diff --git "a/train_logs.json" "b/train_logs.json" new file mode 100644--- /dev/null +++ "b/train_logs.json" @@ -0,0 +1,3593 @@ +[ + { + "step": 50, + "train_loss": 6.0813891887664795, + "val_loss": 6.026689910888672, + "lr": 0.01, + "tokens_per_sec": 794214.7164234697, + "mfu": 7.466395208164367, + "total_flops": 1266401163608064.0 + }, + { + "step": 100, + "train_loss": 5.087921142578125, + "val_loss": 5.089700198173523, + "lr": 0.02, + "tokens_per_sec": 736508.4501651466, + "mfu": 6.923899859032017, + "total_flops": 2507970931851264.0 + }, + { + "step": 150, + "train_loss": 4.558597803115845, + "val_loss": 4.58702654838562, + "lr": 0.02, + "tokens_per_sec": 635169.419700884, + "mfu": 5.97121384628007, + "total_flops": 3749540700094464.0 + }, + { + "step": 200, + "train_loss": 4.243805885314941, + "val_loss": 4.244970810413361, + "lr": 0.02, + "tokens_per_sec": 740626.4277880366, + "mfu": 6.9626128767526305, + "total_flops": 4991110468337664.0 + }, + { + "step": 250, + "train_loss": 4.045148968696594, + "val_loss": 4.084274005889893, + "lr": 0.02, + "tokens_per_sec": 729753.6828782216, + "mfu": 6.860398439251636, + "total_flops": 6232680236580864.0 + }, + { + "step": 300, + "train_loss": 3.926572263240814, + "val_loss": 3.9035560965538023, + "lr": 0.02, + "tokens_per_sec": 587099.4249975935, + "mfu": 5.5193088756376385, + "total_flops": 7474250004824064.0 + }, + { + "step": 350, + "train_loss": 3.7291558384895325, + "val_loss": 3.7622887253761292, + "lr": 0.02, + "tokens_per_sec": 789797.325586338, + "mfu": 7.424867413354072, + "total_flops": 8715819773067264.0 + }, + { + "step": 400, + "train_loss": 3.615461826324463, + "val_loss": 3.6281559228897096, + "lr": 0.02, + "tokens_per_sec": 302900.2753542377, + "mfu": 2.847558875062059, + "total_flops": 9957389541310464.0 + }, + { + "step": 450, + "train_loss": 3.5205406546592712, + "val_loss": 3.5723757147789, + "lr": 0.02, + "tokens_per_sec": 607157.4423502305, + "mfu": 5.707873858821802, + "total_flops": 1.1198959309553664e+16 + }, + { + "step": 500, + "train_loss": 3.4816104769706726, + "val_loss": 3.532227611541748, + "lr": 0.02, + "tokens_per_sec": 703708.9583817514, + "mfu": 6.615552552922427, + "total_flops": 1.2440529077796864e+16 + }, + { + "step": 550, + "train_loss": 3.3761273622512817, + "val_loss": 3.5005128383636475, + "lr": 0.02, + "tokens_per_sec": 722074.6273645067, + "mfu": 6.788207806032169, + "total_flops": 1.3682098846040064e+16 + }, + { + "step": 600, + "train_loss": 3.4218554496765137, + "val_loss": 3.471066892147064, + "lr": 0.02, + "tokens_per_sec": 621526.2112982491, + "mfu": 5.842954341973545, + "total_flops": 1.4923668614283264e+16 + }, + { + "step": 650, + "train_loss": 3.4142528772354126, + "val_loss": 3.455700898170471, + "lr": 0.02, + "tokens_per_sec": 760036.6436618323, + "mfu": 7.145087892378344, + "total_flops": 1.6165238382526464e+16 + }, + { + "step": 700, + "train_loss": 3.377539098262787, + "val_loss": 3.436308777332306, + "lr": 0.02, + "tokens_per_sec": 722331.7487977484, + "mfu": 6.79062499901207, + "total_flops": 1.7406808150769664e+16 + }, + { + "step": 750, + "train_loss": 3.3675426840782166, + "val_loss": 3.4219574451446535, + "lr": 0.02, + "tokens_per_sec": 710436.7041358028, + "mfu": 6.678799943862251, + "total_flops": 1.8648377919012864e+16 + }, + { + "step": 800, + "train_loss": 3.331727921962738, + "val_loss": 3.40491441488266, + "lr": 0.02, + "tokens_per_sec": 809757.6478994827, + "mfu": 7.612513967605048, + "total_flops": 1.9889947687256064e+16 + }, + { + "step": 850, + "train_loss": 3.3138641715049744, + "val_loss": 3.3908426761627197, + "lr": 0.02, + "tokens_per_sec": 742484.9786147004, + "mfu": 6.980085072494384, + "total_flops": 2.1131517455499264e+16 + }, + { + "step": 900, + "train_loss": 3.314534604549408, + "val_loss": 3.372792398929596, + "lr": 0.02, + "tokens_per_sec": 699743.0740374597, + "mfu": 6.5782693607931, + "total_flops": 2.2373087223742464e+16 + }, + { + "step": 950, + "train_loss": 3.2452561259269714, + "val_loss": 3.362290894985199, + "lr": 0.02, + "tokens_per_sec": 739832.2356642535, + "mfu": 6.955146693937383, + "total_flops": 2.3614656991985664e+16 + }, + { + "step": 1000, + "train_loss": 3.262390971183777, + "val_loss": 3.3680756092071533, + "lr": 0.02, + "tokens_per_sec": 736792.1990218526, + "mfu": 6.926567375838519, + "total_flops": 2.4856226760228864e+16 + }, + { + "step": 1050, + "train_loss": 3.3225939869880676, + "val_loss": 3.335637629032135, + "lr": 0.02, + "tokens_per_sec": 729562.5144698353, + "mfu": 6.858601269218378, + "total_flops": 2.6097796528472064e+16 + }, + { + "step": 1100, + "train_loss": 3.300910174846649, + "val_loss": 3.316732621192932, + "lr": 0.02, + "tokens_per_sec": 802942.0066579218, + "mfu": 7.5484402731062605, + "total_flops": 2.7339366296715264e+16 + }, + { + "step": 1150, + "train_loss": 3.2660632729530334, + "val_loss": 3.3214237689971924, + "lr": 0.02, + "tokens_per_sec": 753021.8553813706, + "mfu": 7.0791420209150395, + "total_flops": 2.8580936064958464e+16 + }, + { + "step": 1200, + "train_loss": 3.2546098232269287, + "val_loss": 3.314812517166138, + "lr": 0.02, + "tokens_per_sec": 761949.7264553182, + "mfu": 7.163072741949536, + "total_flops": 2.9822505833201664e+16 + }, + { + "step": 1250, + "train_loss": 3.200866460800171, + "val_loss": 3.3129363536834715, + "lr": 0.02, + "tokens_per_sec": 759970.4128061022, + "mfu": 7.144465257549713, + "total_flops": 3.1064075601444864e+16 + }, + { + "step": 1300, + "train_loss": 3.130142331123352, + "val_loss": 3.3093976616859435, + "lr": 0.02, + "tokens_per_sec": 763491.4059958718, + "mfu": 7.1775660376491555, + "total_flops": 3.2305645369688064e+16 + }, + { + "step": 1350, + "train_loss": 3.259128153324127, + "val_loss": 3.3080609321594237, + "lr": 0.02, + "tokens_per_sec": 737350.9208228615, + "mfu": 6.931819907290637, + "total_flops": 3.3547215137931264e+16 + }, + { + "step": 1400, + "train_loss": 3.2849743962287903, + "val_loss": 3.3059651374816896, + "lr": 0.02, + "tokens_per_sec": 523486.77008353075, + "mfu": 4.921287695713131, + "total_flops": 3.4788784906174464e+16 + }, + { + "step": 1450, + "train_loss": 3.06045800447464, + "val_loss": 3.3061444997787475, + "lr": 0.02, + "tokens_per_sec": 735041.2803046936, + "mfu": 6.9101070272081735, + "total_flops": 3.603035467441766e+16 + }, + { + "step": 1500, + "train_loss": 3.2446532249450684, + "val_loss": 3.291266882419586, + "lr": 0.02, + "tokens_per_sec": 724220.7811134176, + "mfu": 6.808383750566429, + "total_flops": 3.727192444266086e+16 + }, + { + "step": 1550, + "train_loss": 3.244351804256439, + "val_loss": 3.2926158547401427, + "lr": 0.02, + "tokens_per_sec": 779334.7750054352, + "mfu": 7.326509203782974, + "total_flops": 3.851349421090406e+16 + }, + { + "step": 1600, + "train_loss": 3.2528223395347595, + "val_loss": 3.2792186975479125, + "lr": 0.02, + "tokens_per_sec": 732232.7445099982, + "mfu": 6.883704043523984, + "total_flops": 3.975506397914726e+16 + }, + { + "step": 1650, + "train_loss": 3.2681846022605896, + "val_loss": 3.2697198152542115, + "lr": 0.02, + "tokens_per_sec": 727952.4135197375, + "mfu": 6.843464745341551, + "total_flops": 4.099663374739046e+16 + }, + { + "step": 1700, + "train_loss": 3.257525682449341, + "val_loss": 3.2683892846107483, + "lr": 0.02, + "tokens_per_sec": 727843.2189831685, + "mfu": 6.842438209887426, + "total_flops": 4.223820351563366e+16 + }, + { + "step": 1750, + "train_loss": 3.1968674063682556, + "val_loss": 3.277073395252228, + "lr": 0.02, + "tokens_per_sec": 715307.3754703364, + "mfu": 6.724589018731667, + "total_flops": 4.347977328387686e+16 + }, + { + "step": 1800, + "train_loss": 3.2354595065116882, + "val_loss": 3.276608467102051, + "lr": 0.02, + "tokens_per_sec": 745560.8215328843, + "mfu": 7.009001004610087, + "total_flops": 4.472134305212006e+16 + }, + { + "step": 1850, + "train_loss": 3.219004511833191, + "val_loss": 3.267551374435425, + "lr": 0.02, + "tokens_per_sec": 706164.9282135066, + "mfu": 6.638641071686961, + "total_flops": 4.596291282036326e+16 + }, + { + "step": 1900, + "train_loss": 3.205766499042511, + "val_loss": 3.2662167072296144, + "lr": 0.02, + "tokens_per_sec": 682567.1752142893, + "mfu": 6.416799110976101, + "total_flops": 4.720448258860646e+16 + }, + { + "step": 1950, + "train_loss": 3.321563243865967, + "val_loss": 3.2645433068275453, + "lr": 0.02, + "tokens_per_sec": 713166.0692443526, + "mfu": 6.704458645654604, + "total_flops": 4.844605235684966e+16 + }, + { + "step": 2000, + "train_loss": 3.162742018699646, + "val_loss": 3.263501501083374, + "lr": 0.02, + "tokens_per_sec": 736771.9689582054, + "mfu": 6.926377193451875, + "total_flops": 4.968762212509286e+16 + }, + { + "step": 2050, + "train_loss": 3.239263594150543, + "val_loss": 3.260008931159973, + "lr": 0.02, + "tokens_per_sec": 701897.3676516128, + "mfu": 6.5985218280227755, + "total_flops": 5.092919189333606e+16 + }, + { + "step": 2100, + "train_loss": 3.1417112350463867, + "val_loss": 3.2654417872428896, + "lr": 0.02, + "tokens_per_sec": 724962.1222182881, + "mfu": 6.81535308210683, + "total_flops": 5.217076166157926e+16 + }, + { + "step": 2150, + "train_loss": 3.1347965002059937, + "val_loss": 3.259656322002411, + "lr": 0.02, + "tokens_per_sec": 779163.2670647042, + "mfu": 7.324896861376877, + "total_flops": 5.341233142982246e+16 + }, + { + "step": 2200, + "train_loss": 3.1822943687438965, + "val_loss": 3.253624069690704, + "lr": 0.02, + "tokens_per_sec": 760755.4181808244, + "mfu": 7.151845075937621, + "total_flops": 5.465390119806566e+16 + }, + { + "step": 2250, + "train_loss": 3.1740533113479614, + "val_loss": 3.2457565784454347, + "lr": 0.02, + "tokens_per_sec": 769730.274715895, + "mfu": 7.236217506266253, + "total_flops": 5.589547096630886e+16 + }, + { + "step": 2300, + "train_loss": 3.198862850666046, + "val_loss": 3.2464448690414427, + "lr": 0.02, + "tokens_per_sec": 720703.044629105, + "mfu": 6.775313586683879, + "total_flops": 5.713704073455206e+16 + }, + { + "step": 2350, + "train_loss": 3.1597362756729126, + "val_loss": 3.242673170566559, + "lr": 0.02, + "tokens_per_sec": 738083.676306573, + "mfu": 6.938708525593981, + "total_flops": 5.837861050279526e+16 + }, + { + "step": 2400, + "train_loss": 3.1868541836738586, + "val_loss": 3.237907588481903, + "lr": 0.02, + "tokens_per_sec": 769817.1624990003, + "mfu": 7.237034336418113, + "total_flops": 5.962018027103846e+16 + }, + { + "step": 2450, + "train_loss": 3.153138518333435, + "val_loss": 3.2368793964385985, + "lr": 0.02, + "tokens_per_sec": 767427.2208379519, + "mfu": 7.214566547044729, + "total_flops": 6.086175003928166e+16 + }, + { + "step": 2500, + "train_loss": 3.247396469116211, + "val_loss": 3.2282278895378114, + "lr": 0.02, + "tokens_per_sec": 780143.0316463324, + "mfu": 7.334107606816523, + "total_flops": 6.210331980752486e+16 + }, + { + "step": 2550, + "train_loss": 3.122944414615631, + "val_loss": 3.2200260877609255, + "lr": 0.02, + "tokens_per_sec": 742767.58170015, + "mfu": 6.9827418179303224, + "total_flops": 6.334488957576806e+16 + }, + { + "step": 2600, + "train_loss": 3.1630945801734924, + "val_loss": 3.216521847248077, + "lr": 0.02, + "tokens_per_sec": 704446.6323657001, + "mfu": 6.62248740993911, + "total_flops": 6.458645934401126e+16 + }, + { + "step": 2650, + "train_loss": 3.0957342982292175, + "val_loss": 3.21864767074585, + "lr": 0.02, + "tokens_per_sec": 191314.34620024994, + "mfu": 1.7985419914594185, + "total_flops": 6.582802911225446e+16 + }, + { + "step": 2700, + "train_loss": 3.173651099205017, + "val_loss": 3.224730205535889, + "lr": 0.02, + "tokens_per_sec": 709829.4426204753, + "mfu": 6.673091091615642, + "total_flops": 6.706959888049766e+16 + }, + { + "step": 2750, + "train_loss": 3.1663557291030884, + "val_loss": 3.2234278440475466, + "lr": 0.02, + "tokens_per_sec": 724416.7606760697, + "mfu": 6.810226150155891, + "total_flops": 6.831116864874086e+16 + }, + { + "step": 2800, + "train_loss": 3.147324025630951, + "val_loss": 3.2159315824508665, + "lr": 0.02, + "tokens_per_sec": 734666.1504854651, + "mfu": 6.906580439968213, + "total_flops": 6.955273841698406e+16 + }, + { + "step": 2850, + "train_loss": 3.224787950515747, + "val_loss": 3.2105109453201295, + "lr": 0.02, + "tokens_per_sec": 699516.5324807317, + "mfu": 6.576139648564623, + "total_flops": 7.079430818522726e+16 + }, + { + "step": 2900, + "train_loss": 3.2367559671401978, + "val_loss": 3.231448769569397, + "lr": 0.02, + "tokens_per_sec": 662728.9877131225, + "mfu": 6.230300743425831, + "total_flops": 7.203587795347046e+16 + }, + { + "step": 2950, + "train_loss": 3.2191177010536194, + "val_loss": 3.2184744119644164, + "lr": 0.02, + "tokens_per_sec": 672224.1362785365, + "mfu": 6.319564427771635, + "total_flops": 7.327744772171366e+16 + }, + { + "step": 3000, + "train_loss": 3.131461203098297, + "val_loss": 3.213130033016205, + "lr": 0.02, + "tokens_per_sec": 717274.5614815141, + "mfu": 6.743082491471089, + "total_flops": 7.451901748995686e+16 + }, + { + "step": 3050, + "train_loss": 3.1307908296585083, + "val_loss": 3.228156554698944, + "lr": 0.02, + "tokens_per_sec": 699103.3039631982, + "mfu": 6.572254896293749, + "total_flops": 7.576058725820006e+16 + }, + { + "step": 3100, + "train_loss": 3.102234125137329, + "val_loss": 3.2135570764541628, + "lr": 0.02, + "tokens_per_sec": 708978.8172659309, + "mfu": 6.665094381230177, + "total_flops": 7.700215702644326e+16 + }, + { + "step": 3150, + "train_loss": 3.1781490445137024, + "val_loss": 3.206849551200867, + "lr": 0.02, + "tokens_per_sec": 710505.0044441634, + "mfu": 6.679442033569876, + "total_flops": 7.824372679468646e+16 + }, + { + "step": 3200, + "train_loss": 3.020294487476349, + "val_loss": 3.218398690223694, + "lr": 0.02, + "tokens_per_sec": 700330.5278482502, + "mfu": 6.583792001241841, + "total_flops": 7.948529656292966e+16 + }, + { + "step": 3250, + "train_loss": 3.1484155654907227, + "val_loss": 3.195547950267792, + "lr": 0.02, + "tokens_per_sec": 803595.1037902995, + "mfu": 7.554580024987982, + "total_flops": 8.072686633117286e+16 + }, + { + "step": 3300, + "train_loss": 3.2074788212776184, + "val_loss": 3.202925610542297, + "lr": 0.02, + "tokens_per_sec": 725059.8870553945, + "mfu": 6.816272167205863, + "total_flops": 8.196843609941606e+16 + }, + { + "step": 3350, + "train_loss": 3.1220479011535645, + "val_loss": 3.2095597267150877, + "lr": 0.02, + "tokens_per_sec": 730116.0178445905, + "mfu": 6.863804742359225, + "total_flops": 8.321000586765926e+16 + }, + { + "step": 3400, + "train_loss": 3.133055806159973, + "val_loss": 3.2004257798194886, + "lr": 0.02, + "tokens_per_sec": 684447.9707085511, + "mfu": 6.434480428352253, + "total_flops": 8.445157563590246e+16 + }, + { + "step": 3450, + "train_loss": 3.142967641353607, + "val_loss": 3.2058743834495544, + "lr": 0.02, + "tokens_per_sec": 674727.5005119084, + "mfu": 6.343098500270955, + "total_flops": 8.569314540414566e+16 + }, + { + "step": 3500, + "train_loss": 3.111777603626251, + "val_loss": 3.2019073963165283, + "lr": 0.02, + "tokens_per_sec": 733620.1779752404, + "mfu": 6.896747275237419, + "total_flops": 8.693471517238886e+16 + }, + { + "step": 3550, + "train_loss": 3.0948435068130493, + "val_loss": 3.208818733692169, + "lr": 0.02, + "tokens_per_sec": 752006.8155076118, + "mfu": 7.069599653224275, + "total_flops": 8.817628494063206e+16 + }, + { + "step": 3600, + "train_loss": 3.1297584176063538, + "val_loss": 3.20062255859375, + "lr": 0.02, + "tokens_per_sec": 757551.0811660548, + "mfu": 7.1217211736254615, + "total_flops": 8.941785470887526e+16 + }, + { + "step": 3650, + "train_loss": 3.097672760486603, + "val_loss": 3.195528542995453, + "lr": 0.02, + "tokens_per_sec": 753421.1065528493, + "mfu": 7.082895372460834, + "total_flops": 9.065942447711846e+16 + }, + { + "step": 3700, + "train_loss": 3.239434063434601, + "val_loss": 3.1988398909568785, + "lr": 0.02, + "tokens_per_sec": 663844.9946939754, + "mfu": 6.2407923006858645, + "total_flops": 9.190099424536166e+16 + }, + { + "step": 3750, + "train_loss": 3.1775227189064026, + "val_loss": 3.197325074672699, + "lr": 0.02, + "tokens_per_sec": 770276.3805974862, + "mfu": 7.241351435735383, + "total_flops": 9.314256401360486e+16 + }, + { + "step": 3800, + "train_loss": 3.133751928806305, + "val_loss": 3.198042118549347, + "lr": 0.02, + "tokens_per_sec": 701977.7272873463, + "mfu": 6.599277287203729, + "total_flops": 9.438413378184806e+16 + }, + { + "step": 3850, + "train_loss": 3.067799389362335, + "val_loss": 3.1975177764892577, + "lr": 0.02, + "tokens_per_sec": 681804.8338186503, + "mfu": 6.409632356160707, + "total_flops": 9.562570355009126e+16 + }, + { + "step": 3900, + "train_loss": 3.1261569261550903, + "val_loss": 3.199464535713196, + "lr": 0.02, + "tokens_per_sec": 700615.0151335543, + "mfu": 6.586466459999472, + "total_flops": 9.686727331833446e+16 + }, + { + "step": 3950, + "train_loss": 3.1253642439842224, + "val_loss": 3.1979977607727053, + "lr": 0.02, + "tokens_per_sec": 764910.951011311, + "mfu": 7.190911149344868, + "total_flops": 9.810884308657766e+16 + }, + { + "step": 4000, + "train_loss": 3.0701352953910828, + "val_loss": 3.1961832523345945, + "lr": 0.02, + "tokens_per_sec": 763747.9314474192, + "mfu": 7.17997762781816, + "total_flops": 9.935041285482086e+16 + }, + { + "step": 4050, + "train_loss": 3.127575993537903, + "val_loss": 3.1980692744255066, + "lr": 0.02, + "tokens_per_sec": 711749.536967804, + "mfu": 6.691141856651423, + "total_flops": 1.0059198262306406e+17 + }, + { + "step": 4100, + "train_loss": 3.1469481587409973, + "val_loss": 3.1901233553886414, + "lr": 0.02, + "tokens_per_sec": 780514.0241170248, + "mfu": 7.3375952989332305, + "total_flops": 1.0183355239130726e+17 + }, + { + "step": 4150, + "train_loss": 3.2410582304000854, + "val_loss": 3.195158803462982, + "lr": 0.02, + "tokens_per_sec": 747323.0486882784, + "mfu": 7.025567663621373, + "total_flops": 1.0307512215955046e+17 + }, + { + "step": 4200, + "train_loss": 3.158669590950012, + "val_loss": 3.2082596898078917, + "lr": 0.02, + "tokens_per_sec": 762441.6464148208, + "mfu": 7.167697270748211, + "total_flops": 1.0431669192779366e+17 + }, + { + "step": 4250, + "train_loss": 3.171230673789978, + "val_loss": 3.193644082546234, + "lr": 0.02, + "tokens_per_sec": 789190.9245840942, + "mfu": 7.419166650772175, + "total_flops": 1.0555826169603686e+17 + }, + { + "step": 4300, + "train_loss": 3.1072226762771606, + "val_loss": 3.1926248908042907, + "lr": 0.02, + "tokens_per_sec": 798558.8382577032, + "mfu": 7.507234202678174, + "total_flops": 1.0679983146428006e+17 + }, + { + "step": 4350, + "train_loss": 3.197103798389435, + "val_loss": 3.1995687246322633, + "lr": 0.02, + "tokens_per_sec": 797254.8530887661, + "mfu": 7.494975466576287, + "total_flops": 1.0804140123252326e+17 + }, + { + "step": 4400, + "train_loss": 3.1892601251602173, + "val_loss": 3.1967344164848326, + "lr": 0.02, + "tokens_per_sec": 755555.0768235743, + "mfu": 7.102956780382684, + "total_flops": 1.0928297100076646e+17 + }, + { + "step": 4450, + "train_loss": 3.1196552515029907, + "val_loss": 3.1996314406394957, + "lr": 0.02, + "tokens_per_sec": 799516.4178640184, + "mfu": 7.516236387649291, + "total_flops": 1.1052454076900966e+17 + }, + { + "step": 4500, + "train_loss": 3.112544298171997, + "val_loss": 3.1933774828910826, + "lr": 0.02, + "tokens_per_sec": 728543.854373343, + "mfu": 6.849024867892742, + "total_flops": 1.1176611053725286e+17 + }, + { + "step": 4550, + "train_loss": 3.0423641204833984, + "val_loss": 3.1891826510429384, + "lr": 0.02, + "tokens_per_sec": 750392.8790864372, + "mfu": 7.054427072167598, + "total_flops": 1.1300768030549606e+17 + }, + { + "step": 4600, + "train_loss": 3.125426948070526, + "val_loss": 3.1997652888298034, + "lr": 0.02, + "tokens_per_sec": 661473.8626517142, + "mfu": 6.218501340128009, + "total_flops": 1.1424925007373926e+17 + }, + { + "step": 4650, + "train_loss": 3.1154977679252625, + "val_loss": 3.191564977169037, + "lr": 0.02, + "tokens_per_sec": 670619.464242622, + "mfu": 6.304478940998509, + "total_flops": 1.1549081984198246e+17 + }, + { + "step": 4700, + "train_loss": 3.0956615805625916, + "val_loss": 3.1927812576293944, + "lr": 0.02, + "tokens_per_sec": 703655.0511273266, + "mfu": 6.615045772000593, + "total_flops": 1.1673238961022566e+17 + }, + { + "step": 4750, + "train_loss": 3.169421136379242, + "val_loss": 3.2010963797569274, + "lr": 0.02, + "tokens_per_sec": 498538.56643634004, + "mfu": 4.686750177946495, + "total_flops": 1.1797395937846886e+17 + }, + { + "step": 4800, + "train_loss": 3.1868945956230164, + "val_loss": 3.1932270765304565, + "lr": 0.02, + "tokens_per_sec": 697798.546786271, + "mfu": 6.559988902561603, + "total_flops": 1.1921552914671206e+17 + }, + { + "step": 4850, + "train_loss": 3.1466269493103027, + "val_loss": 3.197691464424133, + "lr": 0.02, + "tokens_per_sec": 702094.5409157231, + "mfu": 6.600375449573584, + "total_flops": 1.2045709891495526e+17 + }, + { + "step": 4900, + "train_loss": 3.1853439807891846, + "val_loss": 3.190361273288727, + "lr": 0.02, + "tokens_per_sec": 673496.7381930096, + "mfu": 6.331528130583528, + "total_flops": 1.2169866868319846e+17 + }, + { + "step": 4950, + "train_loss": 3.1415867805480957, + "val_loss": 3.18849755525589, + "lr": 0.02, + "tokens_per_sec": 775784.3591892282, + "mfu": 7.293131822214809, + "total_flops": 1.2294023845144166e+17 + }, + { + "step": 5000, + "train_loss": 3.132832705974579, + "val_loss": 3.1886184215545654, + "lr": 0.02, + "tokens_per_sec": 705061.2508035857, + "mfu": 6.628265424454055, + "total_flops": 1.2418180821968486e+17 + }, + { + "step": 5050, + "train_loss": 3.13687264919281, + "val_loss": 3.192574620246887, + "lr": 0.02, + "tokens_per_sec": 766733.9904356612, + "mfu": 7.20804950316883, + "total_flops": 1.2542337798792806e+17 + }, + { + "step": 5100, + "train_loss": 3.1279391646385193, + "val_loss": 3.1825401544570924, + "lr": 0.02, + "tokens_per_sec": 783440.1023898001, + "mfu": 7.365103296887419, + "total_flops": 1.2666494775617126e+17 + }, + { + "step": 5150, + "train_loss": 3.1994088888168335, + "val_loss": 3.1858301997184753, + "lr": 0.02, + "tokens_per_sec": 732295.3383307715, + "mfu": 6.884292486666397, + "total_flops": 1.2790651752441446e+17 + }, + { + "step": 5200, + "train_loss": 3.2014970183372498, + "val_loss": 3.1864626884460447, + "lr": 0.02, + "tokens_per_sec": 698731.1794154128, + "mfu": 6.5687565615450065, + "total_flops": 1.2914808729265766e+17 + }, + { + "step": 5250, + "train_loss": 3.09488445520401, + "val_loss": 3.183492195606232, + "lr": 0.02, + "tokens_per_sec": 728472.9470492065, + "mfu": 6.848358269686729, + "total_flops": 1.3038965706090086e+17 + }, + { + "step": 5300, + "train_loss": 3.119004726409912, + "val_loss": 3.182840895652771, + "lr": 0.02, + "tokens_per_sec": 717396.2762640028, + "mfu": 6.744226729483736, + "total_flops": 1.3163122682914406e+17 + }, + { + "step": 5350, + "train_loss": 3.2293498516082764, + "val_loss": 3.1889729499816895, + "lr": 0.02, + "tokens_per_sec": 725808.5296261577, + "mfu": 6.8233101396677, + "total_flops": 1.3287279659738726e+17 + }, + { + "step": 5400, + "train_loss": 3.2103021144866943, + "val_loss": 3.179009199142456, + "lr": 0.02, + "tokens_per_sec": 662681.5853182896, + "mfu": 6.229855114546993, + "total_flops": 1.3411436636563046e+17 + }, + { + "step": 5450, + "train_loss": 3.1303423047065735, + "val_loss": 3.188521647453308, + "lr": 0.02, + "tokens_per_sec": 728308.2940295547, + "mfu": 6.846810370244026, + "total_flops": 1.3535593613387366e+17 + }, + { + "step": 5500, + "train_loss": 3.1481414437294006, + "val_loss": 3.18265780210495, + "lr": 0.02, + "tokens_per_sec": 782476.503243071, + "mfu": 7.356044522348301, + "total_flops": 1.3659750590211686e+17 + }, + { + "step": 5550, + "train_loss": 3.103872239589691, + "val_loss": 3.1814547777175903, + "lr": 0.02, + "tokens_per_sec": 693542.3596234568, + "mfu": 6.519976579973838, + "total_flops": 1.3783907567036006e+17 + }, + { + "step": 5600, + "train_loss": 3.1177070140838623, + "val_loss": 3.1881037831306456, + "lr": 0.02, + "tokens_per_sec": 670932.9462096259, + "mfu": 6.307425978125727, + "total_flops": 1.3908064543860326e+17 + }, + { + "step": 5650, + "train_loss": 3.0816861987113953, + "val_loss": 3.1840782165527344, + "lr": 0.02, + "tokens_per_sec": 696515.0198524478, + "mfu": 6.5479224938813365, + "total_flops": 1.4032221520684646e+17 + }, + { + "step": 5700, + "train_loss": 3.09510201215744, + "val_loss": 3.184128928184509, + "lr": 0.02, + "tokens_per_sec": 687919.7057449103, + "mfu": 6.4671181336269905, + "total_flops": 1.4156378497508966e+17 + }, + { + "step": 5750, + "train_loss": 3.1403180360794067, + "val_loss": 3.1881123423576354, + "lr": 0.02, + "tokens_per_sec": 648873.811017929, + "mfu": 6.100048529829092, + "total_flops": 1.4280535474333286e+17 + }, + { + "step": 5800, + "train_loss": 3.130989968776703, + "val_loss": 3.1858367681503297, + "lr": 0.02, + "tokens_per_sec": 727039.472701932, + "mfu": 6.834882208646567, + "total_flops": 1.4404692451157606e+17 + }, + { + "step": 5850, + "train_loss": 3.069033682346344, + "val_loss": 3.17846132516861, + "lr": 0.02, + "tokens_per_sec": 706744.8880096792, + "mfu": 6.644093261067901, + "total_flops": 1.4528849427981926e+17 + }, + { + "step": 5900, + "train_loss": 3.13921719789505, + "val_loss": 3.1820234179496767, + "lr": 0.02, + "tokens_per_sec": 726040.7135660015, + "mfu": 6.825492895816653, + "total_flops": 1.4653006404806246e+17 + }, + { + "step": 5950, + "train_loss": 3.0790565609931946, + "val_loss": 3.201737642288208, + "lr": 0.02, + "tokens_per_sec": 642554.8551086267, + "mfu": 6.04064416329421, + "total_flops": 1.4777163381630566e+17 + }, + { + "step": 6000, + "train_loss": 3.0759347081184387, + "val_loss": 3.1834460496902466, + "lr": 0.02, + "tokens_per_sec": 740776.0605106918, + "mfu": 6.964019570711222, + "total_flops": 1.4901320358454886e+17 + }, + { + "step": 6050, + "train_loss": 3.1061267256736755, + "val_loss": 3.1822536826133727, + "lr": 0.02, + "tokens_per_sec": 699657.8714128656, + "mfu": 6.577468372779611, + "total_flops": 1.5025477335279206e+17 + }, + { + "step": 6100, + "train_loss": 3.151519775390625, + "val_loss": 3.18179247379303, + "lr": 0.02, + "tokens_per_sec": 557667.215608958, + "mfu": 5.242617317799736, + "total_flops": 1.5149634312103526e+17 + }, + { + "step": 6150, + "train_loss": 3.080659031867981, + "val_loss": 3.18075053691864, + "lr": 0.02, + "tokens_per_sec": 726509.7037583116, + "mfu": 6.829901862374601, + "total_flops": 1.5273791288927846e+17 + }, + { + "step": 6200, + "train_loss": 3.0890342593193054, + "val_loss": 3.179175615310669, + "lr": 0.02, + "tokens_per_sec": 699340.2421265863, + "mfu": 6.574482346794437, + "total_flops": 1.5397948265752166e+17 + }, + { + "step": 6250, + "train_loss": 3.1207754611968994, + "val_loss": 3.1796735525131226, + "lr": 0.02, + "tokens_per_sec": 757885.2205318413, + "mfu": 7.1248624105076015, + "total_flops": 1.5522105242576486e+17 + }, + { + "step": 6300, + "train_loss": 3.0962599515914917, + "val_loss": 3.180558133125305, + "lr": 0.02, + "tokens_per_sec": 667449.8589065681, + "mfu": 6.274681550439623, + "total_flops": 1.5646262219400806e+17 + }, + { + "step": 6350, + "train_loss": 2.9786906838417053, + "val_loss": 3.18046555519104, + "lr": 0.02, + "tokens_per_sec": 694050.5334623031, + "mfu": 6.524753911137333, + "total_flops": 1.5770419196225126e+17 + }, + { + "step": 6400, + "train_loss": 3.049709677696228, + "val_loss": 3.1755032181739806, + "lr": 0.02, + "tokens_per_sec": 665420.704705452, + "mfu": 6.255605516099607, + "total_flops": 1.5894576173049446e+17 + }, + { + "step": 6450, + "train_loss": 3.163714349269867, + "val_loss": 3.177147853374481, + "lr": 0.02, + "tokens_per_sec": 671713.0309627261, + "mfu": 6.314759537857205, + "total_flops": 1.6018733149873766e+17 + }, + { + "step": 6500, + "train_loss": 3.1299668550491333, + "val_loss": 3.179553759098053, + "lr": 0.02, + "tokens_per_sec": 750298.6979698391, + "mfu": 7.053541677546817, + "total_flops": 1.6142890126698086e+17 + }, + { + "step": 6550, + "train_loss": 3.1561201214790344, + "val_loss": 3.176179659366608, + "lr": 0.02, + "tokens_per_sec": 680751.6339691349, + "mfu": 6.399731247371044, + "total_flops": 1.6267047103522406e+17 + }, + { + "step": 6600, + "train_loss": 3.204793870449066, + "val_loss": 3.174552595615387, + "lr": 0.02, + "tokens_per_sec": 711036.3609291689, + "mfu": 6.684437304283743, + "total_flops": 1.6391204080346726e+17 + }, + { + "step": 6650, + "train_loss": 3.143680512905121, + "val_loss": 3.178517699241638, + "lr": 0.02, + "tokens_per_sec": 716431.4505102517, + "mfu": 6.735156423638787, + "total_flops": 1.6515361057171046e+17 + }, + { + "step": 6700, + "train_loss": 3.075408637523651, + "val_loss": 3.1751816272735596, + "lr": 0.02, + "tokens_per_sec": 704070.4786472794, + "mfu": 6.618951197045199, + "total_flops": 1.6639518033995366e+17 + }, + { + "step": 6750, + "train_loss": 3.104284882545471, + "val_loss": 3.1782999992370606, + "lr": 0.02, + "tokens_per_sec": 696606.8731664469, + "mfu": 6.548786004880711, + "total_flops": 1.6763675010819686e+17 + }, + { + "step": 6800, + "train_loss": 3.0922266840934753, + "val_loss": 3.1723532557487486, + "lr": 0.02, + "tokens_per_sec": 769887.0884655679, + "mfu": 7.237691708903047, + "total_flops": 1.6887831987644006e+17 + }, + { + "step": 6850, + "train_loss": 3.156929671764374, + "val_loss": 3.17406679391861, + "lr": 0.02, + "tokens_per_sec": 674543.6381873306, + "mfu": 6.34137001457206, + "total_flops": 1.7011988964468326e+17 + }, + { + "step": 6900, + "train_loss": 3.13032203912735, + "val_loss": 3.181139957904816, + "lr": 0.02, + "tokens_per_sec": 711158.9804319512, + "mfu": 6.685590047552117, + "total_flops": 1.7136145941292646e+17 + }, + { + "step": 6950, + "train_loss": 3.0561442375183105, + "val_loss": 3.171230399608612, + "lr": 0.02, + "tokens_per_sec": 657218.3994794192, + "mfu": 6.178495823759307, + "total_flops": 1.7260302918116966e+17 + }, + { + "step": 7000, + "train_loss": 3.1265305876731873, + "val_loss": 3.170287549495697, + "lr": 0.02, + "tokens_per_sec": 705210.176723673, + "mfu": 6.6296654737771386, + "total_flops": 1.7384459894941286e+17 + }, + { + "step": 7050, + "train_loss": 3.104595959186554, + "val_loss": 3.1692514777183534, + "lr": 0.02, + "tokens_per_sec": 674145.6150938105, + "mfu": 6.337628208160338, + "total_flops": 1.7508616871765606e+17 + }, + { + "step": 7100, + "train_loss": 3.152307689189911, + "val_loss": 3.177652657032013, + "lr": 0.02, + "tokens_per_sec": 659974.2035204765, + "mfu": 6.204403077378892, + "total_flops": 1.7632773848589926e+17 + }, + { + "step": 7150, + "train_loss": 3.120695650577545, + "val_loss": 3.166973686218262, + "lr": 0.02, + "tokens_per_sec": 756319.1076230536, + "mfu": 7.110139417246694, + "total_flops": 1.7756930825414246e+17 + }, + { + "step": 7200, + "train_loss": 3.116680860519409, + "val_loss": 3.1675556182861326, + "lr": 0.02, + "tokens_per_sec": 736554.6306746806, + "mfu": 6.924334000993837, + "total_flops": 1.7881087802238566e+17 + }, + { + "step": 7250, + "train_loss": 3.1216148138046265, + "val_loss": 3.176751661300659, + "lr": 0.02, + "tokens_per_sec": 768497.8671271454, + "mfu": 7.224631669432881, + "total_flops": 1.8005244779062886e+17 + }, + { + "step": 7300, + "train_loss": 3.1001328229904175, + "val_loss": 3.1768130898475646, + "lr": 0.02, + "tokens_per_sec": 701846.6496135799, + "mfu": 6.598045028854608, + "total_flops": 1.8129401755887206e+17 + }, + { + "step": 7350, + "train_loss": 3.0882458090782166, + "val_loss": 3.1811372876167296, + "lr": 0.02, + "tokens_per_sec": 664789.1802584612, + "mfu": 6.249668568562189, + "total_flops": 1.8253558732711526e+17 + }, + { + "step": 7400, + "train_loss": 3.0628254413604736, + "val_loss": 3.167879915237427, + "lr": 0.02, + "tokens_per_sec": 674304.7162599304, + "mfu": 6.339123914749206, + "total_flops": 1.8377715709535846e+17 + }, + { + "step": 7450, + "train_loss": 3.06341552734375, + "val_loss": 3.172988545894623, + "lr": 0.02, + "tokens_per_sec": 679328.3677658916, + "mfu": 6.386351152870034, + "total_flops": 1.8501872686360166e+17 + }, + { + "step": 7500, + "train_loss": 3.0606775879859924, + "val_loss": 3.1777786135673525, + "lr": 0.02, + "tokens_per_sec": 782756.6463094545, + "mfu": 7.358678141198712, + "total_flops": 1.8626029663184486e+17 + }, + { + "step": 7550, + "train_loss": 3.0959790349006653, + "val_loss": 3.180156445503235, + "lr": 0.02, + "tokens_per_sec": 710747.656465127, + "mfu": 6.681723199920876, + "total_flops": 1.8750186640008806e+17 + }, + { + "step": 7600, + "train_loss": 3.049330413341522, + "val_loss": 3.1722224354743958, + "lr": 0.02, + "tokens_per_sec": 686499.064916994, + "mfu": 6.453762720222107, + "total_flops": 1.8874343616833126e+17 + }, + { + "step": 7650, + "train_loss": 3.0719059705734253, + "val_loss": 3.167457950115204, + "lr": 0.02, + "tokens_per_sec": 711523.1675032718, + "mfu": 6.68901376225792, + "total_flops": 1.8998500593657446e+17 + }, + { + "step": 7700, + "train_loss": 3.0947948694229126, + "val_loss": 3.1769357442855837, + "lr": 0.02, + "tokens_per_sec": 701649.4700183949, + "mfu": 6.596191347785505, + "total_flops": 1.9122657570481766e+17 + }, + { + "step": 7750, + "train_loss": 3.044970750808716, + "val_loss": 3.183099925518036, + "lr": 0.02, + "tokens_per_sec": 697595.9668251104, + "mfu": 6.558084452769876, + "total_flops": 1.9246814547306086e+17 + }, + { + "step": 7800, + "train_loss": 3.085705578327179, + "val_loss": 3.1682560086250304, + "lr": 0.02, + "tokens_per_sec": 690917.0614373354, + "mfu": 6.495296209047032, + "total_flops": 1.9370971524130406e+17 + }, + { + "step": 7850, + "train_loss": 3.047266960144043, + "val_loss": 3.1687360525131227, + "lr": 0.02, + "tokens_per_sec": 718745.4816144382, + "mfu": 6.756910579524549, + "total_flops": 1.9495128500954726e+17 + }, + { + "step": 7900, + "train_loss": 3.1841835379600525, + "val_loss": 3.169699013233185, + "lr": 0.02, + "tokens_per_sec": 705554.4630634915, + "mfu": 6.632902102140603, + "total_flops": 1.9619285477779046e+17 + }, + { + "step": 7950, + "train_loss": 3.1347849369049072, + "val_loss": 3.1597500681877135, + "lr": 0.02, + "tokens_per_sec": 752479.9128405473, + "mfu": 7.074047230921565, + "total_flops": 1.9743442454603366e+17 + }, + { + "step": 8000, + "train_loss": 3.133264899253845, + "val_loss": 3.169278311729431, + "lr": 0.02, + "tokens_per_sec": 714196.2511140726, + "mfu": 6.7141433629187866, + "total_flops": 1.9867599431427686e+17 + }, + { + "step": 8050, + "train_loss": 3.0986027121543884, + "val_loss": 3.1637734532356263, + "lr": 0.02, + "tokens_per_sec": 717463.4166620161, + "mfu": 6.744857914899511, + "total_flops": 1.9991756408252006e+17 + }, + { + "step": 8100, + "train_loss": 3.2134180665016174, + "val_loss": 3.157232141494751, + "lr": 0.02, + "tokens_per_sec": 688577.3122537801, + "mfu": 6.4733002780586535, + "total_flops": 2.0115913385076326e+17 + }, + { + "step": 8150, + "train_loss": 3.1356531381607056, + "val_loss": 3.156856894493103, + "lr": 0.02, + "tokens_per_sec": 707653.2779380869, + "mfu": 6.6526330149510216, + "total_flops": 2.0240070361900646e+17 + }, + { + "step": 8200, + "train_loss": 3.1011769771575928, + "val_loss": 3.1655483722686766, + "lr": 0.02, + "tokens_per_sec": 705070.7567580029, + "mfu": 6.6283547897239785, + "total_flops": 2.0364227338724966e+17 + }, + { + "step": 8250, + "train_loss": 3.043505847454071, + "val_loss": 3.1673601865768433, + "lr": 0.02, + "tokens_per_sec": 712346.4170891232, + "mfu": 6.696753113639627, + "total_flops": 2.0488384315549286e+17 + }, + { + "step": 8300, + "train_loss": 3.0787187814712524, + "val_loss": 3.167344641685486, + "lr": 0.02, + "tokens_per_sec": 726796.4044192191, + "mfu": 6.832597129027953, + "total_flops": 2.0612541292373606e+17 + }, + { + "step": 8350, + "train_loss": 3.04863178730011, + "val_loss": 3.163574528694153, + "lr": 0.02, + "tokens_per_sec": 684120.0113477479, + "mfu": 6.431397289562027, + "total_flops": 2.0736698269197926e+17 + }, + { + "step": 8400, + "train_loss": 3.118406891822815, + "val_loss": 3.1581183552742003, + "lr": 0.02, + "tokens_per_sec": 714684.0160097927, + "mfu": 6.718728830053563, + "total_flops": 2.0860855246022246e+17 + }, + { + "step": 8450, + "train_loss": 3.094644069671631, + "val_loss": 3.1536608576774596, + "lr": 0.02, + "tokens_per_sec": 757979.1997237524, + "mfu": 7.125745906838811, + "total_flops": 2.0985012222846566e+17 + }, + { + "step": 8500, + "train_loss": 3.061720073223114, + "val_loss": 3.147541642189026, + "lr": 0.02, + "tokens_per_sec": 727306.0241525556, + "mfu": 6.837388053013979, + "total_flops": 2.1109169199670886e+17 + }, + { + "step": 8550, + "train_loss": 3.11970978975296, + "val_loss": 3.1589664101600645, + "lr": 0.02, + "tokens_per_sec": 704990.4558965408, + "mfu": 6.627599883078666, + "total_flops": 2.1233326176495206e+17 + }, + { + "step": 8600, + "train_loss": 3.0975873470306396, + "val_loss": 3.1601911067962645, + "lr": 0.02, + "tokens_per_sec": 740458.8535430809, + "mfu": 6.961037514934626, + "total_flops": 2.1357483153319526e+17 + }, + { + "step": 8650, + "train_loss": 3.055800497531891, + "val_loss": 3.158919131755829, + "lr": 0.02, + "tokens_per_sec": 422942.260395141, + "mfu": 3.9760709554275504, + "total_flops": 2.1481640130143846e+17 + }, + { + "step": 8700, + "train_loss": 3.0930920839309692, + "val_loss": 3.166372525691986, + "lr": 0.02, + "tokens_per_sec": 737549.19874948, + "mfu": 6.933683913749565, + "total_flops": 2.1605797106968166e+17 + }, + { + "step": 8750, + "train_loss": 3.1593551635742188, + "val_loss": 3.1717628717422484, + "lr": 0.02, + "tokens_per_sec": 763392.7556223898, + "mfu": 7.1766386276419905, + "total_flops": 2.1729954083792486e+17 + }, + { + "step": 8800, + "train_loss": 3.046756327152252, + "val_loss": 3.166207802295685, + "lr": 0.02, + "tokens_per_sec": 742377.0576086775, + "mfu": 6.979070509473116, + "total_flops": 2.1854111060616806e+17 + }, + { + "step": 8850, + "train_loss": 3.075532078742981, + "val_loss": 3.158578038215637, + "lr": 0.02, + "tokens_per_sec": 709000.5443841416, + "mfu": 6.665298637394083, + "total_flops": 2.1978268037441126e+17 + }, + { + "step": 8900, + "train_loss": 3.062712073326111, + "val_loss": 3.1605077266693113, + "lr": 0.02, + "tokens_per_sec": 713368.5260956465, + "mfu": 6.7063619380931625, + "total_flops": 2.2102425014265446e+17 + }, + { + "step": 8950, + "train_loss": 3.1221720576286316, + "val_loss": 3.1749680519104, + "lr": 0.02, + "tokens_per_sec": 703279.3709544374, + "mfu": 6.611514010897886, + "total_flops": 2.2226581991089766e+17 + }, + { + "step": 9000, + "train_loss": 3.106475830078125, + "val_loss": 3.1677522897720336, + "lr": 0.02, + "tokens_per_sec": 735564.6727283602, + "mfu": 6.915027427955281, + "total_flops": 2.2350738967914086e+17 + }, + { + "step": 9050, + "train_loss": 3.0899603962898254, + "val_loss": 3.160882556438446, + "lr": 0.02, + "tokens_per_sec": 681783.8537221702, + "mfu": 6.40943512273178, + "total_flops": 2.2474895944738406e+17 + }, + { + "step": 9100, + "train_loss": 3.0604807138442993, + "val_loss": 3.156177818775177, + "lr": 0.02, + "tokens_per_sec": 712196.8495005193, + "mfu": 6.69534703200483, + "total_flops": 2.2599052921562726e+17 + }, + { + "step": 9150, + "train_loss": 3.1883918046951294, + "val_loss": 3.1617040753364565, + "lr": 0.02, + "tokens_per_sec": 757768.7291387941, + "mfu": 7.123767277465046, + "total_flops": 2.2723209898387046e+17 + }, + { + "step": 9200, + "train_loss": 3.1464973092079163, + "val_loss": 3.1746654391288756, + "lr": 0.02, + "tokens_per_sec": 689372.4842663524, + "mfu": 6.480775672787012, + "total_flops": 2.2847366875211366e+17 + }, + { + "step": 9250, + "train_loss": 3.0570514798164368, + "val_loss": 3.1565617084503175, + "lr": 0.02, + "tokens_per_sec": 744331.5530949961, + "mfu": 6.997444678865422, + "total_flops": 2.2971523852035686e+17 + }, + { + "step": 9300, + "train_loss": 3.1816789507865906, + "val_loss": 3.161880099773407, + "lr": 0.02, + "tokens_per_sec": 712816.8001758846, + "mfu": 6.701175174767898, + "total_flops": 2.3095680828860006e+17 + }, + { + "step": 9350, + "train_loss": 3.117702007293701, + "val_loss": 3.1647464752197267, + "lr": 0.02, + "tokens_per_sec": 735756.5505862661, + "mfu": 6.916831267507982, + "total_flops": 2.3219837805684326e+17 + }, + { + "step": 9400, + "train_loss": 3.143872082233429, + "val_loss": 3.16401606798172, + "lr": 0.02, + "tokens_per_sec": 717769.7278227371, + "mfu": 6.747737539433432, + "total_flops": 2.3343994782508646e+17 + }, + { + "step": 9450, + "train_loss": 3.0404112339019775, + "val_loss": 3.158752131462097, + "lr": 0.02, + "tokens_per_sec": 742365.9928105028, + "mfu": 6.978966489547064, + "total_flops": 2.3468151759332966e+17 + }, + { + "step": 9500, + "train_loss": 3.1219199895858765, + "val_loss": 3.1590757966041565, + "lr": 0.02, + "tokens_per_sec": 625727.6886516663, + "mfu": 5.882452338837707, + "total_flops": 2.3592308736157286e+17 + }, + { + "step": 9550, + "train_loss": 3.1075311303138733, + "val_loss": 3.161842477321625, + "lr": 0.02, + "tokens_per_sec": 721577.0335884815, + "mfu": 6.783529937808249, + "total_flops": 2.3716465712981606e+17 + }, + { + "step": 9600, + "train_loss": 3.1048957109451294, + "val_loss": 3.163857471942902, + "lr": 0.02, + "tokens_per_sec": 806568.6400450236, + "mfu": 7.5825341743942705, + "total_flops": 2.3840622689805926e+17 + }, + { + "step": 9650, + "train_loss": 2.975189983844757, + "val_loss": 3.1734387636184693, + "lr": 0.02, + "tokens_per_sec": 745568.3417033429, + "mfu": 7.009071701568373, + "total_flops": 2.3964779666630246e+17 + }, + { + "step": 9700, + "train_loss": 3.0660216212272644, + "val_loss": 3.171565854549408, + "lr": 0.02, + "tokens_per_sec": 654774.0789957069, + "mfu": 6.1555168202613695, + "total_flops": 2.4088936643454566e+17 + }, + { + "step": 9750, + "train_loss": 3.099682331085205, + "val_loss": 3.159890115261078, + "lr": 0.02, + "tokens_per_sec": 748786.3525892427, + "mfu": 7.03932415164447, + "total_flops": 2.4213093620278886e+17 + }, + { + "step": 9800, + "train_loss": 3.0244967341423035, + "val_loss": 3.1675970196723937, + "lr": 0.02, + "tokens_per_sec": 728149.514021465, + "mfu": 6.8453176828547875, + "total_flops": 2.4337250597103206e+17 + }, + { + "step": 9850, + "train_loss": 3.16311913728714, + "val_loss": 3.1630403995513916, + "lr": 0.02, + "tokens_per_sec": 762955.907496443, + "mfu": 7.172531827948133, + "total_flops": 2.4461407573927526e+17 + }, + { + "step": 9900, + "train_loss": 3.0646278262138367, + "val_loss": 3.166551649570465, + "lr": 0.02, + "tokens_per_sec": 699854.4342206167, + "mfu": 6.579316255443516, + "total_flops": 2.4585564550751846e+17 + }, + { + "step": 9950, + "train_loss": 3.0938143134117126, + "val_loss": 3.170255482196808, + "lr": 0.02, + "tokens_per_sec": 590216.7351573871, + "mfu": 5.548614640386318, + "total_flops": 2.4709721527576166e+17 + }, + { + "step": 10000, + "train_loss": 3.0690277814865112, + "val_loss": 3.1615793347358703, + "lr": 0.02, + "tokens_per_sec": 753254.4144864467, + "mfu": 7.081328303984148, + "total_flops": 2.4833878504400486e+17 + }, + { + "step": 10050, + "train_loss": 3.0185793042182922, + "val_loss": 3.160174882411957, + "lr": 0.02, + "tokens_per_sec": 696998.2878720823, + "mfu": 6.5524656859822175, + "total_flops": 2.4958035481224806e+17 + }, + { + "step": 10100, + "train_loss": 3.115284562110901, + "val_loss": 3.1621490120887756, + "lr": 0.02, + "tokens_per_sec": 571145.0607667634, + "mfu": 5.369322245852169, + "total_flops": 2.5082192458049126e+17 + }, + { + "step": 10150, + "train_loss": 3.1012810468673706, + "val_loss": 3.160170817375183, + "lr": 0.02, + "tokens_per_sec": 703584.5720835248, + "mfu": 6.614383200048655, + "total_flops": 2.5206349434873446e+17 + }, + { + "step": 10200, + "train_loss": 3.054893910884857, + "val_loss": 3.1524683117866514, + "lr": 0.02, + "tokens_per_sec": 733004.6487976534, + "mfu": 6.890960698878425, + "total_flops": 2.5330506411697766e+17 + }, + { + "step": 10250, + "train_loss": 3.079162895679474, + "val_loss": 3.1491398572921754, + "lr": 0.02, + "tokens_per_sec": 737557.201350517, + "mfu": 6.933759146027185, + "total_flops": 2.5454663388522086e+17 + }, + { + "step": 10300, + "train_loss": 3.1038126945495605, + "val_loss": 3.1571540594100953, + "lr": 0.02, + "tokens_per_sec": 720052.58965927, + "mfu": 6.769198673714534, + "total_flops": 2.5578820365346406e+17 + }, + { + "step": 10350, + "train_loss": 3.0655335187911987, + "val_loss": 3.155700421333313, + "lr": 0.02, + "tokens_per_sec": 689044.8419944294, + "mfu": 6.4776955149424245, + "total_flops": 2.5702977342170726e+17 + }, + { + "step": 10400, + "train_loss": 3.1391899585723877, + "val_loss": 3.154104471206665, + "lr": 0.02, + "tokens_per_sec": 731350.8642784627, + "mfu": 6.875413506722346, + "total_flops": 2.5827134318995046e+17 + }, + { + "step": 10450, + "train_loss": 3.118773400783539, + "val_loss": 3.1520774245262144, + "lr": 0.02, + "tokens_per_sec": 719848.7605407503, + "mfu": 6.767282480621748, + "total_flops": 2.5951291295819366e+17 + }, + { + "step": 10500, + "train_loss": 3.1492371559143066, + "val_loss": 3.152772569656372, + "lr": 0.02, + "tokens_per_sec": 367770.51127847325, + "mfu": 3.4574025466051, + "total_flops": 2.6075448272643686e+17 + }, + { + "step": 10550, + "train_loss": 3.1020652055740356, + "val_loss": 3.1449167728424072, + "lr": 0.02, + "tokens_per_sec": 751263.41527877, + "mfu": 7.06261096390438, + "total_flops": 2.6199605249468006e+17 + }, + { + "step": 10600, + "train_loss": 3.041691780090332, + "val_loss": 3.1420316219329836, + "lr": 0.02, + "tokens_per_sec": 741753.7205190589, + "mfu": 6.973210531103592, + "total_flops": 2.6323762226292326e+17 + }, + { + "step": 10650, + "train_loss": 3.117504835128784, + "val_loss": 3.1396547675132753, + "lr": 0.02, + "tokens_per_sec": 773011.1429140648, + "mfu": 7.267060876562572, + "total_flops": 2.6447919203116646e+17 + }, + { + "step": 10700, + "train_loss": 3.0745006799697876, + "val_loss": 3.1497706532478333, + "lr": 0.02, + "tokens_per_sec": 733934.356207618, + "mfu": 6.899700857940215, + "total_flops": 2.6572076179940966e+17 + }, + { + "step": 10750, + "train_loss": 3.128191292285919, + "val_loss": 3.1601159811019897, + "lr": 0.02, + "tokens_per_sec": 714878.0629070438, + "mfu": 6.720553060697789, + "total_flops": 2.6696233156765286e+17 + }, + { + "step": 10800, + "train_loss": 2.9438751339912415, + "val_loss": 3.1561400055885316, + "lr": 0.02, + "tokens_per_sec": 698684.514108107, + "mfu": 6.568317862009919, + "total_flops": 2.6820390133589606e+17 + }, + { + "step": 10850, + "train_loss": 3.1292802691459656, + "val_loss": 3.1578634858131407, + "lr": 0.02, + "tokens_per_sec": 732158.2530349453, + "mfu": 6.883003750793463, + "total_flops": 2.6944547110413926e+17 + }, + { + "step": 10900, + "train_loss": 3.072293758392334, + "val_loss": 3.1496147394180296, + "lr": 0.02, + "tokens_per_sec": 744635.3570570046, + "mfu": 7.0003007332789915, + "total_flops": 2.7068704087238246e+17 + }, + { + "step": 10950, + "train_loss": 3.1238223910331726, + "val_loss": 3.1552544713020323, + "lr": 0.02, + "tokens_per_sec": 773699.7554751125, + "mfu": 7.27353450821377, + "total_flops": 2.7192861064062566e+17 + }, + { + "step": 11000, + "train_loss": 3.1469348073005676, + "val_loss": 3.160993766784668, + "lr": 0.02, + "tokens_per_sec": 721327.0859497564, + "mfu": 6.781180185513967, + "total_flops": 2.7317018040886886e+17 + }, + { + "step": 11050, + "train_loss": 3.070757746696472, + "val_loss": 3.1459980487823485, + "lr": 0.02, + "tokens_per_sec": 711585.7360416645, + "mfu": 6.689601967721225, + "total_flops": 2.7441175017711206e+17 + }, + { + "step": 11100, + "train_loss": 3.065983474254608, + "val_loss": 3.1586222529411314, + "lr": 0.02, + "tokens_per_sec": 715606.969297743, + "mfu": 6.727405493761767, + "total_flops": 2.7565331994535526e+17 + }, + { + "step": 11150, + "train_loss": 3.1281445622444153, + "val_loss": 3.1559403777122497, + "lr": 0.02, + "tokens_per_sec": 697343.4573019198, + "mfu": 6.555710616255669, + "total_flops": 2.7689488971359846e+17 + }, + { + "step": 11200, + "train_loss": 3.0846157670021057, + "val_loss": 3.151843547821045, + "lr": 0.02, + "tokens_per_sec": 698598.7589954758, + "mfu": 6.567511680068455, + "total_flops": 2.7813645948184166e+17 + }, + { + "step": 11250, + "train_loss": 3.005524456501007, + "val_loss": 3.1599945425987244, + "lr": 0.02, + "tokens_per_sec": 644463.4567446798, + "mfu": 6.058586885602221, + "total_flops": 2.7937802925008486e+17 + }, + { + "step": 11300, + "train_loss": 3.1120718121528625, + "val_loss": 3.1506922483444213, + "lr": 0.02, + "tokens_per_sec": 703597.1449334478, + "mfu": 6.614501397136266, + "total_flops": 2.8061959901832806e+17 + }, + { + "step": 11350, + "train_loss": 3.0928125381469727, + "val_loss": 3.171697747707367, + "lr": 0.02, + "tokens_per_sec": 790605.0705730701, + "mfu": 7.432461006337987, + "total_flops": 2.8186116878657126e+17 + }, + { + "step": 11400, + "train_loss": 3.0579270720481873, + "val_loss": 3.155070924758911, + "lr": 0.02, + "tokens_per_sec": 674691.2002795199, + "mfu": 6.342757242578852, + "total_flops": 2.8310273855481446e+17 + }, + { + "step": 11450, + "train_loss": 3.107338547706604, + "val_loss": 3.154102897644043, + "lr": 0.02, + "tokens_per_sec": 749453.6631067256, + "mfu": 7.045597523249228, + "total_flops": 2.8434430832305766e+17 + }, + { + "step": 11500, + "train_loss": 3.034080922603607, + "val_loss": 3.1530386328697206, + "lr": 0.02, + "tokens_per_sec": 711996.9935045289, + "mfu": 6.693468190150191, + "total_flops": 2.8558587809130086e+17 + }, + { + "step": 11550, + "train_loss": 3.0200106501579285, + "val_loss": 3.163648450374603, + "lr": 0.02, + "tokens_per_sec": 699272.63712686, + "mfu": 6.5738467936681015, + "total_flops": 2.8682744785954406e+17 + }, + { + "step": 11600, + "train_loss": 3.127845346927643, + "val_loss": 3.164807939529419, + "lr": 0.02, + "tokens_per_sec": 705079.2456450074, + "mfu": 6.628434593565359, + "total_flops": 2.8806901762778726e+17 + }, + { + "step": 11650, + "train_loss": 3.0393056869506836, + "val_loss": 3.160170352458954, + "lr": 0.02, + "tokens_per_sec": 716858.8806831492, + "mfu": 6.739174685361659, + "total_flops": 2.8931058739603046e+17 + }, + { + "step": 11700, + "train_loss": 3.0907124280929565, + "val_loss": 3.155959928035736, + "lr": 0.02, + "tokens_per_sec": 749071.7197254659, + "mfu": 7.042006881861382, + "total_flops": 2.9055215716427366e+17 + }, + { + "step": 11750, + "train_loss": 3.0880022048950195, + "val_loss": 3.150728130340576, + "lr": 0.02, + "tokens_per_sec": 694510.9299758312, + "mfu": 6.529082088708685, + "total_flops": 2.9179372693251686e+17 + }, + { + "step": 11800, + "train_loss": 3.127938985824585, + "val_loss": 3.1598729610443117, + "lr": 0.02, + "tokens_per_sec": 775348.4312721618, + "mfu": 7.289033673384543, + "total_flops": 2.9303529670076006e+17 + }, + { + "step": 11850, + "train_loss": 3.1025193333625793, + "val_loss": 3.1526888370513917, + "lr": 0.02, + "tokens_per_sec": 299671.53479248524, + "mfu": 2.8172055555375426, + "total_flops": 2.9427686646900326e+17 + }, + { + "step": 11900, + "train_loss": 3.1594666242599487, + "val_loss": 3.1516989946365355, + "lr": 0.02, + "tokens_per_sec": 833475.199704364, + "mfu": 7.835482154272261, + "total_flops": 2.9551843623724646e+17 + }, + { + "step": 11950, + "train_loss": 3.0817718505859375, + "val_loss": 3.152645695209503, + "lr": 0.02, + "tokens_per_sec": 749447.2775898139, + "mfu": 7.045537493144163, + "total_flops": 2.9676000600548966e+17 + }, + { + "step": 12000, + "train_loss": 3.145545721054077, + "val_loss": 3.151441919803619, + "lr": 0.02, + "tokens_per_sec": 747379.828585372, + "mfu": 7.026101450194214, + "total_flops": 2.9800157577373286e+17 + }, + { + "step": 12050, + "train_loss": 3.094115436077118, + "val_loss": 3.1536296248435973, + "lr": 0.02, + "tokens_per_sec": 615306.9817176713, + "mfu": 5.784487500477572, + "total_flops": 2.9924314554197606e+17 + }, + { + "step": 12100, + "train_loss": 3.088757812976837, + "val_loss": 3.147271144390106, + "lr": 0.02, + "tokens_per_sec": 704596.9580772287, + "mfu": 6.6239006186709695, + "total_flops": 3.0048471531021926e+17 + }, + { + "step": 12150, + "train_loss": 3.090001881122589, + "val_loss": 3.1499510526657106, + "lr": 0.02, + "tokens_per_sec": 698254.1863021854, + "mfu": 6.564272359702253, + "total_flops": 3.0172628507846246e+17 + }, + { + "step": 12200, + "train_loss": 3.1208192706108093, + "val_loss": 3.155332291126251, + "lr": 0.02, + "tokens_per_sec": 720833.5932935321, + "mfu": 6.77654087182775, + "total_flops": 3.0296785484670566e+17 + }, + { + "step": 12250, + "train_loss": 3.0598087310791016, + "val_loss": 3.1506024837493896, + "lr": 0.02, + "tokens_per_sec": 652741.7328201098, + "mfu": 6.136410778238953, + "total_flops": 3.0420942461494886e+17 + }, + { + "step": 12300, + "train_loss": 3.106735408306122, + "val_loss": 3.146341931819916, + "lr": 0.02, + "tokens_per_sec": 709632.3556119555, + "mfu": 6.671238280951758, + "total_flops": 3.0545099438319206e+17 + }, + { + "step": 12350, + "train_loss": 3.059627652168274, + "val_loss": 3.147243654727936, + "lr": 0.02, + "tokens_per_sec": 721662.8154768226, + "mfu": 6.784336371467576, + "total_flops": 3.0669256415143526e+17 + }, + { + "step": 12400, + "train_loss": 3.074007570743561, + "val_loss": 3.1499706149101256, + "lr": 0.02, + "tokens_per_sec": 635126.2628783196, + "mfu": 5.97080812993343, + "total_flops": 3.0793413391967846e+17 + }, + { + "step": 12450, + "train_loss": 3.0849021077156067, + "val_loss": 3.155875825881958, + "lr": 0.02, + "tokens_per_sec": 378615.5670563863, + "mfu": 3.55935667918165, + "total_flops": 3.0917570368792166e+17 + }, + { + "step": 12500, + "train_loss": 3.0780173540115356, + "val_loss": 3.1482291340827944, + "lr": 0.02, + "tokens_per_sec": 674222.0916978031, + "mfu": 6.338347163044679, + "total_flops": 3.1041727345616486e+17 + }, + { + "step": 12550, + "train_loss": 3.05746990442276, + "val_loss": 3.146533954143524, + "lr": 0.02, + "tokens_per_sec": 718523.8440063655, + "mfu": 6.754826969210305, + "total_flops": 3.1165884322440806e+17 + }, + { + "step": 12600, + "train_loss": 3.0032397508621216, + "val_loss": 3.15649356842041, + "lr": 0.02, + "tokens_per_sec": 715966.0417387835, + "mfu": 6.730781125939977, + "total_flops": 3.1290041299265126e+17 + }, + { + "step": 12650, + "train_loss": 3.0083073377609253, + "val_loss": 3.1501465797424317, + "lr": 0.02, + "tokens_per_sec": 726098.2613042835, + "mfu": 6.826033900847742, + "total_flops": 3.1414198276089446e+17 + }, + { + "step": 12700, + "train_loss": 3.0275267362594604, + "val_loss": 3.148030436038971, + "lr": 0.02, + "tokens_per_sec": 692495.3758321449, + "mfu": 6.510133908211622, + "total_flops": 3.1538355252913766e+17 + }, + { + "step": 12750, + "train_loss": 3.120089828968048, + "val_loss": 3.147168529033661, + "lr": 0.02, + "tokens_per_sec": 734716.3837822445, + "mfu": 6.907052682094441, + "total_flops": 3.1662512229738086e+17 + }, + { + "step": 12800, + "train_loss": 3.1195261478424072, + "val_loss": 3.1519615650177, + "lr": 0.02, + "tokens_per_sec": 734447.1703783188, + "mfu": 6.904521812762156, + "total_flops": 3.1786669206562406e+17 + }, + { + "step": 12850, + "train_loss": 3.092920422554016, + "val_loss": 3.1470771193504334, + "lr": 0.02, + "tokens_per_sec": 722598.539243233, + "mfu": 6.793133090164952, + "total_flops": 3.1910826183386726e+17 + }, + { + "step": 12900, + "train_loss": 2.993108630180359, + "val_loss": 3.1456865310668944, + "lr": 0.02, + "tokens_per_sec": 697838.9152378561, + "mfu": 6.560368405493547, + "total_flops": 3.2034983160211046e+17 + }, + { + "step": 12950, + "train_loss": 2.9819764494895935, + "val_loss": 3.1511509299278258, + "lr": 0.02, + "tokens_per_sec": 731947.2335040426, + "mfu": 6.88101996079086, + "total_flops": 3.2159140137035366e+17 + }, + { + "step": 13000, + "train_loss": 3.0769078731536865, + "val_loss": 3.1605732798576356, + "lr": 0.02, + "tokens_per_sec": 712998.6797673736, + "mfu": 6.702885020836316, + "total_flops": 3.2283297113859686e+17 + }, + { + "step": 13050, + "train_loss": 3.067193925380707, + "val_loss": 3.145203471183777, + "lr": 0.02, + "tokens_per_sec": 741974.4137104355, + "mfu": 6.975285262977095, + "total_flops": 3.2407454090684006e+17 + }, + { + "step": 13100, + "train_loss": 3.037872314453125, + "val_loss": 3.14087290763855, + "lr": 0.02, + "tokens_per_sec": 719613.4782381974, + "mfu": 6.765070596832616, + "total_flops": 3.2531611067508326e+17 + }, + { + "step": 13150, + "train_loss": 3.0800358653068542, + "val_loss": 3.149751353263855, + "lr": 0.02, + "tokens_per_sec": 806705.9466868389, + "mfu": 7.583824991136015, + "total_flops": 3.2655768044332646e+17 + }, + { + "step": 13200, + "train_loss": 3.0346257090568542, + "val_loss": 3.149294650554657, + "lr": 0.02, + "tokens_per_sec": 759897.65491187, + "mfu": 7.143781262174636, + "total_flops": 3.2779925021156966e+17 + }, + { + "step": 13250, + "train_loss": 3.115667760372162, + "val_loss": 3.141120672225952, + "lr": 0.02, + "tokens_per_sec": 760821.1147015138, + "mfu": 7.1524626874942, + "total_flops": 3.2904081997981286e+17 + }, + { + "step": 13300, + "train_loss": 3.044420897960663, + "val_loss": 3.1427155256271364, + "lr": 0.02, + "tokens_per_sec": 652259.2522328668, + "mfu": 6.1318749887727915, + "total_flops": 3.3028238974805606e+17 + }, + { + "step": 13350, + "train_loss": 3.0882224440574646, + "val_loss": 3.1480340123176576, + "lr": 0.02, + "tokens_per_sec": 612838.8821541268, + "mfu": 5.761284950369327, + "total_flops": 3.3152395951629926e+17 + }, + { + "step": 13400, + "train_loss": 3.0521661043167114, + "val_loss": 3.15398633480072, + "lr": 0.02, + "tokens_per_sec": 689524.5090688368, + "mfu": 6.4822048546358495, + "total_flops": 3.3276552928454246e+17 + }, + { + "step": 13450, + "train_loss": 3.0448684692382812, + "val_loss": 3.148828995227814, + "lr": 0.02, + "tokens_per_sec": 685100.4687505038, + "mfu": 6.440614548197968, + "total_flops": 3.3400709905278566e+17 + }, + { + "step": 13500, + "train_loss": 3.056766629219055, + "val_loss": 3.146988809108734, + "lr": 0.02, + "tokens_per_sec": 613608.7624056519, + "mfu": 5.768522577804301, + "total_flops": 3.3524866882102886e+17 + }, + { + "step": 13550, + "train_loss": 3.039847433567047, + "val_loss": 3.150434172153473, + "lr": 0.02, + "tokens_per_sec": 690282.3138646273, + "mfu": 6.48932896097512, + "total_flops": 3.3649023858927206e+17 + }, + { + "step": 13600, + "train_loss": 3.047086179256439, + "val_loss": 3.1425138473510743, + "lr": 0.02, + "tokens_per_sec": 703774.056657341, + "mfu": 6.616164540389883, + "total_flops": 3.3773180835751526e+17 + }, + { + "step": 13650, + "train_loss": 3.163117825984955, + "val_loss": 3.153716003894806, + "lr": 0.02, + "tokens_per_sec": 562963.0033541002, + "mfu": 5.292402902763263, + "total_flops": 3.3897337812575846e+17 + }, + { + "step": 13700, + "train_loss": 3.0215771198272705, + "val_loss": 3.1455086708068847, + "lr": 0.02, + "tokens_per_sec": 726098.8486954446, + "mfu": 6.826039422899221, + "total_flops": 3.4021494789400166e+17 + }, + { + "step": 13750, + "train_loss": 3.0746082067489624, + "val_loss": 3.1506766438484193, + "lr": 0.02, + "tokens_per_sec": 387258.1708015886, + "mfu": 3.6406056082871583, + "total_flops": 3.4145651766224486e+17 + }, + { + "step": 13800, + "train_loss": 3.1453649401664734, + "val_loss": 3.1505769014358522, + "lr": 0.02, + "tokens_per_sec": 740524.9813837727, + "mfu": 6.961659181321131, + "total_flops": 3.4269808743048806e+17 + }, + { + "step": 13850, + "train_loss": 3.104449152946472, + "val_loss": 3.152713358402252, + "lr": 0.02, + "tokens_per_sec": 687207.8565221422, + "mfu": 6.460426054626306, + "total_flops": 3.4393965719873126e+17 + }, + { + "step": 13900, + "train_loss": 3.1265454292297363, + "val_loss": 3.1454100966453553, + "lr": 0.02, + "tokens_per_sec": 769994.2727430232, + "mfu": 7.238699345955137, + "total_flops": 3.4518122696697446e+17 + }, + { + "step": 13950, + "train_loss": 3.035269320011139, + "val_loss": 3.1418202042579653, + "lr": 0.02, + "tokens_per_sec": 677319.992196272, + "mfu": 6.367470457991035, + "total_flops": 3.4642279673521766e+17 + }, + { + "step": 14000, + "train_loss": 3.102819263935089, + "val_loss": 3.14750554561615, + "lr": 0.02, + "tokens_per_sec": 705157.8708164443, + "mfu": 6.629173747085332, + "total_flops": 3.4766436650346086e+17 + }, + { + "step": 14050, + "train_loss": 3.034712076187134, + "val_loss": 3.1495706915855406, + "lr": 0.02, + "tokens_per_sec": 736450.2024026074, + "mfu": 6.923352273088169, + "total_flops": 3.4890593627170406e+17 + }, + { + "step": 14100, + "train_loss": 3.10324889421463, + "val_loss": 3.149154031276703, + "lr": 0.02, + "tokens_per_sec": 680789.2004978532, + "mfu": 6.400084409487294, + "total_flops": 3.5014750603994726e+17 + }, + { + "step": 14150, + "train_loss": 3.1322348713874817, + "val_loss": 3.1461058974266054, + "lr": 0.02, + "tokens_per_sec": 749168.8854789592, + "mfu": 7.042920334988449, + "total_flops": 3.5138907580819046e+17 + }, + { + "step": 14200, + "train_loss": 3.097643196582794, + "val_loss": 3.151403570175171, + "lr": 0.02, + "tokens_per_sec": 681189.7392139593, + "mfu": 6.40384986521202, + "total_flops": 3.5263064557643366e+17 + }, + { + "step": 14250, + "train_loss": 3.064367115497589, + "val_loss": 3.147836744785309, + "lr": 0.02, + "tokens_per_sec": 756080.3892174037, + "mfu": 7.107895230727371, + "total_flops": 3.5387221534467686e+17 + }, + { + "step": 14300, + "train_loss": 3.1396385431289673, + "val_loss": 3.1502734065055846, + "lr": 0.02, + "tokens_per_sec": 724403.9458191768, + "mfu": 6.810105677966042, + "total_flops": 3.5511378511292006e+17 + }, + { + "step": 14350, + "train_loss": 3.0647523403167725, + "val_loss": 3.149309492111206, + "lr": 0.02, + "tokens_per_sec": 722727.958460705, + "mfu": 6.794349757402658, + "total_flops": 3.5635535488116326e+17 + }, + { + "step": 14400, + "train_loss": 3.0957531929016113, + "val_loss": 3.152403235435486, + "lr": 0.02, + "tokens_per_sec": 766725.8902000515, + "mfu": 7.207973353030722, + "total_flops": 3.5759692464940646e+17 + }, + { + "step": 14450, + "train_loss": 3.0434666872024536, + "val_loss": 3.145326030254364, + "lr": 0.02, + "tokens_per_sec": 722017.6919229714, + "mfu": 6.787672557189428, + "total_flops": 3.5883849441764966e+17 + }, + { + "step": 14500, + "train_loss": 3.0627703070640564, + "val_loss": 3.1456138014793398, + "lr": 0.02, + "tokens_per_sec": 738478.2262390595, + "mfu": 6.94241768089466, + "total_flops": 3.6008006418589286e+17 + }, + { + "step": 14550, + "train_loss": 3.0802829265594482, + "val_loss": 3.1429310202598573, + "lr": 0.02, + "tokens_per_sec": 702269.6436384912, + "mfu": 6.602021586447123, + "total_flops": 3.6132163395413606e+17 + }, + { + "step": 14600, + "train_loss": 3.1189873218536377, + "val_loss": 3.1505560874938965, + "lr": 0.02, + "tokens_per_sec": 696383.4380146917, + "mfu": 6.546685495897557, + "total_flops": 3.6256320372237926e+17 + }, + { + "step": 14650, + "train_loss": 3.066819429397583, + "val_loss": 3.143692100048065, + "lr": 0.02, + "tokens_per_sec": 732030.1378021343, + "mfu": 6.881799342287081, + "total_flops": 3.6380477349062246e+17 + }, + { + "step": 14700, + "train_loss": 3.0843369364738464, + "val_loss": 3.145070564746857, + "lr": 0.02, + "tokens_per_sec": 719142.5045767998, + "mfu": 6.760642983725084, + "total_flops": 3.6504634325886566e+17 + }, + { + "step": 14750, + "train_loss": 3.039639413356781, + "val_loss": 3.1410356044769285, + "lr": 0.02, + "tokens_per_sec": 766536.2919992043, + "mfu": 7.206190944484257, + "total_flops": 3.6628791302710886e+17 + }, + { + "step": 14800, + "train_loss": 3.0864169001579285, + "val_loss": 3.1432830691337585, + "lr": 0.02, + "tokens_per_sec": 765598.1552002148, + "mfu": 7.197371540920293, + "total_flops": 3.6752948279535206e+17 + }, + { + "step": 14850, + "train_loss": 3.0957528352737427, + "val_loss": 3.149162566661835, + "lr": 0.02, + "tokens_per_sec": 776568.60407019, + "mfu": 7.300504491217478, + "total_flops": 3.6877105256359526e+17 + }, + { + "step": 14900, + "train_loss": 3.0988165736198425, + "val_loss": 3.1451326608657837, + "lr": 0.02, + "tokens_per_sec": 681868.4795820845, + "mfu": 6.410230688593106, + "total_flops": 3.7001262233183846e+17 + }, + { + "step": 14950, + "train_loss": 3.0074263215065002, + "val_loss": 3.139087998867035, + "lr": 0.02, + "tokens_per_sec": 678105.8915915034, + "mfu": 6.374858681046231, + "total_flops": 3.7125419210008166e+17 + }, + { + "step": 15000, + "train_loss": 3.054157257080078, + "val_loss": 3.150736141204834, + "lr": 0.02, + "tokens_per_sec": 719883.2601829135, + "mfu": 6.767606811004456, + "total_flops": 3.7249576186832486e+17 + }, + { + "step": 15050, + "train_loss": 3.0841880440711975, + "val_loss": 3.144093596935272, + "lr": 0.02, + "tokens_per_sec": 761035.6944468117, + "mfu": 7.154479946994598, + "total_flops": 3.7373733163656806e+17 + }, + { + "step": 15100, + "train_loss": 3.1251383423805237, + "val_loss": 3.148873770236969, + "lr": 0.02, + "tokens_per_sec": 746738.0365748014, + "mfu": 7.020067977515766, + "total_flops": 3.7497890140481126e+17 + }, + { + "step": 15150, + "train_loss": 3.043576240539551, + "val_loss": 3.136397635936737, + "lr": 0.02, + "tokens_per_sec": 638864.8231680673, + "mfu": 6.005954253589394, + "total_flops": 3.7622047117305446e+17 + }, + { + "step": 15200, + "train_loss": 3.1045437455177307, + "val_loss": 3.1389421463012694, + "lr": 0.02, + "tokens_per_sec": 733851.3261809672, + "mfu": 6.898920294472552, + "total_flops": 3.7746204094129766e+17 + }, + { + "step": 15250, + "train_loss": 3.0337718725204468, + "val_loss": 3.1406340479850767, + "lr": 0.02, + "tokens_per_sec": 673028.6294144924, + "mfu": 6.327127450177403, + "total_flops": 3.7870361070954086e+17 + }, + { + "step": 15300, + "train_loss": 3.143120229244232, + "val_loss": 3.1418453574180605, + "lr": 0.02, + "tokens_per_sec": 691795.0422515095, + "mfu": 6.503550087511195, + "total_flops": 3.7994518047778406e+17 + }, + { + "step": 15350, + "train_loss": 3.133650004863739, + "val_loss": 3.143596124649048, + "lr": 0.02, + "tokens_per_sec": 687503.2952059626, + "mfu": 6.463203467242276, + "total_flops": 3.8118675024602726e+17 + }, + { + "step": 15400, + "train_loss": 3.029381215572357, + "val_loss": 3.1452716112136843, + "lr": 0.02, + "tokens_per_sec": 641473.306409728, + "mfu": 6.030476547590464, + "total_flops": 3.8242832001427046e+17 + }, + { + "step": 15450, + "train_loss": 3.0394997596740723, + "val_loss": 3.139203870296478, + "lr": 0.02, + "tokens_per_sec": 701028.9046245916, + "mfu": 6.590357426067825, + "total_flops": 3.8366988978251366e+17 + }, + { + "step": 15500, + "train_loss": 3.0841792821884155, + "val_loss": 3.141300177574158, + "lr": 0.02, + "tokens_per_sec": 690827.9719891226, + "mfu": 6.494458681089567, + "total_flops": 3.8491145955075686e+17 + }, + { + "step": 15550, + "train_loss": 3.0710747838020325, + "val_loss": 3.1462770342826842, + "lr": 0.02, + "tokens_per_sec": 692112.1702584228, + "mfu": 6.506531400979995, + "total_flops": 3.8615302931900006e+17 + }, + { + "step": 15600, + "train_loss": 3.0950469970703125, + "val_loss": 3.138612949848175, + "lr": 0.02, + "tokens_per_sec": 770514.9559267653, + "mfu": 7.243594277196867, + "total_flops": 3.8739459908724326e+17 + }, + { + "step": 15650, + "train_loss": 3.068549394607544, + "val_loss": 3.1408026933670046, + "lr": 0.02, + "tokens_per_sec": 753367.3281070498, + "mfu": 7.082389802466077, + "total_flops": 3.8863616885548646e+17 + }, + { + "step": 15700, + "train_loss": 3.131138026714325, + "val_loss": 3.1432745099067687, + "lr": 0.02, + "tokens_per_sec": 768443.0158710635, + "mfu": 7.224116013972079, + "total_flops": 3.8987773862372966e+17 + }, + { + "step": 15750, + "train_loss": 3.085524320602417, + "val_loss": 3.143316113948822, + "lr": 0.02, + "tokens_per_sec": 724972.6384412758, + "mfu": 6.815451944889525, + "total_flops": 3.9111930839197286e+17 + }, + { + "step": 15800, + "train_loss": 3.0742884278297424, + "val_loss": 3.1385276079177857, + "lr": 0.02, + "tokens_per_sec": 696762.4148946902, + "mfu": 6.550248249271819, + "total_flops": 3.9236087816021606e+17 + }, + { + "step": 15850, + "train_loss": 3.0977269411087036, + "val_loss": 3.140676200389862, + "lr": 0.02, + "tokens_per_sec": 687534.4738464371, + "mfu": 6.46349657696062, + "total_flops": 3.9360244792845926e+17 + }, + { + "step": 15900, + "train_loss": 3.0782747864723206, + "val_loss": 3.140700113773346, + "lr": 0.02, + "tokens_per_sec": 715785.0217764467, + "mfu": 6.72907936122645, + "total_flops": 3.9484401769670246e+17 + }, + { + "step": 15950, + "train_loss": 3.100230097770691, + "val_loss": 3.134903037548065, + "lr": 0.02, + "tokens_per_sec": 705794.6056878131, + "mfu": 6.635159677708562, + "total_flops": 3.9608558746494566e+17 + }, + { + "step": 16000, + "train_loss": 3.068934679031372, + "val_loss": 3.1448294639587404, + "lr": 0.02, + "tokens_per_sec": 734678.9750939112, + "mfu": 6.906701003832202, + "total_flops": 3.9732715723318886e+17 + }, + { + "step": 16050, + "train_loss": 3.0025221705436707, + "val_loss": 3.1464810132980348, + "lr": 0.01943255903154212, + "tokens_per_sec": 699908.4735491821, + "mfu": 6.5798242779914595, + "total_flops": 3.9856872700143206e+17 + }, + { + "step": 16100, + "train_loss": 3.070408046245575, + "val_loss": 3.14631609916687, + "lr": 0.018881217525718468, + "tokens_per_sec": 643071.6331094927, + "mfu": 6.0455023819969895, + "total_flops": 3.9981029676967526e+17 + }, + { + "step": 16150, + "train_loss": 3.0675635933876038, + "val_loss": 3.1374461889266967, + "lr": 0.01834551870779559, + "tokens_per_sec": 617864.6048335439, + "mfu": 5.80853165954659, + "total_flops": 4.0105186653791846e+17 + }, + { + "step": 16200, + "train_loss": 3.045469880104065, + "val_loss": 3.1379233717918398, + "lr": 0.01782501876267491, + "tokens_per_sec": 747194.9668446358, + "mfu": 7.0243635690058355, + "total_flops": 4.0229343630616166e+17 + }, + { + "step": 16250, + "train_loss": 3.113093376159668, + "val_loss": 3.1413995742797853, + "lr": 0.01731928646720131, + "tokens_per_sec": 718972.9093617701, + "mfu": 6.759048622811544, + "total_flops": 4.0353500607440486e+17 + }, + { + "step": 16300, + "train_loss": 3.0531409978866577, + "val_loss": 3.1344855070114135, + "lr": 0.016827902832903904, + "tokens_per_sec": 564062.8449018005, + "mfu": 5.302742489139186, + "total_flops": 4.0477657584264806e+17 + }, + { + "step": 16350, + "train_loss": 3.0165595412254333, + "val_loss": 3.1350361824035646, + "lr": 0.016350460758873, + "tokens_per_sec": 708849.9905872872, + "mfu": 6.663883284437057, + "total_flops": 4.0601814561089126e+17 + }, + { + "step": 16400, + "train_loss": 3.020295739173889, + "val_loss": 3.1313610792160036, + "lr": 0.01588656469448563, + "tokens_per_sec": 717004.3346030491, + "mfu": 6.740542094486789, + "total_flops": 4.0725971537913446e+17 + }, + { + "step": 16450, + "train_loss": 3.117129921913147, + "val_loss": 3.1293462038040163, + "lr": 0.01543583031170025, + "tokens_per_sec": 519429.62306803424, + "mfu": 4.883146545204445, + "total_flops": 4.0850128514737766e+17 + }, + { + "step": 16500, + "train_loss": 3.0911515951156616, + "val_loss": 3.130696713924408, + "lr": 0.014997884186649118, + "tokens_per_sec": 712461.4443617964, + "mfu": 6.697834482518515, + "total_flops": 4.0974285491562086e+17 + }, + { + "step": 16550, + "train_loss": 3.041957437992096, + "val_loss": 3.1251978516578673, + "lr": 0.014572363490264556, + "tokens_per_sec": 733594.2115164933, + "mfu": 6.896503165125691, + "total_flops": 4.1098442468386406e+17 + }, + { + "step": 16600, + "train_loss": 3.0678828358650208, + "val_loss": 3.1350746273994448, + "lr": 0.014158915687682759, + "tokens_per_sec": 579977.1700169533, + "mfu": 5.452352712072391, + "total_flops": 4.1222599445210726e+17 + }, + { + "step": 16650, + "train_loss": 3.059101641178131, + "val_loss": 3.1194016695022584, + "lr": 0.013757198246176152, + "tokens_per_sec": 723278.225219649, + "mfu": 6.799522803188933, + "total_flops": 4.1346756422035046e+17 + }, + { + "step": 16700, + "train_loss": 3.0314047932624817, + "val_loss": 3.119983184337616, + "lr": 0.013366878351372293, + "tokens_per_sec": 706245.4855250673, + "mfu": 6.63939838921405, + "total_flops": 4.1470913398859366e+17 + }, + { + "step": 16750, + "train_loss": 3.0364831686019897, + "val_loss": 3.1191796541213987, + "lr": 0.012987632631524226, + "tokens_per_sec": 716420.5737692728, + "mfu": 6.735054171634326, + "total_flops": 4.1595070375683686e+17 + }, + { + "step": 16800, + "train_loss": 3.0998443961143494, + "val_loss": 3.1202172160148622, + "lr": 0.012619146889603866, + "tokens_per_sec": 714708.8932903456, + "mfu": 6.718962700824868, + "total_flops": 4.1719227352508006e+17 + }, + { + "step": 16850, + "train_loss": 3.0862051844596863, + "val_loss": 3.113366889953613, + "lr": 0.012261115842996415, + "tokens_per_sec": 692450.4445954027, + "mfu": 6.50971151063604, + "total_flops": 4.1843384329332326e+17 + }, + { + "step": 16900, + "train_loss": 3.0278157591819763, + "val_loss": 3.10579172372818, + "lr": 0.01191324287058021, + "tokens_per_sec": 729828.1945023012, + "mfu": 6.861098921402715, + "total_flops": 4.1967541306156646e+17 + }, + { + "step": 16950, + "train_loss": 3.1150811314582825, + "val_loss": 3.1045249462127686, + "lr": 0.011575239766982413, + "tokens_per_sec": 759203.8006162018, + "mfu": 7.137258353090717, + "total_flops": 4.2091698282980966e+17 + }, + { + "step": 17000, + "train_loss": 3.0266743302345276, + "val_loss": 3.1035353422164915, + "lr": 0.011246826503806982, + "tokens_per_sec": 756266.9043518367, + "mfu": 7.1096486554337925, + "total_flops": 4.2215855259805286e+17 + }, + { + "step": 17050, + "train_loss": 3.0746100544929504, + "val_loss": 3.1017999410629273, + "lr": 0.010927730997637084, + "tokens_per_sec": 694065.5937923228, + "mfu": 6.5248954929710195, + "total_flops": 4.2340012236629606e+17 + }, + { + "step": 17100, + "train_loss": 3.0298961400985718, + "val_loss": 3.101312255859375, + "lr": 0.010617688884619766, + "tokens_per_sec": 720785.0277298539, + "mfu": 6.776084308023996, + "total_flops": 4.2464169213453926e+17 + }, + { + "step": 17150, + "train_loss": 3.0325640439987183, + "val_loss": 3.101479244232178, + "lr": 0.010316443301446114, + "tokens_per_sec": 729438.6692286865, + "mfu": 6.857437002810383, + "total_flops": 4.2588326190278246e+17 + }, + { + "step": 17200, + "train_loss": 3.0670490860939026, + "val_loss": 3.1049660444259644, + "lr": 0.010023744672545448, + "tokens_per_sec": 740081.2137324706, + "mfu": 6.957487331320491, + "total_flops": 4.2712483167102566e+17 + }, + { + "step": 17250, + "train_loss": 2.9903318881988525, + "val_loss": 3.1029152631759644, + "lr": 0.009739350503317262, + "tokens_per_sec": 671767.1005807547, + "mfu": 6.3152678451557325, + "total_flops": 4.2836640143926886e+17 + }, + { + "step": 17300, + "train_loss": 3.0684115886688232, + "val_loss": 3.1053053736686707, + "lr": 0.00946302517922961, + "tokens_per_sec": 734440.326670526, + "mfu": 6.904457475214622, + "total_flops": 4.2960797120751206e+17 + }, + { + "step": 17350, + "train_loss": 3.0212934017181396, + "val_loss": 3.1042049407958983, + "lr": 0.009194539770617445, + "tokens_per_sec": 725588.0457060201, + "mfu": 6.821237375148546, + "total_flops": 4.3084954097575526e+17 + }, + { + "step": 17400, + "train_loss": 3.019513189792633, + "val_loss": 3.105720841884613, + "lr": 0.008933671843019263, + "tokens_per_sec": 712422.1938231154, + "mfu": 6.697465489061375, + "total_flops": 4.3209111074399846e+17 + }, + { + "step": 17450, + "train_loss": 3.0266613960266113, + "val_loss": 3.10250746011734, + "lr": 0.008680205272894878, + "tokens_per_sec": 695055.7171985689, + "mfu": 6.534203621494723, + "total_flops": 4.3333268051224166e+17 + }, + { + "step": 17500, + "train_loss": 3.087554156780243, + "val_loss": 3.101519727706909, + "lr": 0.008433930068571646, + "tokens_per_sec": 713302.2077206185, + "mfu": 6.705738480497522, + "total_flops": 4.3457425028048486e+17 + }, + { + "step": 17550, + "train_loss": 2.9932942390441895, + "val_loss": 3.0988452196121217, + "lr": 0.008194642196270831, + "tokens_per_sec": 782921.0291510923, + "mfu": 7.360223500703804, + "total_flops": 4.3581582004872806e+17 + }, + { + "step": 17600, + "train_loss": 3.0073635578155518, + "val_loss": 3.09357488155365, + "lr": 0.007962143411069945, + "tokens_per_sec": 736169.624928723, + "mfu": 6.920714570382327, + "total_flops": 4.3705738981697126e+17 + }, + { + "step": 17650, + "train_loss": 3.1045629382133484, + "val_loss": 3.0909008145332337, + "lr": 0.007736241092661044, + "tokens_per_sec": 679335.1463116886, + "mfu": 6.3864148778310685, + "total_flops": 4.3829895958521446e+17 + }, + { + "step": 17700, + "train_loss": 3.031808853149414, + "val_loss": 3.0918583154678343, + "lr": 0.007516748085768884, + "tokens_per_sec": 387732.62828782527, + "mfu": 3.6450659727559356, + "total_flops": 4.3954052935345766e+17 + }, + { + "step": 17750, + "train_loss": 3.082665503025055, + "val_loss": 3.0935746788978578, + "lr": 0.007303482545096754, + "tokens_per_sec": 728887.4120509602, + "mfu": 6.852254646118762, + "total_flops": 4.4078209912170086e+17 + }, + { + "step": 17800, + "train_loss": 3.0321786403656006, + "val_loss": 3.0965789198875426, + "lr": 0.00709626778467151, + "tokens_per_sec": 691484.3255468745, + "mfu": 6.50062904655513, + "total_flops": 4.4202366888994406e+17 + }, + { + "step": 17850, + "train_loss": 3.062150776386261, + "val_loss": 3.0917367458343508, + "lr": 0.0068949321314629875, + "tokens_per_sec": 699998.4053803874, + "mfu": 6.580669725173044, + "total_flops": 4.4326523865818726e+17 + }, + { + "step": 17900, + "train_loss": 2.999358355998993, + "val_loss": 3.0902578592300416, + "lr": 0.006699308783156554, + "tokens_per_sec": 694246.9714867986, + "mfu": 6.526600620716574, + "total_flops": 4.4450680842643046e+17 + }, + { + "step": 17950, + "train_loss": 3.023828446865082, + "val_loss": 3.0946322441101075, + "lr": 0.006509235669960918, + "tokens_per_sec": 715732.6380369696, + "mfu": 6.728586902835374, + "total_flops": 4.4574837819467366e+17 + }, + { + "step": 18000, + "train_loss": 3.040572166442871, + "val_loss": 3.089223766326904, + "lr": 0.006324555320336759, + "tokens_per_sec": 665364.2394382448, + "mfu": 6.255074687355458, + "total_flops": 4.4698994796291686e+17 + }, + { + "step": 18050, + "train_loss": 3.0870242714881897, + "val_loss": 3.090847408771515, + "lr": 0.006145114730534894, + "tokens_per_sec": 766372.5203509134, + "mfu": 7.204651330794475, + "total_flops": 4.4823151773116006e+17 + }, + { + "step": 18100, + "train_loss": 3.063161551952362, + "val_loss": 3.091911458969116, + "lr": 0.005970765237835919, + "tokens_per_sec": 688490.1831549242, + "mfu": 6.472481179302693, + "total_flops": 4.4947308749940326e+17 + }, + { + "step": 18150, + "train_loss": 3.0751683115959167, + "val_loss": 3.0886308670043947, + "lr": 0.005801362397386308, + "tokens_per_sec": 734212.2270090751, + "mfu": 6.902313115277664, + "total_flops": 4.5071465726764646e+17 + }, + { + "step": 18200, + "train_loss": 3.027254581451416, + "val_loss": 3.089142549037933, + "lr": 0.005636765862528908, + "tokens_per_sec": 688938.4922131193, + "mfu": 6.476695722970377, + "total_flops": 4.5195622703588966e+17 + }, + { + "step": 18250, + "train_loss": 3.0408639311790466, + "val_loss": 3.090257740020752, + "lr": 0.0054768392685287225, + "tokens_per_sec": 692402.7830808776, + "mfu": 6.509263445778634, + "total_flops": 4.5319779680413286e+17 + }, + { + "step": 18300, + "train_loss": 3.0395489931106567, + "val_loss": 3.0902520656585692, + "lr": 0.005321450119597621, + "tokens_per_sec": 719781.5263763262, + "mfu": 6.766650413709992, + "total_flops": 4.5443936657237606e+17 + }, + { + "step": 18350, + "train_loss": 3.0594890117645264, + "val_loss": 3.0901105999946594, + "lr": 0.005170469679124381, + "tokens_per_sec": 701543.6814319686, + "mfu": 6.595196831594308, + "total_flops": 4.5568093634061926e+17 + }, + { + "step": 18400, + "train_loss": 3.026129126548767, + "val_loss": 3.0864153623580934, + "lr": 0.005023772863019161, + "tokens_per_sec": 685278.1300899048, + "mfu": 6.442284738570608, + "total_flops": 4.5692250610886246e+17 + }, + { + "step": 18450, + "train_loss": 2.992745041847229, + "val_loss": 3.0890154004096986, + "lr": 0.0048812381360839605, + "tokens_per_sec": 728753.9542899799, + "mfu": 6.851000012621711, + "total_flops": 4.5816407587710566e+17 + }, + { + "step": 18500, + "train_loss": 3.0637913942337036, + "val_loss": 3.0905563712120054, + "lr": 0.004742747411323311, + "tokens_per_sec": 721667.1139879038, + "mfu": 6.784376781676397, + "total_flops": 4.5940564564534886e+17 + }, + { + "step": 18550, + "train_loss": 2.9693461060523987, + "val_loss": 3.09128623008728, + "lr": 0.004608185952111692, + "tokens_per_sec": 662050.8319155766, + "mfu": 6.223925415579997, + "total_flops": 4.6064721541359206e+17 + }, + { + "step": 18600, + "train_loss": 2.983288586139679, + "val_loss": 3.0884334087371825, + "lr": 0.004477442277136679, + "tokens_per_sec": 720536.761625188, + "mfu": 6.773750363794689, + "total_flops": 4.6188878518183526e+17 + }, + { + "step": 18650, + "train_loss": 2.9768336415290833, + "val_loss": 3.092844784259796, + "lr": 0.004350408068039046, + "tokens_per_sec": 729005.233665835, + "mfu": 6.853362284547675, + "total_flops": 4.6313035495007846e+17 + }, + { + "step": 18700, + "train_loss": 3.0071200132369995, + "val_loss": 3.0864765763282778, + "lr": 0.004226978079673293, + "tokens_per_sec": 719243.7284677255, + "mfu": 6.761594587313548, + "total_flops": 4.6437192471832166e+17 + }, + { + "step": 18750, + "train_loss": 2.9823778867721558, + "val_loss": 3.086198890209198, + "lr": 0.004107050052914292, + "tokens_per_sec": 715917.9550265683, + "mfu": 6.7303290638083695, + "total_flops": 4.6561349448656486e+17 + }, + { + "step": 18800, + "train_loss": 3.015293300151825, + "val_loss": 3.0859832167625427, + "lr": 0.0039905246299377604, + "tokens_per_sec": 700826.0866038192, + "mfu": 6.588450738282884, + "total_flops": 4.6685506425480806e+17 + }, + { + "step": 18850, + "train_loss": 3.012860596179962, + "val_loss": 3.0853739976882935, + "lr": 0.0038773052719044145, + "tokens_per_sec": 758256.9238838439, + "mfu": 7.128356785603984, + "total_flops": 4.6809663402305126e+17 + }, + { + "step": 18900, + "train_loss": 3.057746469974518, + "val_loss": 3.0853593587875365, + "lr": 0.003767298178979601, + "tokens_per_sec": 702816.8260277645, + "mfu": 6.607165636141467, + "total_flops": 4.6933820379129446e+17 + }, + { + "step": 18950, + "train_loss": 2.973491668701172, + "val_loss": 3.0860115766525267, + "lr": 0.003660412212622112, + "tokens_per_sec": 717484.8124776696, + "mfu": 6.745059056495312, + "total_flops": 4.7057977355953766e+17 + }, + { + "step": 19000, + "train_loss": 3.0762404203414917, + "val_loss": 3.0846158027648927, + "lr": 0.003556558820077846, + "tokens_per_sec": 684364.6308949208, + "mfu": 6.4336969525839605, + "total_flops": 4.7182134332778086e+17 + }, + { + "step": 19050, + "train_loss": 2.918803095817566, + "val_loss": 3.086686110496521, + "lr": 0.003455651961015727, + "tokens_per_sec": 408454.8850148041, + "mfu": 3.8398754557952426, + "total_flops": 4.7306291309602406e+17 + }, + { + "step": 19100, + "train_loss": 2.989955246448517, + "val_loss": 3.0845702648162843, + "lr": 0.0033576080362451204, + "tokens_per_sec": 685496.9734419179, + "mfu": 6.444342080144644, + "total_flops": 4.7430448286426726e+17 + }, + { + "step": 19150, + "train_loss": 3.026902973651886, + "val_loss": 3.083245599269867, + "lr": 0.003262345818455677, + "tokens_per_sec": 696411.1376441265, + "mfu": 6.546945899509078, + "total_flops": 4.7554605263251046e+17 + }, + { + "step": 19200, + "train_loss": 2.9645996689796448, + "val_loss": 3.0823713302612306, + "lr": 0.003169786384922227, + "tokens_per_sec": 708031.469734792, + "mfu": 6.6561883877743835, + "total_flops": 4.7678762240075366e+17 + }, + { + "step": 19250, + "train_loss": 3.022516131401062, + "val_loss": 3.0823891878128054, + "lr": 0.0030798530521189845, + "tokens_per_sec": 704160.6039027537, + "mfu": 6.619798462604113, + "total_flops": 4.7802919216899686e+17 + }, + { + "step": 19300, + "train_loss": 3.0703765749931335, + "val_loss": 3.0858085632324217, + "lr": 0.0029924713121888677, + "tokens_per_sec": 634319.5635431412, + "mfu": 5.963224367096942, + "total_flops": 4.7927076193724006e+17 + }, + { + "step": 19350, + "train_loss": 3.0681645274162292, + "val_loss": 3.0809340119361877, + "lr": 0.0029075687712153235, + "tokens_per_sec": 740793.2418918696, + "mfu": 6.964181092500533, + "total_flops": 4.8051233170548326e+17 + }, + { + "step": 19400, + "train_loss": 3.0543947219848633, + "val_loss": 3.0824201107025146, + "lr": 0.0028250750892455092, + "tokens_per_sec": 747138.0373387374, + "mfu": 7.023828375963887, + "total_flops": 4.8175390147372646e+17 + }, + { + "step": 19450, + "train_loss": 3.022664248943329, + "val_loss": 3.0830381870269776, + "lr": 0.002744921922015124, + "tokens_per_sec": 763554.831876657, + "mfu": 7.178162302969573, + "total_flops": 4.8299547124196966e+17 + }, + { + "step": 19500, + "train_loss": 3.0661675333976746, + "val_loss": 3.0806783080101012, + "lr": 0.002667042864326648, + "tokens_per_sec": 738727.4103861994, + "mfu": 6.94476025562122, + "total_flops": 4.8423704101021286e+17 + }, + { + "step": 19550, + "train_loss": 3.0482770204544067, + "val_loss": 3.0791345119476317, + "lr": 0.0025913733950340394, + "tokens_per_sec": 706756.9039727594, + "mfu": 6.644206222874462, + "total_flops": 4.8547861077845606e+17 + }, + { + "step": 19600, + "train_loss": 3.0468355417251587, + "val_loss": 3.079674780368805, + "lr": 0.0025178508235883345, + "tokens_per_sec": 786332.9609115301, + "mfu": 7.392298996687352, + "total_flops": 4.8672018054669926e+17 + }, + { + "step": 19650, + "train_loss": 2.9020655751228333, + "val_loss": 3.079472291469574, + "lr": 0.0024464142380998632, + "tokens_per_sec": 707581.8591161177, + "mfu": 6.6519616081650526, + "total_flops": 4.8796175031494246e+17 + }, + { + "step": 19700, + "train_loss": 3.0088971853256226, + "val_loss": 3.076182794570923, + "lr": 0.0023770044548740367, + "tokens_per_sec": 663726.0970363328, + "mfu": 6.2396745464023855, + "total_flops": 4.8920332008318566e+17 + }, + { + "step": 19750, + "train_loss": 3.035418212413788, + "val_loss": 3.0742117404937743, + "lr": 0.0023095639693789167, + "tokens_per_sec": 749332.4079574648, + "mfu": 7.04445760623852, + "total_flops": 4.9044488985142886e+17 + }, + { + "step": 19800, + "train_loss": 3.022461414337158, + "val_loss": 3.075487720966339, + "lr": 0.0022440369086039274, + "tokens_per_sec": 753592.443440084, + "mfu": 7.084506106796748, + "total_flops": 4.9168645961967206e+17 + }, + { + "step": 19850, + "train_loss": 3.0326911211013794, + "val_loss": 3.0735740065574646, + "lr": 0.0021803689847702556, + "tokens_per_sec": 610836.9087012644, + "mfu": 5.742464441650178, + "total_flops": 4.9292802938791526e+17 + }, + { + "step": 19900, + "train_loss": 2.9602994322776794, + "val_loss": 3.07362140417099, + "lr": 0.0021185074503545782, + "tokens_per_sec": 707356.9151921803, + "mfu": 6.649846915247556, + "total_flops": 4.9416959915615846e+17 + }, + { + "step": 19950, + "train_loss": 2.967044711112976, + "val_loss": 3.0712722659111025, + "lr": 0.0020584010543888563, + "tokens_per_sec": 721988.8068136637, + "mfu": 6.787401008907505, + "total_flops": 4.9541116892440166e+17 + } +] \ No newline at end of file