + "epoch": 1.2417777777777779, + "grad_norm": 1.983569860458374, + "learning_rate": 6.773024435212678e-06, + "loss": 0.39767181873321533, + "loss_sft": 0.39767181873321533, + "step": 350, + "token_acc": 0.8709590517241379 + }, + { + "epoch": 1.2453333333333334, + "grad_norm": 2.1270880699157715, + "learning_rate": 6.7547886795278136e-06, + "loss": 0.39403557777404785, + "loss_sft": 0.39403557777404785, + "step": 351, + "token_acc": 0.8743725231175693 + }, + { + "epoch": 1.248888888888889, + "grad_norm": 2.004920721054077, + "learning_rate": 6.736526264224101e-06, + "loss": 0.3962416350841522, + "loss_sft": 0.3962416350841522, + "step": 352, + "token_acc": 0.8727111799524258 + }, + { + "epoch": 1.2524444444444445, + "grad_norm": 2.127906084060669, + "learning_rate": 6.718237466753206e-06, + "loss": 0.4101051092147827, + "loss_sft": 0.4101051092147827, + "step": 353, + "token_acc": 0.8699705593719332 + }, + { + "epoch": 1.256, + "grad_norm": 2.0145328044891357, + "learning_rate": 6.6999225649675955e-06, + "loss": 0.39425092935562134, + "loss_sft": 0.39425092935562134, + "step": 354, + "token_acc": 0.8738479600821838 + }, + { + "epoch": 1.2595555555555555, + "grad_norm": 1.962022304534912, + "learning_rate": 6.681581837116331e-06, + "loss": 0.38062620162963867, + "loss_sft": 0.38062620162963867, + "step": 355, + "token_acc": 0.8765354149925382 + }, + { + "epoch": 1.263111111111111, + "grad_norm": 1.9428821802139282, + "learning_rate": 6.6632155618408335e-06, + "loss": 0.3802978992462158, + "loss_sft": 0.3802978992462158, + "step": 356, + "token_acc": 0.8801621778163915 + }, + { + "epoch": 1.2666666666666666, + "grad_norm": 1.8854237794876099, + "learning_rate": 6.644824018170655e-06, + "loss": 0.3802940249443054, + "loss_sft": 0.3802940249443054, + "step": 357, + "token_acc": 0.8775688675120245 + }, + { + "epoch": 1.2702222222222221, + "grad_norm": 1.960248351097107, + "learning_rate": 6.6264074855192385e-06, + "loss": 0.4112996459007263, + "loss_sft": 0.4112996459007263, + "step": 358, + "token_acc": 0.8687289629748357 + }, + { + "epoch": 1.2737777777777777, + "grad_norm": 2.038078546524048, + "learning_rate": 6.607966243679669e-06, + "loss": 0.3667299449443817, + "loss_sft": 0.3667299449443817, + "step": 359, + "token_acc": 0.8806123063942627 + }, + { + "epoch": 1.2773333333333334, + "grad_norm": 2.123028039932251, + "learning_rate": 6.589500572820428e-06, + "loss": 0.4327188730239868, + "loss_sft": 0.4327188730239868, + "step": 360, + "token_acc": 0.8645311381531854 + }, + { + "epoch": 1.280888888888889, + "grad_norm": 2.051753282546997, + "learning_rate": 6.571010753481135e-06, + "loss": 0.43652617931365967, + "loss_sft": 0.43652617931365967, + "step": 361, + "token_acc": 0.8608236964901603 + }, + { + "epoch": 1.2844444444444445, + "grad_norm": 2.0510408878326416, + "learning_rate": 6.552497066568282e-06, + "loss": 0.4028615653514862, + "loss_sft": 0.4028615653514862, + "step": 362, + "token_acc": 0.8758127082991859 + }, + { + "epoch": 1.288, + "grad_norm": 1.9311106204986572, + "learning_rate": 6.533959793350974e-06, + "loss": 0.3606717884540558, + "loss_sft": 0.3606717884540558, + "step": 363, + "token_acc": 0.8833404815881379 + }, + { + "epoch": 1.2915555555555556, + "grad_norm": 2.0313987731933594, + "learning_rate": 6.5153992154566445e-06, + "loss": 0.41649389266967773, + "loss_sft": 0.41649389266967773, + "step": 364, + "token_acc": 0.8668627111803033 + }, + { + "epoch": 1.295111111111111, + "grad_norm": 1.957159399986267, + "learning_rate": 6.496815614866792e-06, + "loss": 0.4067304730415344, + "loss_sft": 0.4067304730415344, + "step": 365, + "token_acc": 0.8677676970717662 + }, + { + "epoch": 1.2986666666666666, + "grad_norm": 1.944205641746521, + "learning_rate": 6.478209273912675e-06, + "loss": 0.4043155312538147, + "loss_sft": 0.4043155312538147, + "step": 366, + "token_acc": 0.87167295185003 + }, + { + "epoch": 1.3022222222222222, + "grad_norm": 1.9901310205459595, + "learning_rate": 6.4595804752710475e-06, + "loss": 0.42966562509536743, + "loss_sft": 0.42966562509536743, + "step": 367, + "token_acc": 0.8676847391616276 + }, + { + "epoch": 1.3057777777777777, + "grad_norm": 1.8627445697784424, + "learning_rate": 6.440929501959844e-06, + "loss": 0.38356274366378784, + "loss_sft": 0.38356274366378784, + "step": 368, + "token_acc": 0.8764193583948848 + }, + { + "epoch": 1.3093333333333335, + "grad_norm": 1.8013744354248047, + "learning_rate": 6.42225663733389e-06, + "loss": 0.39349913597106934, + "loss_sft": 0.39349913597106934, + "step": 369, + "token_acc": 0.87350287956781 + }, + { + "epoch": 1.3128888888888888, + "grad_norm": 2.1345925331115723, + "learning_rate": 6.403562165080594e-06, + "loss": 0.4102644622325897, + "loss_sft": 0.4102644622325897, + "step": 370, + "token_acc": 0.8715480111477071 + }, + { + "epoch": 1.3164444444444445, + "grad_norm": 1.9053229093551636, + "learning_rate": 6.3848463692156396e-06, + "loss": 0.4053412973880768, + "loss_sft": 0.4053412973880768, + "step": 371, + "token_acc": 0.872847140593857 + }, + { + "epoch": 1.32, + "grad_norm": 1.9415433406829834, + "learning_rate": 6.366109534078667e-06, + "loss": 0.4315723776817322, + "loss_sft": 0.4315723776817322, + "step": 372, + "token_acc": 0.8626920047262702 + }, + { + "epoch": 1.3235555555555556, + "grad_norm": 2.105576276779175, + "learning_rate": 6.347351944328958e-06, + "loss": 0.3831115961074829, + "loss_sft": 0.3831115961074829, + "step": 373, + "token_acc": 0.8778711855431257 + }, + { + "epoch": 1.3271111111111111, + "grad_norm": 2.0241587162017822, + "learning_rate": 6.328573884941107e-06, + "loss": 0.3638756275177002, + "loss_sft": 0.3638756275177002, + "step": 374, + "token_acc": 0.8817470913599085 + }, + { + "epoch": 1.3306666666666667, + "grad_norm": 1.9856815338134766, + "learning_rate": 6.309775641200695e-06, + "loss": 0.3899883031845093, + "loss_sft": 0.3899883031845093, + "step": 375, + "token_acc": 0.8756500113045445 + }, + { + "epoch": 1.3342222222222222, + "grad_norm": 2.0613181591033936, + "learning_rate": 6.290957498699949e-06, + "loss": 0.38393428921699524, + "loss_sft": 0.38393428921699524, + "step": 376, + "token_acc": 0.8742220562608912 + }, + { + "epoch": 1.3377777777777777, + "grad_norm": 2.0069167613983154, + "learning_rate": 6.272119743333416e-06, + "loss": 0.4109063148498535, + "loss_sft": 0.4109063148498535, + "step": 377, + "token_acc": 0.8669866490124682 + }, + { + "epoch": 1.3413333333333333, + "grad_norm": 2.089128255844116, + "learning_rate": 6.2532626612936035e-06, + "loss": 0.4027180075645447, + "loss_sft": 0.4027180075645447, + "step": 378, + "token_acc": 0.872361230946218 + }, + { + "epoch": 1.3448888888888888, + "grad_norm": 2.0553579330444336, + "learning_rate": 6.234386539066643e-06, + "loss": 0.4322267770767212, + "loss_sft": 0.4322267770767212, + "step": 379, + "token_acc": 0.8657913146424151 + }, + { + "epoch": 1.3484444444444446, + "grad_norm": 1.9676539897918701, + "learning_rate": 6.215491663427936e-06, + "loss": 0.4113406538963318, + "loss_sft": 0.4113406538963318, + "step": 380, + "token_acc": 0.8682260795935648 + }, + { + "epoch": 1.3519999999999999, + "grad_norm": 1.9962681531906128, + "learning_rate": 6.1965783214377895e-06, + "loss": 0.42172589898109436, + "loss_sft": 0.42172589898109436, + "step": 381, + "token_acc": 0.8625356523684358 + }, + { + "epoch": 1.3555555555555556, + "grad_norm": 1.809495449066162, + "learning_rate": 6.177646800437066e-06, + "loss": 0.39803946018218994, + "loss_sft": 0.39803946018218994, + "step": 382, + "token_acc": 0.8717743076850498 + }, + { + "epoch": 1.3591111111111112, + "grad_norm": 2.070378541946411, + "learning_rate": 6.158697388042807e-06, + "loss": 0.43218693137168884, + "loss_sft": 0.43218693137168884, + "step": 383, + "token_acc": 0.8662752772341813 + }, + { + "epoch": 1.3626666666666667, + "grad_norm": 1.9867347478866577, + "learning_rate": 6.139730372143877e-06, + "loss": 0.43019747734069824, + "loss_sft": 0.43019747734069824, + "step": 384, + "token_acc": 0.8639094147041625 + }, + { + "epoch": 1.3662222222222222, + "grad_norm": 1.921633243560791, + "learning_rate": 6.120746040896572e-06, + "loss": 0.39434194564819336, + "loss_sft": 0.39434194564819336, + "step": 385, + "token_acc": 0.8735297476304671 + }, + { + "epoch": 1.3697777777777778, + "grad_norm": 1.8112812042236328, + "learning_rate": 6.10174468272026e-06, + "loss": 0.4182533621788025, + "loss_sft": 0.4182533621788025, + "step": 386, + "token_acc": 0.8663542642924086 + }, + { + "epoch": 1.3733333333333333, + "grad_norm": 1.9014766216278076, + "learning_rate": 6.082726586292982e-06, + "loss": 0.41383665800094604, + "loss_sft": 0.41383665800094604, + "step": 387, + "token_acc": 0.866530286398563 + }, + { + "epoch": 1.3768888888888888, + "grad_norm": 1.8586009740829468, + "learning_rate": 6.063692040547083e-06, + "loss": 0.39000970125198364, + "loss_sft": 0.39000970125198364, + "step": 388, + "token_acc": 0.8746875520746542 + }, + { + "epoch": 1.3804444444444444, + "grad_norm": 2.284881830215454, + "learning_rate": 6.044641334664812e-06, + "loss": 0.3975995182991028, + "loss_sft": 0.3975995182991028, + "step": 389, + "token_acc": 0.870817843866171 + }, + { + "epoch": 1.384, + "grad_norm": 1.9486490488052368, + "learning_rate": 6.025574758073925e-06, + "loss": 0.4103585481643677, + "loss_sft": 0.4103585481643677, + "step": 390, + "token_acc": 0.8694224235560589 + }, + { + "epoch": 1.3875555555555557, + "grad_norm": 1.9968544244766235, + "learning_rate": 6.006492600443301e-06, + "loss": 0.37671855092048645, + "loss_sft": 0.37671855092048645, + "step": 391, + "token_acc": 0.8803168589633201 + }, + { + "epoch": 1.3911111111111112, + "grad_norm": 1.906306266784668, + "learning_rate": 5.98739515167853e-06, + "loss": 0.38691234588623047, + "loss_sft": 0.38691234588623047, + "step": 392, + "token_acc": 0.8737645508455963 + }, + { + "epoch": 1.3946666666666667, + "grad_norm": 1.865398645401001, + "learning_rate": 5.968282701917514e-06, + "loss": 0.4030868411064148, + "loss_sft": 0.4030868411064148, + "step": 393, + "token_acc": 0.8727792833483891 + }, + { + "epoch": 1.3982222222222223, + "grad_norm": 2.0121121406555176, + "learning_rate": 5.949155541526057e-06, + "loss": 0.39079463481903076, + "loss_sft": 0.39079463481903076, + "step": 394, + "token_acc": 0.8758878707368649 + }, + { + "epoch": 1.4017777777777778, + "grad_norm": 2.0630757808685303, + "learning_rate": 5.930013961093454e-06, + "loss": 0.4191130995750427, + "loss_sft": 0.4191130995750427, + "step": 395, + "token_acc": 0.8657507035259063 + }, + { + "epoch": 1.4053333333333333, + "grad_norm": 1.9211196899414062, + "learning_rate": 5.910858251428077e-06, + "loss": 0.39444518089294434, + "loss_sft": 0.39444518089294434, + "step": 396, + "token_acc": 0.8749327884718787 + }, + { + "epoch": 1.4088888888888889, + "grad_norm": 1.922277808189392, + "learning_rate": 5.891688703552953e-06, + "loss": 0.42281603813171387, + "loss_sft": 0.42281603813171387, + "step": 397, + "token_acc": 0.8677702236673563 + }, + { + "epoch": 1.4124444444444444, + "grad_norm": 2.016336441040039, + "learning_rate": 5.872505608701354e-06, + "loss": 0.4200964570045471, + "loss_sft": 0.4200964570045471, + "step": 398, + "token_acc": 0.8661417322834646 + }, + { + "epoch": 1.416, + "grad_norm": 1.9104807376861572, + "learning_rate": 5.853309258312356e-06, + "loss": 0.4054337739944458, + "loss_sft": 0.4054337739944458, + "step": 399, + "token_acc": 0.871870321614785 + }, + { + "epoch": 1.4195555555555557, + "grad_norm": 2.1013143062591553, + "learning_rate": 5.834099944026422e-06, + "loss": 0.39102301001548767, + "loss_sft": 0.39102301001548767, + "step": 400, + "token_acc": 0.8773395505472443 + }, + { + "epoch": 1.423111111111111, + "grad_norm": 1.9902065992355347, + "learning_rate": 5.814877957680973e-06, + "loss": 0.4022772014141083, + "loss_sft": 0.4022772014141083, + "step": 401, + "token_acc": 0.8740838639913493 + }, + { + "epoch": 1.4266666666666667, + "grad_norm": 1.8875541687011719, + "learning_rate": 5.795643591305945e-06, + "loss": 0.37720972299575806, + "loss_sft": 0.37720972299575806, + "step": 402, + "token_acc": 0.8802598852850109 + }, + { + "epoch": 1.4302222222222223, + "grad_norm": 2.043931484222412, + "learning_rate": 5.776397137119362e-06, + "loss": 0.41655582189559937, + "loss_sft": 0.41655582189559937, + "step": 403, + "token_acc": 0.8653957584888168 + }, + { + "epoch": 1.4337777777777778, + "grad_norm": 1.9095847606658936, + "learning_rate": 5.757138887522884e-06, + "loss": 0.4166144132614136, + "loss_sft": 0.4166144132614136, + "step": 404, + "token_acc": 0.8677405451998706 + }, + { + "epoch": 1.4373333333333334, + "grad_norm": 1.9745413064956665, + "learning_rate": 5.7378691350973835e-06, + "loss": 0.4126484990119934, + "loss_sft": 0.4126484990119934, + "step": 405, + "token_acc": 0.8647841433613902 + }, + { + "epoch": 1.4408888888888889, + "grad_norm": 1.9448391199111938, + "learning_rate": 5.7185881725984835e-06, + "loss": 0.37950587272644043, + "loss_sft": 0.37950587272644043, + "step": 406, + "token_acc": 0.8795446822969759 + }, + { + "epoch": 1.4444444444444444, + "grad_norm": 1.899501085281372, + "learning_rate": 5.699296292952117e-06, + "loss": 0.3865833282470703, + "loss_sft": 0.3865833282470703, + "step": 407, + "token_acc": 0.8742229367631297 + }, + { + "epoch": 1.448, + "grad_norm": 1.8692617416381836, + "learning_rate": 5.679993789250075e-06, + "loss": 0.38884326815605164, + "loss_sft": 0.38884326815605164, + "step": 408, + "token_acc": 0.8772 + }, + { + "epoch": 1.4515555555555555, + "grad_norm": 2.08317232131958, + "learning_rate": 5.660680954745554e-06, + "loss": 0.42538750171661377, + "loss_sft": 0.42538750171661377, + "step": 409, + "token_acc": 0.8640242224057253 + }, + { + "epoch": 1.455111111111111, + "grad_norm": 2.0576531887054443, + "learning_rate": 5.641358082848705e-06, + "loss": 0.41507261991500854, + "loss_sft": 0.41507261991500854, + "step": 410, + "token_acc": 0.8678167058320871 + }, + { + "epoch": 1.4586666666666668, + "grad_norm": 2.1288981437683105, + "learning_rate": 5.622025467122167e-06, + "loss": 0.42946210503578186, + "loss_sft": 0.42946210503578186, + "step": 411, + "token_acc": 0.8643975575841848 + }, + { + "epoch": 1.462222222222222, + "grad_norm": 1.8328207731246948, + "learning_rate": 5.6026834012766155e-06, + "loss": 0.4089741110801697, + "loss_sft": 0.4089741110801697, + "step": 412, + "token_acc": 0.8729965320086231 + }, + { + "epoch": 1.4657777777777778, + "grad_norm": 1.9151228666305542, + "learning_rate": 5.5833321791662975e-06, + "loss": 0.42090898752212524, + "loss_sft": 0.42090898752212524, + "step": 413, + "token_acc": 0.8661742983751847 + }, + { + "epoch": 1.4693333333333334, + "grad_norm": 2.014491081237793, + "learning_rate": 5.563972094784561e-06, + "loss": 0.395152747631073, + "loss_sft": 0.395152747631073, + "step": 414, + "token_acc": 0.8729997746224927 + }, + { + "epoch": 1.472888888888889, + "grad_norm": 3.1929755210876465, + "learning_rate": 5.544603442259401e-06, + "loss": 0.42928653955459595, + "loss_sft": 0.42928653955459595, + "step": 415, + "token_acc": 0.8654777748203354 + }, + { + "epoch": 1.4764444444444444, + "grad_norm": 2.0094358921051025, + "learning_rate": 5.525226515848979e-06, + "loss": 0.4076651334762573, + "loss_sft": 0.4076651334762573, + "step": 416, + "token_acc": 0.8753428491463756 + }, + { + "epoch": 1.48, + "grad_norm": 1.8888049125671387, + "learning_rate": 5.505841609937162e-06, + "loss": 0.4455106854438782, + "loss_sft": 0.4455106854438782, + "step": 417, + "token_acc": 0.8613653046244188 + }, + { + "epoch": 1.4835555555555555, + "grad_norm": 1.8966654539108276, + "learning_rate": 5.4864490190290386e-06, + "loss": 0.37113890051841736, + "loss_sft": 0.37113890051841736, + "step": 418, + "token_acc": 0.8809103840682788 + }, + { + "epoch": 1.487111111111111, + "grad_norm": 1.8363292217254639, + "learning_rate": 5.46704903774646e-06, + "loss": 0.38348186016082764, + "loss_sft": 0.38348186016082764, + "step": 419, + "token_acc": 0.8773685228151585 + }, + { + "epoch": 1.4906666666666666, + "grad_norm": 1.836222767829895, + "learning_rate": 5.447641960823549e-06, + "loss": 0.4077785611152649, + "loss_sft": 0.4077785611152649, + "step": 420, + "token_acc": 0.8698669474490285 + }, + { + "epoch": 1.4942222222222221, + "grad_norm": 2.0873465538024902, + "learning_rate": 5.42822808310223e-06, + "loss": 0.42955926060676575, + "loss_sft": 0.42955926060676575, + "step": 421, + "token_acc": 0.8674141779147625 + }, + { + "epoch": 1.4977777777777779, + "grad_norm": 1.903991460800171, + "learning_rate": 5.4088076995277564e-06, + "loss": 0.37381500005722046, + "loss_sft": 0.37381500005722046, + "step": 422, + "token_acc": 0.8775025091120384 + }, + { + "epoch": 1.5013333333333332, + "grad_norm": 1.92979097366333, + "learning_rate": 5.389381105144208e-06, + "loss": 0.37762191891670227, + "loss_sft": 0.37762191891670227, + "step": 423, + "token_acc": 0.8768152509111298 + }, + { + "epoch": 1.504888888888889, + "grad_norm": 2.0035507678985596, + "learning_rate": 5.369948595090033e-06, + "loss": 0.39775770902633667, + "loss_sft": 0.39775770902633667, + "step": 424, + "token_acc": 0.8741946422516107 + }, + { + "epoch": 1.5084444444444445, + "grad_norm": 2.1313304901123047, + "learning_rate": 5.350510464593548e-06, + "loss": 0.39650726318359375, + "loss_sft": 0.39650726318359375, + "step": 425, + "token_acc": 0.8754708858034658 + }, + { + "epoch": 1.512, + "grad_norm": 2.244947671890259, + "learning_rate": 5.331067008968462e-06, + "loss": 0.43273594975471497, + "loss_sft": 0.43273594975471497, + "step": 426, + "token_acc": 0.8628724932841881 + }, + { + "epoch": 1.5155555555555555, + "grad_norm": 1.8360341787338257, + "learning_rate": 5.311618523609386e-06, + "loss": 0.38386186957359314, + "loss_sft": 0.38386186957359314, + "step": 427, + "token_acc": 0.8766920170277898 + }, + { + "epoch": 1.519111111111111, + "grad_norm": 1.9270108938217163, + "learning_rate": 5.292165303987336e-06, + "loss": 0.4086407423019409, + "loss_sft": 0.4086407423019409, + "step": 428, + "token_acc": 0.8683509970365162 + }, + { + "epoch": 1.5226666666666666, + "grad_norm": 1.8471815586090088, + "learning_rate": 5.27270764564527e-06, + "loss": 0.4079905152320862, + "loss_sft": 0.4079905152320862, + "step": 429, + "token_acc": 0.8703739585351676 + }, + { + "epoch": 1.5262222222222221, + "grad_norm": 1.9703770875930786, + "learning_rate": 5.253245844193564e-06, + "loss": 0.4053723216056824, + "loss_sft": 0.4053723216056824, + "step": 430, + "token_acc": 0.8706947032206548 + }, + { + "epoch": 1.529777777777778, + "grad_norm": 1.9479751586914062, + "learning_rate": 5.23378019530555e-06, + "loss": 0.37635523080825806, + "loss_sft": 0.37635523080825806, + "step": 431, + "token_acc": 0.8817069172768004 + }, + { + "epoch": 1.5333333333333332, + "grad_norm": 1.9845861196517944, + "learning_rate": 5.214310994713008e-06, + "loss": 0.4136420488357544, + "loss_sft": 0.4136420488357544, + "step": 432, + "token_acc": 0.8720464135021097 + }, + { + "epoch": 1.536888888888889, + "grad_norm": 1.9639101028442383, + "learning_rate": 5.194838538201676e-06, + "loss": 0.41074252128601074, + "loss_sft": 0.41074252128601074, + "step": 433, + "token_acc": 0.8715951300041271 + }, + { + "epoch": 1.5404444444444443, + "grad_norm": 1.924564242362976, + "learning_rate": 5.175363121606759e-06, + "loss": 0.40540391206741333, + "loss_sft": 0.40540391206741333, + "step": 434, + "token_acc": 0.8720066982975161 + }, + { + "epoch": 1.544, + "grad_norm": 1.968327522277832, + "learning_rate": 5.155885040808432e-06, + "loss": 0.42160603404045105, + "loss_sft": 0.42160603404045105, + "step": 435, + "token_acc": 0.8655846202332178 + }, + { + "epoch": 1.5475555555555556, + "grad_norm": 1.9123497009277344, + "learning_rate": 5.1364045917273505e-06, + "loss": 0.4248104691505432, + "loss_sft": 0.4248104691505432, + "step": 436, + "token_acc": 0.8679361179361179 + }, + { + "epoch": 1.551111111111111, + "grad_norm": 2.8548669815063477, + "learning_rate": 5.116922070320144e-06, + "loss": 0.40673306584358215, + "loss_sft": 0.40673306584358215, + "step": 437, + "token_acc": 0.8735209306950383 + }, + { + "epoch": 1.5546666666666666, + "grad_norm": 2.012047529220581, + "learning_rate": 5.097437772574927e-06, + "loss": 0.4006389379501343, + "loss_sft": 0.4006389379501343, + "step": 438, + "token_acc": 0.8690501375079331 + }, + { + "epoch": 1.5582222222222222, + "grad_norm": 1.932854175567627, + "learning_rate": 5.077951994506805e-06, + "loss": 0.394672691822052, + "loss_sft": 0.394672691822052, + "step": 439, + "token_acc": 0.8751551119503642 + }, + { + "epoch": 1.561777777777778, + "grad_norm": 1.9678010940551758, + "learning_rate": 5.058465032153368e-06, + "loss": 0.35740727186203003, + "loss_sft": 0.35740727186203003, + "step": 440, + "token_acc": 0.8873391200239449 + }, + { + "epoch": 1.5653333333333332, + "grad_norm": 1.8318246603012085, + "learning_rate": 5.038977181570204e-06, + "loss": 0.41763466596603394, + "loss_sft": 0.41763466596603394, + "step": 441, + "token_acc": 0.8691725768321513 + }, + { + "epoch": 1.568888888888889, + "grad_norm": 1.9808677434921265, + "learning_rate": 5.0194887388263895e-06, + "loss": 0.41986605525016785, + "loss_sft": 0.41986605525016785, + "step": 442, + "token_acc": 0.8678516637570756 + }, + { + "epoch": 1.5724444444444443, + "grad_norm": 2.161735773086548, + "learning_rate": 5e-06, + "loss": 0.4062464237213135, + "loss_sft": 0.4062464237213135, + "step": 443, + "token_acc": 0.875208199871877 + }, + { + "epoch": 1.576, + "grad_norm": 1.9123862981796265, + "learning_rate": 4.980511261173613e-06, + "loss": 0.40120089054107666, + "loss_sft": 0.40120089054107666, + "step": 444, + "token_acc": 0.8728662969145442 + }, + { + "epoch": 1.5795555555555556, + "grad_norm": 1.951658010482788, + "learning_rate": 4.961022818429798e-06, + "loss": 0.3967818021774292, + "loss_sft": 0.3967818021774292, + "step": 445, + "token_acc": 0.8751601582084564 + }, + { + "epoch": 1.5831111111111111, + "grad_norm": 1.8390686511993408, + "learning_rate": 4.9415349678466335e-06, + "loss": 0.39594775438308716, + "loss_sft": 0.39594775438308716, + "step": 446, + "token_acc": 0.871834039617621 + }, + { + "epoch": 1.5866666666666667, + "grad_norm": 2.1262521743774414, + "learning_rate": 4.922048005493196e-06, + "loss": 0.38420170545578003, + "loss_sft": 0.38420170545578003, + "step": 447, + "token_acc": 0.8713295526149969 + }, + { + "epoch": 1.5902222222222222, + "grad_norm": 2.0551514625549316, + "learning_rate": 4.902562227425075e-06, + "loss": 0.3614252209663391, + "loss_sft": 0.3614252209663391, + "step": 448, + "token_acc": 0.8843389985171667 + }, + { + "epoch": 1.5937777777777777, + "grad_norm": 2.0104339122772217, + "learning_rate": 4.883077929679859e-06, + "loss": 0.42731672525405884, + "loss_sft": 0.42731672525405884, + "step": 449, + "token_acc": 0.8652869972666927 + }, + { + "epoch": 1.5973333333333333, + "grad_norm": 1.9819819927215576, + "learning_rate": 4.86359540827265e-06, + "loss": 0.36734795570373535, + "loss_sft": 0.36734795570373535, + "step": 450, + "token_acc": 0.8803915922833285 + }, + { + "epoch": 1.600888888888889, + "grad_norm": 1.9737902879714966, + "learning_rate": 4.844114959191569e-06, + "loss": 0.4024088978767395, + "loss_sft": 0.4024088978767395, + "step": 451, + "token_acc": 0.8738017486569051 + }, + { + "epoch": 1.6044444444444443, + "grad_norm": 1.9214792251586914, + "learning_rate": 4.824636878393243e-06, + "loss": 0.38559162616729736, + "loss_sft": 0.38559162616729736, + "step": 452, + "token_acc": 0.8761592164217985 + }, + { + "epoch": 1.608, + "grad_norm": 2.003669500350952, + "learning_rate": 4.805161461798326e-06, + "loss": 0.3978857398033142, + "loss_sft": 0.3978857398033142, + "step": 453, + "token_acc": 0.8741655540720962 + }, + { + "epoch": 1.6115555555555554, + "grad_norm": 1.8802067041397095, + "learning_rate": 4.785689005286995e-06, + "loss": 0.4288980960845947, + "loss_sft": 0.4288980960845947, + "step": 454, + "token_acc": 0.8665644171779141 + }, + { + "epoch": 1.6151111111111112, + "grad_norm": 1.943430781364441, + "learning_rate": 4.766219804694451e-06, + "loss": 0.4066818952560425, + "loss_sft": 0.4066818952560425, + "step": 455, + "token_acc": 0.8713506066868012 + }, + { + "epoch": 1.6186666666666667, + "grad_norm": 2.035663604736328, + "learning_rate": 4.746754155806437e-06, + "loss": 0.4036039710044861, + "loss_sft": 0.4036039710044861, + "step": 456, + "token_acc": 0.8720917099615992 + }, + { + "epoch": 1.6222222222222222, + "grad_norm": 1.8568787574768066, + "learning_rate": 4.727292354354731e-06, + "loss": 0.38928794860839844, + "loss_sft": 0.38928794860839844, + "step": 457, + "token_acc": 0.8776773761713521 + }, + { + "epoch": 1.6257777777777778, + "grad_norm": 2.0654497146606445, + "learning_rate": 4.7078346960126645e-06, + "loss": 0.40362274646759033, + "loss_sft": 0.40362274646759033, + "step": 458, + "token_acc": 0.8697181193529189 + }, + { + "epoch": 1.6293333333333333, + "grad_norm": 2.0320656299591064, + "learning_rate": 4.688381476390617e-06, + "loss": 0.40253838896751404, + "loss_sft": 0.40253838896751404, + "step": 459, + "token_acc": 0.8743815283122595 + }, + { + "epoch": 1.6328888888888888, + "grad_norm": 1.8148250579833984, + "learning_rate": 4.668932991031538e-06, + "loss": 0.37930411100387573, + "loss_sft": 0.37930411100387573, + "step": 460, + "token_acc": 0.8785644584191694 + }, + { + "epoch": 1.6364444444444444, + "grad_norm": 1.9612997770309448, + "learning_rate": 4.649489535406454e-06, + "loss": 0.3845377564430237, + "loss_sft": 0.3845377564430237, + "step": 461, + "token_acc": 0.8750652514355316 + }, + { + "epoch": 1.6400000000000001, + "grad_norm": 2.0092227458953857, + "learning_rate": 4.6300514049099694e-06, + "loss": 0.3916059732437134, + "loss_sft": 0.3916059732437134, + "step": 462, + "token_acc": 0.8748483177054606 + }, + { + "epoch": 1.6435555555555554, + "grad_norm": 2.015611171722412, + "learning_rate": 4.6106188948557935e-06, + "loss": 0.4142994284629822, + "loss_sft": 0.4142994284629822, + "step": 463, + "token_acc": 0.870081531371854 + }, + { + "epoch": 1.6471111111111112, + "grad_norm": 1.9973748922348022, + "learning_rate": 4.591192300472247e-06, + "loss": 0.40479254722595215, + "loss_sft": 0.40479254722595215, + "step": 464, + "token_acc": 0.87182932072227 + }, + { + "epoch": 1.6506666666666665, + "grad_norm": 1.9991106986999512, + "learning_rate": 4.5717719168977696e-06, + "loss": 0.41683700680732727, + "loss_sft": 0.41683700680732727, + "step": 465, + "token_acc": 0.8689903846153846 + }, + { + "epoch": 1.6542222222222223, + "grad_norm": 2.0856075286865234, + "learning_rate": 4.552358039176453e-06, + "loss": 0.3779621124267578, + "loss_sft": 0.3779621124267578, + "step": 466, + "token_acc": 0.8759251459173235 + }, + { + "epoch": 1.6577777777777778, + "grad_norm": 1.9581468105316162, + "learning_rate": 4.532950962253543e-06, + "loss": 0.3692684769630432, + "loss_sft": 0.3692684769630432, + "step": 467, + "token_acc": 0.8823826917673504 + }, + { + "epoch": 1.6613333333333333, + "grad_norm": 1.9361141920089722, + "learning_rate": 4.513550980970962e-06, + "loss": 0.40604203939437866, + "loss_sft": 0.40604203939437866, + "step": 468, + "token_acc": 0.8690947479546054 + }, + { + "epoch": 1.6648888888888889, + "grad_norm": 1.9669075012207031, + "learning_rate": 4.49415839006284e-06, + "loss": 0.4372028708457947, + "loss_sft": 0.4372028708457947, + "step": 469, + "token_acc": 0.8647852760736197 + }, + { + "epoch": 1.6684444444444444, + "grad_norm": 1.9877018928527832, + "learning_rate": 4.474773484151021e-06, + "loss": 0.42890945076942444, + "loss_sft": 0.42890945076942444, + "step": 470, + "token_acc": 0.8637859179644659 + }, + { + "epoch": 1.6720000000000002, + "grad_norm": 1.9818249940872192, + "learning_rate": 4.4553965577406006e-06, + "loss": 0.419004887342453, + "loss_sft": 0.419004887342453, + "step": 471, + "token_acc": 0.8697004250025915 + }, + { + "epoch": 1.6755555555555555, + "grad_norm": 2.002903461456299, + "learning_rate": 4.4360279052154406e-06, + "loss": 0.41087615489959717, + "loss_sft": 0.41087615489959717, + "step": 472, + "token_acc": 0.8679330167458136 + }, + { + "epoch": 1.6791111111111112, + "grad_norm": 1.9055359363555908, + "learning_rate": 4.416667820833704e-06, + "loss": 0.36150631308555603, + "loss_sft": 0.36150631308555603, + "step": 473, + "token_acc": 0.8842860090796533 + }, + { + "epoch": 1.6826666666666665, + "grad_norm": 1.7947391271591187, + "learning_rate": 4.397316598723385e-06, + "loss": 0.3952457904815674, + "loss_sft": 0.3952457904815674, + "step": 474, + "token_acc": 0.871913201433551 + }, + { + "epoch": 1.6862222222222223, + "grad_norm": 2.049100160598755, + "learning_rate": 4.377974532877834e-06, + "loss": 0.40265393257141113, + "loss_sft": 0.40265393257141113, + "step": 475, + "token_acc": 0.8727088721466322 + }, + { + "epoch": 1.6897777777777778, + "grad_norm": 1.9173460006713867, + "learning_rate": 4.358641917151297e-06, + "loss": 0.426979660987854, + "loss_sft": 0.426979660987854, + "step": 476, + "token_acc": 0.8654578698765808 + }, + { + "epoch": 1.6933333333333334, + "grad_norm": 2.024441719055176, + "learning_rate": 4.339319045254448e-06, + "loss": 0.41185951232910156, + "loss_sft": 0.41185951232910156, + "step": 477, + "token_acc": 0.8650110912730617 + }, + { + "epoch": 1.696888888888889, + "grad_norm": 1.8032621145248413, + "learning_rate": 4.320006210749928e-06, + "loss": 0.3909500241279602, + "loss_sft": 0.3909500241279602, + "step": 478, + "token_acc": 0.8736356205221907 + }, + { + "epoch": 1.7004444444444444, + "grad_norm": 2.0005927085876465, + "learning_rate": 4.3007037070478855e-06, + "loss": 0.39660173654556274, + "loss_sft": 0.39660173654556274, + "step": 479, + "token_acc": 0.8720379146919431 + }, + { + "epoch": 1.704, + "grad_norm": 2.057610034942627, + "learning_rate": 4.281411827401517e-06, + "loss": 0.3668193221092224, + "loss_sft": 0.3668193221092224, + "step": 480, + "token_acc": 0.8818962774419344 + }, + { + "epoch": 1.7075555555555555, + "grad_norm": 2.0135931968688965, + "learning_rate": 4.262130864902617e-06, + "loss": 0.40049654245376587, + "loss_sft": 0.40049654245376587, + "step": 481, + "token_acc": 0.8739949395558054 + }, + { + "epoch": 1.7111111111111112, + "grad_norm": 1.8834664821624756, + "learning_rate": 4.2428611124771184e-06, + "loss": 0.3981768488883972, + "loss_sft": 0.3981768488883972, + "step": 482, + "token_acc": 0.8758667507152209 + }, + { + "epoch": 1.7146666666666666, + "grad_norm": 1.852174162864685, + "learning_rate": 4.22360286288064e-06, + "loss": 0.4088278114795685, + "loss_sft": 0.4088278114795685, + "step": 483, + "token_acc": 0.8707245567499651 + }, + { + "epoch": 1.7182222222222223, + "grad_norm": 1.8437398672103882, + "learning_rate": 4.2043564086940565e-06, + "loss": 0.3530130684375763, + "loss_sft": 0.3530130684375763, + "step": 484, + "token_acc": 0.885416097481012 + }, + { + "epoch": 1.7217777777777776, + "grad_norm": 1.923483967781067, + "learning_rate": 4.185122042319027e-06, + "loss": 0.45271462202072144, + "loss_sft": 0.45271462202072144, + "step": 485, + "token_acc": 0.8575012319132733 + }, + { + "epoch": 1.7253333333333334, + "grad_norm": 1.9574507474899292, + "learning_rate": 4.165900055973579e-06, + "loss": 0.3632383942604065, + "loss_sft": 0.3632383942604065, + "step": 486, + "token_acc": 0.8842877805753468 + }, + { + "epoch": 1.728888888888889, + "grad_norm": 1.9624892473220825, + "learning_rate": 4.1466907416876466e-06, + "loss": 0.430122971534729, + "loss_sft": 0.430122971534729, + "step": 487, + "token_acc": 0.8635190918472653 + }, + { + "epoch": 1.7324444444444445, + "grad_norm": 1.9981114864349365, + "learning_rate": 4.127494391298647e-06, + "loss": 0.3593122959136963, + "loss_sft": 0.3593122959136963, + "step": 488, + "token_acc": 0.8826551462432133 + }, + { + "epoch": 1.736, + "grad_norm": 1.9912083148956299, + "learning_rate": 4.108311296447048e-06, + "loss": 0.38457179069519043, + "loss_sft": 0.38457179069519043, + "step": 489, + "token_acc": 0.8769977168949772 + }, + { + "epoch": 1.7395555555555555, + "grad_norm": 2.002443790435791, + "learning_rate": 4.089141748571926e-06, + "loss": 0.4178641438484192, + "loss_sft": 0.4178641438484192, + "step": 490, + "token_acc": 0.8677256917406427 + }, + { + "epoch": 1.743111111111111, + "grad_norm": 1.8774305582046509, + "learning_rate": 4.069986038906547e-06, + "loss": 0.3666388988494873, + "loss_sft": 0.3666388988494873, + "step": 491, + "token_acc": 0.8861342694164107 + }, + { + "epoch": 1.7466666666666666, + "grad_norm": 1.9724780321121216, + "learning_rate": 4.050844458473945e-06, + "loss": 0.4046722948551178, + "loss_sft": 0.4046722948551178, + "step": 492, + "token_acc": 0.8712591141582327 + }, + { + "epoch": 1.7502222222222223, + "grad_norm": 2.0387706756591797, + "learning_rate": 4.031717298082487e-06, + "loss": 0.3910101354122162, + "loss_sft": 0.3910101354122162, + "step": 493, + "token_acc": 0.8714958775029447 + }, + { + "epoch": 1.7537777777777777, + "grad_norm": 1.9948173761367798, + "learning_rate": 4.012604848321471e-06, + "loss": 0.41906845569610596, + "loss_sft": 0.41906845569610596, + "step": 494, + "token_acc": 0.8669435390600357 + }, + { + "epoch": 1.7573333333333334, + "grad_norm": 2.0991873741149902, + "learning_rate": 3.993507399556699e-06, + "loss": 0.41146332025527954, + "loss_sft": 0.41146332025527954, + "step": 495, + "token_acc": 0.8679862497227767 + }, + { + "epoch": 1.7608888888888887, + "grad_norm": 1.9059646129608154, + "learning_rate": 3.974425241926076e-06, + "loss": 0.3699517846107483, + "loss_sft": 0.3699517846107483, + "step": 496, + "token_acc": 0.8773111877154497 + }, + { + "epoch": 1.7644444444444445, + "grad_norm": 1.9729499816894531, + "learning_rate": 3.95535866533519e-06, + "loss": 0.4262514114379883, + "loss_sft": 0.4262514114379883, + "step": 497, + "token_acc": 0.8628194396572948 + }, + { + "epoch": 1.768, + "grad_norm": 1.9515553712844849, + "learning_rate": 3.936307959452917e-06, + "loss": 0.38408201932907104, + "loss_sft": 0.38408201932907104, + "step": 498, + "token_acc": 0.8783570300157978 + }, + { + "epoch": 1.7715555555555556, + "grad_norm": 1.931066870689392, + "learning_rate": 3.91727341370702e-06, + "loss": 0.4075785279273987, + "loss_sft": 0.4075785279273987, + "step": 499, + "token_acc": 0.8675392412537519 + }, + { + "epoch": 1.775111111111111, + "grad_norm": 1.9167137145996094, + "learning_rate": 3.898255317279744e-06, + "loss": 0.37822866439819336, + "loss_sft": 0.37822866439819336, + "step": 500, + "token_acc": 0.8760292061519341 + }, + { + "epoch": 1.7786666666666666, + "grad_norm": 2.104595422744751, + "learning_rate": 3.879253959103429e-06, + "loss": 0.4038432240486145, + "loss_sft": 0.4038432240486145, + "step": 501, + "token_acc": 0.8717237308146399 + }, + { + "epoch": 1.7822222222222224, + "grad_norm": 1.9105145931243896, + "learning_rate": 3.860269627856126e-06, + "loss": 0.41195592284202576, + "loss_sft": 0.41195592284202576, + "step": 502, + "token_acc": 0.8666333416645838 + }, + { + "epoch": 1.7857777777777777, + "grad_norm": 1.9904228448867798, + "learning_rate": 3.841302611957193e-06, + "loss": 0.4045681953430176, + "loss_sft": 0.4045681953430176, + "step": 503, + "token_acc": 0.8716422799737934 + }, + { + "epoch": 1.7893333333333334, + "grad_norm": 1.929783821105957, + "learning_rate": 3.822353199562936e-06, + "loss": 0.3734525740146637, + "loss_sft": 0.3734525740146637, + "step": 504, + "token_acc": 0.8800486494704303 + }, + { + "epoch": 1.7928888888888888, + "grad_norm": 2.106736660003662, + "learning_rate": 3.803421678562213e-06, + "loss": 0.41517573595046997, + "loss_sft": 0.41517573595046997, + "step": 505, + "token_acc": 0.8700871567176451 + }, + { + "epoch": 1.7964444444444445, + "grad_norm": 1.9528405666351318, + "learning_rate": 3.784508336572066e-06, + "loss": 0.42377999424934387, + "loss_sft": 0.42377999424934387, + "step": 506, + "token_acc": 0.8674748304910622 + }, + { + "epoch": 1.8, + "grad_norm": 2.0329067707061768, + "learning_rate": 3.7656134609333576e-06, + "loss": 0.40288347005844116, + "loss_sft": 0.40288347005844116, + "step": 507, + "token_acc": 0.8693903349997127 + }, + { + "epoch": 1.8035555555555556, + "grad_norm": 2.0448217391967773, + "learning_rate": 3.7467373387063973e-06, + "loss": 0.4332900643348694, + "loss_sft": 0.4332900643348694, + "step": 508, + "token_acc": 0.8657247747246635 + }, + { + "epoch": 1.8071111111111111, + "grad_norm": 2.0441441535949707, + "learning_rate": 3.727880256666586e-06, + "loss": 0.41990628838539124, + "loss_sft": 0.41990628838539124, + "step": 509, + "token_acc": 0.868348233290762 + }, + { + "epoch": 1.8106666666666666, + "grad_norm": 1.9429893493652344, + "learning_rate": 3.709042501300052e-06, + "loss": 0.3955222964286804, + "loss_sft": 0.3955222964286804, + "step": 510, + "token_acc": 0.8745842991030939 + }, + { + "epoch": 1.8142222222222222, + "grad_norm": 1.955379605293274, + "learning_rate": 3.6902243587993068e-06, + "loss": 0.39930152893066406, + "loss_sft": 0.39930152893066406, + "step": 511, + "token_acc": 0.8724778901657576 + }, + { + "epoch": 1.8177777777777777, + "grad_norm": 1.9959750175476074, + "learning_rate": 3.6714261150588947e-06, + "loss": 0.4393870234489441, + "loss_sft": 0.4393870234489441, + "step": 512, + "token_acc": 0.8642143983184446 + }, + { + "epoch": 1.8213333333333335, + "grad_norm": 2.0057990550994873, + "learning_rate": 3.652648055671043e-06, + "loss": 0.4101144075393677, + "loss_sft": 0.4101144075393677, + "step": 513, + "token_acc": 0.8663834607759842 + }, + { + "epoch": 1.8248888888888888, + "grad_norm": 2.141768217086792, + "learning_rate": 3.6338904659213335e-06, + "loss": 0.4051585793495178, + "loss_sft": 0.4051585793495178, + "step": 514, + "token_acc": 0.8674832962138085 + }, + { + "epoch": 1.8284444444444445, + "grad_norm": 1.9821730852127075, + "learning_rate": 3.6151536307843625e-06, + "loss": 0.40732455253601074, + "loss_sft": 0.40732455253601074, + "step": 515, + "token_acc": 0.8715576357897028 + }, + { + "epoch": 1.8319999999999999, + "grad_norm": 1.820754885673523, + "learning_rate": 3.5964378349194075e-06, + "loss": 0.38349008560180664, + "loss_sft": 0.38349008560180664, + "step": 516, + "token_acc": 0.8765456721180694 + }, + { + "epoch": 1.8355555555555556, + "grad_norm": 1.998852014541626, + "learning_rate": 3.577743362666112e-06, + "loss": 0.43499207496643066, + "loss_sft": 0.43499207496643066, + "step": 517, + "token_acc": 0.8607266003697966 + }, + { + "epoch": 1.8391111111111111, + "grad_norm": 2.17022705078125, + "learning_rate": 3.5590704980401564e-06, + "loss": 0.3998159170150757, + "loss_sft": 0.3998159170150757, + "step": 518, + "token_acc": 0.8774787052810903 + }, + { + "epoch": 1.8426666666666667, + "grad_norm": 1.9512163400650024, + "learning_rate": 3.540419524728954e-06, + "loss": 0.40242066979408264, + "loss_sft": 0.40242066979408264, + "step": 519, + "token_acc": 0.8706586826347306 + }, + { + "epoch": 1.8462222222222222, + "grad_norm": 1.8834881782531738, + "learning_rate": 3.5217907260873265e-06, + "loss": 0.35156160593032837, + "loss_sft": 0.35156160593032837, + "step": 520, + "token_acc": 0.8842680222811097 + }, + { + "epoch": 1.8497777777777777, + "grad_norm": 1.9059492349624634, + "learning_rate": 3.5031843851332105e-06, + "loss": 0.39381706714630127, + "loss_sft": 0.39381706714630127, + "step": 521, + "token_acc": 0.8784588441330998 + }, + { + "epoch": 1.8533333333333335, + "grad_norm": 1.9964344501495361, + "learning_rate": 3.4846007845433568e-06, + "loss": 0.37252557277679443, + "loss_sft": 0.37252557277679443, + "step": 522, + "token_acc": 0.8810984032266976 + }, + { + "epoch": 1.8568888888888888, + "grad_norm": 1.934705376625061, + "learning_rate": 3.4660402066490274e-06, + "loss": 0.36136066913604736, + "loss_sft": 0.36136066913604736, + "step": 523, + "token_acc": 0.881482741030437 + }, + { + "epoch": 1.8604444444444446, + "grad_norm": 2.021376371383667, + "learning_rate": 3.4475029334317195e-06, + "loss": 0.4018491506576538, + "loss_sft": 0.4018491506576538, + "step": 524, + "token_acc": 0.8702962834116078 + }, + { + "epoch": 1.8639999999999999, + "grad_norm": 2.0042481422424316, + "learning_rate": 3.428989246518867e-06, + "loss": 0.3839758634567261, + "loss_sft": 0.3839758634567261, + "step": 525, + "token_acc": 0.8749489408881368 + }, + { + "epoch": 1.8675555555555556, + "grad_norm": 1.798644781112671, + "learning_rate": 3.410499427179572e-06, + "loss": 0.3922249674797058, + "loss_sft": 0.3922249674797058, + "step": 526, + "token_acc": 0.8744547437295529 + }, + { + "epoch": 1.871111111111111, + "grad_norm": 2.006016731262207, + "learning_rate": 3.392033756320333e-06, + "loss": 0.3796369433403015, + "loss_sft": 0.3796369433403015, + "step": 527, + "token_acc": 0.8796517582792762 + }, + { + "epoch": 1.8746666666666667, + "grad_norm": 1.9467095136642456, + "learning_rate": 3.3735925144807623e-06, + "loss": 0.41722482442855835, + "loss_sft": 0.41722482442855835, + "step": 528, + "token_acc": 0.871854885898186 + }, + { + "epoch": 1.8782222222222222, + "grad_norm": 2.1234662532806396, + "learning_rate": 3.355175981829346e-06, + "loss": 0.3958422541618347, + "loss_sft": 0.3958422541618347, + "step": 529, + "token_acc": 0.8745638417410846 + }, + { + "epoch": 1.8817777777777778, + "grad_norm": 2.785810708999634, + "learning_rate": 3.33678443815917e-06, + "loss": 0.3794466257095337, + "loss_sft": 0.3794466257095337, + "step": 530, + "token_acc": 0.878014553014553 + }, + { + "epoch": 1.8853333333333333, + "grad_norm": 2.0013463497161865, + "learning_rate": 3.318418162883671e-06, + "loss": 0.3745514452457428, + "loss_sft": 0.3745514452457428, + "step": 531, + "token_acc": 0.8798875914530743 + }, + { + "epoch": 1.8888888888888888, + "grad_norm": 1.8750319480895996, + "learning_rate": 3.300077435032406e-06, + "loss": 0.3796156644821167, + "loss_sft": 0.3796156644821167, + "step": 532, + "token_acc": 0.8813118933228747 + }, + { + "epoch": 1.8924444444444446, + "grad_norm": 1.9550929069519043, + "learning_rate": 3.281762533246794e-06, + "loss": 0.4177267253398895, + "loss_sft": 0.4177267253398895, + "step": 533, + "token_acc": 0.8660500717588901 + }, + { + "epoch": 1.896, + "grad_norm": 2.1585350036621094, + "learning_rate": 3.2634737357758994e-06, + "loss": 0.40671637654304504, + "loss_sft": 0.40671637654304504, + "step": 534, + "token_acc": 0.8732265724898008 + }, + { + "epoch": 1.8995555555555557, + "grad_norm": 2.0903804302215576, + "learning_rate": 3.245211320472189e-06, + "loss": 0.40713876485824585, + "loss_sft": 0.40713876485824585, + "step": 535, + "token_acc": 0.872414185042055 + }, + { + "epoch": 1.903111111111111, + "grad_norm": 2.013906717300415, + "learning_rate": 3.226975564787322e-06, + "loss": 0.38883522152900696, + "loss_sft": 0.38883522152900696, + "step": 536, + "token_acc": 0.8766483057920214 + }, + { + "epoch": 1.9066666666666667, + "grad_norm": 2.0468356609344482, + "learning_rate": 3.208766745767935e-06, + "loss": 0.3963863253593445, + "loss_sft": 0.3963863253593445, + "step": 537, + "token_acc": 0.8759223994287074 + }, + { + "epoch": 1.9102222222222223, + "grad_norm": 2.0030837059020996, + "learning_rate": 3.1905851400514232e-06, + "loss": 0.39978474378585815, + "loss_sft": 0.39978474378585815, + "step": 538, + "token_acc": 0.87278641127575 + }, + { + "epoch": 1.9137777777777778, + "grad_norm": 1.9387980699539185, + "learning_rate": 3.1724310238617475e-06, + "loss": 0.3911464810371399, + "loss_sft": 0.3911464810371399, + "step": 539, + "token_acc": 0.8736636971046771 + }, + { + "epoch": 1.9173333333333333, + "grad_norm": 2.0179760456085205, + "learning_rate": 3.154304673005235e-06, + "loss": 0.3845818042755127, + "loss_sft": 0.3845818042755127, + "step": 540, + "token_acc": 0.879066478076379 + }, + { + "epoch": 1.9208888888888889, + "grad_norm": 2.1025238037109375, + "learning_rate": 3.1362063628663836e-06, + "loss": 0.43703532218933105, + "loss_sft": 0.43703532218933105, + "step": 541, + "token_acc": 0.8651649062297909 + }, + { + "epoch": 1.9244444444444444, + "grad_norm": 1.828596591949463, + "learning_rate": 3.118136368403689e-06, + "loss": 0.36484718322753906, + "loss_sft": 0.36484718322753906, + "step": 542, + "token_acc": 0.879990094105993 + }, + { + "epoch": 1.928, + "grad_norm": 1.873234510421753, + "learning_rate": 3.100094964145458e-06, + "loss": 0.3943330645561218, + "loss_sft": 0.3943330645561218, + "step": 543, + "token_acc": 0.8811200443581924 + }, + { + "epoch": 1.9315555555555557, + "grad_norm": 2.059091329574585, + "learning_rate": 3.0820824241856377e-06, + "loss": 0.3626614212989807, + "loss_sft": 0.3626614212989807, + "step": 544, + "token_acc": 0.8836625901536532 + }, + { + "epoch": 1.935111111111111, + "grad_norm": 1.881367564201355, + "learning_rate": 3.064099022179661e-06, + "loss": 0.38543713092803955, + "loss_sft": 0.38543713092803955, + "step": 545, + "token_acc": 0.8775531098029178 + }, + { + "epoch": 1.9386666666666668, + "grad_norm": 1.8997218608856201, + "learning_rate": 3.046145031340275e-06, + "loss": 0.41567015647888184, + "loss_sft": 0.41567015647888184, + "step": 546, + "token_acc": 0.8724944320712695 + }, + { + "epoch": 1.942222222222222, + "grad_norm": 1.927754521369934, + "learning_rate": 3.0282207244334084e-06, + "loss": 0.4172213077545166, + "loss_sft": 0.4172213077545166, + "step": 547, + "token_acc": 0.8675220050723557 + }, + { + "epoch": 1.9457777777777778, + "grad_norm": 2.057682752609253, + "learning_rate": 3.010326373774004e-06, + "loss": 0.38135772943496704, + "loss_sft": 0.38135772943496704, + "step": 548, + "token_acc": 0.8785829693016576 + }, + { + "epoch": 1.9493333333333334, + "grad_norm": 1.9360153675079346, + "learning_rate": 2.9924622512219037e-06, + "loss": 0.3865244388580322, + "loss_sft": 0.3865244388580322, + "step": 549, + "token_acc": 0.8745284481422149 + }, + { + "epoch": 1.952888888888889, + "grad_norm": 1.9218275547027588, + "learning_rate": 2.9746286281777075e-06, + "loss": 0.3842408061027527, + "loss_sft": 0.3842408061027527, + "step": 550, + "token_acc": 0.8750063981163946 + }, + { + "epoch": 1.9564444444444444, + "grad_norm": 2.071923017501831, + "learning_rate": 2.9568257755786474e-06, + "loss": 0.4030836820602417, + "loss_sft": 0.4030836820602417, + "step": 551, + "token_acc": 0.8737645036527718 + }, + { + "epoch": 1.96, + "grad_norm": 1.9977730512619019, + "learning_rate": 2.939053963894481e-06, + "loss": 0.4175572097301483, + "loss_sft": 0.4175572097301483, + "step": 552, + "token_acc": 0.8659229635866134 + }, + { + "epoch": 1.9635555555555557, + "grad_norm": 2.014700412750244, + "learning_rate": 2.921313463123375e-06, + "loss": 0.3887366056442261, + "loss_sft": 0.3887366056442261, + "step": 553, + "token_acc": 0.8758249392150053 + }, + { + "epoch": 1.967111111111111, + "grad_norm": 1.9743510484695435, + "learning_rate": 2.9036045427878e-06, + "loss": 0.3705500066280365, + "loss_sft": 0.3705500066280365, + "step": 554, + "token_acc": 0.8771412436468595 + }, + { + "epoch": 1.9706666666666668, + "grad_norm": 1.9782696962356567, + "learning_rate": 2.8859274719304543e-06, + "loss": 0.4092680513858795, + "loss_sft": 0.4092680513858795, + "step": 555, + "token_acc": 0.870730038423075 + }, + { + "epoch": 1.974222222222222, + "grad_norm": 1.971817970275879, + "learning_rate": 2.8682825191101447e-06, + "loss": 0.3677399754524231, + "loss_sft": 0.3677399754524231, + "step": 556, + "token_acc": 0.883681398737251 + }, + { + "epoch": 1.9777777777777779, + "grad_norm": 2.03106951713562, + "learning_rate": 2.850669952397743e-06, + "loss": 0.3869796693325043, + "loss_sft": 0.3869796693325043, + "step": 557, + "token_acc": 0.8771653543307086 + }, + { + "epoch": 1.9813333333333332, + "grad_norm": 2.000079870223999, + "learning_rate": 2.8330900393720846e-06, + "loss": 0.4369879961013794, + "loss_sft": 0.4369879961013794, + "step": 558, + "token_acc": 0.8628226683809712 + }, + { + "epoch": 1.984888888888889, + "grad_norm": 1.8278487920761108, + "learning_rate": 2.8155430471159118e-06, + "loss": 0.3823060393333435, + "loss_sft": 0.3823060393333435, + "step": 559, + "token_acc": 0.8796760313844596 + }, + { + "epoch": 1.9884444444444445, + "grad_norm": 2.2703516483306885, + "learning_rate": 2.7980292422118282e-06, + "loss": 0.38897961378097534, + "loss_sft": 0.38897961378097534, + "step": 560, + "token_acc": 0.8771003360537686 + }, + { + "epoch": 1.992, + "grad_norm": 2.070188522338867, + "learning_rate": 2.7805488907382316e-06, + "loss": 0.41422706842422485, + "loss_sft": 0.41422706842422485, + "step": 561, + "token_acc": 0.8700247248751636 + }, + { + "epoch": 1.9955555555555555, + "grad_norm": 1.9080613851547241, + "learning_rate": 2.7631022582652808e-06, + "loss": 0.37816062569618225, + "loss_sft": 0.37816062569618225, + "step": 562, + "token_acc": 0.8821118739688115 + }, + { + "epoch": 1.999111111111111, + "grad_norm": 2.0313827991485596, + "learning_rate": 2.745689609850859e-06, + "loss": 0.3750487267971039, + "loss_sft": 0.3750487267971039, + "step": 563, + "token_acc": 0.8804845048643773 + }, + { + "epoch": 2.0, + "grad_norm": 4.01881217956543, + "learning_rate": 2.728311210036542e-06, + "loss": 0.35268133878707886, + "loss_sft": 0.35268133878707886, + "step": 564, + "token_acc": 0.8753623188405797 + }, + { + "epoch": 2.0035555555555558, + "grad_norm": 1.8548333644866943, + "learning_rate": 2.7109673228435925e-06, + "loss": 0.24603433907032013, + "loss_sft": 0.24603433907032013, + "step": 565, + "token_acc": 0.9229494198603646 + }, + { + "epoch": 2.007111111111111, + "grad_norm": 1.940453052520752, + "learning_rate": 2.6936582117689347e-06, + "loss": 0.281282901763916, + "loss_sft": 0.281282901763916, + "step": 566, + "token_acc": 0.9164393250479943 + }, + { + "epoch": 2.010666666666667, + "grad_norm": 1.8731716871261597, + "learning_rate": 2.6763841397811576e-06, + "loss": 0.23941609263420105, + "loss_sft": 0.23941609263420105, + "step": 567, + "token_acc": 0.9253800358360789 + }, + { + "epoch": 2.014222222222222, + "grad_norm": 1.8930144309997559, + "learning_rate": 2.6591453693165204e-06, + "loss": 0.24905499815940857, + "loss_sft": 0.24905499815940857, + "step": 568, + "token_acc": 0.9212751752531434 + }, + { + "epoch": 2.017777777777778, + "grad_norm": 1.8181766271591187, + "learning_rate": 2.641942162274962e-06, + "loss": 0.22281771898269653, + "loss_sft": 0.22281771898269653, + "step": 569, + "token_acc": 0.9299920895016386 + }, + { + "epoch": 2.021333333333333, + "grad_norm": 1.9271034002304077, + "learning_rate": 2.624774780016126e-06, + "loss": 0.22228644788265228, + "loss_sft": 0.22228644788265228, + "step": 570, + "token_acc": 0.9278552624208817 + }, + { + "epoch": 2.024888888888889, + "grad_norm": 2.103590965270996, + "learning_rate": 2.607643483355387e-06, + "loss": 0.19711188971996307, + "loss_sft": 0.19711188971996307, + "step": 571, + "token_acc": 0.9356959962756052 + }, + { + "epoch": 2.0284444444444443, + "grad_norm": 2.3506200313568115, + "learning_rate": 2.59054853255989e-06, + "loss": 0.2155936360359192, + "loss_sft": 0.2155936360359192, + "step": 572, + "token_acc": 0.9300638421358096 + }, + { + "epoch": 2.032, + "grad_norm": 2.5372796058654785, + "learning_rate": 2.573490187344596e-06, + "loss": 0.21778741478919983, + "loss_sft": 0.21778741478919983, + "step": 573, + "token_acc": 0.9297760425139634 + }, + { + "epoch": 2.0355555555555553, + "grad_norm": 2.624566078186035, + "learning_rate": 2.5564687068683335e-06, + "loss": 0.22881969809532166, + "loss_sft": 0.22881969809532166, + "step": 574, + "token_acc": 0.9301398326500814 + }, + { + "epoch": 2.039111111111111, + "grad_norm": 2.737687110900879, + "learning_rate": 2.539484349729868e-06, + "loss": 0.23781782388687134, + "loss_sft": 0.23781782388687134, + "step": 575, + "token_acc": 0.9236206350880894 + }, + { + "epoch": 2.042666666666667, + "grad_norm": 3.0376739501953125, + "learning_rate": 2.5225373739639637e-06, + "loss": 0.24722090363502502, + "loss_sft": 0.24722090363502502, + "step": 576, + "token_acc": 0.9218483229608415 + }, + { + "epoch": 2.046222222222222, + "grad_norm": 2.653787612915039, + "learning_rate": 2.5056280370374725e-06, + "loss": 0.238386869430542, + "loss_sft": 0.238386869430542, + "step": 577, + "token_acc": 0.9219492261327068 + }, + { + "epoch": 2.049777777777778, + "grad_norm": 2.241947650909424, + "learning_rate": 2.488756595845417e-06, + "loss": 0.22438564896583557, + "loss_sft": 0.22438564896583557, + "step": 578, + "token_acc": 0.9269450634628428 + }, + { + "epoch": 2.0533333333333332, + "grad_norm": 2.3280882835388184, + "learning_rate": 2.471923306707086e-06, + "loss": 0.21121571958065033, + "loss_sft": 0.21121571958065033, + "step": 579, + "token_acc": 0.9307650372524078 + }, + { + "epoch": 2.056888888888889, + "grad_norm": 2.198077917098999, + "learning_rate": 2.455128425362153e-06, + "loss": 0.23345378041267395, + "loss_sft": 0.23345378041267395, + "step": 580, + "token_acc": 0.9257084128615475 + }, + { + "epoch": 2.0604444444444443, + "grad_norm": 2.296741247177124, + "learning_rate": 2.4383722069667683e-06, + "loss": 0.2094583362340927, + "loss_sft": 0.2094583362340927, + "step": 581, + "token_acc": 0.9330800343179311 + }, + { + "epoch": 2.064, + "grad_norm": 2.189511299133301, + "learning_rate": 2.4216549060897026e-06, + "loss": 0.221592977643013, + "loss_sft": 0.221592977643013, + "step": 582, + "token_acc": 0.927668578428643 + }, + { + "epoch": 2.0675555555555554, + "grad_norm": 2.2223782539367676, + "learning_rate": 2.404976776708477e-06, + "loss": 0.2142745554447174, + "loss_sft": 0.2142745554447174, + "step": 583, + "token_acc": 0.932991937241229 + }, + { + "epoch": 2.071111111111111, + "grad_norm": 2.088916778564453, + "learning_rate": 2.3883380722054865e-06, + "loss": 0.22165116667747498, + "loss_sft": 0.22165116667747498, + "step": 584, + "token_acc": 0.9295495792773469 + }, + { + "epoch": 2.074666666666667, + "grad_norm": 2.1181886196136475, + "learning_rate": 2.371739045364178e-06, + "loss": 0.24546606838703156, + "loss_sft": 0.24546606838703156, + "step": 585, + "token_acc": 0.9203633526526973 + }, + { + "epoch": 2.078222222222222, + "grad_norm": 2.1280834674835205, + "learning_rate": 2.3551799483651894e-06, + "loss": 0.20226508378982544, + "loss_sft": 0.20226508378982544, + "step": 586, + "token_acc": 0.9365336658354114 + }, + { + "epoch": 2.081777777777778, + "grad_norm": 2.1635935306549072, + "learning_rate": 2.3386610327825194e-06, + "loss": 0.23453623056411743, + "loss_sft": 0.23453623056411743, + "step": 587, + "token_acc": 0.9274319856140147 + }, + { + "epoch": 2.0853333333333333, + "grad_norm": 2.0424623489379883, + "learning_rate": 2.3221825495797213e-06, + "loss": 0.20790933072566986, + "loss_sft": 0.20790933072566986, + "step": 588, + "token_acc": 0.9358981776106451 + }, + { + "epoch": 2.088888888888889, + "grad_norm": 2.4770302772521973, + "learning_rate": 2.3057447491060725e-06, + "loss": 0.2242051362991333, + "loss_sft": 0.2242051362991333, + "step": 589, + "token_acc": 0.9290780141843972 + }, + { + "epoch": 2.0924444444444443, + "grad_norm": 2.1961870193481445, + "learning_rate": 2.28934788109278e-06, + "loss": 0.22397121787071228, + "loss_sft": 0.22397121787071228, + "step": 590, + "token_acc": 0.9278102233728701 + }, + { + "epoch": 2.096, + "grad_norm": 2.231003522872925, + "learning_rate": 2.2729921946491855e-06, + "loss": 0.20935127139091492, + "loss_sft": 0.20935127139091492, + "step": 591, + "token_acc": 0.9302918841955387 + }, + { + "epoch": 2.0995555555555554, + "grad_norm": 2.2265968322753906, + "learning_rate": 2.2566779382589788e-06, + "loss": 0.21950757503509521, + "loss_sft": 0.21950757503509521, + "step": 592, + "token_acc": 0.9263214381540112 + }, + { + "epoch": 2.103111111111111, + "grad_norm": 2.224095106124878, + "learning_rate": 2.240405359776424e-06, + "loss": 0.22451752424240112, + "loss_sft": 0.22451752424240112, + "step": 593, + "token_acc": 0.9290100968044135 + }, + { + "epoch": 2.1066666666666665, + "grad_norm": 2.331759214401245, + "learning_rate": 2.2241747064225942e-06, + "loss": 0.22009071707725525, + "loss_sft": 0.22009071707725525, + "step": 594, + "token_acc": 0.9288181593881075 + }, + { + "epoch": 2.110222222222222, + "grad_norm": 2.2441818714141846, + "learning_rate": 2.2079862247816148e-06, + "loss": 0.22062695026397705, + "loss_sft": 0.22062695026397705, + "step": 595, + "token_acc": 0.9281440382941688 + }, + { + "epoch": 2.113777777777778, + "grad_norm": 2.173550844192505, + "learning_rate": 2.191840160796918e-06, + "loss": 0.22160083055496216, + "loss_sft": 0.22160083055496216, + "step": 596, + "token_acc": 0.9286702536510376 + }, + { + "epoch": 2.1173333333333333, + "grad_norm": 2.1427130699157715, + "learning_rate": 2.175736759767505e-06, + "loss": 0.2196313440799713, + "loss_sft": 0.2196313440799713, + "step": 597, + "token_acc": 0.9282794125069399 + }, + { + "epoch": 2.120888888888889, + "grad_norm": 2.265471935272217, + "learning_rate": 2.159676266344222e-06, + "loss": 0.211512953042984, + "loss_sft": 0.211512953042984, + "step": 598, + "token_acc": 0.9315255006524082 + }, + { + "epoch": 2.1244444444444444, + "grad_norm": 2.2194738388061523, + "learning_rate": 2.1436589245260375e-06, + "loss": 0.21722771227359772, + "loss_sft": 0.21722771227359772, + "step": 599, + "token_acc": 0.930486699717469 + }, + { + "epoch": 2.128, + "grad_norm": 2.1614739894866943, + "learning_rate": 2.127684977656349e-06, + "loss": 0.21905425190925598, + "loss_sft": 0.21905425190925598, + "step": 600, + "token_acc": 0.9260593676059368 + }, + { + "epoch": 2.1315555555555554, + "grad_norm": 2.1098690032958984, + "learning_rate": 2.111754668419264e-06, + "loss": 0.21720340847969055, + "loss_sft": 0.21720340847969055, + "step": 601, + "token_acc": 0.9305532663040492 + }, + { + "epoch": 2.135111111111111, + "grad_norm": 2.210557460784912, + "learning_rate": 2.095868238835932e-06, + "loss": 0.2467886507511139, + "loss_sft": 0.2467886507511139, + "step": 602, + "token_acc": 0.9191097795848492 + }, + { + "epoch": 2.1386666666666665, + "grad_norm": 1.9688745737075806, + "learning_rate": 2.0800259302608656e-06, + "loss": 0.23243823647499084, + "loss_sft": 0.23243823647499084, + "step": 603, + "token_acc": 0.9258482653450247 + }, + { + "epoch": 2.1422222222222222, + "grad_norm": 2.1626553535461426, + "learning_rate": 2.064227983378256e-06, + "loss": 0.2096322476863861, + "loss_sft": 0.2096322476863861, + "step": 604, + "token_acc": 0.9330251615275517 + }, + { + "epoch": 2.145777777777778, + "grad_norm": 2.1868526935577393, + "learning_rate": 2.048474638198347e-06, + "loss": 0.20925773680210114, + "loss_sft": 0.20925773680210114, + "step": 605, + "token_acc": 0.9339817430993262 + }, + { + "epoch": 2.1493333333333333, + "grad_norm": 2.2513277530670166, + "learning_rate": 2.0327661340537536e-06, + "loss": 0.22209998965263367, + "loss_sft": 0.22209998965263367, + "step": 606, + "token_acc": 0.9272773666751722 + }, + { + "epoch": 2.152888888888889, + "grad_norm": 1.9988017082214355, + "learning_rate": 2.0171027095958543e-06, + "loss": 0.22582130134105682, + "loss_sft": 0.22582130134105682, + "step": 607, + "token_acc": 0.9288865189781376 + }, + { + "epoch": 2.1564444444444444, + "grad_norm": 2.0918548107147217, + "learning_rate": 2.001484602791157e-06, + "loss": 0.22065067291259766, + "loss_sft": 0.22065067291259766, + "step": 608, + "token_acc": 0.9292802016065522 + }, + { + "epoch": 2.16, + "grad_norm": 2.161473512649536, + "learning_rate": 1.9859120509176706e-06, + "loss": 0.21554358303546906, + "loss_sft": 0.21554358303546906, + "step": 609, + "token_acc": 0.93250531132538 + }, + { + "epoch": 2.1635555555555555, + "grad_norm": 2.1544253826141357, + "learning_rate": 1.970385290561317e-06, + "loss": 0.24688959121704102, + "loss_sft": 0.24688959121704102, + "step": 610, + "token_acc": 0.9194945121337227 + }, + { + "epoch": 2.167111111111111, + "grad_norm": 2.0566585063934326, + "learning_rate": 1.954904557612337e-06, + "loss": 0.21042990684509277, + "loss_sft": 0.21042990684509277, + "step": 611, + "token_acc": 0.9326913481688455 + }, + { + "epoch": 2.1706666666666665, + "grad_norm": 2.3437671661376953, + "learning_rate": 1.9394700872616856e-06, + "loss": 0.2321750521659851, + "loss_sft": 0.2321750521659851, + "step": 612, + "token_acc": 0.9262762145332785 + }, + { + "epoch": 2.1742222222222223, + "grad_norm": 2.3384170532226562, + "learning_rate": 1.924082113997488e-06, + "loss": 0.2124330997467041, + "loss_sft": 0.2124330997467041, + "step": 613, + "token_acc": 0.93165876144031 + }, + { + "epoch": 2.1777777777777776, + "grad_norm": 2.3116421699523926, + "learning_rate": 1.9087408716014562e-06, + "loss": 0.23787975311279297, + "loss_sft": 0.23787975311279297, + "step": 614, + "token_acc": 0.9253708518122037 + }, + { + "epoch": 2.1813333333333333, + "grad_norm": 2.1084234714508057, + "learning_rate": 1.8934465931453378e-06, + "loss": 0.23596671223640442, + "loss_sft": 0.23596671223640442, + "step": 615, + "token_acc": 0.9248116988944193 + }, + { + "epoch": 2.1848888888888887, + "grad_norm": 2.2205004692077637, + "learning_rate": 1.8781995109873929e-06, + "loss": 0.22982750833034515, + "loss_sft": 0.22982750833034515, + "step": 616, + "token_acc": 0.9263758705555847 + }, + { + "epoch": 2.1884444444444444, + "grad_norm": 2.133638858795166, + "learning_rate": 1.8629998567688445e-06, + "loss": 0.19819015264511108, + "loss_sft": 0.19819015264511108, + "step": 617, + "token_acc": 0.9349508349333798 + }, + { + "epoch": 2.192, + "grad_norm": 2.0928409099578857, + "learning_rate": 1.8478478614103684e-06, + "loss": 0.220377117395401, + "loss_sft": 0.220377117395401, + "step": 618, + "token_acc": 0.9277024863941948 + }, + { + "epoch": 2.1955555555555555, + "grad_norm": 2.0886261463165283, + "learning_rate": 1.8327437551085842e-06, + "loss": 0.20606017112731934, + "loss_sft": 0.20606017112731934, + "step": 619, + "token_acc": 0.9327218775289992 + }, + { + "epoch": 2.1991111111111112, + "grad_norm": 2.1370956897735596, + "learning_rate": 1.8176877673325555e-06, + "loss": 0.19370125234127045, + "loss_sft": 0.19370125234127045, + "step": 620, + "token_acc": 0.9389757743023612 + }, + { + "epoch": 2.2026666666666666, + "grad_norm": 2.318445920944214, + "learning_rate": 1.802680126820307e-06, + "loss": 0.23588991165161133, + "loss_sft": 0.23588991165161133, + "step": 621, + "token_acc": 0.9237868586022754 + }, + { + "epoch": 2.2062222222222223, + "grad_norm": 2.203504800796509, + "learning_rate": 1.7877210615753477e-06, + "loss": 0.2063508927822113, + "loss_sft": 0.2063508927822113, + "step": 622, + "token_acc": 0.9343715239154616 + }, + { + "epoch": 2.2097777777777776, + "grad_norm": 2.117696762084961, + "learning_rate": 1.772810798863206e-06, + "loss": 0.20796969532966614, + "loss_sft": 0.20796969532966614, + "step": 623, + "token_acc": 0.9322590271560728 + }, + { + "epoch": 2.2133333333333334, + "grad_norm": 2.1981360912323, + "learning_rate": 1.7579495652079786e-06, + "loss": 0.21159598231315613, + "loss_sft": 0.21159598231315613, + "step": 624, + "token_acc": 0.9320151206746147 + }, + { + "epoch": 2.2168888888888887, + "grad_norm": 2.126840353012085, + "learning_rate": 1.74313758638889e-06, + "loss": 0.21909956634044647, + "loss_sft": 0.21909956634044647, + "step": 625, + "token_acc": 0.9299890035084044 + }, + { + "epoch": 2.2204444444444444, + "grad_norm": 2.1870124340057373, + "learning_rate": 1.7283750874368577e-06, + "loss": 0.18772079050540924, + "loss_sft": 0.18772079050540924, + "step": 626, + "token_acc": 0.9405240627837592 + }, + { + "epoch": 2.224, + "grad_norm": 2.466320037841797, + "learning_rate": 1.713662292631078e-06, + "loss": 0.2293129712343216, + "loss_sft": 0.2293129712343216, + "step": 627, + "token_acc": 0.9263533107781537 + }, + { + "epoch": 2.2275555555555555, + "grad_norm": 2.2439839839935303, + "learning_rate": 1.6989994254956222e-06, + "loss": 0.2336433231830597, + "loss_sft": 0.2336433231830597, + "step": 628, + "token_acc": 0.9249287298067785 + }, + { + "epoch": 2.2311111111111113, + "grad_norm": 2.227134943008423, + "learning_rate": 1.6843867087960252e-06, + "loss": 0.220095694065094, + "loss_sft": 0.220095694065094, + "step": 629, + "token_acc": 0.9290809257453687 + }, + { + "epoch": 2.2346666666666666, + "grad_norm": 2.006974935531616, + "learning_rate": 1.669824364535918e-06, + "loss": 0.20306402444839478, + "loss_sft": 0.20306402444839478, + "step": 630, + "token_acc": 0.9359589743589743 + }, + { + "epoch": 2.2382222222222223, + "grad_norm": 2.182478904724121, + "learning_rate": 1.6553126139536534e-06, + "loss": 0.22814255952835083, + "loss_sft": 0.22814255952835083, + "step": 631, + "token_acc": 0.9277772162134843 + }, + { + "epoch": 2.2417777777777776, + "grad_norm": 2.1115550994873047, + "learning_rate": 1.640851677518927e-06, + "loss": 0.21996191143989563, + "loss_sft": 0.21996191143989563, + "step": 632, + "token_acc": 0.9317068619676261 + }, + { + "epoch": 2.2453333333333334, + "grad_norm": 2.108252763748169, + "learning_rate": 1.6264417749294543e-06, + "loss": 0.20758897066116333, + "loss_sft": 0.20758897066116333, + "step": 633, + "token_acc": 0.9319148936170213 + }, + { + "epoch": 2.2488888888888887, + "grad_norm": 2.140549659729004, + "learning_rate": 1.6120831251076118e-06, + "loss": 0.23084986209869385, + "loss_sft": 0.23084986209869385, + "step": 634, + "token_acc": 0.9250885601166909 + }, + { + "epoch": 2.2524444444444445, + "grad_norm": 2.100417375564575, + "learning_rate": 1.597775946197117e-06, + "loss": 0.19340035319328308, + "loss_sft": 0.19340035319328308, + "step": 635, + "token_acc": 0.9373923341946597 + }, + { + "epoch": 2.2560000000000002, + "grad_norm": 2.188093662261963, + "learning_rate": 1.5835204555597227e-06, + "loss": 0.21983733773231506, + "loss_sft": 0.21983733773231506, + "step": 636, + "token_acc": 0.9298552568792747 + }, + { + "epoch": 2.2595555555555555, + "grad_norm": 2.3939616680145264, + "learning_rate": 1.5693168697719024e-06, + "loss": 0.25748205184936523, + "loss_sft": 0.25748205184936523, + "step": 637, + "token_acc": 0.9202338692837753 + }, + { + "epoch": 2.2631111111111113, + "grad_norm": 2.0136702060699463, + "learning_rate": 1.555165404621567e-06, + "loss": 0.20887431502342224, + "loss_sft": 0.20887431502342224, + "step": 638, + "token_acc": 0.9319585523965086 + }, + { + "epoch": 2.2666666666666666, + "grad_norm": 2.35030460357666, + "learning_rate": 1.5410662751047855e-06, + "loss": 0.2057146430015564, + "loss_sft": 0.2057146430015564, + "step": 639, + "token_acc": 0.9342605259157927 + }, + { + "epoch": 2.2702222222222224, + "grad_norm": 2.2834086418151855, + "learning_rate": 1.5270196954225175e-06, + "loss": 0.2352995127439499, + "loss_sft": 0.2352995127439499, + "step": 640, + "token_acc": 0.9233021077283372 + }, + { + "epoch": 2.2737777777777777, + "grad_norm": 2.142120122909546, + "learning_rate": 1.5130258789773583e-06, + "loss": 0.23747482895851135, + "loss_sft": 0.23747482895851135, + "step": 641, + "token_acc": 0.922663153850376 + }, + { + "epoch": 2.2773333333333334, + "grad_norm": 2.11653733253479, + "learning_rate": 1.4990850383703005e-06, + "loss": 0.22967968881130219, + "loss_sft": 0.22967968881130219, + "step": 642, + "token_acc": 0.9277046746198351 + }, + { + "epoch": 2.2808888888888887, + "grad_norm": 2.314891815185547, + "learning_rate": 1.4851973853974987e-06, + "loss": 0.19626225531101227, + "loss_sft": 0.19626225531101227, + "step": 643, + "token_acc": 0.9342799188640973 + }, + { + "epoch": 2.2844444444444445, + "grad_norm": 2.335237979888916, + "learning_rate": 1.4713631310470571e-06, + "loss": 0.22626441717147827, + "loss_sft": 0.22626441717147827, + "step": 644, + "token_acc": 0.924857324032974 + }, + { + "epoch": 2.288, + "grad_norm": 2.1262426376342773, + "learning_rate": 1.457582485495821e-06, + "loss": 0.194329172372818, + "loss_sft": 0.194329172372818, + "step": 645, + "token_acc": 0.9373802247889367 + }, + { + "epoch": 2.2915555555555556, + "grad_norm": 2.31451416015625, + "learning_rate": 1.4438556581061819e-06, + "loss": 0.23406609892845154, + "loss_sft": 0.23406609892845154, + "step": 646, + "token_acc": 0.9260538673287634 + }, + { + "epoch": 2.295111111111111, + "grad_norm": 2.194739580154419, + "learning_rate": 1.4301828574229026e-06, + "loss": 0.21070820093154907, + "loss_sft": 0.21070820093154907, + "step": 647, + "token_acc": 0.932391744676192 + }, + { + "epoch": 2.2986666666666666, + "grad_norm": 2.0832717418670654, + "learning_rate": 1.4165642911699435e-06, + "loss": 0.2064182162284851, + "loss_sft": 0.2064182162284851, + "step": 648, + "token_acc": 0.9358140052283219 + }, + { + "epoch": 2.3022222222222224, + "grad_norm": 2.144744396209717, + "learning_rate": 1.4030001662473086e-06, + "loss": 0.2185964286327362, + "loss_sft": 0.2185964286327362, + "step": 649, + "token_acc": 0.9284243788002107 + }, + { + "epoch": 2.3057777777777777, + "grad_norm": 2.2188925743103027, + "learning_rate": 1.389490688727903e-06, + "loss": 0.22442124783992767, + "loss_sft": 0.22442124783992767, + "step": 650, + "token_acc": 0.9279408636460328 + }, + { + "epoch": 2.3093333333333335, + "grad_norm": 2.0185303688049316, + "learning_rate": 1.3760360638544012e-06, + "loss": 0.20931315422058105, + "loss_sft": 0.20931315422058105, + "step": 651, + "token_acc": 0.9338386375051367 + }, + { + "epoch": 2.3128888888888888, + "grad_norm": 2.3230834007263184, + "learning_rate": 1.3626364960361282e-06, + "loss": 0.21443793177604675, + "loss_sft": 0.21443793177604675, + "step": 652, + "token_acc": 0.9311932236680579 + }, + { + "epoch": 2.3164444444444445, + "grad_norm": 2.1585381031036377, + "learning_rate": 1.3492921888459566e-06, + "loss": 0.19514310359954834, + "loss_sft": 0.19514310359954834, + "step": 653, + "token_acc": 0.9386795094360755 + }, + { + "epoch": 2.32, + "grad_norm": 2.386367082595825, + "learning_rate": 1.3360033450172106e-06, + "loss": 0.2087172567844391, + "loss_sft": 0.2087172567844391, + "step": 654, + "token_acc": 0.9313333019356662 + }, + { + "epoch": 2.3235555555555556, + "grad_norm": 2.233969211578369, + "learning_rate": 1.3227701664405868e-06, + "loss": 0.2040831446647644, + "loss_sft": 0.2040831446647644, + "step": 655, + "token_acc": 0.9350364119754941 + }, + { + "epoch": 2.327111111111111, + "grad_norm": 2.4440672397613525, + "learning_rate": 1.3095928541610936e-06, + "loss": 0.19524219632148743, + "loss_sft": 0.19524219632148743, + "step": 656, + "token_acc": 0.9357307285509862 + }, + { + "epoch": 2.3306666666666667, + "grad_norm": 1.9441568851470947, + "learning_rate": 1.2964716083749829e-06, + "loss": 0.1906876266002655, + "loss_sft": 0.1906876266002655, + "step": 657, + "token_acc": 0.940142699803668 + }, + { + "epoch": 2.3342222222222224, + "grad_norm": 2.291825532913208, + "learning_rate": 1.2834066284267189e-06, + "loss": 0.23174762725830078, + "loss_sft": 0.23174762725830078, + "step": 658, + "token_acc": 0.9274139616235403 + }, + { + "epoch": 2.3377777777777777, + "grad_norm": 2.1343655586242676, + "learning_rate": 1.2703981128059534e-06, + "loss": 0.22186598181724548, + "loss_sft": 0.22186598181724548, + "step": 659, + "token_acc": 0.9277619663648125 + }, + { + "epoch": 2.3413333333333335, + "grad_norm": 2.2487030029296875, + "learning_rate": 1.257446259144494e-06, + "loss": 0.21604236960411072, + "loss_sft": 0.21604236960411072, + "step": 660, + "token_acc": 0.9310605180624928 + }, + { + "epoch": 2.344888888888889, + "grad_norm": 2.256002426147461, + "learning_rate": 1.2445512642133218e-06, + "loss": 0.22330042719841003, + "loss_sft": 0.22330042719841003, + "step": 661, + "token_acc": 0.9272789682734272 + }, + { + "epoch": 2.3484444444444446, + "grad_norm": 2.103328227996826, + "learning_rate": 1.2317133239195866e-06, + "loss": 0.2091825008392334, + "loss_sft": 0.2091825008392334, + "step": 662, + "token_acc": 0.9309422921711057 + }, + { + "epoch": 2.352, + "grad_norm": 2.1986818313598633, + "learning_rate": 1.2189326333036323e-06, + "loss": 0.2068125307559967, + "loss_sft": 0.2068125307559967, + "step": 663, + "token_acc": 0.9325726704699809 + }, + { + "epoch": 2.3555555555555556, + "grad_norm": 2.073876142501831, + "learning_rate": 1.2062093865360458e-06, + "loss": 0.20707294344902039, + "loss_sft": 0.20707294344902039, + "step": 664, + "token_acc": 0.93637175163658 + }, + { + "epoch": 2.359111111111111, + "grad_norm": 2.1793975830078125, + "learning_rate": 1.193543776914693e-06, + "loss": 0.20341120660305023, + "loss_sft": 0.20341120660305023, + "step": 665, + "token_acc": 0.9338571712751144 + }, + { + "epoch": 2.3626666666666667, + "grad_norm": 2.0330312252044678, + "learning_rate": 1.1809359968617894e-06, + "loss": 0.21796318888664246, + "loss_sft": 0.21796318888664246, + "step": 666, + "token_acc": 0.9299340964273326 + }, + { + "epoch": 2.3662222222222224, + "grad_norm": 2.131748676300049, + "learning_rate": 1.1683862379209747e-06, + "loss": 0.1980917751789093, + "loss_sft": 0.1980917751789093, + "step": 667, + "token_acc": 0.9360713720107692 + }, + { + "epoch": 2.3697777777777778, + "grad_norm": 2.269514560699463, + "learning_rate": 1.1558946907544034e-06, + "loss": 0.23017074167728424, + "loss_sft": 0.23017074167728424, + "step": 668, + "token_acc": 0.9247300875942147 + }, + { + "epoch": 2.3733333333333335, + "grad_norm": 2.316723346710205, + "learning_rate": 1.1434615451398467e-06, + "loss": 0.21987459063529968, + "loss_sft": 0.21987459063529968, + "step": 669, + "token_acc": 0.927415908674684 + }, + { + "epoch": 2.376888888888889, + "grad_norm": 2.1507935523986816, + "learning_rate": 1.1310869899678122e-06, + "loss": 0.20558345317840576, + "loss_sft": 0.20558345317840576, + "step": 670, + "token_acc": 0.9330247753879662 + }, + { + "epoch": 2.3804444444444446, + "grad_norm": 2.180081844329834, + "learning_rate": 1.1187712132386723e-06, + "loss": 0.22541581094264984, + "loss_sft": 0.22541581094264984, + "step": 671, + "token_acc": 0.9270610348248316 + }, + { + "epoch": 2.384, + "grad_norm": 2.0593152046203613, + "learning_rate": 1.1065144020598067e-06, + "loss": 0.21945737302303314, + "loss_sft": 0.21945737302303314, + "step": 672, + "token_acc": 0.9291660203712321 + }, + { + "epoch": 2.3875555555555557, + "grad_norm": 2.048077344894409, + "learning_rate": 1.0943167426427625e-06, + "loss": 0.200881227850914, + "loss_sft": 0.200881227850914, + "step": 673, + "token_acc": 0.9360871768879878 + }, + { + "epoch": 2.391111111111111, + "grad_norm": 2.1359784603118896, + "learning_rate": 1.0821784203004238e-06, + "loss": 0.20189784467220306, + "loss_sft": 0.20189784467220306, + "step": 674, + "token_acc": 0.9340088058988913 + }, + { + "epoch": 2.3946666666666667, + "grad_norm": 2.249748468399048, + "learning_rate": 1.0700996194441944e-06, + "loss": 0.2084050178527832, + "loss_sft": 0.2084050178527832, + "step": 675, + "token_acc": 0.9319084888621313 + }, + { + "epoch": 2.398222222222222, + "grad_norm": 2.2152130603790283, + "learning_rate": 1.0580805235812042e-06, + "loss": 0.20847636461257935, + "loss_sft": 0.20847636461257935, + "step": 676, + "token_acc": 0.9326282804543674 + }, + { + "epoch": 2.401777777777778, + "grad_norm": 2.0967862606048584, + "learning_rate": 1.046121315311508e-06, + "loss": 0.20526213943958282, + "loss_sft": 0.20526213943958282, + "step": 677, + "token_acc": 0.9339563862928348 + }, + { + "epoch": 2.405333333333333, + "grad_norm": 2.124208688735962, + "learning_rate": 1.0342221763253207e-06, + "loss": 0.22753050923347473, + "loss_sft": 0.22753050923347473, + "step": 678, + "token_acc": 0.9296912803638827 + }, + { + "epoch": 2.408888888888889, + "grad_norm": 2.2814064025878906, + "learning_rate": 1.0223832874002603e-06, + "loss": 0.21996313333511353, + "loss_sft": 0.21996313333511353, + "step": 679, + "token_acc": 0.929431253451132 + }, + { + "epoch": 2.4124444444444446, + "grad_norm": 2.370649576187134, + "learning_rate": 1.010604828398587e-06, + "loss": 0.19924776256084442, + "loss_sft": 0.19924776256084442, + "step": 680, + "token_acc": 0.9357241672874321 + }, + { + "epoch": 2.416, + "grad_norm": 2.19769287109375, + "learning_rate": 9.988869782644872e-07, + "loss": 0.19810141623020172, + "loss_sft": 0.19810141623020172, + "step": 681, + "token_acc": 0.9353627541589649 + }, + { + "epoch": 2.4195555555555557, + "grad_norm": 2.2505204677581787, + "learning_rate": 9.872299150213455e-07, + "loss": 0.22599104046821594, + "loss_sft": 0.22599104046821594, + "step": 682, + "token_acc": 0.9286800651819663 + }, + { + "epoch": 2.423111111111111, + "grad_norm": 2.187098741531372, + "learning_rate": 9.75633815769036e-07, + "loss": 0.22643718123435974, + "loss_sft": 0.22643718123435974, + "step": 683, + "token_acc": 0.9291113963771217 + }, + { + "epoch": 2.4266666666666667, + "grad_norm": 2.0827479362487793, + "learning_rate": 9.640988566812475e-07, + "loss": 0.21825626492500305, + "loss_sft": 0.21825626492500305, + "step": 684, + "token_acc": 0.9295883229685974 + }, + { + "epoch": 2.430222222222222, + "grad_norm": 2.102893114089966, + "learning_rate": 9.526252130027919e-07, + "loss": 0.20474126935005188, + "loss_sft": 0.20474126935005188, + "step": 685, + "token_acc": 0.9330246581242597 + }, + { + "epoch": 2.433777777777778, + "grad_norm": 2.2421255111694336, + "learning_rate": 9.412130590469438e-07, + "loss": 0.21682539582252502, + "loss_sft": 0.21682539582252502, + "step": 686, + "token_acc": 0.9328572264824679 + }, + { + "epoch": 2.437333333333333, + "grad_norm": 2.262820243835449, + "learning_rate": 9.298625681928031e-07, + "loss": 0.21688266098499298, + "loss_sft": 0.21688266098499298, + "step": 687, + "token_acc": 0.9302923041790363 + }, + { + "epoch": 2.440888888888889, + "grad_norm": 2.0951485633850098, + "learning_rate": 9.185739128826454e-07, + "loss": 0.19251985847949982, + "loss_sft": 0.19251985847949982, + "step": 688, + "token_acc": 0.9382178217821782 + }, + { + "epoch": 2.4444444444444446, + "grad_norm": 2.2758517265319824, + "learning_rate": 9.073472646193171e-07, + "loss": 0.2285204827785492, + "loss_sft": 0.2285204827785492, + "step": 689, + "token_acc": 0.9248173207036536 + }, + { + "epoch": 2.448, + "grad_norm": 2.0433483123779297, + "learning_rate": 8.961827939636198e-07, + "loss": 0.22125282883644104, + "loss_sft": 0.22125282883644104, + "step": 690, + "token_acc": 0.9301707248810523 + }, + { + "epoch": 2.4515555555555557, + "grad_norm": 2.224929094314575, + "learning_rate": 8.850806705317183e-07, + "loss": 0.2144889533519745, + "loss_sft": 0.2144889533519745, + "step": 691, + "token_acc": 0.9317734405759605 + }, + { + "epoch": 2.455111111111111, + "grad_norm": 2.1798830032348633, + "learning_rate": 8.740410629925744e-07, + "loss": 0.24951277673244476, + "loss_sft": 0.24951277673244476, + "step": 692, + "token_acc": 0.9221635267168435 + }, + { + "epoch": 2.458666666666667, + "grad_norm": 2.1472280025482178, + "learning_rate": 8.630641390653743e-07, + "loss": 0.2345062792301178, + "loss_sft": 0.2345062792301178, + "step": 693, + "token_acc": 0.9254497354497354 + }, + { + "epoch": 2.462222222222222, + "grad_norm": 2.1737072467803955, + "learning_rate": 8.521500655169823e-07, + "loss": 0.2108057737350464, + "loss_sft": 0.2108057737350464, + "step": 694, + "token_acc": 0.9317256162915327 + }, + { + "epoch": 2.465777777777778, + "grad_norm": 2.2921435832977295, + "learning_rate": 8.412990081594102e-07, + "loss": 0.22143515944480896, + "loss_sft": 0.22143515944480896, + "step": 695, + "token_acc": 0.9278946601234148 + }, + { + "epoch": 2.469333333333333, + "grad_norm": 2.152460813522339, + "learning_rate": 8.30511131847297e-07, + "loss": 0.21150238811969757, + "loss_sft": 0.21150238811969757, + "step": 696, + "token_acc": 0.9321623599617181 + }, + { + "epoch": 2.472888888888889, + "grad_norm": 2.103050947189331, + "learning_rate": 8.197866004754029e-07, + "loss": 0.21047276258468628, + "loss_sft": 0.21047276258468628, + "step": 697, + "token_acc": 0.9325780740889268 + }, + { + "epoch": 2.4764444444444447, + "grad_norm": 2.230440139770508, + "learning_rate": 8.091255769761213e-07, + "loss": 0.21073895692825317, + "loss_sft": 0.21073895692825317, + "step": 698, + "token_acc": 0.931117471055028 + }, + { + "epoch": 2.48, + "grad_norm": 2.122370719909668, + "learning_rate": 7.985282233170011e-07, + "loss": 0.19768092036247253, + "loss_sft": 0.19768092036247253, + "step": 699, + "token_acc": 0.937193998420637 + }, + { + "epoch": 2.4835555555555557, + "grad_norm": 1.9905856847763062, + "learning_rate": 7.879947004982896e-07, + "loss": 0.20742535591125488, + "loss_sft": 0.20742535591125488, + "step": 700, + "token_acc": 0.9349551251771374 + }, + { + "epoch": 2.487111111111111, + "grad_norm": 2.1526570320129395, + "learning_rate": 7.775251685504826e-07, + "loss": 0.1919325292110443, + "loss_sft": 0.1919325292110443, + "step": 701, + "token_acc": 0.9377613911512217 + }, + { + "epoch": 2.490666666666667, + "grad_norm": 2.123676300048828, + "learning_rate": 7.671197865318952e-07, + "loss": 0.2076290249824524, + "loss_sft": 0.2076290249824524, + "step": 702, + "token_acc": 0.9339177343223196 + }, + { + "epoch": 2.494222222222222, + "grad_norm": 2.2289929389953613, + "learning_rate": 7.567787125262449e-07, + "loss": 0.21761733293533325, + "loss_sft": 0.21761733293533325, + "step": 703, + "token_acc": 0.9313719744476352 + }, + { + "epoch": 2.497777777777778, + "grad_norm": 2.44468355178833, + "learning_rate": 7.465021036402531e-07, + "loss": 0.19692447781562805, + "loss_sft": 0.19692447781562805, + "step": 704, + "token_acc": 0.936207229847284 + }, + { + "epoch": 2.501333333333333, + "grad_norm": 2.075695037841797, + "learning_rate": 7.362901160012492e-07, + "loss": 0.21066246926784515, + "loss_sft": 0.21066246926784515, + "step": 705, + "token_acc": 0.930864451178797 + }, + { + "epoch": 2.504888888888889, + "grad_norm": 2.2850937843322754, + "learning_rate": 7.261429047548085e-07, + "loss": 0.21127204596996307, + "loss_sft": 0.21127204596996307, + "step": 706, + "token_acc": 0.9324094192260262 + }, + { + "epoch": 2.5084444444444447, + "grad_norm": 2.2993197441101074, + "learning_rate": 7.16060624062393e-07, + "loss": 0.20937559008598328, + "loss_sft": 0.20937559008598328, + "step": 707, + "token_acc": 0.9327612426678253 + }, + { + "epoch": 2.512, + "grad_norm": 2.043705463409424, + "learning_rate": 7.060434270990013e-07, + "loss": 0.20873934030532837, + "loss_sft": 0.20873934030532837, + "step": 708, + "token_acc": 0.9352699866437703 + }, + { + "epoch": 2.5155555555555553, + "grad_norm": 2.557668924331665, + "learning_rate": 6.96091466050855e-07, + "loss": 0.1908925473690033, + "loss_sft": 0.1908925473690033, + "step": 709, + "token_acc": 0.9394826744753538 + }, + { + "epoch": 2.519111111111111, + "grad_norm": 2.200516700744629, + "learning_rate": 6.86204892113076e-07, + "loss": 0.1828106939792633, + "loss_sft": 0.1828106939792633, + "step": 710, + "token_acc": 0.9416185600697092 + }, + { + "epoch": 2.522666666666667, + "grad_norm": 2.3608219623565674, + "learning_rate": 6.763838554873892e-07, + "loss": 0.2285737693309784, + "loss_sft": 0.2285737693309784, + "step": 711, + "token_acc": 0.9282051282051282 + }, + { + "epoch": 2.526222222222222, + "grad_norm": 2.161430597305298, + "learning_rate": 6.66628505379851e-07, + "loss": 0.22755438089370728, + "loss_sft": 0.22755438089370728, + "step": 712, + "token_acc": 0.9240190192594016 + }, + { + "epoch": 2.529777777777778, + "grad_norm": 2.125821352005005, + "learning_rate": 6.569389899985723e-07, + "loss": 0.2217322587966919, + "loss_sft": 0.2217322587966919, + "step": 713, + "token_acc": 0.9283425440022292 + }, + { + "epoch": 2.533333333333333, + "grad_norm": 2.257399320602417, + "learning_rate": 6.473154565514695e-07, + "loss": 0.2207571566104889, + "loss_sft": 0.2207571566104889, + "step": 714, + "token_acc": 0.9280563659382397 + }, + { + "epoch": 2.536888888888889, + "grad_norm": 2.078617811203003, + "learning_rate": 6.377580512440301e-07, + "loss": 0.20790180563926697, + "loss_sft": 0.20790180563926697, + "step": 715, + "token_acc": 0.9337928153717627 + }, + { + "epoch": 2.5404444444444443, + "grad_norm": 2.1453518867492676, + "learning_rate": 6.282669192770896e-07, + "loss": 0.20475344359874725, + "loss_sft": 0.20475344359874725, + "step": 716, + "token_acc": 0.9338645822759111 + }, + { + "epoch": 2.544, + "grad_norm": 2.300065755844116, + "learning_rate": 6.188422048446263e-07, + "loss": 0.21685275435447693, + "loss_sft": 0.21685275435447693, + "step": 717, + "token_acc": 0.9297328086164043 + }, + { + "epoch": 2.5475555555555554, + "grad_norm": 2.3417282104492188, + "learning_rate": 6.094840511315703e-07, + "loss": 0.19980204105377197, + "loss_sft": 0.19980204105377197, + "step": 718, + "token_acc": 0.9341432225063938 + }, + { + "epoch": 2.551111111111111, + "grad_norm": 2.238039016723633, + "learning_rate": 6.001926003116282e-07, + "loss": 0.2108420580625534, + "loss_sft": 0.2108420580625534, + "step": 719, + "token_acc": 0.9316694162098877 + }, + { + "epoch": 2.554666666666667, + "grad_norm": 2.1367430686950684, + "learning_rate": 5.909679935451235e-07, + "loss": 0.21130424737930298, + "loss_sft": 0.21130424737930298, + "step": 720, + "token_acc": 0.9343672602901285 + }, + { + "epoch": 2.558222222222222, + "grad_norm": 2.1364989280700684, + "learning_rate": 5.818103709768519e-07, + "loss": 0.2188308835029602, + "loss_sft": 0.2188308835029602, + "step": 721, + "token_acc": 0.9303019574123716 + }, + { + "epoch": 2.561777777777778, + "grad_norm": 2.224710464477539, + "learning_rate": 5.727198717339511e-07, + "loss": 0.2173798382282257, + "loss_sft": 0.2173798382282257, + "step": 722, + "token_acc": 0.9306829765545361 + }, + { + "epoch": 2.5653333333333332, + "grad_norm": 2.1043975353240967, + "learning_rate": 5.636966339237882e-07, + "loss": 0.21026195585727692, + "loss_sft": 0.21026195585727692, + "step": 723, + "token_acc": 0.9331371753246753 + }, + { + "epoch": 2.568888888888889, + "grad_norm": 2.336123466491699, + "learning_rate": 5.547407946318628e-07, + "loss": 0.22842931747436523, + "loss_sft": 0.22842931747436523, + "step": 724, + "token_acc": 0.9263288427935291 + }, + { + "epoch": 2.5724444444444443, + "grad_norm": 2.1821067333221436, + "learning_rate": 5.458524899197204e-07, + "loss": 0.2321673035621643, + "loss_sft": 0.2321673035621643, + "step": 725, + "token_acc": 0.9270679569187031 + }, + { + "epoch": 2.576, + "grad_norm": 2.137766122817993, + "learning_rate": 5.370318548228886e-07, + "loss": 0.22247251868247986, + "loss_sft": 0.22247251868247986, + "step": 726, + "token_acc": 0.9284617746392385 + }, + { + "epoch": 2.5795555555555554, + "grad_norm": 2.4947710037231445, + "learning_rate": 5.282790233488283e-07, + "loss": 0.22641000151634216, + "loss_sft": 0.22641000151634216, + "step": 727, + "token_acc": 0.9255803102764493 + }, + { + "epoch": 2.583111111111111, + "grad_norm": 2.4761240482330322, + "learning_rate": 5.195941284748879e-07, + "loss": 0.195369154214859, + "loss_sft": 0.195369154214859, + "step": 728, + "token_acc": 0.9357425307557118 + }, + { + "epoch": 2.586666666666667, + "grad_norm": 2.3867130279541016, + "learning_rate": 5.109773021462921e-07, + "loss": 0.21387247741222382, + "loss_sft": 0.21387247741222382, + "step": 729, + "token_acc": 0.929379934846226 + }, + { + "epoch": 2.590222222222222, + "grad_norm": 2.197998523712158, + "learning_rate": 5.024286752741364e-07, + "loss": 0.24190066754817963, + "loss_sft": 0.24190066754817963, + "step": 730, + "token_acc": 0.9215247830485013 + }, + { + "epoch": 2.5937777777777775, + "grad_norm": 2.291250467300415, + "learning_rate": 4.939483777333931e-07, + "loss": 0.1971968710422516, + "loss_sft": 0.1971968710422516, + "step": 731, + "token_acc": 0.9359804990698569 + }, + { + "epoch": 2.5973333333333333, + "grad_norm": 2.0206751823425293, + "learning_rate": 4.855365383609457e-07, + "loss": 0.2089833915233612, + "loss_sft": 0.2089833915233612, + "step": 732, + "token_acc": 0.9330547951784085 + }, + { + "epoch": 2.600888888888889, + "grad_norm": 2.0774805545806885, + "learning_rate": 4.77193284953622e-07, + "loss": 0.19828307628631592, + "loss_sft": 0.19828307628631592, + "step": 733, + "token_acc": 0.9343999143881427 + }, + { + "epoch": 2.6044444444444443, + "grad_norm": 2.192539930343628, + "learning_rate": 4.6891874426626125e-07, + "loss": 0.22322461009025574, + "loss_sft": 0.22322461009025574, + "step": 734, + "token_acc": 0.9286832220367279 + }, + { + "epoch": 2.608, + "grad_norm": 2.224142074584961, + "learning_rate": 4.607130420097866e-07, + "loss": 0.21452000737190247, + "loss_sft": 0.21452000737190247, + "step": 735, + "token_acc": 0.9309251500396332 + }, + { + "epoch": 2.6115555555555554, + "grad_norm": 2.1087353229522705, + "learning_rate": 4.525763028492869e-07, + "loss": 0.23164571821689606, + "loss_sft": 0.23164571821689606, + "step": 736, + "token_acc": 0.9273714402786314 + }, + { + "epoch": 2.615111111111111, + "grad_norm": 2.0967001914978027, + "learning_rate": 4.445086504021384e-07, + "loss": 0.1973828375339508, + "loss_sft": 0.1973828375339508, + "step": 737, + "token_acc": 0.9360559440559441 + }, + { + "epoch": 2.618666666666667, + "grad_norm": 2.2060184478759766, + "learning_rate": 4.365102072361116e-07, + "loss": 0.2151787132024765, + "loss_sft": 0.2151787132024765, + "step": 738, + "token_acc": 0.9307169245596187 + }, + { + "epoch": 2.6222222222222222, + "grad_norm": 2.3538944721221924, + "learning_rate": 4.285810948675156e-07, + "loss": 0.20537155866622925, + "loss_sft": 0.20537155866622925, + "step": 739, + "token_acc": 0.9342716857610475 + }, + { + "epoch": 2.6257777777777775, + "grad_norm": 2.3916561603546143, + "learning_rate": 4.207214337593557e-07, + "loss": 0.2062980830669403, + "loss_sft": 0.2062980830669403, + "step": 740, + "token_acc": 0.935564371640645 + }, + { + "epoch": 2.6293333333333333, + "grad_norm": 2.235715866088867, + "learning_rate": 4.129313433194948e-07, + "loss": 0.22878654301166534, + "loss_sft": 0.22878654301166534, + "step": 741, + "token_acc": 0.927353595255745 + }, + { + "epoch": 2.632888888888889, + "grad_norm": 2.1887261867523193, + "learning_rate": 4.05210941898847e-07, + "loss": 0.21094763278961182, + "loss_sft": 0.21094763278961182, + "step": 742, + "token_acc": 0.933275663206459 + }, + { + "epoch": 2.6364444444444444, + "grad_norm": 2.223055362701416, + "learning_rate": 3.975603467895739e-07, + "loss": 0.20947298407554626, + "loss_sft": 0.20947298407554626, + "step": 743, + "token_acc": 0.9332679663167609 + }, + { + "epoch": 2.64, + "grad_norm": 2.1617233753204346, + "learning_rate": 3.8997967422330693e-07, + "loss": 0.2059226781129837, + "loss_sft": 0.2059226781129837, + "step": 744, + "token_acc": 0.9345860843067547 + }, + { + "epoch": 2.6435555555555554, + "grad_norm": 2.176805019378662, + "learning_rate": 3.8246903936937806e-07, + "loss": 0.20904037356376648, + "loss_sft": 0.20904037356376648, + "step": 745, + "token_acc": 0.9319428073591964 + }, + { + "epoch": 2.647111111111111, + "grad_norm": 2.3655245304107666, + "learning_rate": 3.7502855633307246e-07, + "loss": 0.20194074511528015, + "loss_sft": 0.20194074511528015, + "step": 746, + "token_acc": 0.9335845200842966 + }, + { + "epoch": 2.6506666666666665, + "grad_norm": 2.252324342727661, + "learning_rate": 3.676583381538945e-07, + "loss": 0.23102250695228577, + "loss_sft": 0.23102250695228577, + "step": 747, + "token_acc": 0.9256612952265126 + }, + { + "epoch": 2.6542222222222223, + "grad_norm": 2.3526337146759033, + "learning_rate": 3.603584968038487e-07, + "loss": 0.23694927990436554, + "loss_sft": 0.23694927990436554, + "step": 748, + "token_acc": 0.9282321091076263 + }, + { + "epoch": 2.6577777777777776, + "grad_norm": 2.2598018646240234, + "learning_rate": 3.53129143185742e-07, + "loss": 0.2143687903881073, + "loss_sft": 0.2143687903881073, + "step": 749, + "token_acc": 0.9308655719454495 + }, + { + "epoch": 2.6613333333333333, + "grad_norm": 2.5423829555511475, + "learning_rate": 3.4597038713149455e-07, + "loss": 0.22095084190368652, + "loss_sft": 0.22095084190368652, + "step": 750, + "token_acc": 0.9309197840339375 + }, + { + "epoch": 2.664888888888889, + "grad_norm": 2.2223427295684814, + "learning_rate": 3.388823374004735e-07, + "loss": 0.19406753778457642, + "loss_sft": 0.19406753778457642, + "step": 751, + "token_acc": 0.9363719442540553 + }, + { + "epoch": 2.6684444444444444, + "grad_norm": 2.3748669624328613, + "learning_rate": 3.3186510167784456e-07, + "loss": 0.2002527415752411, + "loss_sft": 0.2002527415752411, + "step": 752, + "token_acc": 0.9347783508519946 + }, + { + "epoch": 2.672, + "grad_norm": 2.15315842628479, + "learning_rate": 3.2491878657292643e-07, + "loss": 0.21869969367980957, + "loss_sft": 0.21869969367980957, + "step": 753, + "token_acc": 0.9320373942721472 + }, + { + "epoch": 2.6755555555555555, + "grad_norm": 2.210447311401367, + "learning_rate": 3.180434976175784e-07, + "loss": 0.22816941142082214, + "loss_sft": 0.22816941142082214, + "step": 754, + "token_acc": 0.9251303441084463 + }, + { + "epoch": 2.679111111111111, + "grad_norm": 2.260847568511963, + "learning_rate": 3.112393392645985e-07, + "loss": 0.22029274702072144, + "loss_sft": 0.22029274702072144, + "step": 755, + "token_acc": 0.9313919348073582 + }, + { + "epoch": 2.6826666666666665, + "grad_norm": 2.362165927886963, + "learning_rate": 3.045064148861282e-07, + "loss": 0.22507572174072266, + "loss_sft": 0.22507572174072266, + "step": 756, + "token_acc": 0.9284106806155343 + }, + { + "epoch": 2.6862222222222223, + "grad_norm": 2.1725196838378906, + "learning_rate": 2.9784482677209013e-07, + "loss": 0.20984846353530884, + "loss_sft": 0.20984846353530884, + "step": 757, + "token_acc": 0.9317151240330754 + }, + { + "epoch": 2.6897777777777776, + "grad_norm": 2.172759532928467, + "learning_rate": 2.912546761286333e-07, + "loss": 0.20438626408576965, + "loss_sft": 0.20438626408576965, + "step": 758, + "token_acc": 0.9336910477266309 + }, + { + "epoch": 2.6933333333333334, + "grad_norm": 2.3090860843658447, + "learning_rate": 2.847360630765866e-07, + "loss": 0.21061569452285767, + "loss_sft": 0.21061569452285767, + "step": 759, + "token_acc": 0.9341996942255675 + }, + { + "epoch": 2.696888888888889, + "grad_norm": 2.0886290073394775, + "learning_rate": 2.7828908664995216e-07, + "loss": 0.20906978845596313, + "loss_sft": 0.20906978845596313, + "step": 760, + "token_acc": 0.9316569954867827 + }, + { + "epoch": 2.7004444444444444, + "grad_norm": 2.121429204940796, + "learning_rate": 2.719138447943881e-07, + "loss": 0.22756122052669525, + "loss_sft": 0.22756122052669525, + "step": 761, + "token_acc": 0.927072082648509 + }, + { + "epoch": 2.7039999999999997, + "grad_norm": 2.2823638916015625, + "learning_rate": 2.656104343657251e-07, + "loss": 0.2025155872106552, + "loss_sft": 0.2025155872106552, + "step": 762, + "token_acc": 0.9327565632458233 + }, + { + "epoch": 2.7075555555555555, + "grad_norm": 2.312992572784424, + "learning_rate": 2.5937895112849886e-07, + "loss": 0.20305731892585754, + "loss_sft": 0.20305731892585754, + "step": 763, + "token_acc": 0.9356452926800884 + }, + { + "epoch": 2.7111111111111112, + "grad_norm": 2.355311393737793, + "learning_rate": 2.532194897544882e-07, + "loss": 0.21853376924991608, + "loss_sft": 0.21853376924991608, + "step": 764, + "token_acc": 0.9290450928381963 + }, + { + "epoch": 2.7146666666666666, + "grad_norm": 2.2817697525024414, + "learning_rate": 2.4713214382128204e-07, + "loss": 0.2163524031639099, + "loss_sft": 0.2163524031639099, + "step": 765, + "token_acc": 0.931403714237912 + }, + { + "epoch": 2.7182222222222223, + "grad_norm": 2.2208986282348633, + "learning_rate": 2.411170058108558e-07, + "loss": 0.226409912109375, + "loss_sft": 0.226409912109375, + "step": 766, + "token_acc": 0.9272852088503987 + }, + { + "epoch": 2.7217777777777776, + "grad_norm": 2.3535399436950684, + "learning_rate": 2.351741671081631e-07, + "loss": 0.2059878706932068, + "loss_sft": 0.2059878706932068, + "step": 767, + "token_acc": 0.9346518105849582 + }, + { + "epoch": 2.7253333333333334, + "grad_norm": 2.386939287185669, + "learning_rate": 2.2930371799975593e-07, + "loss": 0.23683971166610718, + "loss_sft": 0.23683971166610718, + "step": 768, + "token_acc": 0.9236963403121008 + }, + { + "epoch": 2.728888888888889, + "grad_norm": 2.138763189315796, + "learning_rate": 2.2350574767240395e-07, + "loss": 0.20213018357753754, + "loss_sft": 0.20213018357753754, + "step": 769, + "token_acc": 0.9366425407799503 + }, + { + "epoch": 2.7324444444444445, + "grad_norm": 2.208327054977417, + "learning_rate": 2.1778034421174433e-07, + "loss": 0.22335566580295563, + "loss_sft": 0.22335566580295563, + "step": 770, + "token_acc": 0.9293032786885246 + }, + { + "epoch": 2.7359999999999998, + "grad_norm": 2.1569812297821045, + "learning_rate": 2.1212759460094268e-07, + "loss": 0.2326255440711975, + "loss_sft": 0.2326255440711975, + "step": 771, + "token_acc": 0.9258406485345613 + }, + { + "epoch": 2.7395555555555555, + "grad_norm": 2.208338737487793, + "learning_rate": 2.0654758471937098e-07, + "loss": 0.20663884282112122, + "loss_sft": 0.20663884282112122, + "step": 772, + "token_acc": 0.9352021803214862 + }, + { + "epoch": 2.7431111111111113, + "grad_norm": 2.318976402282715, + "learning_rate": 2.0104039934130348e-07, + "loss": 0.20255661010742188, + "loss_sft": 0.20255661010742188, + "step": 773, + "token_acc": 0.9320805976907404 + }, + { + "epoch": 2.7466666666666666, + "grad_norm": 1.997756004333496, + "learning_rate": 1.9560612213462837e-07, + "loss": 0.19720430672168732, + "loss_sft": 0.19720430672168732, + "step": 774, + "token_acc": 0.9358122048507633 + }, + { + "epoch": 2.7502222222222223, + "grad_norm": 2.2332332134246826, + "learning_rate": 1.9024483565957707e-07, + "loss": 0.22286754846572876, + "loss_sft": 0.22286754846572876, + "step": 775, + "token_acc": 0.9301641199773628 + }, + { + "epoch": 2.7537777777777777, + "grad_norm": 2.276001453399658, + "learning_rate": 1.8495662136746916e-07, + "loss": 0.21966683864593506, + "loss_sft": 0.21966683864593506, + "step": 776, + "token_acc": 0.930837639706766 + }, + { + "epoch": 2.7573333333333334, + "grad_norm": 2.1044652462005615, + "learning_rate": 1.7974155959947614e-07, + "loss": 0.21146978437900543, + "loss_sft": 0.21146978437900543, + "step": 777, + "token_acc": 0.9335238381711536 + }, + { + "epoch": 2.7608888888888887, + "grad_norm": 2.2231340408325195, + "learning_rate": 1.745997295853985e-07, + "loss": 0.19222068786621094, + "loss_sft": 0.19222068786621094, + "step": 778, + "token_acc": 0.9357403311614562 + }, + { + "epoch": 2.7644444444444445, + "grad_norm": 2.304358720779419, + "learning_rate": 1.6953120944246503e-07, + "loss": 0.22022399306297302, + "loss_sft": 0.22022399306297302, + "step": 779, + "token_acc": 0.9308119125806817 + }, + { + "epoch": 2.768, + "grad_norm": 2.1242711544036865, + "learning_rate": 1.6453607617414603e-07, + "loss": 0.18581785261631012, + "loss_sft": 0.18581785261631012, + "step": 780, + "token_acc": 0.939911061700945 + }, + { + "epoch": 2.7715555555555556, + "grad_norm": 2.068387985229492, + "learning_rate": 1.5961440566897913e-07, + "loss": 0.19189482927322388, + "loss_sft": 0.19189482927322388, + "step": 781, + "token_acc": 0.9358648448257599 + }, + { + "epoch": 2.7751111111111113, + "grad_norm": 2.267868995666504, + "learning_rate": 1.5476627269941925e-07, + "loss": 0.20797067880630493, + "loss_sft": 0.20797067880630493, + "step": 782, + "token_acc": 0.9308566097788509 + }, + { + "epoch": 2.7786666666666666, + "grad_norm": 2.229050874710083, + "learning_rate": 1.4999175092070716e-07, + "loss": 0.22151583433151245, + "loss_sft": 0.22151583433151245, + "step": 783, + "token_acc": 0.9292866436612547 + }, + { + "epoch": 2.7822222222222224, + "grad_norm": 2.1798455715179443, + "learning_rate": 1.4529091286973994e-07, + "loss": 0.225885808467865, + "loss_sft": 0.225885808467865, + "step": 784, + "token_acc": 0.9293732970027248 + }, + { + "epoch": 2.7857777777777777, + "grad_norm": 2.1661946773529053, + "learning_rate": 1.4066382996397898e-07, + "loss": 0.23626591265201569, + "loss_sft": 0.23626591265201569, + "step": 785, + "token_acc": 0.9254885119175434 + }, + { + "epoch": 2.7893333333333334, + "grad_norm": 2.110128402709961, + "learning_rate": 1.361105725003603e-07, + "loss": 0.22975578904151917, + "loss_sft": 0.22975578904151917, + "step": 786, + "token_acc": 0.9269548324429334 + }, + { + "epoch": 2.7928888888888888, + "grad_norm": 2.185422420501709, + "learning_rate": 1.3163120965422494e-07, + "loss": 0.21508333086967468, + "loss_sft": 0.21508333086967468, + "step": 787, + "token_acc": 0.9323255228017133 + }, + { + "epoch": 2.7964444444444445, + "grad_norm": 2.1552278995513916, + "learning_rate": 1.272258094782741e-07, + "loss": 0.21280154585838318, + "loss_sft": 0.21280154585838318, + "step": 788, + "token_acc": 0.9312762280918292 + }, + { + "epoch": 2.8, + "grad_norm": 2.1382768154144287, + "learning_rate": 1.2289443890152852e-07, + "loss": 0.21895045042037964, + "loss_sft": 0.21895045042037964, + "step": 789, + "token_acc": 0.9317877842175658 + }, + { + "epoch": 2.8035555555555556, + "grad_norm": 2.2897143363952637, + "learning_rate": 1.1863716372831513e-07, + "loss": 0.19704243540763855, + "loss_sft": 0.19704243540763855, + "step": 790, + "token_acc": 0.9369601176254365 + }, + { + "epoch": 2.8071111111111113, + "grad_norm": 2.0622072219848633, + "learning_rate": 1.1445404863726806e-07, + "loss": 0.21027334034442902, + "loss_sft": 0.21027334034442902, + "step": 791, + "token_acc": 0.9313714960799054 + }, + { + "epoch": 2.8106666666666666, + "grad_norm": 2.389130115509033, + "learning_rate": 1.1034515718034321e-07, + "loss": 0.26240217685699463, + "loss_sft": 0.26240217685699463, + "step": 792, + "token_acc": 0.9195796810645264 + }, + { + "epoch": 2.814222222222222, + "grad_norm": 2.1406102180480957, + "learning_rate": 1.0631055178185523e-07, + "loss": 0.18593335151672363, + "loss_sft": 0.18593335151672363, + "step": 793, + "token_acc": 0.9398170438131921 + }, + { + "epoch": 2.8177777777777777, + "grad_norm": 2.208811044692993, + "learning_rate": 1.0235029373752758e-07, + "loss": 0.21565991640090942, + "loss_sft": 0.21565991640090942, + "step": 794, + "token_acc": 0.9292321032854803 + }, + { + "epoch": 2.8213333333333335, + "grad_norm": 2.264435052871704, + "learning_rate": 9.84644432135623e-08, + "loss": 0.22617554664611816, + "loss_sft": 0.22617554664611816, + "step": 795, + "token_acc": 0.9273122838335379 + }, + { + "epoch": 2.824888888888889, + "grad_norm": 2.219986915588379, + "learning_rate": 9.465305924572566e-08, + "loss": 0.1961258053779602, + "loss_sft": 0.1961258053779602, + "step": 796, + "token_acc": 0.9366288684148358 + }, + { + "epoch": 2.8284444444444445, + "grad_norm": 2.1045148372650146, + "learning_rate": 9.091619973845056e-08, + "loss": 0.22305172681808472, + "loss_sft": 0.22305172681808472, + "step": 797, + "token_acc": 0.9301952929394092 + }, + { + "epoch": 2.832, + "grad_norm": 2.2457568645477295, + "learning_rate": 8.725392146395728e-08, + "loss": 0.21335893869400024, + "loss_sft": 0.21335893869400024, + "step": 798, + "token_acc": 0.9316475265017667 + }, + { + "epoch": 2.8355555555555556, + "grad_norm": 2.234353542327881, + "learning_rate": 8.366628006139243e-08, + "loss": 0.22679826617240906, + "loss_sft": 0.22679826617240906, + "step": 799, + "token_acc": 0.9277570769145209 + }, + { + "epoch": 2.8391111111111114, + "grad_norm": 2.3157103061676025, + "learning_rate": 8.015333003598025e-08, + "loss": 0.21770483255386353, + "loss_sft": 0.21770483255386353, + "step": 800, + "token_acc": 0.9293159990630124 + }, + { + "epoch": 2.8426666666666667, + "grad_norm": 2.201233386993408, + "learning_rate": 7.671512475819876e-08, + "loss": 0.2211962789297104, + "loss_sft": 0.2211962789297104, + "step": 801, + "token_acc": 0.928027989115344 + }, + { + "epoch": 2.846222222222222, + "grad_norm": 2.176928758621216, + "learning_rate": 7.335171646296436e-08, + "loss": 0.22525876760482788, + "loss_sft": 0.22525876760482788, + "step": 802, + "token_acc": 0.9286589440855424 + }, + { + "epoch": 2.8497777777777777, + "grad_norm": 2.1503818035125732, + "learning_rate": 7.006315624884296e-08, + "loss": 0.2188667356967926, + "loss_sft": 0.2188667356967926, + "step": 803, + "token_acc": 0.9330058528428093 + }, + { + "epoch": 2.8533333333333335, + "grad_norm": 2.3525588512420654, + "learning_rate": 6.684949407727015e-08, + "loss": 0.19675451517105103, + "loss_sft": 0.19675451517105103, + "step": 804, + "token_acc": 0.9376433397384244 + }, + { + "epoch": 2.856888888888889, + "grad_norm": 2.303893804550171, + "learning_rate": 6.371077877179277e-08, + "loss": 0.20101016759872437, + "loss_sft": 0.20101016759872437, + "step": 805, + "token_acc": 0.9360420022671678 + }, + { + "epoch": 2.8604444444444446, + "grad_norm": 2.3082034587860107, + "learning_rate": 6.064705801732962e-08, + "loss": 0.20633909106254578, + "loss_sft": 0.20633909106254578, + "step": 806, + "token_acc": 0.9356571428571429 + }, + { + "epoch": 2.864, + "grad_norm": 2.2950446605682373, + "learning_rate": 5.7658378359443104e-08, + "loss": 0.20719939470291138, + "loss_sft": 0.20719939470291138, + "step": 807, + "token_acc": 0.9325678383794617 + }, + { + "epoch": 2.8675555555555556, + "grad_norm": 2.5380027294158936, + "learning_rate": 5.4744785203635355e-08, + "loss": 0.21504728496074677, + "loss_sft": 0.21504728496074677, + "step": 808, + "token_acc": 0.9327440880355888 + }, + { + "epoch": 2.871111111111111, + "grad_norm": 2.3624236583709717, + "learning_rate": 5.190632281465713e-08, + "loss": 0.22169551253318787, + "loss_sft": 0.22169551253318787, + "step": 809, + "token_acc": 0.92899569505749 + }, + { + "epoch": 2.8746666666666667, + "grad_norm": 2.135430335998535, + "learning_rate": 4.914303431583389e-08, + "loss": 0.22522825002670288, + "loss_sft": 0.22522825002670288, + "step": 810, + "token_acc": 0.9264089671066416 + }, + { + "epoch": 2.878222222222222, + "grad_norm": 2.120131015777588, + "learning_rate": 4.645496168841468e-08, + "loss": 0.2010488510131836, + "loss_sft": 0.2010488510131836, + "step": 811, + "token_acc": 0.9365325077399381 + }, + { + "epoch": 2.8817777777777778, + "grad_norm": 2.2362990379333496, + "learning_rate": 4.384214577092927e-08, + "loss": 0.19556254148483276, + "loss_sft": 0.19556254148483276, + "step": 812, + "token_acc": 0.9392440833627693 + }, + { + "epoch": 2.8853333333333335, + "grad_norm": 2.263963222503662, + "learning_rate": 4.130462625857257e-08, + "loss": 0.21070370078086853, + "loss_sft": 0.21070370078086853, + "step": 813, + "token_acc": 0.9336381685819144 + }, + { + "epoch": 2.888888888888889, + "grad_norm": 2.5001983642578125, + "learning_rate": 3.884244170259732e-08, + "loss": 0.2125704437494278, + "loss_sft": 0.2125704437494278, + "step": 814, + "token_acc": 0.9333896542760592 + }, + { + "epoch": 2.8924444444444446, + "grad_norm": 2.329131603240967, + "learning_rate": 3.645562950973014e-08, + "loss": 0.2282361537218094, + "loss_sft": 0.2282361537218094, + "step": 815, + "token_acc": 0.9251103915219311 + }, + { + "epoch": 2.896, + "grad_norm": 2.223870038986206, + "learning_rate": 3.414422594160527e-08, + "loss": 0.19045020639896393, + "loss_sft": 0.19045020639896393, + "step": 816, + "token_acc": 0.939187530682376 + }, + { + "epoch": 2.8995555555555557, + "grad_norm": 2.3049705028533936, + "learning_rate": 3.1908266114210054e-08, + "loss": 0.20488542318344116, + "loss_sft": 0.20488542318344116, + "step": 817, + "token_acc": 0.9326743415104575 + }, + { + "epoch": 2.903111111111111, + "grad_norm": 2.2952005863189697, + "learning_rate": 2.974778399735423e-08, + "loss": 0.23333868384361267, + "loss_sft": 0.23333868384361267, + "step": 818, + "token_acc": 0.923388199267239 + }, + { + "epoch": 2.9066666666666667, + "grad_norm": 2.1744277477264404, + "learning_rate": 2.766281241415203e-08, + "loss": 0.19846579432487488, + "loss_sft": 0.19846579432487488, + "step": 819, + "token_acc": 0.935785466353152 + }, + { + "epoch": 2.910222222222222, + "grad_norm": 2.2673325538635254, + "learning_rate": 2.5653383040524228e-08, + "loss": 0.21949008107185364, + "loss_sft": 0.21949008107185364, + "step": 820, + "token_acc": 0.9283810973951819 + }, + { + "epoch": 2.913777777777778, + "grad_norm": 2.2093045711517334, + "learning_rate": 2.3719526404717975e-08, + "loss": 0.2213304340839386, + "loss_sft": 0.2213304340839386, + "step": 821, + "token_acc": 0.9314456792002842 + }, + { + "epoch": 2.9173333333333336, + "grad_norm": 2.124161958694458, + "learning_rate": 2.1861271886840508e-08, + "loss": 0.21505405008792877, + "loss_sft": 0.21505405008792877, + "step": 822, + "token_acc": 0.9307557782294276 + }, + { + "epoch": 2.920888888888889, + "grad_norm": 2.0560755729675293, + "learning_rate": 2.0078647718415058e-08, + "loss": 0.21757104992866516, + "loss_sft": 0.21757104992866516, + "step": 823, + "token_acc": 0.9300579491055682 + }, + { + "epoch": 2.924444444444444, + "grad_norm": 2.273923635482788, + "learning_rate": 1.837168098195119e-08, + "loss": 0.22025007009506226, + "loss_sft": 0.22025007009506226, + "step": 824, + "token_acc": 0.927333705980995 + }, + { + "epoch": 2.928, + "grad_norm": 1.922160267829895, + "learning_rate": 1.6740397610533477e-08, + "loss": 0.19530706107616425, + "loss_sft": 0.19530706107616425, + "step": 825, + "token_acc": 0.937991891247317 + }, + { + "epoch": 2.9315555555555557, + "grad_norm": 2.183375835418701, + "learning_rate": 1.5184822387426246e-08, + "loss": 0.21241140365600586, + "loss_sft": 0.21241140365600586, + "step": 826, + "token_acc": 0.9346082665021591 + }, + { + "epoch": 2.935111111111111, + "grad_norm": 2.1544220447540283, + "learning_rate": 1.3704978945698888e-08, + "loss": 0.20677995681762695, + "loss_sft": 0.20677995681762695, + "step": 827, + "token_acc": 0.9312607247400828 + }, + { + "epoch": 2.9386666666666668, + "grad_norm": 3.1114187240600586, + "learning_rate": 1.2300889767866697e-08, + "loss": 0.21935826539993286, + "loss_sft": 0.21935826539993286, + "step": 828, + "token_acc": 0.9291467192340186 + }, + { + "epoch": 2.942222222222222, + "grad_norm": 2.2450428009033203, + "learning_rate": 1.0972576185547256e-08, + "loss": 0.21265751123428345, + "loss_sft": 0.21265751123428345, + "step": 829, + "token_acc": 0.932667076098487 + }, + { + "epoch": 2.945777777777778, + "grad_norm": 2.3317325115203857, + "learning_rate": 9.720058379138474e-09, + "loss": 0.22713464498519897, + "loss_sft": 0.22713464498519897, + "step": 830, + "token_acc": 0.9262736056214317 + }, + { + "epoch": 2.9493333333333336, + "grad_norm": 2.1594772338867188, + "learning_rate": 8.543355377512164e-09, + "loss": 0.22452396154403687, + "loss_sft": 0.22452396154403687, + "step": 831, + "token_acc": 0.930357416781715 + }, + { + "epoch": 2.952888888888889, + "grad_norm": 2.0905351638793945, + "learning_rate": 7.442485057722048e-09, + "loss": 0.21176943182945251, + "loss_sft": 0.21176943182945251, + "step": 832, + "token_acc": 0.9325927796011109 + }, + { + "epoch": 2.956444444444444, + "grad_norm": 2.1530587673187256, + "learning_rate": 6.417464144736208e-09, + "loss": 0.2164577841758728, + "loss_sft": 0.2164577841758728, + "step": 833, + "token_acc": 0.9303258688933471 + }, + { + "epoch": 2.96, + "grad_norm": 2.2747607231140137, + "learning_rate": 5.468308211179496e-09, + "loss": 0.21495883166790009, + "loss_sft": 0.21495883166790009, + "step": 834, + "token_acc": 0.9302575748520887 + }, + { + "epoch": 2.9635555555555557, + "grad_norm": 2.289250135421753, + "learning_rate": 4.595031677099293e-09, + "loss": 0.22347143292427063, + "loss_sft": 0.22347143292427063, + "step": 835, + "token_acc": 0.9289260517376535 + }, + { + "epoch": 2.967111111111111, + "grad_norm": 2.1104331016540527, + "learning_rate": 3.7976478097451196e-09, + "loss": 0.19463887810707092, + "loss_sft": 0.19463887810707092, + "step": 836, + "token_acc": 0.9373419917711793 + }, + { + "epoch": 2.970666666666667, + "grad_norm": 2.1791234016418457, + "learning_rate": 3.0761687233682445e-09, + "loss": 0.2004217505455017, + "loss_sft": 0.2004217505455017, + "step": 837, + "token_acc": 0.9340432649420161 + }, + { + "epoch": 2.974222222222222, + "grad_norm": 2.33919358253479, + "learning_rate": 2.430605379035722e-09, + "loss": 0.2124345898628235, + "loss_sft": 0.2124345898628235, + "step": 838, + "token_acc": 0.9309055015844979 + }, + { + "epoch": 2.977777777777778, + "grad_norm": 2.200502634048462, + "learning_rate": 1.8609675844655228e-09, + "loss": 0.21903391182422638, + "loss_sft": 0.21903391182422638, + "step": 839, + "token_acc": 0.9308282208588957 + }, + { + "epoch": 2.981333333333333, + "grad_norm": 2.1386334896087646, + "learning_rate": 1.3672639938777655e-09, + "loss": 0.21041738986968994, + "loss_sft": 0.21041738986968994, + "step": 840, + "token_acc": 0.9326586220319191 + }, + { + "epoch": 2.984888888888889, + "grad_norm": 2.19281005859375, + "learning_rate": 9.495021078614885e-10, + "loss": 0.20342174172401428, + "loss_sft": 0.20342174172401428, + "step": 841, + "token_acc": 0.9352552552552552 + }, + { + "epoch": 2.9884444444444442, + "grad_norm": 2.223222017288208, + "learning_rate": 6.076882732625189e-10, + "loss": 0.22914628684520721, + "loss_sft": 0.22914628684520721, + "step": 842, + "token_acc": 0.928556263269639 + }, + { + "epoch": 2.992, + "grad_norm": 2.1534955501556396, + "learning_rate": 3.4182768308577143e-10, + "loss": 0.20344775915145874, + "loss_sft": 0.20344775915145874, + "step": 843, + "token_acc": 0.9371043072132849 + }, + { + "epoch": 2.9955555555555557, + "grad_norm": 2.0710599422454834, + "learning_rate": 1.5192437641753376e-10, + "loss": 0.21217861771583557, + "loss_sft": 0.21217861771583557, + "step": 844, + "token_acc": 0.9329498211350672 + }, + { + "epoch": 2.999111111111111, + "grad_norm": 2.261275291442871, + "learning_rate": 3.798123836162848e-11, + "loss": 0.20855969190597534, + "loss_sft": 0.20855969190597534, + "step": 845, + "token_acc": 0.931514415111013 + }, + { + "epoch": 3.0, + "grad_norm": 4.708669185638428, + "learning_rate": 0.0, + "loss": 0.15930986404418945, + "loss_sft": 0.15930986404418945, + "step": 846, + "token_acc": 0.9476702508960574 + } + ], + "logging_steps": 1, + "max_steps": 846, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 2.3456028590195343e+18, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +}