From f929feda83b649aa637cec13e3aa7aaffaa8bfc1 Mon Sep 17 00:00:00 2001 From: Lars Bogner Date: Wed, 5 Nov 2025 09:23:37 +0100 Subject: [PATCH] Add comparison against OLS --- notebooks/01_simple-tests.ipynb | 109 ++++++++++- report/main.end | 1 + report/main.pdf | Bin 618885 -> 630852 bytes report/main.tex | 320 ++++++++++++++++---------------- src/easynn/feedforward.py | 7 + 5 files changed, 270 insertions(+), 167 deletions(-) diff --git a/notebooks/01_simple-tests.ipynb b/notebooks/01_simple-tests.ipynb index 7cb13d3..d90baee 100644 --- a/notebooks/01_simple-tests.ipynb +++ b/notebooks/01_simple-tests.ipynb @@ -2,7 +2,7 @@ "cells": [ { "cell_type": "code", - "execution_count": null, + "execution_count": 1, "id": "2dd1fa3f", "metadata": {}, "outputs": [], @@ -10,6 +10,7 @@ "from easynn.feedforward import (\n", " Layer,\n", " ReLU,\n", + " LeakyReLU,\n", " Regularization,\n", " Linear,\n", " FFNN,\n", @@ -82,7 +83,7 @@ }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -117,7 +118,7 @@ }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -143,7 +144,7 @@ { "data": { "text/plain": [ - "0.0006306650385315765" + "0.0019756351174760136" ] }, "execution_count": 6, @@ -179,7 +180,7 @@ }, { "cell_type": "code", - "execution_count": 13, + "execution_count": 8, "id": "00f52ecd", "metadata": {}, "outputs": [], @@ -197,17 +198,17 @@ }, { "cell_type": "code", - "execution_count": 14, + "execution_count": 9, "id": "a3e32a73", "metadata": {}, "outputs": [ { "data": { "text/plain": [ - "0.944" + "0.748" ] }, - "execution_count": 14, + "execution_count": 9, "metadata": {}, "output_type": "execute_result" } @@ -216,10 +217,98 @@ "accuracy_score(y_test, single_class_predictions)" ] }, + { + "cell_type": "markdown", + "id": "1cec1615", + "metadata": {}, + "source": [ + "# Runge Function" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "id": "74212e01", + "metadata": {}, + "outputs": [], + "source": [ + "import pyoptim.optimizers as optimizers\n", + "import pyoptim.datamanip as datamanip\n", + "import pyoptim.plotting as plotting\n", + "\n", + "from sklearn.model_selection import train_test_split" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "id": "9651d1c2", + "metadata": {}, + "outputs": [], + "source": [ + "x = np.linspace(-1, 1, 100_000)\n", + "y = datamanip.noise_data(datamanip.runge_function(x), 1.0)\n", + "x_train, x_test, y_train, y_test = train_test_split(\n", + " x, y, test_size=0.2, random_state=datamanip.get_RNG().integers(0, 1e6)\n", + ")\n", + "\n", + "x_train = x_train.reshape(-1, 1)\n", + "x_test = x_test.reshape(-1, 1)\n", + "y_train = y_train.reshape(-1, 1)\n", + "y_test = y_test.reshape(-1, 1)\n", + "\n", + "X_train_scaled, X_test_scaled = datamanip.scale_data(x_train, x_test)\n", + "y_train_scaled, y_test_scaled = datamanip.scale_data(y_train, y_test)\n", + "\n", + "def get_regression_model(\n", + " n_hidden_layers: int, n_neurons: int, activation: type = LeakyReLU\n", + ") -> list[Layer]:\n", + " feature_dim = 1\n", + " target_dim = 1\n", + " layers = []\n", + " layers.append(Layer(feature_dim, n_neurons, activation_function=activation()))\n", + " for _ in range(n_hidden_layers - 1):\n", + " layers.append(Layer(n_neurons, n_neurons, activation_function=activation()))\n", + " layers.append(Layer(n_neurons, target_dim, activation_function=Linear()))\n", + " return layers" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "id": "ec000a54", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Hidden Layers: 1, Neurons: 50, Test MSE: 0.9177768024329133\n", + "Hidden Layers: 1, Neurons: 100, Test MSE: 0.9167870523476759\n", + "Hidden Layers: 2, Neurons: 50, Test MSE: 0.9136592028639727\n", + "Hidden Layers: 2, Neurons: 100, Test MSE: 0.9135554326148551\n" + ] + } + ], + "source": [ + "for n_hidden in [1, 2]:\n", + " for n_neurons in [50, 100]:\n", + " layers = get_regression_model(n_hidden, n_neurons)\n", + " network = FFNN(\n", + " layers,\n", + " AdamScheduler(learning_rate=0.001, epochs=1000),\n", + " MSELoss(),\n", + " )\n", + " network.fit(X_train_scaled, y_train_scaled)\n", + " y_pred = network.predict(X_test_scaled)\n", + " test_mse = mean_squared_error(y_test_scaled, y_pred)\n", + " print(f\"Hidden Layers: {n_hidden}, Neurons: {n_neurons}, Test MSE: {test_mse}\")" + ] + }, { "cell_type": "code", "execution_count": null, - "id": "74212e01", + "id": "23e6f86c", "metadata": {}, "outputs": [], "source": [] @@ -241,7 +330,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.13.7" + "version": "3.13.9" } }, "nbformat": 4, diff --git a/report/main.end b/report/main.end index 0b84c30..a18e5c0 100644 --- a/report/main.end +++ b/report/main.end @@ -1,2 +1,3 @@ \@doendnote{endnote17}{\protect \url {https://github.uio.no/larsbog/FYSSTK-Project2}} \@doendnote{endnote18}{Solely used for data splitting and performance metrics.} +\@doendnote{endnote19}{The $x$ values were expanded into multiple of features containing polynomial multiples $x^n$ of themselves. For optimal performance $n \in \{0, 1, \protect \dots , \geq 10\}$ were required.} diff --git a/report/main.pdf b/report/main.pdf index b08c6f85222acf7b8898ee07e42131a7d4f2f5df..6161ee809e4301e586d015969cd44a69a125c694 100644 GIT binary patch delta 87371 zcmV)HK)t_(<0ZtvCXgqS!DuOwPy~M*x3PZLuh^GUwXD*N2e%m@Z<1vtI>)hNSx)Md zqKDOTL{83fm)VOI<=3CD(YVcWNRgE7Dra3X2irh5fJQei-I(cSb}?&ae|*v$J)ggM z^3_Z2WQ*H|JyP2 z!)D8g=i7d<+jIKs#d_HX_=|quFZaV{%|GMw+u>^1_vifM<(bp-ez8An`(3}^oxOYW z-(S6ydgf}^%GS58>uPD*StV8N1UK&6#TJb)R?MGoF4m~eqnT!Af!1n5t3*?Gw)0Qj z2E3ov!_WP8H|#$F+cope2cm!C=HKsDn?OU9G&(Tmh{WZ7f4%#|StU9>|LQAxs`%|E~+$(1WrRBqZ@wJuc7 zb^pCF&5H$ihx)bY=HG43q?`YIChfexqU8Itiqgb?t>)$_KV>5gPz8Tg0(wf3zc)fJ7h$MLNWyb} z>1j6ezYV+PX1yB%JV^}FNn((GF)S|Do85j`^6&G-egUyy8V8_Xa35SPg&B{_I>@VT zCDMCp>PPy+@=8!@GChC$e_b!u=afdr7_^EaNk4prlu6P-fa^=*@Tc9dAASP1!!(kN~6Us)~D6)rqSYhTr`3`Ya>(wXUwPrNn%s(sV+r{~CAic9{?i)|Q z1-Qs5YAIzcd#aTnYijagx#|ZJ0wSw+Dc$-isbmyMG?g&vVWt z?rpbz-oY>5YZOsIdrYA{q7}7p3f21R{xG3nvNp!W0>VrY!4!h|u!-8IVoiExFfnv~ zeloSew4`d=I4*xTrOdh(m|3c#6lB1Fmq7-o8=J$v+I*-!Y*wMSfvv1n#HBcd)c5y$ z9z(!~^?85Invawp`F+rJt7{2i8nHf;1^1;~mUcn<69!=8($$`<@sZ7v2@c>1;NFCc#uJMJ<{V|t;=c!ue>Pp3pF zOSYYdIcn&6QY2u z>Zh+6h>36;TFXu7=FMaa+FEFHZvl2ra3dN&?Sq%g&DHf{yLi6}Y+!A*;PLLT+V2>6 z2m=W?jlO@E8mU}uum+bi5zl9|nvg8en2_e+4|uHCt-FoSz?9BW`XKq=LG3I%)J&JyMxkW6`Um zfNF=nz4C-(b${;nm!YcNWGzggtcA$gS_|4Lh`N83hVz~UifJo_?$cHZk_nmZ!qOy^ zfOIe&@Y}Cz1(D~0$TM{N=QC^>(lV>Dd9v7VtfR&WkZO7QnN>|KlpB-!I*|Hyv)gfN zKtPoQl*6uq2?$+lKXViWL?G}zL}kkqzCoMJe6c#DW+~#}c36U*ZDg38K}XjX8atx* zBG7+(-CrEAM;iVehzQ-dh}hItjj_Iot+?J}+5UG?@w3&%W{aid3Uxc(nnXf&M?>>6 z(0D<@>~Vj*UJd&_D+7MMAA+l5H?0kXf>i%8s?XK+s=o@$LS(Qq z80b)ct@M%1{VAaSwClNEKT)9z&tCCwno4LRu<}AlU&4+Aj*jE;e*=0YQxt`t7jp7lCZqBdH=J3lgk$x*0pJ0>?W3Q z91I;6Qph1rij8!6HO8(;&0l@@<~1$k-@l+8jhMfA_5F7f6;&;o){a;79~Xa^i~J~J zLgwA3>})u-5GbEoIEUsPUv${*4xx1aFC1T%Ua2T6Y+xv7!4MkobsrCkuI1%J8ygRb z{W)!kq=EGBu7@oS)cotic5}rkZWV+z_ZSOQmi)e{z4v`DRc10)NQrd!N(BGyQtF2u1nuikz#(3cy9!@cTSw z)LZcmqlMv||Lt;E^;4A>jcuFxMVQZhBhDEXr<)IFWc0bBQ#xl1JAN)!V9vSkq?_+9 znKx_B;{%<~pu^^cKSA%ou}Mz&4*O%L)P=F3o8xHll*z1y{|h^Msj+`E>z6~EnsA=p z_Ov`x@~wRLG_;^7zW4pqY_*{0aD^(Uz|U06B!hzK0s{$oYL^2P>J#fFFNP|YI43(7 zslt9Z6KT4;-ZDFVP9Jw%LtJw%hg~SQxaP6n`u+4Qi)@-WqM^VUZ?M244#F6}G zQCHedS|xNN=esPQHJJb-ETUPF4)SxtG$ublnC#&O2)*7CD0d_hx!E>)@D|>(Z{5FRE;7210jM11gEbMRC z$vF}iQBzs|zf3Q$coY+%5&c{0inTy^;T0f5gXyqohCvQBdDq6sh!RW1LcH!y81Zm8 z#X&{a#>#AJZ<5g*%GdhiYd-Ve(!fG>jw?qCf5?2Q0;8(TsUK!Z1Aq9BSEg!Go2n$5 zQFyCk0k3}w36DyUU`bChp#n6eP*qmENbH6NSEUN4h9zKbeWEqd&9P=xk5?^ngh_0z ze*@jZ0ib2syhFD_p?m0-`n+z*2kKUC;e>8wMxB;w8M5En#*9_VDA-ty&MQ@=EsWSM zi+T86DpmkxbxBkTP^@mGS1lYPwjrI2v^sR(c1?d#VSRgKr6}cWPMh6P1HTKIw!pkmk zLm_{IAgJV4YCDBS=7O|PwJq9`sU=XV*5f9qs#`@S?MzO>sv`j`>P;ZRjqS_ja@P{I zo4sY-$}~$hJim>{W91T}!|rFM+FH7M>Q*2$qJK+WS+^YYz9SKZsBZg3q4G? z#OL;d6AdhN>!|W0-3pAVGN(>OxoGr9dewi{je)#nwSptk;l7S3#5;jTYLyEnZAruu z#7kSLP*pAy2~`tSR--y&+lrE@1DW8clgH{5I{q+C!X*Ob^lza_nC&%7;vJe43jH#g zBp;|rxrGy&l=yTidQ?rXQdKd-N12^iFRIakkrg2q&H7TTKp8fC3AO*RrhAkk5#OC{E}5^RX3gcEi{ds zQ`DVcBR$I{b6+yWiEiJ}vyveX)H8pWn7j73SkZ8Bt%dx|iWUp~GJ2*Tre{S9C-f{c z>TW#~jU&+oA!Ad$H=$?(1k9p0f=nff?mChqG{tifPvP6iiI@?QRYaDjz(-7gPBtxV8qjM2Lk}|&n`zu`f;kU2A{)%@ zB_oY=ji08qrFR5kB4$M7E+T(-Cn1n@j|i9kXZ{u|BrB@5){THM5692OjI|nbx8~a< zuFkMp9U-anb46%0?RB8bNN)* zl25RL+UrJ-4~0_m_OYQGYWc|-5myCVoOXY|jQrgju^bdclP!N^WG+t7Y+iK=*A76s za8_Ep-eqlfPS3VCnkR{hNPT+wSsjCV_vHe^Y*$fAhHV zTeR(D9GqF?Mh?yboIH|Qij(6qi)HEI%$ku3W(CUh3Rvmk%;4;g%oMZVGS@xwaOPI# zqPaJjl+1cj!#M6=S z4hO{yO(%ca<1v~jdjf_4`cjm`LTIql)Iv(O1HcNcG3zKgVKw-|Lum(qA}eo+>TGDW zPsEIfD+8{ur@=-za~+v96&XhG+8y%|);YWl?(gp6y8=+A(!l3@b8O>ZhmVN>52z9G zG-2I&vT3Kgwtt<8RIQX9NdCZ3Ct}6*+|b2>+D{+u=3+xVqre6Mnaa9V>mHvJ}O} ze-VERvmkv^qKb)_5s@RVJ`FzN3ZPa1RPceocLv+KXC$_*PWw+l$r?`KMm=DCVO*PV z&FOHGJ}am2H5IJWCNdJjsmQ;MJ+C8c*ocvaMG-q8ui^2@84+71Y;`)yt);&tV{vC8 z>NY%;{!%twZ?j{+FeJ3otkl}S-WZ>p5wU+0a(7A#4Vju6vIoX#vZ14zcaul(-G`T> z!Y8xQ5p>0f4xEZ^uZ67+nLhaSSL_5Ic7S1Im5}A>?7ry*WL!sWa1XGJn{U{Tyq5vr zp96#Jc=xU8%T%=MHioz1{N#*?Z4!2OI)_0+*AD2Wi%4g|uJtcE;~tH%#|w;!3ygoA zf#6_ppx@=F|VE3m*Pmd46#|B)xH+4uJliG!7twiwb}JrS;xeEZjI? z-`j%!rIp~WHw+n#Kb?E`YX(HGEzW;MwQ000+{}`0+Z)lUk>?rTe1waU_N%#|vShw9 z1+@eTKHv;Ue&S51ny7N*dCrY1il#1Pilc6_nLbl2JkOK-92@AdIR{n!0MB!xG^*g1 z%JP>JB;A!XpYc4$QjhaItGn-UOg6Bv&mUKQWS@_m`V!Cc?@WtE-qjKu9mIcyz$Qm> zA={)IpJZ%cDjBl!h2 z|BT>>sRQPT@DOt7{Jm;)ACsO{Q4;ZbkiR7i1U-7Q3h$_T_TpnynkCXFVn)Q35%)CQ z5V&!P)l2#N>5wNan2&VZTG0z=f{roW*kaiV|GDa#3>*CFsZYd=h%J8-c6%zU#LHEf z)?n)p{tcACO1Nl)`PPwdVZ=$xx86GoGtn$RIU^!R+}53na^vOGhKw1w>JNuP)NS1S zq0i^T0HTgm+kaM{RJxVjEI&CTqH0Ek4c(s5LgSAnbTNiPy~L#k-^pWRuEL68jusm> zqBoK~jPc1C5j!EP(-D73OY@`;=tg1Tf=qVg?;=|+tUk@yR-}C&jcyw6vE@X72UL2= zD&YR!7^$RRJ<%!OD|Dd~%)!YVaPg67kHI<<)#9JB&O=t6=LybP{$c6AiK8+fm88|YQk9|?o6o+rOh>(AU&{qfy^ohUp9*!0d zZ{z6Wj@*FU$kDeQnax$^^Mu$)zpsGsHHreG)onQD+yF;`c?iP8N{@vE_cw8aXJ(*{ z+TQPYIYIQ+7C*3USuiXaDd<%?dM#6_TO=ctiiJ4PG+;BZYdQq?>3J7$z-Nj?( z0~%CH!BJ&LVi12AS!GU6h=G6S!yke0eR#$79>Jmj+^^_;%NpkQ5_lPGRp5|15@X;E;jmj9tX7#4nJ( z8Ef;cem9PSa1_wSr~Lc<#7T~4vtMjO$S~0%d<;eZxZ)oieSx4k(4`1L2hqqB&r^hW zE(8qmJUxG@l1>W6A;&QwilC~pbt$MSrDIT4d|-SORF(2lP*wWzRy^f|ddH$L%i}hra&N-L}_?Lr;&>G<- z>Z1?;!;(i4W_^kRAfa+G4|414!V@u9ciez1ze!T3;urZiL=>^j%dNH?8i$@dgef2` zfy#f-PL$(E8}M_M&S6JEaS?y!Re~o zD8O47xE+^R_lT3VTTmn&kFeqLLkgPLF)bZwAym#SQH)S|80Pgkixy(=q-o2ib9ko^ zMjKF8e}~9196^bpz99fcsSZ$t$Em04e&c@|*Qz48C!#+w=~i)6T?8u15%v6U3gg&w z6kYAZIuWp7Vjs{s(5xSm=?Njl&3WZR8HM4h8TN8mBRtV33R6dO4T;IY$Lh{^PG6=L@>(iN3{KUBA_9qSYkPhh6bm~ZIZX?3R zK=`8*MbqYA#X8%B@Z>}QQ`5emp1nBAByYZ+vbN=j*P69;G)OcLFU~`B%0(y@xxzm` zgkY-tYrnZYvKW@!u(4wu6 z1r;KGVadN7cp{3krqglNN^C$dxI@qdj#=gxv;kpF2hVtCiutbOV3+jexFMH|5Qfdh z$eTqZIUlKtutT`B8-cB2;7Z;eFSa4b+FKt28SwU9j;Y29;9|wkNJeE5_&kQ?>WaYU zK9bm!%5ok#17E)U?mO=f^bi~pNDfH0V;xv9JRRUD&KqX6a-Ug5-THsJkA5AqIU8|J z0X2~k69hFjIXRPYP$++`Tw8P8wvvALub7upHL8tx2m-w2l()*evCr|Ao>?pr}zrKK%5yqRW^`Ze9ATAA{(fxHp@#0|-FaG}G-)>&~^-bayWmINK zwz#=pq-h=*V;4o7MtPbpZq|$2WqNgY^B*r_{b+HW#+EBxCux6F##SnI&0*hNT_?^h z5BKz&E)SpXuo%eTFJ)M%w@+Z{q(cYS+W^{2L8qh6LRn?3!e%Sy7l1BPEL*QagO zN&z#YqPaX?6>j+*sAaiC<6OVrRK3*eWWF^%MVn^%pcSg_Qx_V@w)sRY*H_nR;+A#Q zZXYRWF~Y;~$gHMZwO>n(_2a(UH7lWlHSK<=(^kDUQ>}mh=hXEDlt*! z3@US_b$x$N{KTF=QlI~f)3~d*_umOocYoh~k3pblAn+-Cee>ei7t~U`FpDHBqr5bW z97C{ixmfL9+}_2DH3~7Z$d<+83s>DOaAIwN)OPXV#ed1cpN}w4qB1GDDPwdpk4GAu z;fY=+rr;X0`Ck9)e9%wFXiO5B%q+69jEXe*mN9=ClSY;s%2GQH@q1zTOpYa4WYU4* z;H)t?U0)ik>C@G5E{c=fSN;mSlRPS2@f=f0tcl_xe>xdq=S%r{*L3}@?6Yf==lsQt z{WmA<>f8M>+OMm&tsZp?jfA_q0CP-+B^kFPPL^|YvPigD?DiEM3`%_D&0kZd#7RI} zPo#ezlQuLP)0ejC`J zS(;xu_`AB_3}-YcakwiHe4S?T87GEICFk=`B)+Gnbyf$u5h!(k+OK-O(U9eh!>P6o zp8HNjNWG%(u>v&jB+0btgynp`tec087eIeRWa&%QJ~ohFouqkGD6A`yVnW?k!YyV$ ztyo+xU(ahmTIof?m86v>cso=bi7>jPEU}t?FlO3JLZLA*iJFr8S_rhMcp%+b#f^SB zy}LR_aW+o^pQ>i>j65L)fO9GMV8H?*rjKyP0WCD0p+cZ zxE>9`p2=EjzQ7jWRkS+_+R$Fx>!gH6LchIKg&?j^ZM7BRzV@dnCymrBKS_U`N;HPW zb)e6g)F?FN`cTk>9*||31_i z-;gH8tvGbTIc#H@C3VCUHETG{y_wiT^K2hZqcR%O7e&|dr(@i|K`o4i`Mq5Iq+34K z!Qw0`ty!cdi&FQ;^_Hm?7e`r{|LGP&N}P=fn{rFm882>|km_im?-hUUTZU~*=z-5M zYL`SwQOrkurnqmbN7yP}YiHf<9g*(7lcg&PZep%Ctkg>JpMCHI-{?VB{){sv&Zxt3 zz)X(|`%N<|v7rj^D0Wc+W{ZH>57R*ZZ&Vrbw!oRZC>@l*Y}6)YpAzS-35ZY=7x7%3 z>6i$a-Rf;fnX{vWZ{&X{1%z9XMAqc3?lDdD$!N#18WPbpzrQ0)! zW3F(fh5GSTIDnr;qYH>h%R<2Aag>{Hv#_BZC5hptQZyyPK##xY+N)tKUNCq8m+;54 zDYW$ifn93a`6hlkQm)$G{yZzb^VV*B1m38~e$`oGpPJ{UR4{YDgy{6@*fQ z3L;t3C{k0fI@^CqQq&0k>L9;fzikg&-<&FJ+V5oIa_k{%?gU*{b@@$9@SvQof2dUh4i7{s-4<%*-~tc16xst zj)}Rq*)rK&Y&EFKKs2@skeDx^(8TJ?N6RAwKu#q*s$+lE0RyiZ2msda$Tlpk8c?pB zT54jU6H%>`axjpo8L?){GbR${jBDINTZiqtrf;fk^Lx$sC%D)ic9hCV9VNnP-;kvz znHB>%XnKu>YS_By>Aq?{o`f+}*64440j1<^aX9VQYH-}jZno1iXd$z3dBk33#<6bs zd{PO5+6{l|VfQzTCd-#y2UGeL2Xm4U5P7^gw5v^kiO{~w5jfd4A6xQ!vDj5E!3rzS(1tfufq^xvSP><3JD>Ipl7C_N z+zCiYiH7epXf}dl)-7X*`&BK}M(e|~S6BK8;O`p4l#Ev=hR;HfE?%>R3%u~8C+)_ODyF*J{Ds%O|(5DEi;58$k z6cY~j7rZ|Nx0(TKuD6{Kl%JG5<0p|6Tg@ReV60Iu*YJ?QQ0VoWRaOw8G5Y|3B@CC+lp1ri2|;Ltgf+C8SKhIT!g)I!sDXrXH|zl-zZ2b#uq2kR=yov6 zA3wa71e>(@4OuA#{N%+TYJz@xFmMoCWa_XXlUbvU@8!3WpK<8UFD_kJYay^+7LdPS zYUa+Ex~o3b7m+`c0R4YR(r*D$pdDmT<*6l`WCIkUi%m*<2*`o3jq_*e-{3+>%-H}s zMg+Q-28{N&S18KCALgetn(5pe)=@MV^4K zIG^7ZdV!(fh%a!?IEX{#zj$?<&-F{|S#tnZ@iLCg1=NYW6GeYoP7M`6W*(4e=rV$; z$4Vv6*PvTd9XgmkgaW}QMSA8>n;+WD6c_YhvT5DusSECDhhV)FRvyl7l$PJLpbl;% zI!V-&Ez=3y-rP4b+6z0|l`*yQEQ26-!+Z$yeLabm$udHc#nP%NYRh9)e7e6>QQon7UaPjEg>(9 z`NKC_iPKHCOeUI=BDK(!0hHx=KntRn&cL;e08onbl=6Q{2*AUyH-}zm&;nI=;z?0C z+pQ>4%eKG``92u4dI7SQtn8%shk)948o# zgzKtpOobdgKmoV3d&Hu&5=ctFgZDiHRzMo47H`NuPpR$%K4w9sX63__NU0=+)Nd4f?}9BJM>PO`nSnE#B1qN@3iC^RcI6h`g8!stF-_>JHLJRvVRFF z8pCm^#yzc%he<}`D!G8=Sg(*%f~G(4J;LwGc0d;r{a zpi{5kK+M=u1IHL4PPbG3s|FLXR%BybeZNH=Aq zGgIB4>v5x?)4>U2 z#NVB{9HxUA+RA|;#Y3MIhPC_``AO2yJ%xW;JUO2wVyB}rryJSNGT}%su9O5?NjcOD zkaPMKe*|2`@#qd>e?)fXk04kG*EmoGFZ9R`CTP7{ujQn><=?9lP6LM?qky$aDcX`ICQA z#R)wrmy9?w;k$z?GB^e3Qe$vU&|i|UraN=Tf!koaJv(d0+t+MpJMUkmYpURqZ21@H zCQ|1wQowPo*Q707cSnFZqBOa|VJo$!M$bqL2`Rx13Xv@I8#<%Vn*KOyP{La#O>waK zr6$P=`V(#D#dqol)(gBCV*SQ2&YFMS^)#7%EeT;Ljb)7WRM`hL)}8sKHjR4x%QN3~ zCp0-Hl&TvM)oC0r(^zN+&GWJnVRUav{y5-yl+(`sFIRmePhG?_1`jI~n-#)EH6sq4 zvLoi~nQ1mZY3cB0jEaBOq%d3f zIsKy|t@8 z{BPyizA5$99Ic*_KyjjZKUCLzB4XaaLK-UE5EEd!`#`1ymW{eht6d< ztR>TzRzIlA5YMo@kJ^9m!=-^L9_(1JZK52jQJzx1YI`x#8*yx#WuvCccH5|1MyZ#S zUzMiGQX-8gB@#4&OonUcdRy%uPSu0fnz=`aK+4v;cRzNjrGm1lv`9A{KdVU-1pz`< zxPRS4wVnu3B2ovGGH9ISiofE0FZsNeVM6jvrFm>*W*pKZpHqMHHIEQI)mcf*+S)$^ z5-h_RxBL-;i1sPHrjbbw5I+PXwnuWOJ2t~hHRdc_ZQef&_@pQLo?0VV&y zUilp_tfXXGuUvn=YWe0Hwrilrw@)^&!&%)_qpwjYla1vxP$iP58n66wUTawuq1q}{ zoBgDWwE%WiAu~Hg;fGjHhL!0?$%~!!PfyfUewpP$BsK2TArT8C^bsNxH+9nr?VYH#ToK9x-0yEPmiL+^B7t9mD&90k4Yg6e?c zS1>Q{y`F`MQ7DgEu-+t{FqQ~(fZAH?KpU<}S)zZ?TUB+XhL-Q@YOl>)5N4y{P7;4G z-1%0~H=C)yhmk;ooxBpGoS&dZn}IT}4qq@~;jR7vk}#ccv#SZ(?mgHvC?Q~~lU;iv zOMzn|63T19&b98cLVQmU!Ze(v_ zY6=)KHaQ?LAa7!73OqatFHB`_XLM*FG%+)ip*tym#adgB+_n{d_pg|jqFTVCNTRL= z^57uZ)`gRz@wRzr`k)<|Tf_39MvzU`m=4L{qp$8`JEd`zVQO`tN3 znM$k|q<#>DLf~z4ybSftJ_$0D#6wSn8eIev;n)K_JtkqF2p04bA;! zYWi-AQnsxC9_n(!^rN&eCr$%|?!&;3Q>B1^=QDZgRpnHU^-bta&jBbhZ>-PdP)>D4 z{fT#f6Xo7VTi-{8S9i^^lfh-DSf{cDumm4CKhH^O8z>b5& zdv}v$*ukZiChaN4o&sqe`T_enWBJ_cyJ{5Z)Cgd$ClzK^xi)+awUYCLS*c>W(gC8U@Pdgt9IV1b&p(`H?SeYz}2B5M>EzG~;{-p&uuiEPTesbaJ3& z%s5;!tHfEZtTN$7^DOo51e`rxEn72xpv-e!9DD9tvBtUP=?2qh;0Zg{LvyO9df4`cUOCjI*6>KKZXMj&_B@fCUSvrc=B)yM)Mz<9 z;-Ts=IWe80?vB$VogSFZl20AuNha^7N9Rd&^++!?Ho5n*sv24mqd(r{#^VhYc0=?A zy91x;yJixZV3Bf*lzPlRmlOT#V;E#KzoX48^jA9>Q7CSd(REkrtsnZ+xjaln{M@Tu zNq0^@DVG{Xhcf zSH7<2By%GX#nd^!1jl-^KpcDMCx)z~#ZnEWo}b<+%1}-hyk=3$p}+7Bfe|RE)4s>` zDNC0W^roD=AP&;|wn^C4vdDPUPAoFT}o%k{~-SOIE z0QjrhG+2qoF!K}W^gwXls4#d#hcyWLMv`W!U&N6`KeXjY#AGVOGjnB?#P~rn4;v57 z+h*Fgbvbm7HBxB#d`51Y`=JE0fgM=J@Agp``vGM6W_K@%8Dj63ODQ+lOgP>7YErz8 z<*tTzP5oZ*g%e|C-jtMo*MbI7r&tftxl#c$G`3s*-sd|6 zsObdd=OB#5iXo^W-Y5d)4FKM;Im3pGO9DUgmf!O-p~??b9^{eZ5Sd6#sr#n25|g1sPGoe`1mMp1^iF!|IS}@>^6)C8>ym!SQ?7mh(9Q%J~D>!>P@|0{v15es(l zU6kB0`;+!G0dd)RAa;@oE(elP++ptzoqUBE8^VE=V_faK_M>2z0%)64Baa1*kq1)_FrOmtbS_K`Ia&mr1IfUJJhg8cwg2DvwQ7fzf%2wv@|Z z$v9qRem=Z^0637utyH_^;%k8%pi`}~CvA62n_$ihtf50r*JWZ73HV+mIn{ZaR4Jus z#SFc%rVyiCQG$?{&^pAx2KtjU1Q~)X^W?y@wX%zSSQ=Il9UWJa_+e031hQ)#vlu_m zVvF;ubCMv7R+Eq)grol2F;|%Tp^1KLvd|)s>6pQPX!mP!5ly7x;ghX1?BUWKmR3K= z6+`0sS-L$$R!?wGR7M#%Abdalz}^~~r6A0w4bU*Z1fnI2y*sw>q*FaF86^xrUXPl$ zPE+cE45N;WwkT79pQ-$$%E@wE`Z@`b=t%sQ>!!l-6Q4gaa2rrI;5*r zalWSMaxb*6q}ail%uG&*5n=_vG=3T;cE-SeBAOKW6s-alFiAy~BMJ z6|fwinB?_A&Bix>qyS!)F5D8h_y9;jKC_hbaf_YsJKuKf;G-1k zFWkKCp95o@GX=?dLARg)_GU5;GG053!cV{KERs*o!hG&5!YljyZ*4P15VqJ%MrHw+ z|2;bspDqqN)A-R-rN5$JJ$nHuFxwj=e7c5&iZ!>@%#K%4Mgp&cAI7W6|Q>YRD*@}!cQg7mnZTCjp)NC;$Q5Q23m1!uOZa*OmydTeSLMCxM z2)z-9H}fLIenR+1e~{TlBf8Qa1_2cuz<-eYh(5`E*f~F+0GU^IWkC#%CVoLiV5af` z0CDzEsp!%SxM4{ym?h3w%yGtlQF{2kP+IqZ>STl$i+jMZM*G%sYLyZg*1i!!{9SFm zbOio~w-Vl$h(WAQN+C`PwIxBo!ZB7-ZYX@m%?&kQkQzv0vk8`SApVB^O=?htW)Qc( z5(pn1-UebetOF);V7nuE^K8~NqjXD&V9!ZIu7gf%BaMPqDjEdPTA=5D>Upq-o3pau zlDwzPyKl--Qb-hsfpTDjjaqU+Hs+JB%-VJofYmV)<55GJI(lc^=F*LIt*@0*6xI(8 zR{WQ$S~dx6KBb((nPGrX)t{l&64J~laH#_|X61%=b$0{9@B-r#^tc?YeFAF}5b}ND zaY@K;l&CuAYj6|9NE+IIs-%ZQE-1%ZJsWDg;;p}u_x#Os0m{9j^13eHew6Bu_3OV& z4Z-mJ;_*h?v$=^=s34eDR$^huHFGGDodW9S$RXi_UOVJRM&K236gHiP4utcG z`>r^&UEl3>p=Nwk<2|O_45Z$i$6_nK+xR-Mr`^q!FN zxfXRI21OW)-vUP!&qG6w%5zO4K9-EZEcsDzXBGtiN#6+jre%K--7Gu+>@Nzig50sb z$m!VF8$ji?-X3C;=Rsn>O~uDUy`>bVeo%DiC$<9#pY_3viqTH=WAuSs1T8@j9<}8g(B7x0yyoPQS zYHU^qz=>5eufoJ8S*C-`yR@nb(G;Zt1dR$i#MD}gNK_X=oLP%xtVJ@_4w_S^a;6*g zO(;k4L<4-$RI5OiKr4+}y zL$;vR>2Uy+R5`p*kkUFJRXNtzsEQFziYtexp;fXg^G&^xIE%SU#f7*6(4Ew4xo;s4?EHg8#8`y$W(j z<3#dvP68!M)AQ{F9A2hEEhF|1SIOQWlblejnt{qySdZH(Wa-ZRn98XUv)g`>363(PhN|*b+q^B5THZW3B zTH~Uv2v@E*CEE(unOthI`cqBcrok*sTy`$pwF|7XIRX_SQAIZ}%~@Q6SjnlMNl@i= zw0DiQ|9-fX>dkmvBX{@J!|V|v=QZQHiZPusSQY1_8# z?rGaLX8U`8_wL==-G5T2QhA@GDpg6Ib6(F0WkPxz*x!+7sQYA~!0Gz_9WC9*SiEYi z=m0=^7Z_ji;9Q+H)UN@rd=};i2)b#b7N!Q)mf}j4U6r)9-pYioF`NM8EHITdAid-Y zld01KG^L*0i0kf0L$+}&Ccz_(?3Ny!gk z39sj>*5Mx(F%OrKYs=z_5l9aAcOnH`*=(^E;WfP(O9Z)_+!mw!k=II%-jbqyygGmY zWl*AR9`Z@vSjUg#r*~V&7J?%{{O8yFtM zre*@uCHnFMvnL7ovf;3^{Zt~9i0H)_iI~{`Q6o3P>AfGrn>AJZ36y{o!6-2Y^h^8g>BBb$LUbbav;kWju>xgHJ)6(> z{qMVr@IpJ8bn)aJ2ajkW`U4-M9tG1Zih0JdWI{pDSmFQ!t7L#%2}&((B;B2;8$}dJ zLI%^H+cBWpaf@;7^NDhiM)bfvnZTA zV}<8C(L$F1w;NPbc)c_jhJgHuTeL)bAYsHM)lw8$nq-&sJkun(rswwIGTUtoF_jcBk?)vL0CbL) zapE9J+@81_8Aua+Bm)xz{cEn8LFWw5H(Svh#siHr0T{e5GF`d=HSedvEXxDs568he z3H2xbQ2cqKpF0vvffGwT=wS|qjs<%~A%PRV4e%8WM-@Xc>o3|cR2Qe2oNs>|$j0VqX4_+2~e*wbsJ{uMbI%)E7oqZ=eB27K(H*(y{g&PMdeangP zFaGM)+S#F2TlU%xeP)bsr&+I?;`lf-T=RDRV);x}|=8QLMZt~@2Z09Pyc3|eVl zGr-)9;{7yxghxU+;YOM*2?3@4^m&2Q?1kUDU2fW-f~$3(73PK4pUtO6r)UhsHZbzW=6sdr6QnoUo{ zh!#o_hDjwFEV9)S939I>o)k1o%LSgaR)Dbv>sMXRhy&U7V#Vdklb53wUURA?M1Nqc z6fc(c1W)d)YeIP!aG?Tubf{>{3^{G6>{(AgZ^BE0vwa8 z0{(^VVpm6X-V~1-hLtqQy!bm#ic*wRj+AxWV~S^{wMIad1S{C>%#2(aL6Eey5}@LP z`@>4YC3eO`2~yV4-5fju8ne5Uc6H`3A zfY-Dle-gRPq}Q3G$2zOUyPnEiwQ$tEFji_w7_S@0p<_H(s!BFQL)F zp@Q7GsD6fe0sVo|z~X--Wf)b`$N)qB0OFL8T&=NbR2D48^9KzBX+GB?dZwNr$?y|h zZf`|F0{^!I$_Nk5ZrnumFp}pk9ToHRbFV)u!ibo}eT76%S1nnpzPC6+%Z*lgX*(@9 zi9}v2KZP9u(!~Mcy4*B;0Wy=-A1uN?uK!F3ncPdkn5$6YpQTJxdaO#p0|0G-CaOHz zvMP-()m27TLO(~$)8Esub(aB+>Wj!Sa%tf47zyI&vwN<0;T$dZgf$l7$BzIO8`YAd zuOhqH>%f%0 zjAQ|oqF4K6e-znb!#xdud;jC zX2N1*vtBwi-C5G{cj+u=*-9#v{7$(~rVh;kVF7{X&e$C29XsVi0n}ec6*d$>cU=t3 z**cW}7Mo&v`!I4E6D<9;dty8arJLK|g}bH>Z5{1vZuQr9oi77~kLEPkKzPE#KcJ3z zrC@i~(*5^K=6qV5V=dyl;Y+rY{S{1G%Ld^YTsXK73#P_A{sr7*2n;A6ZNt8VxdWOo z`-a6ON9^t;?{P?V17Jk!!uVkkdkEvZ(5*jrdp#q!x@@)3q+%I4ne}A&6gk=KI?gqe zQ+zH^u!+U_dh&HD%#m)Bvnj`Lv+;LBCV^b+k7B7g$vI2PRn z+xujs%9oi|iCl>(%&ndJr7*jDb4>Z@jt&)inibGTJyPZhT#K?Xpk%pK(pFGjd0l2z zZ#z-*O^cO6$S&L+`6e#wd|n5|HV1`CY8Fu5h&pa=4@_GOQl7T z{>Z}(qKUcY9xM#2!RQLtP{tMO7u~4;>%P%G*RItF06B8au-0h5uuDzC#)wXnVfY*C zeow_8www|Op-l1p0Yrq92{P3l@0FP*FE&+Pe6pHaUkOKET1B zc0f`cfP-*cN{V?I@$l&e7FMCdOTYTikf)wzZ1&0Pm@YNYbL*gJ9ePjzkm0B+LguO= zjzT3VtTz&q3-4pVC2%DeG*O=Jw3>fDiqib*Ai%kv*x4+|b3*)cx__I_ygRN>aEw3f z+#cfKHalUs*d(l+S2@b>a%84CTbH{I&5IujSn^Dcm?p^10rS(Gc`&T9c1#=pXus5N zrLIfmsir2~BjT>Ci#y@GbMWePQh_#zVww~dkHNxXoYrcQt#YgkuU`;-i2SlL$RaMEJojqY}aj@9{s2#{u z4hYG-S#5Mf;Ks76t2-76U(-e!-w___4wG<{61hhtRWi9DR{ASq^0*fK0&0VqXreSZ7+lRwB<`=wOSl0_uYydUhJu<>WZkj(6Cql*k zAOuMELb>b%g`#4#KxqfZZ5Tv1;7ZZk7&&$izm7*P7FFz%g$O!gADnZPT6348<<2c7 z<~7lvOA>wCupLWTZyI6c3%X9k;lIL$aqT+aK5{xkrzVQ z6%>}%&T=zpDD3_~jEBN?NQ!sdeZ(ta#rC1DA_(->d?_j!nn@DYmEAUc2~=0B=!hfQtN?^uTX zUl%R7BT$`w{U`Ds5IbQv1u5i6Zp3}zhzlD>7=eBf4!1Z*pbk=i+$@jn)GxAT8S=ci znG4F{?%MTqbo<~p4ffDswtCkgVs-x}K#5+nUWnv@9S8aEqmFUb?Q`-yf}V2wP&ggj z&|4`B=tx5HmS(?m&R{R`S1K>8x}P%_YEys6K@?E9tpu3g&XPYGxj@{N} zOiXkhln#boGm!ouTY5HC_0|W?;c7ir%t~A4rL1GAKAVCP06e8J)$x-lZ*6>Xr5i7z zYL?k%9MS+iAj(M1y|Hv-G8W&+FOUa=bWFn{EIq3^tI`hXs#Gg)Fs2Fu85J*XCCwDj z99SSy#hZ#dBj&O!s@#3idcw;k2QCvmTzc+JP|ao}m7e*(!TC;)miQ>?QsC5Y-k_6?ADSE0rdv`4W>l|a9ixLt%x$5a(yzwZrq$spV@4^ zV07OO>?3Zq>ogqS8qo*_#fd^5NAA3PdN{`6f$=8IZ^Aw7YA1jU!4>YVMhJxwmBX}k zzyKvaS0z7x5{+5hrB*c5?dJliRnNid=+L3hv`Q;qO*7J^4J=z{H|-6c?Klflvsry2 z{{!g{Snlj7E7B$1X^QevV{=Fp4RP%uaRqdoN{MX^>zBy{T?zA6hPnwKgBML*th3Qz zdDv4CMA#3z@7KTT@gVaD>+bFh{R=PK&v@;8Rpi0^N9~6~KO}vJcn8=wzl(Y;d&Qi$ zTuoo@Lx*qW7^&!MMAZs6kzL1d-g2|XduLGv5WUy1D18-LZql&?|JC5EJTIkCrE`0$ z-@`-2?GM(`X?qGRSzTtRIj_Om>G!!a0I$_NzBf>%2k-t)IJY0qn3P>35D=Q+q$DP? zt~b-oWj9UdFN~b#c=?_N7F57n?**U^limNT!GmU}$ArBdNJr|GT~C39h#jh0&y|+~ zi1^15pMI4FW2`|#)@Iiu2z`|-p|bx&{C%z<8lu`>PusH``{~c&LPelHtq_?u3DYY- zx%}3J*1NRtNvpTSL5Kct?a>wV+M`FrR6mMPT>*A~Ur1*u9@!+G2s|ain%DFjK6+A` zVa$fxfP4ahXr>c}wbE%s2r`;T))~wKAO#$%EhsgMJ5_>0`z`hicf1{ed2^iO>d zWSXxeQ{S0I5iQE;*R)axd%SxAi$Q1#7{|w7Qcj>pNRlIMV|CiM4I+}kw*Tr#U>_c{ zq8)GS+Mua)H8E#2y4PNAF4Ps)?XexMyR3kb1_pNr-U=u8gG}fc;$9BNDjf6&h~hjO z{AFa9BmVW60KIKfP){6EVD&rP>6NUytrkG>`)X-P4ps423pEo9W_32WQOI@CKFWy^ z#W9F8Dr`Mk^J}XIElxd(^?YQE2PxNll%UOUi3jy|1od6k9}ly-V0b9#fw`!L&6-Rn zSW#_u!G(T+@74{5v%I|z3imT6Kv<{X=?&_bfZ>#75>n6yX$VrVOFr;bJjC3w8jUdj z6*bJ<_ioRi7(kD8_BCx#oK_YUc~0~3$Z_}TJ5i35kVioI>_*QMbx?x@Ica2oK)gm` zE+7;ppaX?|tE}kPJN%k_0$CMpN3vApF^yY5IDJtEg-^UPf=lNUryf`ZP%7BkL)G)c zy=3a4^lXQ|f}+$Q6dv#oDRIec`!km&D0Z?u-6#FZ31#jpWOHry3jN6y%Jg~#L5GUw z`5H><7=YlJALh7su@$Utjuj<*m_lfHA;2+;Xv_+1fD$SqKm&%TEFug{PRWYQ5E0OL z)}EU*4+n;=S421pf)(@z&|W|=#9K&FG(zoRU+E$Wj_bi5U?g_3c`GoPG5<8f%lj9C zTy3%&aSMR}L`d@F|GYcR@9}`R{W;k3E{H@3 zb|;9`>ra$AZ5ENT&m9G1N|3K&Y!0nv3O$RoyP{!nhco|#QYLE%2w%o1B;8&%QjW4) z5A@|+Q?MImLkN$m2XXfRk1T`_5<4Z8iC`z!PdVsJq#wIT=Z?Pt!`6*U+R zqANJgpKY+!*UtlLU}3s-!E)Dt;iW3DWxffc#@i57a`%+Dbj?ldq*e^9g^fSd1xE4B zDKoi|5*%4Jjr(C zY}pex^6zWlUuwW=j6Ffy@_?@W-#-kA?)v8(-5Bq=3;d zS$`^A|1WNbnT?bA|F}xEHWN2TkbPIxFSYAbrE%Q-{gUvhj*^YlX-YnkPJpEi+J6Or z#YCx1te#|F69D}MM_0|>s1TK(^)0P6Ki$Ez`^@U?dcXdvwBHWBo3`te9(+9h8k9{m z*(3_LbK+k&lXQCM>3SKe;icq=oeVy zj4HJgl}xxYjXOh8!={9e>hi+KBVx%dd6yS4Sj>O6Nt z74J#&@$Qr?>hu1(W1H%(phi%;Yjw952ls`cXP-LWO&c#6Nd313Q88G1C<*(-Y1xWLxJW)tacsmZz- zh3e-lPwoz86G+;|Y*(4xcY@ZIdP=c%-SF$A8BDV7i8yYa4PGgSrOs@9|9OFS$+FSn zF^f=?dJrI(ophv8(5xvX1Xr-B3pWyS+=GXJ*<~MhQds%%Y|?3aWWy$yY0F$=>?!ga z9$4;vDL3q0txT1V3aEFzdp)&zbD)iGko~1HxL1A6Mp<|!{3c94$2mmCXa| z?TJ%YLuN2fI=zScJSo8k|L7F#16@7>m?0CK6h+-RPTJ5+&JN;Xda0TuTY!f|1 z8yNwPu!aN&yG|Fz%MWx9ecp?k(AO7*mLArQR;U9~I=MA!Mz3gtaYo6CY6oD2gq_TX z_!RFmbmjWo&(9)73E(Z4^;FFozD&3cgvkXf;K7Zw&DZ*BD5gE>U@QDhF-an{?pKQ1 zXcMOw2kn~Z-8EK`f?cB7!`sT$P6sNGLlFQq$G_~VUuAlnx{52t^ogmSwOPlSkm0%6 zXk!mWSVukERkGb>=b@f0w~jPh7iNU8a#eoH)O8UZR2HNUOKMVVTphT0$)u@F?#Jjc z=3iRc;>)guc4p~sIX0Je#Dk7KG50wVj{Ct|FX|z@aFik8Jrnmk)yo;Eau)2t?hS_|JPm4Y0 zp&%n*LQG4^jZl1oOKdPgwfsVNq9qFPZB!1q_;5)GiDv3z_{3Anb*A8=uwG;-d1B*hoa z&r1l-%s6w9Yfu|z(Q4YN)uZ4qUGV##nm!C9ZZa>fC6JYSFp za3TAVEN|wDLn)vnOQ;#9pa>;N%fN55av!Wl)%6VHcO<$>{IG{H-{8)8KC$$MRR%;= zzA??rWr5+Gr{$X>E-t4%Ocu6~z%>f>(liG$rvi1fVQTJ_^TK~b>p?nYVji7+5N~p= zIbiKJS^99J;hDu&v7K^5F(Cktzk1fX7IZmWe;RNt`LsJnJoh#K(OL$UQsA~tC|PFj zm-?n-^)IALfacQlnsIPN36J`!%Awz1ReqgUuPYetei7;3@%ztvP0AmKpottKRkt|a zi)1EPgdf_Y=|T?D2=K@J(Nx)O_|r!Xo83bp7hvnj`W&CNb2SvTN=6HaC8;%tH15*V ze1a|%lO7)XxUwbiOa^aJ$Pi6JFNaKPsA~zNd^k-mEX({TI*d-#j0jfr)U?FisECmG z6rI2?*^K6+U#D1~p@B}tdcB+ZAkZAO4cu&;A{|h~gbar(62|ZgezX9k{Wk!@f!rh` zvD}mv3M~mYWk8`1aL)i-Qr6#;ylax7(@Zdu-{B2Smqsp6l0_W7v$%Is%wg1ur>gHS zl#z+-4<<o}#E>47XV-Jd_fR8aB1g&HqqD`M`CuoDK z(ZeE-GP-e8OduQZ7v$)#sF8^ac|F=DCzjm_(Me8`LiVh#_9f?Ak<=xzkObYJ)Kp*;FyEXh{*EI$nsFizw7y4 zw0w1WNg==Ekx(mxrs%Y%Fp{mOt4}jNm{mfdjF6R&J)^Y&$3DgYarO+WiE+B9#e!xhTlV5rG=Zqad7ehuTfT7K%Dd-Y_%H_MBLu7AZpaHN3+7fiRD_2S#E zFi3ESy}+=9z@ly^Y3updwdWwMws8gBbw1(Fq^T(x*v`)|Z?o`6J&9jr(1k9jFCR5M z;sBVcB@Sl^FUQ|BHJh)6XW;z<71E;1EK4KIlGH7*9#JBNHcE)^As_MCZFJO?RF94J zTH_3WLYjxA^!UxzX%G0fdEP4>zV;{M|u1;VdCQ zV*$5Y(OQfDbk#{Sks+CsrU14%LOKfgq86%?hwCa(Qd)9$Vm8F%|6;*oFKMAZ*ssPxz5hE1x*R7UH6W6!U^2}5lMAQ-I%P<*+v^3?dn60+54MRQ!f{Z}8X|AJdnrJ;AK&D+ zt%bbcP;pwjyt=lV@qQ3Cs+9-TZU`j-G{bC?Q zfpTXeAdj74l4NGLYR_|k8=aoiM@o+C6wV1ur6SXLF*t)~sF6QS1Xq8Z;d{;)4oaPF z@77J3>a}6ZRb4nEouy^mMABsE_x}QqYgc>mVo!~L45j3aO!-^4tOcaP&x+U7n`aN6 z(1snGALlomu|o|ghhTUm0;3yRZt&r@+7pH@OLJi`h!XqFA6y3lsPL3@Ts3zGk^P8l zFc-86LjGD>y3W9de0D>jf+}@Uuwgc>{yw9u0sr2c8HOjCF=lVaIQj*5y6FuA<-JBR z-i4ECCb;BWWNY|4;vj@FzycOHsfXtSfS5BcV8qZ30(s;7(8EiWJ z_UgwQP)q&LzBZ2t$bJn-Om~j#BH7VrV;%$P)u)asQc6p}+3Y0L>hee8*2q7YTpRM^ zeC}91i5u|4V})eBKY&(rUXk}+HtNOH;Eugan7!7}2iHDQP#LQ3(6<=A3r=|%R6XcT zkC*@tqgp{`2$y#8fptt-A$$)n5q@}k3Qxe(L(L_iMoy>3=U4wkBo~^d`~N{d*jWFU z-#EYrmIh&B{13oE#X`i)#QfjT2J=t+zc2_h6UTq=XCh)|W=#8mL9)_pbbvKm0ffN2 z&_C?P|0Cs#6fj>}fHp821q;J}ulmW%@;_(@<4^oQXvoi5SlIr9mZ%^Cv!(@LL8F2( zvomx2mw;~8+O#|TF9AIz4Foec*QF14XxKOdfxC7=mYM{)iHH`KNjyc2uGd8k@Rife zyu@yfb`2Ep@96AUUp=o-E~8w-ckjQZX9sljzU}KKUrO~(723G}Ym;95i#X_^4pyF5 zc6R+;_t8&n2k7{BA)8N_@i_a!PR!I#_qSvmfUxr+nS6czFx_i^Yne)>V`lb7x-hd}?R^G};)=NIAtM?U9N_KlVukWg2+1alpjZ`oSU7^~v_w)3rq@ zdD)HsK5A+emTP5nlVwG)DZgOlfwu0mFneB;S^7LmB$KIFK@g@`(Ik5rGyOBu^%@Bm z5ZM0Xfstp*8LFF-pS(0a<+d-MX{s-_B^~3e-%R?GYL|uld6ksE3?j!6gpCMW2I%z z=_wDlTn}WGSlN|XTGrKiDtNq<*$x-vrE~6Tq}t_P3I^I5VfAvcI~N(txGqb><`|yh zZ=97^TFjYmk&hM2Zc3vn!Ot@#ruJCl+}K5Fq$d8^Ad{?sbT9qaVQ1U_GY|q5pwhyN zm#G)|EHWZ#q6&z&e>cMp^Xoum9mFb%4Loq0`crs@F?U^xL+)L}*KE#L-Uwy#^9HL4 zbSv4Q|0=+Kd1At8mduafhp@a2o!IFWWlKI*=8c{aPJmSj_6mqEuatDI@|S27QPNFC zvk*&Wqpg;|Pa#v@B;;NUJIgf#JSL>FgzqjH2_t&1W}74(IBXtfvt=|GJE zC%thz<3h%-z<$UpnR{Nn3caH#j7w*W4$ziYhi!;km$@~Q3Sf6iz$15&$OvbMKVO~lWnwD72JeJM4!OjJ(t8;KFjh&Is@)xpi>Y5tO zRZgQr+XELwG^ZdXEpa3OTy!j}zoJam$*d|ey?7$hZR2caUuobiR8^WYh<@%I76ntp zf~&I**X2TOtv%{-bWT;hzVq<07Y1}~rj-ZD;F5(jS3H(o)DoDr8D9xk||GXdO0Zlqovkmb6z3a zE;NW@@}aoKKPTy-Z_HTM$U|GU5>lf@u9~shKg3CYV}xyae}BT7mrcK~PW`Ddc}O2a z51Vs_J(7fVMJKIL?S*if#R{7&p}SfXrXo@&Z8KGhLGh6WgL%EcCuqlN+s zhkbohW1ttxHUcJmr-zXH+lYsKH02qOU;DoCOH|?AFkWXSPrg0an9yuoI!ASp{2x!1Qx%uW|B?oVVgt%G?&25rslclmxN~4J~~Qd z-CJ>E^~rHQ;Y``Y*``TA0u4j2&d)fwQiz)Lwf=(FyCF7;#1!GmI3=9-YfY!ayP~0s z4QKAguaIbreJm|I47Dwl2NFIsjZ}M%VwUTwK6Q$Wp|qJuB*r55@kT$TqS1njAz~c< zSEFfkUt>1_U+k4Y;c+LuKteJT1wG6x{$%S*AR_FjUqbRM_}Kx|3>eZnJ%fXYHpfu1 zQADJBs7M9>V+hD>_GQ=U^fsJ*5pzuRpFK@dQUY~TjO`E>%3zhl?HvpD&~{|V<;1k2 zb1N@CP{X+5rpG(sNy;mSm^hHQ`kIQYHfN32pfh{GCL2WCoKju_KLZe&*c=i4JX-+? zObPv>FImLphKIDpJbXN-T{QCzPjnlR3mo-NCzc$p`(>`+*Y?~hndp$a(Rv97TMqhgLNS-UK;%R;>*-^=Hm{VAGJglmh{DW!` zJ+(Pqs({;%M*oFR1Xt=`?#~?ohwCuwV4M)nSj1u3&D0hU^mo-#4`Eh)#W`#g3p5*z6S7$Jc|a;sCDwlvJ3%H4%p~8j0`CYj68;jx_&~+RI<*Y| zZ2mMYJ4&oyu7<_Ba{k)T{hTx8F)7lqETfHddOvG`55~@i0!&PN{@NIBwOE8kt(R&D zPlPml7`I+}M`@0e;)0uhS8%mEIDhm&+w)<|M-m8pq-E60;Mj`?Lru`_t64Y%!Rv|Rb!ZSle+tfgTDlsGP==!H$ zdJM+f!7fNdf*kIRLuOG=c7=oI4)0~F@3|hWQS5BM3M&Pi5LU86Z(+|8g5XV6oHC-V zyAyU`s!)uPk?`Y%|F&oy)^F+YZIkSbt09RBTvo}SLiXS>kb<@_hC8{0gz%zvaE3Ls3@ae}fDSQ2!CRHwSG*O^YXhkry!>&M!HBK@|_Gh#; zN4_^9?ieT`LrKZMv4)KjQ%4XgnrDGNd{;KJv2xq(I5_ZQNI=bCjKq2daW;Sr7!+HU zen+z@1AA9d7fep++lBM3SvCQz?EZX_Coi(YYK1_VqK#P9FnkK$N;lg(@Kg zIwJ&@=L^=!2?@<3>G5_uRo9Gh+)eJk)CCnx!P4_V#XV@~8u{(r2H?@@7|9{Y`G+hm zfVY5Xr_>xueJ5&pLIW!yU1SxBVp?jTSdn3lzKvR$7=XhHq-SijZd7H+6i4O<3rSRn z_-}{jVQ(d;D=;3CLlXc@>mgjA+ja?}IiHfSD&m9B=Z($hFrqs15r<4ihjGG^4q9y! zrzsxsM8Vzu5I}9x7>+A-i ziz>IvBxzj_YD7L*cP zutEW4Hrh2&P@v`ON1_FW_t2R;YYbnGfmfGo7p=Zz>KX~PKz8YoIcq*FF!zL zi+HYL;TlH7VS>I222S`$5b)GNg#sN#7K zm3kXB%ZK+eW@zSP-iNhtJqMj*PRLLGS_FCGyHk}WixdNFjXt$g>@7}JLuQ+J*SCA; z>QSlo)g;Hn^56W&(Y7a0?pb+3?Btt z(W%~MqBaBARNh=r5KVT~rnv(mV{8wfH|}8Ba)UQFyB9$@Ug;Hm4KZnTMqAyKX6mgE zsVclZkx@e2!i1%}>!pL`wxH4b6Mj9V6OR~II*u?01j?pCPVNc*~IEZq_ zPNG;S#+cButI?!?1kK)INU+uUZtv5eK8?nDMe25-?`#;@!gK0Cf9}GgfUEi@` z(66D2C0;Yp(pTs!2@DEM-oK_NKlN()9vs3C@vv@n2QhN%eRSfPCW$qzUXGpugf^LKr{f81;X zKKez?Okde1)2CpJb2timiX~jD%e&3HvZ(#3mDX#Q##!2lf}29Vd{tB4LV_K`JB&c^ zyUpPK*#SX~qkRgR5O+@Pwl9r1(_fDCU#|e34*PG&KAqt`hqZSkU)iyze zyyxcpz9w|}+r2_Ex>3LNw|N(%C}(6^iG0c9qm1~+BL7juA0)i8GEb}&w60EBcuxb8 zB~Rt4PT7k8jS9r-wV)T5?!iuzq=IXE1DN3PJT~6D%ue!8dEnIxSCJRWoAruFJsXVl zEPSDt^m~Y#<+wTyliJbC-d4)~QcPrQ9?n5MWT6s~7EKFU(n(8XA5o6m%w*y0w#rf0 zllFcF-a3-?q~PcyrINL^N!6U-_aU-A`l)TQ)H^c$<+6E%@(Uz2f9Hj}kSFf#L`iVi z^nZcJMUtTN`%JrU8AI&R;h+3OY|DD#{ZC&Q6@=lxj1B+sR`8~YYXfusI3|7sHVYy% z+y4mP9}&#A6t9O$Od!CU#+8Ih^iqpcRibwU!z7i=c$WvPr}W0ty;#NaW{l4#?7q zl^0OQ#L-=D_E8mJ(3A}vpKOfUY55I z?f7v{6CXy=_={nPWicTSJT#+aF++}b6F`^cmb4m@GZW3Iw|fE5uevh&bXrmu zs^o`7-Zi3%_%~y}tfB8>wWYZ(-}HW-WMKYQ#TrwX*yhXY$LXto=sV-?d3U1VW~$G| z+UR$ukG z?p?|%t!c;$u6W(qFh;#0=g@$f^&)KDNjYZM)#)mUGMbR~;sd_5D|_ksCRY0H)J2`y z(Rh{Cpf@h;qnTFO%&^-+jl_6!Yj?aMtp3feV3?3JVMct(P9Iq?9ra;}Vm+@2I}V3Imm0~r&XdZw1L6Y2O% z-$7IMnB6$mCLqYc)-u?fLx8D6Q@AW*dv&QNA}#0+%60%D##_U`25rI(Eohx_ygl+? zsK&+HS{~S{y-OdMtCQS=&%xYoGGVGrjX|3#{z#R4_I$wxgCa#l?CR83WhZQaTfM)-Z$FBgj*FM&oR zZUWDt%aUcgeSBEQAx2nmp#>$&yG`>p{?`)fcHl_h*aYOZ(-yIomi=`b<@$R7S4kAf zJ1xzA`*q4)+Kw$Amx}fq9ze78J^$N21Cw>@j#i7jf!8Pn3tuZr9ABgSZipaZ=l7TAuO zsp9h~%?l+#a?t$F3tD1&dpU#)_=t z8V3z7=Uk>k>Lnbs1-g&;#=H4!CpB?`!`cDX$UG|a>I0?>4ydsqag8I5v|IfL8I2zm zW@-_(mR#{r^ywED8-Q`#|WXBFaibBV(_xmT`S z>6<0q5nd;ZqnO3c6CEnVv#@gf(CiI-+?mEE`8(1KB+2TYdaWg(* zowBp7T4aZBjMCaBT#X|e=c>r59 z09$a~{|ub5>IdE6Ch+5@xmK*-1!FY1)Q{K+*$4$;j{tDc2hD}3L2rcd66M-DiAgra zAuWJ7O&-E-a)=F zD;%7QQv!WiHt~DT3leRL^et)}#gH5@X@&BK;0LfoSSK)q;wbaCNTd#bh&DYF^R!_| zqVUTIX^|rer+hjf$`PRe`i@WeQ2cV#!37jM z?Sc->l089%Qbe%dXivuT#F#SO`K4ZjS;B;BN;X)j4~+bQJKmWvstx(aRjclq)@rH; zYYotZ6@W!(sU!tI3B=i5@cJISdiy*+AMWLHJ9y&-`-1XU3X{O0bzi%cA!R%O$Kqu` zgKE|`sxD6tEU$#ezK;X?1JphnwXg#2BoW@%t*o*_akt0;8U`w5WQ8B%R6KrheXyrTe|1>6LJzC;ZdY1uLvpRE^wL}Dd6;o9CNJP)8v zu@&&xc!sNliYL8EQCU@0t)VgGC{9eg4KSas2-SLptsWrbvog${43!EOaOft?t4Nz` z0j67jt<_K(`$F%ku#=j}fSFN0Gd>X53bJE8e^v>i`;HqaA82*LSas?ll=5)wF(mQO zp&>}CALVr+ED_=lsUsS3pKtICDhD_TLO`Er(uwWdzY4FTUsB+s#|sEq=HB2So6sXp z>japZoXyOLf{v5`=UDJBpGCk77^7!P4QM4T@2mxg6lB7=K|#1dvWkrs^}ac2`+aYP zVXvAA(noM^R3ZFKy`al^H}o*NOL%CjIcn2W(bZ?xS}khCxluC%VDlYeMFFx;N{G|V z4_fqY|MY*j*NiNokIlDphB9|XZo$#XUGyf7#}$2;RU(TJ@>oHefGa1s zTvMs|L-ut|JmeVH?c6&U@c<>F?1-*FZ_KiV!st5sTle1fbtZ|jX9B}=-69ql$0GC* z%Jj1e1o(9YeNTnKshcb0XdV(Mms4L9qvvr#o|oO5+Td?gw`4pYm{39fDw=UnLC zLFzLa3?HT4z||IgW=RlK(3d6enJ=Gwo}BZw!BExF^~3%MGN7lPNnTkf!{0mz${EW0 z$gF=2K-MMoYuVl;hX6RL&Iz5)()(GwEZ$Oq$11$xe7OzLg%DS?9$ir~4?TvT$)sKf z1nd?kSB8zF1xaXiF4=h0qU1IL(N;nlazDQ51<2Q7usW0_u+h>2kg`&}I2a(Pel@uB zb`fYm5c`z*!=}Cvy6Dv7%rzHzPSNroBZ0qOn45Qa_ycE1eSqHYTu^l?6&kG^m_hn$u0D+wR+RqODdkDDX0_|yK&G|OZML4+W1Q%;TaJ_)_2`?=TL zv}rTr&3Z|fAUk^I8)=(g>vy!XA<+pvGbMv5h)f?sX8itf_ncUbn{%|?nG7z=O_ZHI zSOLiCF#sIpZNO`oTTM4-^7aLC{>wgG3VkyB+uA7)?>bQt98nlXNdhXtTSlZ@1rRFv zCkE$fpXZ!{fQg~k<@!WSyLfaxIN=Ppo;R_o_p%N9TF3Ms@X#G6n;u6lw?Cf-gsyfL z)z~!7JtdC;!uJlO&T&9AP}7w6-*(oP^(-GaNAuwrqJX*~Dk&AaVxI`zWi(m1{#%fa z9N91nUZsea#N*@_f@%6RM2lN&qu1YX@*#VsS@uL6PeOk5ht_BBYmpR2zgurj=LT%s z@k5tdg?yoNn~l3M_C=SM=FURD^J(}X^fz3hwzj;sXvW*_+TEE%{Y)ooyqCgvCKLDx z+nymHwE(v|^?`aNo|5G0HwVffNo7hb(AzBKU6=jphAUp0jG z#VtqBo~nq4jnpPDLhY}~6XwgI`!T1)L0_K)O#5PX`yP*QAd?f^ z^E$7E-60Az(7~tAg)9v>no=BxK8Uli5H`#ZYyZr%Fl@?Y=T&~o_xL?ccDXO$pDF;HV*O3X#QAFZxcmHMKKi>1o(~#_5wo6!77b zlLxT6lfJ0xn-#i82TMkI74%x@-e@f`mu7`E(`1_r3$!BuiGgH1DUy==zxaBmAjz7r zU9@f6wr#toZB5&@QEhA5wlQsc+O}=mKK+0D-~001R6La{^UYi+qG~xG1a%(XOioEn+gkQRpYBWQ z7YHGPg09);*hm}f%5bGv@0{T`TyH`OuiXRe!-8}H_*kzXtJ!6=T+Hg6U_H7~jOiyr z9Eoo?$8%*ZH3%xBd0^YGIHYw}c?t1m}gHz0>r6CoFsl-DtsM-9r{?Fhg^Hf117w z8er-z3f3?Ml_08HC|=%+$E4=s1zIS|CtSj!m@k67Xk1^YgU6b(=wHv)R6fVQ+}g!$ z+}I9IYe(rfU+k^S4xRt}NBRa84*&!E-yJz?>ZcGi6+GMj^Aup^W=`z0paxv(>NKo3 zBl`*V1lq@V6GMRz^+uPnb`RLC6*>B5QIroOCW%Q$Q!7)H8+ZH|EJ7yKAM87)8!w*9 zo-!lYjSpTKxkDpi4o1Kf5Xrj`B_kM(H;U4MqPGEQ6{p95CJzNH2B>2oam*m7DMPB~ z)1(AAkETbe5fyTY3!_M-lL4f-c4AcXlHiJ#JRol7G^2uH_NdsUundyu`aqTIz#ODS z_&AvAntE*l#dtWtX-Fgu6jK8!z>B5?i}!mye>EXw#on~agP7k)TH>}2P*2O^NE*nO z3R@YwmdJ`_ut&iuFQK5mgeQkVt{VKy??QtcK@$sNv@Ad~MJ_TkQU=iWl`6$YMQU0X ztc)vUh@G;MqfWE#(kaVJ1eg}H()~&YF_IiD+JO}@h`hkzOCRVhLRp{~3&4I-Gomt6 zPT1K_7s(wRP?|x?Y`j1Yc`9;^bO@!M)HZO#P*yxKOs9Y_#V{?G3&8FiXH%mSWt2pw zhRAm`$>LSvfH$Hh{RgltA+z2m3gsha9pQ~7)=@;nSN;S1s7wMbRRmN@WoW8tMp8l~ zgCw@4X=WR}1!R`|rrlUcdQ$<%WDTtq=j%G zX7}BiKZ3743#~R-hYtBDrF9VghrE1Q`Cn-YKZGSvu}KY+xF;Y%fV8iPjQL00M==@} zph_&B7i3nH#HLMp(1ZHY4K~=c!^rYX0q4CXgjp-?&9l{^>*)* z192dzqN0P%deV<$M`X zTQw;D84Bp+#9DwkzQ`=_BT*|r5a~F;L%XL8^p@+TV$fM!_fvrFIZ*KX74BOw7bbaa zCLvlX41*_64Mka{kuBeGl=4&)qf$39>@7ZWxKDB0Af!`cWLsKBl9@m zEY~B&Rd-8O(JJ@zNcZ`nbG)4|zI#WOue@>!_Cgx&8A$ZrLBU{L8L?U*AEg|47gZEK zRTSbEG$gZGUXz`eQ0mnlGH(0)B{C7{_88m=hTu2TdosTaPWR>v)H{!fL4u(z}%YgKsxgeb{41 zuXU!OU>`IWeg45rfw>kanuom@f9vuxM$CLnA{mb(p$zB2@SQ+U>4j3Fx$Y*IzHhGU zk3e9|2;+B|qD20uaqU&R?RyharmwChXQg+}`XInAhfuItmo7+)1CXr|g*@1m54^H! z{B*7n*Aq+1g^Er6_WSY+X|N}%V0)7n1Fhq;_f3;3A5MJUEpAB*Rm^;^!$xX+B_|>` zQoP0P(|Tn*2#VnNvXi>mgrT{e5s=o@wp_jfPrj{s@1>+l`;A~W8Udx0OC~l*h#I&& zOb!s4poV7FkQ&fSkaCJZ^8O`~KueaOmk|D$hJ+py+!>6^auY&RyF(O!*e^s}TdO_J zj5dZ~4BEiqfz5n_4dE2~>|s!FqD({(!&#?*=#eIWyO;du0yieC7ls6Hzb@8ZA0p0# z4j1Q3wNNAEC69h@)}w*d<7k9QvYrg3kPrZu00Ni8^flKVYmCS`BNI^UgoJYcH;&#b zIPjKBcA9S}+AUZ51%xnMAo@9a$B?m_cuOX^=38mZ5%C(_RA8auw;@w+OFZV1;*obX zA>p;Rh{SM$Cpdkr2WbhlL`NilQ~=MK6IK!z+?CI&u&e|<#2+{omXhy&V%#v! zkSw0z`-6XkB|=gDB9>p>fHBi2YLRL5Fwf=N&iF8k%fQQoZT|`a=q*;~JQg3|0NeMl zeBgbcZ(B#xsluCqtppJ|OFohs)-*I<_ki}J-$@9BRoaG;E-v1fAI|?7dz?DFT}^jol9`Tlx@TWt1IOLjLcLU2sP|c zHD4#J`<%zLd}wjy84Z?{~18`JM`iJ7;BUKOnOCs zbYw4RaE=~6<)@3L6J%++v0{Wn-V#9P#Xl=9cj96F#eF)jr+5*OpL3W7MqJcY8_I&x zzp{+T9}pnwLxpW697!U7h~<3eilvY4WeYMR=dU4f@%@_t+tH6|71VD;3oa8|0QkVX zc_R9n&P8+b0@kh=$P3{?e<5I}UYS zI&~rKr5yYA`isTI9CYu`q-asj;XZ+vE0YDbBuZKh+PG*;8d9 zid$l#Ognt???t5TtDXT~P{tE`XpeOr6+@(WEYA6ME#x^3e0y|{oX{ds-WHbQ#y3YzIsxJ`%cnLDY6oR2T8 z@ljsLq+K@w?#(Ikceu8RoJtSk2309@{TemWw?-L3r5S9}l{2WQ75h z@BAOX8~Z%~?`_wQ7+07N^&4q=u+4+m5A7RGFd}!5aZjJD^lz`)Fu|j~#2ZOX1F4o) z6ILR-$0oblI=bmqNcZ%kvVMx|<_k>j8SJsA62?ysW9}8LnL(B0zKZCu8tswZvF>)| z`^631i?|S?pJIAcp^KxXo{{pSEWMgl!xB-pXsH5X$aPwfanm0EzE<)u=y)E-RyZ!)yf92{T;Rrr8aM_x?^PXG&jL zPm=du%i^dweh1YyEun93+a3QM+qwGI&sNJ_5&ue0hVgSXE~p;9)ikDE;d1q!@LAo zytYJKNhBcO0ax?13XHL>?6yba8*)3G@TO;4h-#Dmn%{BG<)3j-Qf~dg= zTs%~h$T&RyjItg+M4ECZ_oT1o1LIKqNF@kxiqKB&rB+Fu^Z0(rQv>0O!B74@m8rde zzHzv&?Q%{rJ&W$(AoC*Oa?6{x0G>1I{SO9c%6hCeWbdgw2mE@*o7|%zXgC!8t$>-d z{z_i&cIG=;d99nOMFbV$E_`Lh*< zj{L9+t+*qAkgy~JrDztMBg?=j(*FWd(GPfyn;YuiIUe8&>m4;a zeB{KZn;JKbJjBEl?e8!-9PK!>+kNN+9xJ_Gnojm`q&m=sNSU!{sd+&e1 z*EaT^?sQg(^avs|4>W4doVwDCa+%Y~b^)F|jBap+vn~CFW?jv&OPpy56@I$0$Cq+>EPMp5s(X{1nT=PH1zx{VrUG}o!B0ApX zLl2V9pGM7Xh@?tf-z_)4EH?eLtFzUuJ?ZDzg0jrw=i!{*l34%uc|B#;(U)kIZbjP) z8nZU@9|Bx|Ep7V3up;rw8J18_u!W?TFuhaDvc+VT{F!HS$|zy_-=`g+Ug&ttt?u)R z!I|IhTE%Q9cEp=HKEM3erzE%G4%;!tP9bQZdYq3ZKauSq4Xx7ocgThDF(o-ZVl}&!v;=CGy>TWPgD%cau>rszOpAN z9tD3-&rSZ0k3El!ty2WwUDTN3W?hC_UNp#<;rA_jXDeS{_M%_-i9^$~t4KZNhJ2na zR15_24Am4B$_3-j1gLIk-<_*!E9w?a>=Nnev6#W;-7v6-9z~qdZ}x9}(mhRmheUwO zNgRS<5cD)9&CrK$2;0Er8>J%V->u98w+^`2oxiK_4p3qAG#|-fSZB8OKwXZheaKAB zjAR+lcf3~|FDxB!cJ2+Ym+sg%r3e?Tw4AjReB2{W%H zvKQLx`+V;PzReR9bYk%H`J(UF*q7{4o2_~??DRAyF)bY2;Tzptv<7CBNO{qQ5!6 zrR~syqJmU7x}BWhG$x$^0^73Ry5P`oCxgVh3q&($fs95Drv@fBau~;2S=^BcRC;rL zb__CF@abN+twl%b1uDc#2?y0ymZFAK>v`=Cc*Qsjc?M{PZ?O%F4=#rJ!ERMdQoc`? z*bGfzw)DeI<_#KFoG_5$!{6WarITi6Mx6CDC%8Yrz&LZWP4A}%M-U;-oQo6;eZ1&70 zqoMY8@nVd`cs?^EriG;$Yz8xhTc}+i!4WhvMpsJz4$^GmwB9KY?M7dscYO->8gd=q z>ZEHqWw51sOMz0*N#NBAHR})x;#kug%9s4+%NHP5^t&Kp47b#$GY0F}a#foK-NQai z{5{7#42MOt;||R%&0vHnLN@d&RwPOf2T`hSan$K>6Hg?TgR-oQ>O`XltPWaZaYA9SU%f zm}g)zAruc1*SeCC+y*wcbB1cMt5)NFl?1#MBtyH2imoQPFie7zev>*HdQ2GCW!( zG(x?og-CI z1B5cwmJtM|rE(t_?WY@8OWy%7F%<~&e*~5PdhxW#!-FXPT;p!ZLja*50%lHS`k}P^ zPXXo@YDN%8U=UWOmMkU^S&G!#T9E%U?SJK4e#(pgTxCxE8Hm_ha`H2YMsW zUBXo4>mNuK2paW6*qZ?5Rv6AQ(SCbc)xY)PR^4aK)@u z40>6jRv8CNw$_f04pgEUq5Z8N1fIHcbtUgF6ffjJ$pz+4ke<%V`uqi0^j=ik++K5A z5-`71`BX}vsbmgMTfq(>yR^y1(!@&l1U7K+ zK_poll1+czLo$FTEL5ri$mE(iWmoK*!XP@gfj2M&t#!DocerI?1=UPX_xwV&I39sT zf^64J146?Bu6JuDVge~n%kJtZ3`uXBKj-=O1XIMF2h#NLXg&Fsf(&2r$D@x9Xa)k$ zOXMCoauqkuV;7p7S^+sY_|hjYd!5YAy6{X-`>m^s9l&8A!7IrM5XSNHuMA1#7D5Mu z;BNPq1?tg(paQWK0yGVU3qdP#wg7#fDmJjXI)U>90@XlR!y3oazpxLkE`c3DdFbNj zQIH}k)<9A?onGDz_5F+I(X;x!>moMkFCyM2~DgFt&QF~;8D%J zDd2TrACrE^eNnPX6!97yIDee;`bCOl}}p0TXmK$Y14ORuJi5 z5&F~Hv;W}p#y@BTF@Wx`r?(Z}YG7FF8k*d`DL>*!&6G7}W>XgP-zyJ$vI`2_?0^&? zIMjhaX(3qw{vqkvxV?{WJD=OGNdXvu+ZT<$(beVnJMMMcKKh(zf?2v>D8P^Us3M>* z*G7?UHFMy=W#3Be1gww^x96DS-wj*eQm@~Muh%r+8pYp?q|)Zr=Yk�EiF3hY&8T z^>Yu2Uh(gdr!W0o}jq~u!+&G&3z9*51Vr5oGW}Kqw7~jKuGqV^iPEj zaLu`Y@FoQ|yCybYr8ZxU)*l31*04J9_4q4$&*=9VgA*Tu-OJRh&0bv+y)|ne^Wb1T zv03wkkgvEwdSPcho>D6~>**gk0_q6zKr^$*C$TqPFE!zyJEM0zB@{$0f(!5lMvYC* zUUI;I&Zi4_rY2{x7vGtz-#`>~+a-cXn7IDZM}!0L2Ff3}jy)uoJ;>A61UL2%G_DkR zOP-C-4 zL~pe9i}oNndc9}ITJY*Su!eX47V?9!3yYVRryJ%hP8fhb54a)J;DkDWV3E)Sq&2V} z!PZy7p0!pZ6FZo$r0;|Y$aO86nWj;_KpnGDG27Pu$xBIYqtc*gW9lV&ovqhVTG-|A zcK!P2+>uk%m;IR$Am%h3JwRo5b@(=EQ*aDVyozT9E0&`kNyThmw@(zzqC}>R*cl(O znukMPvHaftMHul{Rg*m7cYPl|AX1l~ojm<+C73TmGP6RyT7vY97caH-+L^=F2(+c( z;h!LNN#f|O`zA42t9oHIVq7d;AtKVF#M^i~*hQ97Fjv0JO$<_GL%+H;13$id3LodTfZzgUvHaOC3#Tkgj1l);E2&6}{t${Bz;$ABJBCSX z41qkriU32l6WIExR7>i)THOvsYcv#Myp3_{NugShXWTboHX9G7phQCo$>Z;ez@1&u zMYJ0;e1MI&eqvxZy>M8zwDXQF&WmUGKOXDX3?2{lD794H#ZFP(T8nhbvPlKp?3f?I z8cEYH(2(-mV18 zxKP7U1~=~BWdIj7qHg#|Hc$-YUrNtm1PWpCr$FLh?+}P5EQjtsfQ5K|xc|Z6 z$>7w^W;#{AS>M_xK(QCgX1)yrdz6{rz)eM6Vw>m2XZtFMOp`i?f~wsAZXx)wbCLa5 zU6+%D_RX)2_cJ&k=F6nep0jzV?-o)RJX2 z%9{nv)W}tq(<*t~!12!|IAGRAJcVlR7gqGf8qUy9*8Yb8!0Co0(2d~~kYRjp#-^zA zx6!ovqfuAp2oVn1Z(Rl0)7@m0+$z()l?hgPRMxB%Vw?hu1_-mv=$NLl;%_X|qXYgddfkd%VBPTkSg(~O>cZB56-=tu&^VyD&?Dpd*E+Wn}6PMtgU&m>7UaHHvvmcGZIo>VM^<#QR`wYH7a z`*~b4T3v@A=?ZXc*0q#tKY>+4>5<&tK12V~r>n(K|AppMD}d>usy89Na2C53Sch)4 zw;8w?dA-}puqc|un_?OMQRej0jHoQb0dU_gA{JigU^K#+Xc61d?&&sYO!d}o7wi6A zWbDz5oH4`7kPE9wT3qT=!TdbHy{}OcULt}%t zx@(ZJ1>jynAv)iN zb@!hMZ@!w~tnYX-2*;vhSDG4Jhtso%>FdDB&8p8n&ZL?ky$|-=Gwkj|_t^y&f1BF? zqlr{sFBU^f+LY#QvO);qh>9}C?HK%N8Ba%c>$4$4)q5lJxB6_n)CGxC|3)9AsWAB< zE5F8}`xiqTEz1}0)33r}k3d1r4@jVjZ2w^MV;U6}(VEV5YUi#xM(OJyU1;-KhLDnB zU7T0m`iZB~Ay*eXTWDWz^R=$8;(0wt(Jx&}hbcbZk}{4WT}Y{J!z5Y{mh^!Xm1Gey z0Ee4e9DLFw;FPO=?0;dirxES=IjE-uD)$eV@brP$ku`9 z+cXZu%H5dAZA^b5SEP|zhZW57HobfgDI*L;VIHE~R`4a!<%dyWb_?0^*))pa&Q??r zSMDv(G=*(Sv?eZooQ3DmHb6f~;ehoimz&y|C zLXHfGYXx$cqf;Z~c94c0DW6=_CsJwbAZiQQER0rJvB^+v8p|?n2tb>(%f2Tn3M46% zm?CKYp337))HoxD7$Ko{A^2&+2)f4r8&S1JZYc?napbU^aDR7`*}d}Sa@4eP0H$L= zZBM2Dw1G(6@aI~K=xG$0UePfFwa0}yR45?DFusO zKU$)nGG)-BlT4M9PnHPy^Nz_=0Lpodjjyq5l^4!P<|2cg2cXHxi2JMf1S0dn&!ggI zgkPDbgt$~n4I@5|bV!^|vgdc#9B!l!e58K-lZOUXUrMQ&#hw1x7y1PSojR&_Y8$c8 zv~t01J-G00Uf`+#ccPAeS-mPstsa7#ig@WXpvLis_Fp9gILWp9vA00uL0RUV=Ah6POGgLAI4ulvu^fN!tHdNVKyg%F;CQL80QF3sFlW*j=k6H=j(Q6Tq3eXF*|8zhj) zFZn40w&J2FW;bI720gxJgYXZjV7#fufA!?#`!4&H8UT2r64at9KvwHlFOG(_gi0lL zcUG`H6iVsGSytXz1NCwf8&`lYPEYV1aZm$ah7j_6(YEG6*lI(PAk_MDHRNVonO+YT zhfiKdTPYUaQn)Duqua5(XEop$lCKVE4Pkfxg{7E#!vHNZvKcU2frf}_1s1t9!h=oX zOj!Zrx&T7vGtOegL)pPui!;$r6E$&a2&cakf}lg)>}d&KsdsaS(S1jZl|euW$0q@J zGn#7VjoQ7=xRV3hFhPQQ8|~6vOhaM4`2>43o54kZaK5^#t%Lz({i~NySp{G|_yG;w ziZ;kh3ibockC}#B)@Gx+%Hs9M5Lo5VlFdV z=G#)8ihX%G{;p8mXu-=JK@DlBc< zbO0(TNgJ8P0_+>iy2t%V*^Pzf^0*`6Cv9mrE%~x`jjW$DCQ&Ed>q#Fm@?qSu@GTyb z8NjD=jL)nX!;o;D+6lr^?*d8!?h1}IIOQ|r#Rh$KmWP4dP(Y@M@_~T~D?YAGPuw6a ztBHI9RXW3}l#(tNO7**_<}BFbM|Y|FNs!z8ED^*QENlb=pyZnUtGj98T%2j#(cbo6 zm&jf&Q~Mz`H5pLdTI5ugE2&eE30F6=| zOrsf-Lq>+pKlB`LtxC-#Gg8N2wwrz z4O6R+k(*-7XLvn__+viGtNMrS!8|SB5HpBF*s_{`EPjT!k)EMj67fNo0Rrvq;Rfxs zt>fzk2971p(zQrjjX?+s4Ajz0JIa~XCJQJ;8YY0L{W^==`N~mjr@l(H0`_n<+&`PR zyCzJTDuX6R8x%SA<3!p@-mL|uD`_9qpMV$}- zK>7m)4MX0b-(rCtg9+>p0L`fB{$cxYOX!_6P6z*HS}dm_gW7M{bCJ3!v3LdBTCB?6 zbgD%MyS_**@!%<`zR$MtQ4SXKtNAi3SGlFSuIqSQ4fapxu;*(|#4=gGXg9IoMfj@S z@C1Qp{ijE*+>(h&&;7H*>swo)wd)VA*x(UQ-40t3zTuO?l@VOk08lScmh|M0cR`=Q zF9~eewCFB0LB3!G8|4aet+PwyxaIy010iF6bCg_XqKP#53 z$0bdWe;^u=$_Hv9t@@FNM5`L&Z0}>SqKRpBes2#N2}DL5*@;~(8bDWZ4as4bWEq=k z^QtO0oYlCLQU^lS0}h@N(#|7dgCk2XwUq7ri2b7F`6c_j!tBZ}3ZFzsh+6JQ$Q)Z} zCFB1Ht0}l6=UkUb_Q5FR^AZjoaF|aofjhRA7yi-MMS|Ltg0)OLhob627Os+A*df3n zXpAasf+R)WB9=kZ)6qf4g+&!{DYWZ!|9 zW*%&{&=WyereKJ1k8kMZ1F_sVnO~2j*9mV#pL06XA349l@U|G((6uYx)k=udKPErc zxhYiWS@!h22jG0r@f;QNBpN@fAWJ`%kND3D4?Fnv_cyzO7dH))JM+vyey}S2){_0B zD_>)I=%XPO_!C%o`5?!#+82uc`fJKGuvYu3E0vC9a=f^4O9upt&i<{)5)+=-(GRDm zE=lDxFuF7_@!+W;IC65)x7@z}q$^?uY4Y?8&6@>g8Gw8B2kCTU9%^szcn1p*Y%t>L91%Ly8pOM9V zqD)ev1#o_84x&iE0Lv{%0!?S!QP)`EfWmJJ|0oQh97dnU)U^f9A5w zA=nFjmsIrv<_s#^a)Lj3TFGhRlP7Ogl$jl*x63!k&&%cAo4K4tx(#k;{-M}O!*14c zXp2J}Kui4TO+{nWz;VIucU>6sgC&}Q7bP<*9ZY{taE;+Q~NzP$wRom?uFm(i`eQkoM=qcB)f1A1;6~pz( zCBaw0wJ<6+oo?FS`O7mYDBHOkO64#be~R+-q-2FM-XT%?IMB~!`&93h5HJLCml|`l z0l;w$@$nivod3RZ=oznXpJF#?yqk#N_c z7oT;xA^CfKaEENi100-un|||(lC>_jq#c|8!E`a*0#gX_@^)ruDR8Z#TM*)!~#TaBM! z7UFB$05g_Sfq3noikU?Vo1GucXN1-m_)T(Xghv{toEByBp=c*)N_Y0&E5MO|0%RWD zPr3=*@B&Cf#t$DU&=oCEx+4$^RX;U>1r0(YVeNbmb+ALyaMiK8?ddD9BZ-B<9Ju`# z=3#52USnX8FE6vtGBir9vlg0?;6$NHr&q9(H>fsHHv9^{Bik13`o^oVSoc#B|FLix z{-u1(ZRmKQnuFoPC2ZKi51l>B1MmXgrT=ypzB|f)R-aER*NT>TiKlu3mG1BLwfT^s z*QCe~2uW)58H`s|w7iR3JK%^^vaL5% zNF=EsM!6|X`jvYJB4>^0fc6WpmEZh=hB_?nU;B%vpFT3g(OEV3t4!Df4ItZES%0Xj z98P~8G0?BkwHiDDqv3XN^rUk3o&E8Yb!~zbG?5C`RZMRnH7u;p2aL_CF_gfRM6$j0~xK`a3Qvz(5SlpWGK}}6L zAI!6B`}+00bya;Lz&Gx64}eodzrZSnIgA8bAuTR!{ZY5kOC0)_FVH&*)F;S70ad)` z#=EmtY06i*y5BW8qSZUXT~3sxx-_fh0g>gb3oUmzf}ht24EMVK-zBtahH1) z0`)G-WO3~&dYq=9_1Lyi7Pq!?iCme~#%lqMz1{ABWEwOa{ZeCe_IRi!lBQs3pLr0Q zczl`>0xld47v*NkTIW4&S2QZipgSemL)y%NLK1T}O2sw-r+r&j3ab-7EXtGy%(s{PXB_K_~4FW02 z1FJW++46lu&h;W{T31y8nAay%mLA<0mIrcV>JkX!|BVcM3S!S4O-I>F-$roq0rqpc z0EMA1JRXp$VF)VYc4C~y;|bwE4VIp5d1Y9V+N77p|P&koi=kj~YJnn@k z{9{b}TH)h-MQ@L|NM2kGBa?Ffr=o|5YjU+XH1DsJdZeA*ME{8O!Zl;i?>de7MFuXO z^WgZ<8z&Tin3rOmn*u@8%WW8SmjO2rHmR1{-$m1ueuiF*q!Ne3<3KMkU_{3-i%ZJ` zm-ul_3kdISMuO0$t|G5MG}I}#%*S`cY`ODVwM?^WF>xoE83|>%AH!n;ZN8by+)l%K zPzKP~7#%xGWSR=nuNXp>(Ez03m-X_%hEKb?- zamkvFxBi%Lt@8e=8Be#~ZIdm$EV>yT7UOq4UBV0Hd7iiM;Bx=5y1MF_Ta6@UfZDj3 zhOo;@Q4JAblzR`ZWV<gm8v1%5U-zKZCc$x1}tAr}>GK?}ckx|KYSz*}pg z0TiWt<+X=#o#`Gf4wg8>aAC<>E&E_XQs>Ga^0mN7E9k5RH$oh`g1O@3dSnzv2WCOt zLM6)z_jq7%V9xUgPc55mhu0}6;4eV7AyE#*O9N}T+Ig~XscnD-i4Hgj(|+}{;lA6D z@e}F$`2Z(>_=QyJh`bKe2vN{V0T`kw`Qxw0J&PUnSza=&?&g`mKPN6$#Dx9F z=PNi}ov7f;J7337%WyN#J|-9|#Co_N!M=-XSwz=aG-JK^5ZYUtYf=!V{7DvGUkV!| zbB`X0k0_>`W5fwU!RHNNKMon>LMO45wGRp+ek*)iMs*j|IStf_Ukv?nh26}BAqVP^D(eIkJkVxLZq)R zE*YzB`JJc_3x#OdwAFp)4Vo-wCv$vscRCovyGD1j4hfZum{c0!tqN(r5ntztuVCJP zvX&FC9pRN0-@q+pzB_dztF%K70WNqf2kX;+Ed=~|DZqt(xGQDm8*b9H&mXA|Q&bu9 zcItW7fk1;Jxn=5pkU1vrKv?l?f<>Vw!piX>5Le3Ad1|`6D$0I^#(E2GC-Qr2%;pF> z2c^n=&V0u5%-+Z76O=3#f9rdWce34Q5pMK27a;YM#Jyzf|5DXeN47e!2e<*fGpPYb zJ`F^ea*>19nS5E6V+U3ewjNwuV#Js@N-%RmZFHkil_7rbUC1V5S$J)KmU-FkE-yo` z`;GQ^vFqNz5<}V>JcB|pLhOo%O^@qAwf}Wxe*p{kPx&aK+BWKUA-H}k2>kVu2-YX? zMa}Q!&Q%Sv>gV;R2q7@G4S<>><%pXc0}gzco!$-@IxymD%u6$6-}sGz<>Sr1s&<$Y z`{H&F9ABs`{S13Ixy$X?kcc~|t14%sV%)QHAv=-SqtM*C?o3lxu{a?RU1jt2kI0wv z7sE?J%cQ^fnk>~gXO_TQwoihU4Z1O6`r23Vi&k)u1PyfuBx1qTkAdUDOl-?RC$VzA z<@tCqVX8oqLffb>N^1g44^8LZ!J{{`NIs|ae*j^)z>L+S;`)2AJtdR+{rQ0tLi4MR@ny#$s^9qHES4?-Nk};R z96L*=Fm8xOpkVx?Qp#u&O|a2e+#QAmH9jO#NyPYNad8C(2#km z)-6q#nD(`Nzas4}$hT<=utso};&YO-`;Did6xurBqa!1J)yr!-}_}Qk8|H(^;jnHK8BZy z(1Z@6GJNGanE~DT*Wn3Z_BbhUrxphp>2ZYOhwC|RBr5iO);f`0#G$O(2dzkj1)&3& zjeH7--dIxtD4oyVl2U^dzQHgLmNMwm?YqHp<0gm+zhmcSeXsi+{>rav*wWq4Hz!NOG+5xWVj15ua2qb-$P*Qngtr3cH z+j&4C-Z*(X=;2baEJ`I@3Z6}Vuj!}RQpV#lR;2TGMCe9pM|`LTqqy8uOmV12aKJCZo|=F$_L`#|LGK!|{D>Fjcp&@P!+G4;y~A z2_U|xet}}R0l~m^BC29jvwq$cYAf1ASQ5IVx&o>cfj~W^dsYRSzRQ658!l>@iusQUF8;p-0Ljn6W(LfREL z;Q^a{WZl~Y|K1$z#jPQM^s_}KNpBz&x_8dQrHA>QHZ5@1{9GtZR4Wqfc`J9rZ^WyH ztrN-Q!f0@9KD1MHiDax)t#-qV0d^Xhc<4x7Xm0U?K4~}b02F*EVg7t-YxWhs^AboA zQ?oQ@Kca3)v75aho3zsN;hO?2JOB^0Js?Gp<@NS!!h?3X2DSz>8z>Bzz4mcE;dN?9 zZ;CPqU%uS_aUG(rL9crQMb|x1c))XA6P@C5Zlcdw(aDm*UFMFaTJR6MYU8i$t-sM;iy9LvOUNvCSShlu9y&$m7zGoD)B+* zF*L4FCcI`GIGD6Qo^$DxEywtqX6mk5PV5PoN$t8Ji5G#hWIQ@6xpF+LQ#k2A@oY-2q>K<-h#QLlkh*wqmrAhv$aN5N@S zMFfLKvXqZO{l3YV;F3A;Eb+VDydT#7glKO)Xr_Rr5Vi`sY9{eo&uolvtgtw+!z&qY zn6eF5NX3+nDB$x`4>^=s${e!B-?A6aqhvHjpC{dU}{&h^`@hmt5_cG z<#lP^$U{uigFi`Z&4|;ujNnTqK$lnBQnTPN9M@#sSKUeWLkgKcsskR zYpfJZi6Be8Xb&tf0y5Dck-NP*2s^Jphc+&saQc<-?a|rkm9*|`qPK+WJ#0?ke8f)2 zd-#V7)B}0l6VPq0k1?l_^sYJ5GvlHnK(U-u34Gh3nR$kO|8u2$Zfw-zWo>kBa_EoeW>Lxn@C1N2(y9Zf*21-y*-Q?{NF$AB$S`v~%yAKgrJ97X=< z(mF2XQUDH+GuY=8AFq}5*7CMVFbRhUeYZrT#U6{KBw1 zc$znHIa;CopNjVMP=(3wa9~&xYe{1{yYSKjSHPFDmN&g)C}U`9p>=HNAm$yP^Gp#7 z-UivyYOnv|>O>nsxbo9t8?|ybQ=i|#KUt!aRsr8+N&!5$MP;y$ z=HVAwLKc3`vAEAe1Ge)+NYix+A9)wrwN#Ap0(dOwe*$Af5=)3__0VRHBLf0+;qLQr zwE$tp7lvFbpR0KINbA;|aC4ab{;vGx3S{&q6|iluvy?yGmqb}Q3v5@&)a;jaQdH$n z_Giambp(YN?zR}_b1`UK`lECA9B2I7L6+wJ%&U44Yqu7OY4lSEFK}u#wkBI=kaR)0 zr(^!4@N^{wi`io11)#|*!dgz(N>hj1_HrvB9i{phg#wVc$a-^vSTHk|;sHss%Y zOO9a=;Al92^+LOn99zPJeJfADq5%JqgrcMnNx$nB8alB}wz7VR^}RQG1>ac5(Xlqt z=Z@?z)aq3wB$FeCS&w6uAdm1Pr)8ls^YS(lDUb2A>yG~qe?Wl0Z)P}Sw-~HID{IL} z7c8ZJ&$tTKkw?DF4g}j3e+0?g1!(Zy5P!3p(*fx^yQRJjaoW=M8yaXVcit5RZm)iN zPP``Tk{U9xH*S+)DxR^FjvusOG}rmc=R0xbtf;&WibOo^`hd^TBA*}!ILHv~kP?Q| zLfPpI3ws!^REQ2&7N?7k>$vATzU&YeEtXh+>*clXjQ&IzeheGIR?c({tBAh)^KKj%5bPuj3C-$JZpx=ntTazVTIXY(Du*Qh*7%j zG^(OpO_JJIr*n~oxB|bpWaG(gcaCTh6v*z^Bn>f`-+wl4<*6rHjZxJ&-dB=|1lsv- zNmO=yffL97L{J4GEx&O&l75mfV{yJIt7M3{{lmm+XS#O@R(TuKdRpBQV@!;HKXfy% z3^JJul8RNqx3Kk``AWzakBUamay1@Z7ksKC9!CoMKYDuSY@f#8&Rg0#{j>ymaYXK` z4T(9l4%O;Zb8EbzGEPa5z0YZXQ?jHHboe1nkpr>r)423wQmQ%ME|`Z{hF_-@MVTZ; z-wELWSJF2IzkO)(%vQ=~_T8SS@A~TZZ7G3M^(PacFq_7pCWoXy#E zFid`+SqGo%h{v8n7RgFp0b%($Gc^#qq;8j>I6v{55@x)qW)oHT=Y6n$CaNggS0+t2 znoO{|dZ!&=r5;Wtxq3)E;JR)KqAw$MpliF67jkaFr4&Gs(uI&~iY0&DWQGWUFRa=C z<=-Lq;-s&0`vY<3-{`)(LgK*&j`vOMcEOGfLI2GZENC4f0{E7qateawqvs%JC?R=0ltZUJMBVwOHca!dUl;Y zUq@nO!24pM(@ktR8Cly8d*^KL;w&&Yl`1)Dy3JC$G!^d2!&Q0I8i^$aOt>egaDJuc znzbv+GCkb3)v&nu>Rr#lXlf*8u#$Z_09W;`WmM)?E7Fw)sft~cQcnu?bJ2Okpp$bI zTv%JGhvnla;F4#5ZypK07s#`}ff1m{&hYJRyNYB>-JEB-c8)!bzK4$4e_Ou870ByV z!6EiRCbwnFmcq(N!y5y*-aVCwfye0}6WGGLxNy!?Eh_3+1lME(EySdmp(oWIyt2#A z+DKvVlRs(*(E}H(Zq8b=EYQqj5FFm8Iofbq2Vtf|GQu-|f{RG`rEd_ostYlvqE9^Z z;e^~p>aEN>vICBgZ252wl#&rBkXx9xBe!=En7nkCFg#_{wiUeV$2%FX2oiIufsf z^gPRiWzFOiv=dZ^k4Ssp&N9e5!VH(`nzdW{-A?8g(5R}=!{0__GlZSuCEabcSOTD%O zY4Voh7wquBPh_F0X^lre(=s_B0 zf^5J{4;2}o8_i<-Yz?S&KrRnVNC*xR^hR zIfq8CJPdm{x8Rja>x8Q}$C9_BW#kf?h5mDYA+P}=2X)|K_!#Fr%hjjuwxt7U+8LX2 zpb)-q9Wc{xj%)JuJw>9xG3CvOFy^vkZza&6M6z)MrfikD0cyYJOkublP50<>Ba88R zRoqQ_YH)wD7oDbvJmi2Y?AnU2N&anFHB`_XLM*XAT}^Imyy^86$CdhF*B2KP$++mbOlt?Z~Hf;v?$#SX{pgE zk|SgYj1o{7+kg=pjM0Kp3eq5*f+I#r3Q|&nNSA;hp_Cv9k_z&l@jTD}_q^{p?|XL6 zclXuTbzh&)Ex>CG5!ZmD9S|Ti3M(!p0aONPfx-In03c9C0tl3*5)d#$V%-pb)l>o& z2#gmJjZ%O9Uk+Lr1QbiqX+yCDKQI~vFz|K*NXY=CQ2;O$>x^(G zRD?m@01z|`iNN~*BZaW4GZyQiEGg;h>nj0u_mY1=W1Q4P!~niXtTVtA;f28XAmD&s z0|Shp?ub8=Nl*y@%$$*4fAkQvBi0v+K>!E}HzW*!@*+fdqu>Y(fKVI&(KiI#@<5>e zI2-70E&YD za)f`nd7%mMP#-AL4eCH}_}w`a0Mal4KnVl>9nTAfL3&`lB)pJrzebe&m4Q5VqLBI&t?l1Xgv#uz#FDme_;D|)Q9e)i0_x6x9M6@`cHO(8sYdyCM+L=^aI!c3D}bY0Dt}d zeY!({7aWap^Z$?g@BK>NGSr0Vf<^yK{BN74CfW}WNT?|;Ee8ZhNy|zB6bOf)e`bF% zh9duV0sK#>KFSddkOKbcmT*n~>e%Nm0EGXNLj>^8WJYKLY7qe8|A5>UCc3qr5r0Sv zM!=EY?*DDo$3h82&_Fp6z$`8$D*=C${i8>EfslR(xG@q7bN&OFKYH_DRC7b35XNXP zsUYLJ3216+P z8xDk`|B`=`8-egcz^G=X&@h#7mzwaF_sw1D=IV$JG4Kgp`O5zg zdqqd3k5gNgyvky(;L>ScvCn2EeV~IbgBCLc{FK*}m8wsx;7QXCtL=X&AGwuFc9e56 z)g~)lYeD83&Kdn6?LI)5K_sq|JJ5e5Oxg%<0rvb(63JUhLoz01(y$gSZ}9iqrYNAeGqy|N-g}MQ zuH2;REGiX6dv@PsQfV%JXYw4*qqZ4YC-JqR@4UZ&*=*Go2`i23zeciQvNzoPo??z) zXP0#wH;*fceehxkc+@Uep-zR*dY_Impn*Aj$m+aXHb#+*Hd%kmBu}pk6Y95-GSIwJ zcXcL#`*N5DG75|*y2c=^@W7R(;GxgPcZFZ*|&Y+wzJWK zl2jUcZz?n$Del~o!k4OBBp3A0ETg5c%7g5ct@heAC!b<>1n44Gh{9lf-|pBJ@*D;` zId4`m!Kp+}ZoPjO8;~jBG)XDl_yMAcinmj(z`dL=q^Y0=r0b-fYeH317g!sPg`1)V zV!ZBj1#)5yMdks|c30k=C5?Rm3&A<&wyqZ{$4dB8^GDnjc_mDj77!`IB>M6RX*`Ag zP#q0l<^@X=b?G-L9J<8n?6btPGJ6hgd%m3pwv+S6X6Y+gM02_jYqw4>)kysQSI?Jwn44m z&%~p6l-J?5L8UQ6eO0p6+e&GA+I*ZmlhHTg638iIYm#FhO%AYE8gO>x$R0j801vTk zkL`GAU!Q+XO{#z(4u$NlrxF@%jI&mWVBUEn!Y|Gs9ptZ~#L~Ux(S6}UY?gv=^wSQJ z0L~GcOQKTro&5N>mx-@yH3x^bJ8<4qF-D{n2GQNcgy?+$vLv05*$*w% z^vxW3$s0YXUsck>hQelxi5Z#FPnBXnOZ%&OzO8>-iMU>Oh`M`d^cw92?p}!IpPMjw zl4#O+RZ+Pp#_rgsoiS4-WPtbS71CVn_NcV^@^hvak{TOS&gYSS#=eG61J+|6cKqz- zPanGV`E-dy{cfoEG(*)EnWwZx&GA0d2+YMdd~mxUl-StX>W17-4B6GsQD-fX*OF)L zeo238z8I6NL$cy<MghGfYBM5lbY9etxs|(*y?EFweK6HxF&pa)cP7}dgV46(Y=?0A zTvbDx6i*F?w$~gw${UrJYpvhYAYy|!R7ro{#J-q$JsKPf%K3(?@2{}vXH5181Br2h zZwd^PlMHD2mSi|Pka@EnHMNUvuq>fJ_T6pWD~h*27%9pq~yl#Igo-Zj%u|9q-`iSHFi-cKX(TbR{7n#@w zJMX2}l!=_J;k6aT@z}k5aQZk`<(_BIO)mM?Sy!xY)y8LNj*?B`NN}{8QTg}mB)6@> z18#q}ix_FpkuzLhYRE|M@OpRi!Y4v zWEh5=v43ge@V;LMlV0_vpEm!w$@bLs=H4W5zP8veGIfBKyLlmlF<=N!QviheK8%@^ zGe8qJq%aO=@!Os3e__W>Wms14fvrQFe+bm*suw&l43^ZDev^G4{?bVjh+n{ z$_a3Mch(7aVUc~k$k)tfeyxA-ObGs_dB5L@BlEB}BDbbso5z{YV)3PImW=y?ji=H4 zB~^JK?|IH!=`yuoM;@C{l3j6|>8y5>tr%a|MUONur58^pwG2U!gNW{zD@UGvkgg)S z4!(p+OwlDLvMYgn+1!r#4vnkt*(QpEtzxwGQv}MD?YZ0+J27C%Y5jjim!V$LOXyc| zuUn#@^q2a7H{5d{)gP%ROOUkM&>L8jZXIt~6g=at`dZH>(tT!wC>9U47^{7`QOtki z-4Z{#gOADLi<;iTlNkU`)sIz^4n4|_EM-<#z#Xrd_I1=wlV^X7>_U_z)1qGUG7(iV zrrlbs5V@KG-llwjsy9eW9sQ=lQJ|?r*B`dH&EEcGh3Q%5a!Gh#DC4b{oWUuiEtWjTT4%`}8ho1gF?zhYXk)*u zrfu(2E;17HrYC=0;*+*ScUPu%O;^+gdTDH?GA-XGu3g;AGdPtx9Zcl%V6g~UKs5Ry^gM)llv{L!SCy&lzCqTedw%i~R)NS@hHqhY097D6SDeYCYs zqbIg_b&G#m*5y$oH@P{5D_tNix87ssPzNg1q|yN_Woe*2NDf!3bnl>b~S`m%`YL7J?NI~RYI$(NWrG}y+%fA!>NewW7jC__<} z<$&l#EXYa8)hlmC(x@oaUDLiiPQ>i(iqJqq<}(893sFv8a;jX$gV z$VGn*>o`#NoY<4s&3YHY7m+ybExyav7sm4sMsptvWf+iHokWWOM=VP1O+H-2V~st! zAD_-WQFn9h_nz(o6cPiO?i*aVN5uWUG&eR}tidkF&zmpF-D_KTT7iA2hFiyBakF@TB z4>v(N-^D)L)dEKuLKsc$hp$M4=}+nG^_!F*J4ukEx|+4rop?*{f9v-dJU4&P zVf5-uu8HPe0_o;rtwM)JMHR;j^5`Bu+{CB6sgZrdLS9)O%8*>bu8f|`&6nC`2yfrJ zG8AwsO%iX@$Ch#Zb1S}(amx=9mBq-FyG{Azgp9OsC58T@7Z>6Oj2bOkoU3DKg6DQIK^01zL%R zKbXKXO7%N-hJ0M)kssvF$7=0yE(%ywR>Xgl{-Td+mR~J@Q0+oMr^sm)> zpeQn0fA&X2O3VrGRLKjG*F_Gyx12w7$L%LKnXp^mEd*9ZxYj$deGM*_iK&F&6I;gJ zJ_@5@|HiOdk&?M$Kk;#kWz!m1z&(DALjO7^>e1dIE@Z!PMqzTOWgPT-CTUlpxsgCAZVufmw%(gMC z!590e`Xqe1a+}E1PX-QCj5y4%ldkza;Lq_pME)dVy6-%>sl)oXt_#39hfl4eRG{tV z8EHvx%k?LTXCh3#jdR|y8t%g7u6eXcAK`5{n#NmQgCC`=Ro}iAI&^=rg1Y7F%orG^ zIgGJ5Kj0qD?YmgOm*jV2@A+5j&{(w!u_Cy?2Ho*Q$Gr_I9;_vGbQ4@_0Xf}tny0Ig zNMc6Z6TWkrxEAzw%1)~Z?G{-bi=2M_Zwlc$E!7KZOH_ zn2)w+oQywvIN_H2@Ruqre?S0m+XHgyu>r+nB>-)knsb??xM00ZG@GKaQD zeq;@3_5C?R9tJ??&$<^EyC+8y*AXoKR0Z0FGfUEN@d}eq^Phjowup9w)VR{`S1WN# zto!gX4-P3TRYy);JEHd&3|xQr!t#nz<~Mf#&6MMOaK-?Us^a0f0BoQCJ?)QZ+(%aH zk+w8tlKqp0fDvxj9F&KArv*DC`SMDS$}o)Ro)1p#ems+EPKpWN?w-T|=H)}2n;W2m z>wDu*BdUSyi{gI|PoHo#X&N*fF_ZZ4g>3yjo-qMHJ2xiP5$>Kn)AQ&}JO81R~sYd1_-hFV~ z1lnGKvkIB-Tpta~zuJ(+p5^BU?dv7K|CTmDl|+kUn{R*p3yVN|6pf3~(}-c`Moe6y zy!=A3ZrK2r9WyyAw|wloP_jO+`Ja5eo{F`Cy7gtm>1RL2F0}OYb{IC})|5k+GhCwq zt=BA^IX^ryN))_s6qDYaSe!P*x)B;AFSDdkh#E7g2Y z`B2c?ow=({NK}06__f@H#4CF)LY!t!tOZ3YVCTl(2slmTN2)TDNzks!sXOyUC7mHOo$J{wfXeU5kG)`ZxNbn88tAvDjG&*gn0ef=sa- zQHp+mN8z-{TnbzG%v*&EAr8)M12#18xHv(tMVk)kg^sFPZr@2xY^66jFfZ?wk4kWw zYv|bTI3~_Y?Y`h(O8Vld?$4mK0R**erUo0W)X1cRR6g=8E%)-k6Q7Wq8b9-kp9dB@ zIdXq{(;YT;$DS0~l%iaMU~QxG&(B8WO>Z2wjB;7Kno~3{k}gGM3{KgdJ;@8S*_-Y! zKR$?^A+mkSKA@}ywx;r#?@W(>mQpt}NTLDK{nYZ7#++ERM#%;V6S$3C#`Yx4ZWO2K z}M=k>u4f+H3oE&Gl5V@L198en{Biz|W@ zmU26nx%Bv9yC_Fdz9GWppl!C88#5GYxK1+5+5pnAx2We8eRS)ZSv`)2#9^#JGr~eQ ztEbQD+cb9Ic6h&#mwl`_b>xPIwW0Qf_B#w)vNrNnxGw<|jF~;Ad-sj+H7h!cc-wzx z{V1X#L&~q1#cBpDt>&0m(A0Scf8DoxrFWX@BiUK!aZ1T=VY`#*a@HTs+YAQRn zx?Oqt_Ly4lpu&;voz~^!G70zf)5m|(8je#zh^1BCPP)rWj$xnf@l^d>SW{8eHQ;%$ zFgODKQ~+-N}k_b2<=BUHBqogQg5-}fs@l8 z`H=-H3f~h?ED|6sNt+w=$)g5VQdFfCl;SVN`r602lv<7fqnh#RsSGRv%7>6oAoj2)BiZ}6l~7aHGmh`CrY34iz9c3!6Dhu$hZ}rU zWWsGhp`s{mQ!`-Lvo|F>)b~Q*@=SJ$XFO5u`ssE?ZPoDqWH*7>bZ{g zk$~brMWcqS)-YWmd#GGf~xk zhF8C^E9>pR)@4*b%a77siK4gfO0~_nL6v&3rmXZG^kUW@pL|o1?Ly6psqZ|#(yJjY zoX|B9!kXAe@s*qxZ#Y3y$W%RRd&?=r{Z(vLFBC_dM$2%=&r0r#Czyy;W4N4kp_{=nr%NkR%^^p{&2$KQ)7gSPt9flzKwY^nHozH zY2bO483$Kx;s|*YW3u=kphSF9K^Gix`=_a{CO^GQFw`y7luc)nt35ag)w%&iZ#z$07+Q_fi^(||1RAEd9 zKhH9CB)pi-FVc0i?*-OhR~(Md7-cngj`5BiiHFV1x|!Z~|9LbitFbHzEEA-E$n|I* z@}ifB`C<7Khiw!68LC2^N6UV?Z|@j|ZBeJCG)UjqW2-590~$5s_J!mcmB}iPX~*?j zb#aHC=E#4O7lxmhI%RJZ^mdas-;4%Z_LR-kCXy}f6ucN56I7HHd{^`HYpdOO*@6k2 zCSCo0)S475wJlbCzFY=@z2q8>2yZ2Q7pXzAw`mAeb z+8T%>m-%Jx1$#c^8^7bY&)v?w=g+_Ts85+DqrX%%DXOl+IRCxNa!+wG^adm(Nj}l_ zv7GhE4a#q1`oL1)&8M^<@+ss-j)Bh_)W)7yn(XU#u+{q3mQi>sna9Q;@6162utNdm zt@|1e5;l^e76-PNgQ6x$$~{QEQ-+H{mlauA(tQ#1{b#| z?*o~8mneb;6%jEtGBOG;Ol59obZ9alH#9JpFUSHF12{A?lc75)e~onoRFv)3Hr*{C z?I0mY4J9Gn9g-5_zzi_N3@}3p$WThB(jgMkjdZ7^N-5F|0)n)pz&Gl9&U@bT|7(5U znzd%0eeHeid+&Q+`x!2F10w-NFx(cR28SU9gav`p0A*cWO(_5nC@Kg9ijZ(|8AFjS zkiTFOE>nmH0t$yofB#2-vIhi %5QBqmN54g+X;x&VYl0mApBg(akcK!6AkDD|&E zxQ8@A1>^+<19SxeT5uQyLBgdBck}gtIyfRR$NcLNz+=Y?5SEgX;QvhrD7r#Cpmrb_ zKo^8`gt%f(v;(;SjNo=q2-5c-DR^WYkw`abAt7&XZ$XeNe?k!M;ULG$5AcQ}9RUv@ z2#ALl1Pu7qFhCFF3i;ESAPE=1*b$2OgB!u^k=`H=2mk}PK80Kve&e*!@+2snlh@&Z9!K(?5G-<5*^YKn#c5T?Pu+C$iRK;4iCK?Kz0SBpZw z(qL{`6$Vy@yShSPNCe5R`c$AE5IfAZ`wIP;tTPPm4fFfUu!q9H_P?3{d%6jkz@YA) z5KWc8LNE}?-xA4(698yf4**z-%jCQFs6e5KR37=z#h{C zBmim;!TgZ;AwXUb0Mf$~65#irj(=Jt!omPB)D8)-g*ZTAB!6eez!3XCHfH!9P#=IL z5TiX|0Pxr6-%r*Ud4b_D7vI0bf6rG)Tg%8uSBLLU$N%;zDZzaJegYy=009wkAV3%> zAr6qhe>?>IGmQZV`d1mizqy()dpL&ir(DcE{i|fJzZAgp*LLs%{+Ub;de*L(L$6wpJGchI+dGw^tJh!fb*f%mJfje*s}JL7>wi_5E%Bqdz-O4-d=*{FWuA`G48J-!c%055$h-^EBMBO+v<>PMU9e?+6K%vtRcwa>l6YLA>c^pxCBvlnI#{m`|y z>U_nH1#YQrINY+0pT0iEo=oyPVYxj`*@%^EmaN1k{l=pY%YZeCs*_kA?9gY#Md^R1 zrsq?VH#cpOnuJzZt6qC&RZ*r6J&Tm-_R0fflkK4Kn)ngJotzcdxRBLotB4f6fArhh zGVNz1iQbh|O9MlLbJV@8T~R?=Ucqa7oX?Qds`m$3RKy4?P3K-y{LFddwUtieXRCfg zVGG<>y2zV+TBCR1h(6e|X20v+o87?BOSD~}qn`LQJvVkJQDHQp>M+4-y6ib@f!0TO ztV7;87v)pQIAREWW9VrGIdMPPtpW^mZor0$+>|j0pe!j#ZZ)e&X>36@B10fBbcM^9N8Q z(Z8?eQna#);4Q?@-&9*kzWZSxR4+q5P_L8Ha}GO*O{~j~0#R$$Ltr|~6#odgh$HVT zdp2rHGZARG22FVz!VdpA0wvjC;i1-VuuX{e9*x58OrTFbp4Ix8FNT&X;q)J#h2ocd zk&gVLaZByBtiXXs?<)ckf5QehXTGA{saL9f&r_+Y;wa-6A@2ySvXuhk!>=J9*ANsu zH`O$HuSoAUK2=P}kU5L;i9shto7aEwEe_YN#9@nkxujUeys5C!vrAL%OL^z)tsc3& z0++5uQ&w6Q9rMD7wNNYhYy}Ey@R&PL^QOI8rO?AkgY*e0UKfJhe~fiq`3**Y$iv|~ z3xZFJzeQ)?w%fBC;-NuNgQ9A#r?)d{y+Apvp}x?M2PK|Grk_y%msV+4XQMS%(&XIrg+$v5!`PM*1dAb%BOE=N}Q1BYbMEfp`X&pNOdoDGE zP1+E?y+3dpiz?AHY3tn6vyD5ILI#w;AUKsAdFfLVupRfd%Vr^V?;^{1>c_i9Uy2Dv zeAQ+i)xOf}1=klQ=k0tOVVUtAU0mY6VJ=rh9QFN{?6aLnEgGXRg9+o-Zz!Inn!RlH zd?AH3f8F68YrTE}K@VqKC=*zY;@P}6eJ!MzlE%!Sh;nFxkwD9Qrg#;1?nRfKVVapLM3XwnnuZE#$-H~ z;rj|E$f!2~bYL4|wby|y18X!Vk{5!qHst34vjO?qnpbaf+;Z4?yVQdS2f-FyY>Mxu zfAI@$AA&ps7j@JcogM~REN;Rx4WEnG&5@e2JF*PBo5WBWq?X`Bm}Xh$5$vKW*J{?C z)}g|4)m~?hCo?sde`r(=KmS@Pv$<|=Fyie}p0J!5!&6{TH~9muw;klgm&6Ie?pCaN z>>k-yy0DV05$TsRx~aP_Dix$0UZsSdf29hYR;U>7ZcS*pR1UpgLPnS`;OLig8q_#o zk4raPM11}f+$2)53fVGCF2ZKt)1lq&HgSW8BnkxxbiXQ*$E2Q zkSB~i1} zmmV(hO-8_&m97O-%=R(8wDI*(ou|lKQVZNRu>*H7F$5+p$ zpFXiY^n|AY&qPluqy>9o?-xsJe-;9R7{2+dgsRQj3Ox1?DDRD5uvFr1{_yT@V{(Id zojU7sjGAMei+~KuinzbgPAty?)gF`+_G!UkwWgMGq9o5S66=W=gs{#sm%nkr+^508 zyQ7Nb>6N)i-IU0l5N8^_k^m*?0|7$^`jee6JzhJ+!OuOz`FHV0-THBfe@+qr*LpEi z;w#M1h!R5$eLJUw)uU79?3uR8AB06DVJ8aBH$|j;a6&>nKm)a@(iyj?>5YRq3AS`E zzCQ}&&TR8HI8J>#a(ap<$AZWB@#mAESU9l@V$bE9^`INskplALy3_+{J{!IsruZB} z>198pwrt7OC^Zl^-$7AiBZ3>jV%q!=7@kM}aV$IWcvZI!QqQI2m4Ws zZ6GRT4ip_hDgYC?F1zBu1C9wBIsv*%#4$Brfe^ufRzb`~wIKa^k{@1v<^@m;pN;j2 zwlv{7LtE*cS57#le*+MEG-qt9o!LNBoP)Vt`Dg8Iv--)RnOov5l8Ht%&8XE6zOyf# z7~3iO9kTN^@}o$PZgC=nI`s6WCgclwGZLj0N=6vzDi^5F+_huQY3&$y?K3Fq4$>Ky zPVJJo`J3uL(tKBlcP_h*7sKDxTgh3!pKW7947JERJ8^tEf2pu4eeSt`6r&aHJ{Bh} zgwLmu^n?1l+6Y&;V?kr|y~S1yi-A6dyVk^`&T>Cf+d1vkk3Fn*Jv*7XepK;acL;K4CxM0Se%zC-Q1gJUf6AG)@06eBc0P8 z!3yTFq4-WFR@*FfYWX2mEw3hU$t{cNw&~zDr0u77C88QuBug-GSSROv0 zx?X?HRg&+|CAE-oRZv|^T#YwwTo~h&vRNJ5&CZg%fBZ}={gbJ%P4o$IAU*`{_AMT~ zsG5+m2__P1Rw=Rd_e-Q>T=+6?CP`oKsTPU2H&-P(Mx;ZMciZ{XdjPi_c(=8sF+9_8 zW~RVrns__<9JTCC8C3mvjUxOyWwj9@5Ge00Vgn@q$#wRr91>sOrfz@wy+TJ+q!Vq^|7}r>C~3ifltPuf38IFdX6;L zoHOHE7{xi=f}f#sgHuRWEh@}#;CK%<@7hUYL8Hb-jz2f1Y4tO_TAN!Y5kK)9+05*+$wsqAxz+ zE>+ViD~{Wz5xeYb1nH6{-oF3Sx+1TQPPyckxB%{#-tN<$fn;I8&Ue}1WV;2<_)nTP z?2ST|qBtxAuZomX@=<8@N4R>PvNeW1_tR^QoXvhF%S*IS>!`XmBME4wj}ZG)=zws& ze<#G@rW&A}McSZzXWprwsy`dPqJF5fzIa<8s%UXqTv!SoP}|Xc5+ntt=w^NmDGQdq z^)~Ia%TB944d3L3#W6bl#cQx%ks(>j7G8pUGwpWp<#TFuwA832L-n3;WiicHdVxoW za9&tPnkX%v#Kbgcy>6VfMz){2|Go&jf3)nqOQyc;pD5zn9~?9dziv%|a3lmQX;rDd zzEvoGME6)|RS~VNrJ4V9P~PkrT{}yqqK*gc5z)Ed3+QefYs^c~xg(u=bpezy*n8xp zV#yUm`!LSc%P@8v*0FZtnx&fm+`8S}+-boK$9H*T5ic@*X2RZ98EYIz&-|fGf7}ez z^e+4USpaB9LhAkKUf)o0n3zBc`@}H=wR`x3^S&u~1i+R2J*jh_M7;p2a|np9Ve;-p z@s()*O#5iu;sq^d47B70q6gntNPq>5XgkLA zm7EBT$TDn95>RK)~b|~9`jO$9CbUdYo-KzbMK!Ip8ZxP&*(gn>uJvdu%u72I;t+v z7O#cXQP2>h*{dIpNzmYDvrTD%CXr#IFXVq1J%!5Tp-+3Un_6!hXz?YADV*BIo1C-Z z6!{O82M=Zxha~fSo^Aw)f4Ju%8K3nT&42(K9^8KazyZ(4oxM4enw$f2ZDBa6D} zw7FEzh@kVDI}uc0H1;#fKWmM>{%CPxg5Dg)ojQGWcL?$*Kl%u-vt7R5mBYG&MQZ37 z^+a=|$l#c)VylVK%PxF5LNc6AAA38xn*sHT67{zLZD_t2A9j2gUd{XMe2|^+EHz}8nEPZTME~+C>7w>!yBH}wJ z@kyoNKPwcQP9Y9naG@Tt%T1LMKh*??1RlJ8_YCZh2O|^HYp=YZfBj;{gI_&BI<3m! zA{)v=TbZ2?dl(=gf0W2!F)7t_=N{bsFx_K5DQKyU5mvjsv)(LS7rA8-SNP6iiFATi z?o*)lL|7*OTh=i0yJ&K^dS54x7!ifQ!5ml8wJ2@l&2SvSL@^IqT@~JYuu~sAolpJa z_+KvG@B0~`(*deC%&8{T39Nh5EFSNC9*ZTVy^zG)cPhXie|Y}kTOj{-jjy#B2`)7Ztl88O3G) zDK541+_C$Nd}Y;;1V4={M(%9P=cATLsIgA{{6S?EJZx@eM>+zhsW@*)?nIgEqARE0 zve&i!>%DRV>dM3TdFC!Pm#hcNX&kop^v9ytGcFecO-=a+S#(mDps#$-WH2o zm*Z-7S81T#k&*b;(Ju;B=?do=xU5ox!P%U{9VwYyf?UC##YRvn#?o8NMyl8&KeA>* zfiiX)IQj8clxuLsOFu+t!ik4bfD%de#?2C>ePTY3f7<4ZrCZ`eR4Z4i%U8QxbUiuZ z#(1)%HXd$XcB#OuZHBd?h2&bpw$^Q{hZkMxdocyo;K1q^!hj4`ZGSw-%a5b&A#-2M zi<+G|M>_Ci(gLFmMbYYZ9MRq~k|n{q-}QM~xaB>DdY!aMU6dvP5b~SIsfjS?`^)Wg z+@D`qe@Ni|I;7jxFC+AqiZ-erd`Q$4K~Y_k7(-~+((G(r)L*~Y>M|aQz)h#@Al-e4 zeDhK4s5RYclKyeyc?2wB={>t&=x)UNBAO=FN68wfyt5z0-XxSgxL|ahIs8wD{KX6%5Y(wV7exNOt}&)=1MH_cTM4978?ENG45P&te}#z-**ZLZTnTG<@>2rW>#Kq*AFSb27V}Oa zZQR6I%9z3CEZAEl zG>~p8*a2#wY4D%9KX2z5!ozZ(rXQtWsQqa`#r573e-`lz_(NMdWt9)AXR7&&e=_68 zcgtcU?-{(ZjZ9Y2=|8$NKIh78nXIs_zs_{w{bi2UEwW;*5<7HYb&ok^JXT-v*(XqJCGRt> zco0IXL%ILc3#iF-0L0!6SDTfKTluc{ndx=_<<8n?1BQScfuz;=bN_xfO zz&AK2M!kl}sV@G8EW{yHn6R8c#$?O4=qcD?l1!ByHr#zVdZxz}9`!5-INrvs&~dV< zqr=VW+l+m}W~#xr$=R!NwRfQIkmujQAaqV;lG}01sPnDcWy#=ZkU>oIvc1&?qV>v} zm?R=}69v3-W}hPHf9=DvXd?5SyS(y@|~OBk{Drg(`pJTf7wht|n@3#^A6AT;6B1JHS-*DXI@ zf|qxXRPzx}CH7twQq`koXKD*eTwDQDL7RuiG5yHvN9~!tf04+bKK<_ylJ5bXmqv`F z=hXtM*Sm+bTM_J1S|ub$OMc13vy{du>y$pbL`}?0N_y2-y&TW4NCvs@JTE)D@Hf*^ z0Ka+>{*1ANvebX0O=^4&e|UPC+4{cb{=)vx+HOTqQ?30M?2-jK=D}~I9BAc0?jNte z#>c^#^vf06e@o-jCkKjCstFOw3E;OK;w~sHL|ptMDQIJ;S7H}y9d*bdAIGW?HTF~| z%Wz?a6LaALGby4%OLrW5uv*4+r%UD8t?36<#F6ii`WhvPnZRA9Q9Ns}biJe+zjKB}PB#1J9-ois&((0e12m_ckIxf)RGH-L?9cx zIhl5qS1j)^4;P~5W^n*WJ~w%qc{8{X$8cvO3xAc2JlkfVnUl=vW@UX1Ua_#TdSq8L zAo%4adY^{) zF>0lu7&m{sqka9jmFz=aG3^L7aEujdW@MX5r5DH$O)8X*`^I{7CvA+8LP#0)pg~5| z1AkR@z?fLbz?q^i9$V&fNWvb&!)1D>)BjM|CbcIP^&(C=yu(W8H8JTMwhY@kx9m`f zsAsy$Ukx~mWHi>?HB{Jl)=Es#5ju0XJF@@zP$~@@J|}$A3L4w=p!9BOCWB;=3a2=y^hEeWD2;M0z(@ z3Ul10C@p*q{nX8xNWz~w&P!Ai(|6O2>SVS6?U%%V%Nq6e<<;p)Gtd3FhcE2o!W^LW zqNfDH>3<~M zQX@X1T8G2KUa+J++6dYCy%}z9&*+5db2d+){5M~%~f3>zHH$AB?3Lxy1R58V(gfI`SHd{m7Qu@Wbme=PgN+ogmD(p z<-7*WJY(~dD1{5_RpOVMlfPOA#D9s2XewTF-R3eR53=>4N7Grp=veBCE-}AXG3F6| zeTqV5$=PBF`lr{F=OHZvUJ_DdewqMb%hVehhOU)UlZx-Pu?Mg?2dO;%mQzy}y<^Q> zxM}8AX~ysCF2d+RCd(ZhOU`7WMvZIzrb*qWK2T+$Cr;T{D^}-QFRP0=uYZ#~A3%a@ z92I#~t+|?iY^WV0z{D3mLiWJq5mrf|lBBo5h5@_ly}peNF5L^OY3vM!@sua55+g1p zlEdQdj5_aeR*L5c9=Np`r&hKZIS~2NOY@H)#4Y#~L<3DHDYgw1!i1-jmusw(#a*q_ zH<40fuk~()N>L=Xc|>X38Y)^uY?J5z$W=orvG_(4Z;xj8t_vqX?+>YK)l2%aJbR(T zhE)9@dyP+cm+=S%6PIh)1{JqgiUh5G0yj06v6}-Fw;Blrif@;``34rZx_AXdaRN6t zm$91z6t@)W1q~;c=Yj_nw>A+5@^At;FqZ*Z2o$$If(9*CmlX{N7Po=e20bB{kSPcj zx77ItIUAS%SqK%kA{7V3Z2~wqm$91z6u10>2NfKbu$u!FwFw6`v zFwDTr&>@H*-6_)DC4zu}q=2-Fh|(R>Dd;=K^E~hG`M$NjZ{2IneV=Pzd!N0}zV`WJ zI9Uyh`INv22Z%ZXj^-2M7nBC5=o(u}0R#m__yq-pi8wh;plBH6FHFQ~20@~r2)Oir z|01X$AwV<+Qw5?iJY57FpzR3*2#ElM#HEEKqy+^5!h(WQ{~{ug(g0PU7ZeQ8_{v!mBtTP(zE-fJ7?d{DEbVKofBalw=cXDEzlG5ClTFxdY+8P`DGo5ekC<^wo9v(LQKi01yuTMFhfsPza1a z&Pm(HAf~{-<)J`Gs5=_PkAlK}l_>Bl4Ca*8;9wPmn;Qgh=|A~%)A&!66nC2s)J^*V$jP`^8g1?@Bf7xQ>1xCPO zzW<^B?yrEFiiVcC$^AbS|J$aljPL>Y^9f4<_@u-n0YXBe;s6N=2|(a~Kf@RRp?}j5 z{Ex2|+z|ng`jai@nEsWq*Ixnf{2h+FfPco)Lts=30r31sl-u&(P{Zz)(-O z|F&wOftW#1f;(Zv%qJv&$}cGT2Zy56p*|3>0Tc~#{v(+`xaqG^gF)dC0|W~C>u$k# z2@3v~4KrULSIqr^!szn134$4)f9|UW2O+?}=1f>j8~{Wjfxbk77&!@xi2?kDFarsO z`21EfK!6{PKx140m>L5CjtC^tueJ&a3IhZjkU$Uw2606Hvi*gBMgGEnn)ojqlZF6> z`D^t4jR^|@1e|`|dk`eV!xPiSUxt4q3GzfDF&@9QgsI?P>)$s61mXh$5zTx=fMlMy zG(2fPuTx_1=35=XpI_{VOt9>W5JPiK=lMJ9+7i^7uBIJMH~3T@KJhVt;oi6N zQ6HN(&qL2^1VbuCQ^HATx;%R^N&|y z65S%}HETuJ9jJ9HscnE8pyZT{HH`rstCZ8RjU8>t%CU$Zva}(Kqd;0`c$#RLj$LC% zL*ver+hgc1e3z}`GKao?efzw>FJ0uY;6YEVcy%}>-5o!FZPuLOqyajr#OK=0s2g0$ zVy7zQsR?X!#cttTb2#=AbKj>5`r8r*L_C|E^i)EK^@HBM^lXGhWr_0#A{6szd03ip zx~CesHBu_lyux}9shUl<6V&-?e-5fF4A}%rBdIZ-7 z4&skIS#ZREojlE5YMd7Uo>zNIw>zCIUMv#miS9Qfz0NDCnts$Pt;y6JC81OD3MVOz zNwV4HQ|ZyS?|eOXzC_Tf$89UGG$-;LY-P~<`AgZSRw&;fV#C+Nu3sG1+{+#S$=^rG*z z!jvtnm$i+{3}-@5sYLauTdB5yyV9Nn|n4_HS) z=^#IUy6E&Cd$i;XPED9oY~duakmAt~2mr%W2cRRnc6Sh?L+(u*8 ztnrM!GW)U{T$#~)3iRGXR8B)a2aEF$7>|4cJVS;K7;{o-x5qaaqNmH!<`3BKWa3Ob zppE9a1}Y4dUsu1TXtTeqz^v-Z*|fz~o8S_E^W{$5puFpGhd}`KZr$|cQ|nJnXS-cv z^c0>j*@HU06XN(g6e4>fOOFmiQ#VKOK_WL&hSD_1A7#Gi%xeKGc}pRt)s# z6tE|tz^}@JG1{(VafGw->CMn|qgPGc185urr0Q*WQIK+LtwNQXimMJ}pfk7CR8FaX z)4=z6IBxX<{f5Kg`R`;~R5l#UmV2$P1ZI_5CzI7Qo(_?Cs_}Md4|O&PQm4x6Kp^sW z_FsTQZzppT@sfG$drHjOU5I;%x+f=1Y}_!IxOv#~aJ0mBk!uCjP$irGhqpb*NG)%aLMvbhGg_i>}Z@Ryin0X%&<_b#9oO z4+hk6YRBxSQ&5T}H%v(|BE#0%NAvdQoTsE)SXyYyL_iaB@pWdUw@IvP$fr5_ZMV;- zzYqrA6)|>naQD^yIkRm3Hu{P%8V z*??DPhaIPsSeSi!lCp+X&NpX&AY0n;=b|7lFJ(qYjF@f90dY_+LcJbVVnyarHNXd_ z@rx*vs`aLN2DlBU|4~F?m2g9Oz_N6p*`zFzF>1qZ+G2VgzcA6dGlGEYsu+A>~D zzK%y{!=z{DO$pnN^Rq`pmrO;*%7c{j*nptSv?F5ok49|g1HGX)pY6Xl+HCEow+z}u zlErF~EHNAw2YXr0U4@8$9r#YzlpW451nhJp(nV}sx>lG?#GX7q<`+;c95N~!LoV+4 zcD#NYJB&R)&-QLKWbTrv$y8^&&XT^bMOQVFHNV(8Uwz2U`?g(x_?$#&XpQ{NyTixY z>l<%_s&*|bLQRR4m06y=D(NsVCp>@fQoGDNGfYgA{a8!NseXWeEz4c9>{E=N-)bzr`k$Q`RVf4pX=NVI-!IZhstBmtWpEvTf^ZGwEUz`8X4Q`8onw>>n{tm`+z+HIx>Hgo@AW7Eu^k~+6t>vY zA_MQWVx<(D$PCUI-Xt4UXghU1bj5mrI&2_f+cpV;%^W^^cKr0qr4v1OqpVLz7CDDL zRzYZIv-aop9Lc+2&1ohPfh^(*F>RT&ZYni0Ce%HuR$V)Plm+R?E2VX9zK3Gee1Sw& zmFDxI!f1QZkC5(+J#PoZ;-yL^Y1OEJ`qVHov@mG?B%$wP1fJ0GEfw}VD{S#m{)6w( z`#V}r4mXieskigW$Ew`Ezkm!hW4T0d5BQMf5Jyv8V9S?{+T5+?n0J=L+)Z%77fVYj zNMudsbyOOED6Oa7r_gO9GimsQ+D8l9D_@Dxa{FbC3Sn)a08|iFI5@4-=q5_qt;M`j zAq}zaYR_a>Cl@i^CC2?WCfa=Y5LqS>*BnW3;PPsuC{PLnJ2I&_2Cz zICA1vnDVyXh|<*b1$y!#%4tx#)!2~nUO0)So~xvP1Tm*5*|WDU(kt?UMJA*d>F0-s z`d}4rX~u>{>izyNyHw{ZyM2rR_XZ>edsNqnIu@JX?>a>Vpo* zSc0;D_tVe0lqyVLQR+450G$RmJC*R8E`LOhz+?&2pAA52uXE5dFa zJ_sa`hYc>;WinX1{!Ef>q+j+}S$p~*e0hV#k&KFgL*l}p#WA}CtjRWK;UXU>mK8UD zU3WyrtpsXJ+q;ZRo_z*`**7;8u|EnyAbCYjHbqPx^BL0HzgTM9U5u@qJ2x|aIvK2e zGbC3jo`lnVmz|V==*`T9oki|rHTrROZ#vp-Q5h5y-RP7mPP|6=J!!Zsd>Eiq=-tFjHm-zy@2{&Zo^ZvI*>S!6*uX0EMjt^p6Zi{UfOAC zyeOJ>b+U?$TZ=X*FpRdqg5`2jdf5B!ysIz|Ckf-2+!-e#JbWlgY=g_F` zQD~99-jkr9Xm)lHFT*p($2YF^$7`k&@10y#jirtHb`zMk$)lNhMOqPa`b`N#F(^oJ zN%xcBTc8z(h9&8sUXPN@OI}?;$&HZ6c$&Jwo52SPQ`-Hxdu^diKR_qxcdmL`T*(&% zk5NAI6ARvh!VWSW_B6?VKO}2DdV3OBXY!T$6I<;dQfjMj&f)k{Mdg*yedmBDe8e7; zV~+5Fz5W(S+jhI?D}mA8q;Y+(6YcC_0xAh%njt$&E~GKbq$BIF&@m^8vlGmG6m5HB z^f7A<*d-UO2-=8h;NQ1cy55>reEZ2INrjBYb#toT@^K(*{$k~S`1m+%qjB{6%OY7% zS~2mv58z13pm`x(U#<0;z<#`)_D6oF zz?YV0#jhyP-Ww&i&*647_@8wE(FHgI+Q57cy~IR&XB7f}!5p_jZswZeTj;dXphIQE zc^_Kye3+&pUCIX4B5rX^ges`cDa0c9Cd;##m7EkAwCV7wGSVkz2`lv~pooH8nthq{ z#%Gv@-p;wgv&^jKaj8vz@OOsk4dKPNMPJhB*@1mcFZD_~2!P{_=arFUQYL_?noWoLiu!8 z-iQ)^Ba!>`%aYS3aV?gk=fndVaa2?KZ2n@lehS}B3lbdce_j)6@_f`hj~luZ;kgxf z6_ypBn^LY{_bobs{Ef7R$jW`AS){VL36`y=>(xZ0qwpRjwG)RiiNCe;XYvoPT>ZL4 z9!q^MgvNX>s=SmxRZ89w)ZRya5l=k^`+K*4X}?i^7h5F%I)wE=!@QZabzZt+&C_a= z`LSBbzOgna|?IJNC-%&hwCq4Wk+!G9Zmb5Klf)gBb+JJtcI*5(xQ8vzKnf;H;C9) z$tiCypgHEEHboDI_1&Cm7HY)FXs7I;m&VO?nj9Kc;J>E*Heu_sa97w#VRNe$xi>vG z5Gff%6wj62+%3d2@gV`Xjkxy#a0hRb_?=ghmYkYjOOn-@oD|#Pw%47rpoiuSo|}=Q zy_P#6#5*xirZ)^plBSQvDaToV#<#lq^WzF@@y~@g)f>M0y*cJj(Wg}VQkFTWT z%phBCvnC{AdK2W(L8pVs#8g~E0sGNM8sS|{x{ z|M;9jlei%31UIZZ^aom4BS|WExsNJ+2T}fsW694t8q}$b*AawBx+vedFAznwK*mVz zk?Z1*vk7gr4)~X}Y@kJ7O%d?3i%!yI)w++Vq+3j6X1?jY{5OJsm|Pk4>Aq+52|Y=B zs1?*Uf+z3DT@k5J!p7f%!OKad521_mBry;vqz-BrK&**LiNo;FjBI)(b`= z2slR~6nF&rKHNzcQ^Do&QJ50huimghdJI-j>x*gcN11M+g+YApx!U+i9d+@3$|dG| zf9HuU)bA5vQ0?7+6<#kPDU}Z;+jbR9t*cWZKij(X{jFqI$k$1U$4O)D;99OyJ{y_J zg<9}-@t%cDXC#)@VRw)ZP<+J&B0&ii)W=vFR>u>TorW zdxCYc*s!rjl{d}(9 znMwQXS=YlM!D6c!z1LYpyaBK#_euxIH`g>DA-BwZ&#`!aT&Ho=mB|!I>|HO-@6XSQy5WkR z&3Z%oM@8Yg1)`$XRhyrj`(E1iQjNEpv0s;eH-kB$}N9+u+uY5X0^!H==(U z@Wi)&uFq^((h1EHU`qq4qK)Ynh#6Ok_Pq-&DParmcnDUPNH)4^G6QMS+q)(P zT;fvkGPSD+21<@H6eQ|qgWKstb6m2(%kk4L6yCU39~zz59<|xTXo#LQch8-2RhR!r z;mW?Rmc~TCEV{tSU|^SP**D6OEgr5^^9+uE`bPNuL9R*br{KcqGg;xX#oKvA?LaB5 zCDEW(j|2G497iePtvkCX@9y?+>%AR$Q$*U9%> zH(ok0hz461%@h(~1&?HZ$jh!D#%T^1nYopq5ZGq)s`E=_YmU#%3b!3D!JPen%NCtG zvc2k}xg%XBR?0W{d#y=L^r+hUZb19hbRYG1=6>N+zB`*>D#sPbyW+daof1`azN_i; zWsv(=O+}TBYKqP$I1&1s_{&SS_d(Ke%0+rbZ`(3?=<<$^-V30JHnW-Vgkf}4OQjD* z(Yb!AgF0>}JXd){I;)0Ta@1{q75oP&F7f4GX{hi|FA2KE4I8au1h6DLwVmR~3{IX` zlWgj;mU=g@#+q-=UACBfUVN(HW@l1j$KPnbaGOqiORhmjLi4#w+Q49oB&ub~OVME5 zSV)@sQ$L}7@p$Nd7rl-u+&e~l{qDzf)jOH#Upj@iN8EIEhpP!ltrev_1Qwt>f^|1 zBzQV;vbLKpC|HX!da{RrZ2eVG4x!_8#iOrn?K_qIy-91c?4}{7T65KRS{Vbd7ORv{ ztEQMgoP0Z?ZgYqFIt&Sa`P?^>3WJ`}K(owDy(F4u!a1>;a#EwZ;*5jpQG7ZIux9-6 z{mZiB+c`dT48GX4QXyiVnVR{iRjl0w;Rz0E9y?7guf($vA0MM^wj4vBGDV>`-$FJf zU(uwW&pf6La}@*nMX55z5qHGedgy#j&ek0+Hs2WHTHhAFToVg_E&pPA#-nIK_U)O| zPyduRI|&j2IB{>~LhpwGU;&N~7PXO{HD(RRC2kI+Y=eDq%wOT^K~KS@)b+F2C+9|s z2Qq2ow=GyY9N#N66G)%sQRRYKctDboWAEy(lGuAnobjG|#)VlP7SBrdhXtYyy|eBncL` z?hFYa1C}TW7nko039`4>5D9Gpmo+O16}MRv2`U1YX)XyBx0n?P=K_~LG6@y8lo$#0 z0+$gs2^E)}8ws-kC$~`?2`K^uqU3_@mvLMP6PM>X2^6=T8wv9Qm%BX)6_=eG2|c&s z9tpStmq%F%6t}J+2_FLjF))`A5D62Pt|19Umyjz7ZMT#q32XubGdVGr{XhvP4>dFj zFHB`_XLM*FH#jgclW|Zef9+jaa~nC5e%G(y+swplvknxFn&_B__SNm%ZQ1RkqaP@$ zBypM|Ws>qSzka_j6IDde|zBNUb-L?40a%A$`t zubGEFYM+x-ZcqXt<^{w`q$`lR%!L;YYy#%_LFK?FqKQ1ze<(=Q%ArQ#OSIxqqo~2T z67(wZ*&~)S25G9iK@7|tlmV@j7W6DVV~(w$era;&AwdOWV4?w?OH(>XT`4gNh(~Y6 z7U%=qEsY0jK~BU1SsGZkMxRwcdmtE~f&`b9w`>pPs@jGF0}(>Z1Gj+o&>W#+%mhG% zeh{Oivvq6He?kTMFKwXcK{fy~VF3+GSFsi1>jJ55iB(aWg902Zed)jzOBk!dk{5_k zh=n$u{4I!)58#R+tsrWs!4Vh=a=v1WRAdU2Ga?O$*^nRn_~Su#XYjiLDe}SdLH7F1 z+X0wrOff{g+2wTl$H71Uxwlx64hbdb@s+#tI>pUxMD7o%}A zFk1ZCXtin84RL3VFGTMd5vagg0BE5TdJ-J=NGA|?z+N@%~3 zJQ0|CkTrub5{oGbW5NlIV}e4XtpU{R*gOY z0T7z4f1259iMA?MKggap%lYME+$=>&NY>-#crv;*|2X)a{n^?MOX#Oi6~j>ff!;<7 zGy%C5Xxd(GFPhP6z8L)V@o051ov)^oqrrOv+7JDI4bE1pi{(GE?7aInoG(uQCXgRr zj+=h9i{q2Q(P;ch{@aF~|?HN23^rW;@7UO^#P*OHubi3N~=3S&kQz3nEw6G5^+R z*$AhpEhOjeY&<`n%uYoSCv^E%K#JXf_-Yf7P;CKH-y|UAPC%&f`hZXy_5q=>x+W zpkIVMa1`#xz@xB(t_-kX4z|X^%2>KWJ~}qAc9zZytc>w=SxTwrVApL~%6C^lqrkd2 z+H1H!^n4stnS;OLppzV(B#bg=2umqte+{gg6A#911tl#S9*Z`v^R)E>h{tsMwt&Mk zA9C&Cdj!DfVdb3A!3#a;UBH6D?*ffV0s?gb>n4PAa20}K13;o85dvJcAa8hEXvLG7 z7MY+Joz3{HaiWiA;jPfek-jQgbI^jcgO5tO8hRz5xuxwb;QE2a4z#jp$)bMcfBDqn z1(={cT9osSd<)Kf`*pV)OP!#YvfR!G?tZV_QuvA2sdFk7td26jatiy+c zzyMU6=UWfkp^GT-*bd`$UaI@ej50p_zy{lQ`;9WeM;dAKV#m8W-S?D9Uu;J_A|H_> z!jO3Z65f~t&m9c8hpOc_^1w!Te@U>qh(uyK=$qq+gpoF^A`f}-UBVSb2Xy-G5V2ON z72h3Ik?yEmYIz})U4Li3p#ao6h}t~i8s20m~TF8M_h2mV>^ro=%^ZfF)j5g%_kh~yZvV9f8cSDFu%>C z3m)rS^=bRzBXDp!V&_Q`Hk~n2N$F! z%Q;KUYc&v3xrmKW5ql+;G>E4t%XPwIhTTG)zB^1%S?Cqt9bJ7wf9J|52Beg?-FN0I zrMyxr<`Eda`Lw(MP-RLQF$oY-@ea!99Z)mEQhErIa%7T-O)YP-V_T7GK4QXr6r9)2buIV1M{dDdDGzuRig_kDL%^B%utYphgxJF68a7MwGtyJt|Eo^reD(D@(AQ z)3&+J=?bn>9;*a=b`&yLzLP}jwtx+#2At=$x{P1gO{n#0e*s~uHD>Ff;IL*sM-??0 zO`xOtqtlE674-lrbJ2>6f-tsLWFvU-3K_UUoHI|Ug7fA>3md)g(&*c4KP`S1mC~mr z4%gE7kZWxo%kwRqq$cHB1-%kWbru_L(M%f^l=~d|8A`+AADK7TvoKL2sMK-#)Sj<1R#pi5KduS6$ z59gQTv$g|?x9zu0XQRbp{$bjjtTeWmoSwz#<%PbF=O1PoJDP*Q5}VBztFs2Y&?MH5 zVlo1$f6!_5qv^VNB$6JG>N+`^uKPh^qpJ^1uPaFsU3|k@kPu>W-ei;6$z(QJeaaf} z6{3@&db7w@i^*tqI&HehWcEIrHK($Ww49EXXIPTDSgz*TK$H|gwWD~@v z^V7+AG@U^j8u0lOsr&q7F&eKf7uo$bGR2zUe`MUO<`)!R2+Hwfd?^fr?44ylWw)~1 z*`4feb}#!m`z5=dJ;)wrkFv+vlk92sEc-Qko*iZ{vX|L!*{kez_9lBvaON{$wIWBy zgjTaeW?QgEv*R{82dW|RHbV3+rBpwe&gTpHATiAxH5m|2+e{uMwG&(;z z9%Zj5*|SM@I61}Vmlw;)6w>kR44KPG_VgTTAv-zAP9_*5h~R}cs7zQVZRB5oc=R3u zGQM1a6Qc#1I9fEwIUkLgSO6l7rsK&J0t|sW$0$9T0+%M6<4?u^r@Prx{AmJ2e*xyp zB{a`?Iy%qZUCvicr!Qxh=SQH~3i9px{oR8^Pj_7*g+1r%}KrC^(*z zhD!*p#uNon=qd#7E}^k83MH#ee`KT8S82sm|Fl%<)KsU5I4Ca9DbQrvXe`8%;;Oi^ zqimcVXTU8x&Caq(_OI;EY?_^CvuvJSWbd*?w#-)9W%fS%kbTTPW&h5g(n+7@TwZ6F zv(wSy@_ahFgrc2HWu!y?mXl@2%tUW|l{hX>R*^%qBW>pYfPl7GejkY6e+T0Cf%ttO zejkY62jcgE_ungRT^u8J z9p}jHQXIkS1V(a~z=&SQIkLNyjPNeSk=`Yu!cZkbfs48$t}$fG6;%`#d9bv13~~6X zfqsJ3J}ggReT~Z=*o$+4UM?3$4OT29HaeO^qNlnI=c|VF_PF`be?YvKtHlU7y~nZ^ zm-{9S#p4?Xv#d4!uZT6^@9oE|O|l=8KR^^ihaVWW_HFjcaa)5K8dLwr_HFjIY9<>p zx$W}x)1!ZZ8B**1xtAaRzPu&#ckk+<$8TxeG4cKR?DL&UH%v|J&>rGfyeowKIdB-- z*Zwcq05z;**#FY2f7;K(#+YFd!_QR0hVr55>ICdVj$w`NTnuH8vg{-*;#UR6S0yUP zFackC4P`+HJX^3R!=LsI(Luc8{)Qp<%;ji@I{>XKhupg3ST2<4!{?p|mt~DaK&67) zBH0x}?qVY3*y=arE+96EJw60qz`^E5LA>xo{Niu;8ITj!e;$@Q!h~)(_b?1Q!h|7U zI8JJC5IMqZ1L+t-0bD{|4Y`@j4Y_Ylz80j&zL5IZ=x#*AI=--^g@myL=)&K_xg&%n zaP|P-vY9ue6Nj_{t1r z7XT$namJn|fBf1{HdDvRoa7=l(4RY z$nrYJ71C2e*OV$LPMSlW3E#R@q<5E>){&Dl!a3zol$?Yv(q36~9E;Lp+S<+$CY)X& zH%Sl3U#LqdDT6}dIG3_Bew7f1ctTh0x|OfgbbRgqf0Fjy|I(}4&%=%>k)`nYN=5{t zt%K3Rjd<8gI^8K@j z97!005Kls}1|q*2YKgejm?KQX1W4`8W1L}Js|GC{^|CKGF4Y_rfvrGiax+gg82Frh_U%FpD&`gvaHgtZ9zD@nVRKf3S%qtAd8xa*}}h6?gEioe?d zAcK!55@PdOMnWD5T~||Oj1g;lG}w-(Ic(U9>#&Ytk0KRl_dV8i+wS^E#(_vL*rkKQi+vJp5VH!kWS>ptQlXY%u1#pG`@*7cwJ% zZq0EAxnWOQBo{3y{*MPZ?2x*V1C_gme+$u@mFV9a)rz^3L-)!QHsM(GKsk_ zWu9PiHgD+iS6rU}NpZG`-8l7w{m{3H7ej*w38?U^n)ak}wG3-%)k_aL+;SPVe>kF{ zrLE89q>dH)!#r*&&N)15F%~cx0NWNamno%XCtsNpWNyIv#?h!aQ6n3+RDv?`D6Vs2 z#%|Qz<&BV1bv)uirR57U1LvniNejcEJ`3)kADVV<0ZP(pi?CJn6me{VL(q$e!ULyQ zEjU%&R4vkE>*x`anw89>SVAdVe+wr_5zF&hLpim^IZ0v$HC$7r5<_iE8knk7PD_ho z(ktO}#k0Z?A0PQ>bYNmUYdLN`4`RSVtDu7jV{XJqtFkVbFXBkys5+lv?vt`u9Q7KXhl;csS1pY9oEp?mJR}O?z$YDKIqKhy0n(4WY&lK2*l)ajLKJP z1SlG`BVjFwYKSWUILiXBf1Qj~zc-8ux}*#^T~4snk##BhC(_jOm_|7WD=UO#BE~94dFcEIv5&$cY^^6@a`7m2f6`V;)pYXbrI5m} zv`s0KVh1@P{iW&1L5)|Y$9v69Wx|P>)g0wkteKS8GPw6a=HasDqx??3D%CWup};|B zK8QY%P#q8kunAGEK3iU@jwDTnjoCdsZ9c z3-?kCZ4A?8&2I_QHg)|mu1+R#D^iU!6A}_zZ>U#vy66>WOG|@$LPQ(nVn=BEtk`~> z;Sy~tT1(|tWSDD}MO(KJi|M^chgj;Gc8s+OiC1E4#rpDWe>dq`lQTyAw>0e2tt!=e zh-_AQoOYrOBNk1(f~pqBtInmbGD@9Xtiv2*yLG5KmkCQLca#VxM4GJFVn<5R<>T3v zB5JpahfX=GH%?@yj4oE1e^UX8UuA*%rUWyEEfvg4W&|3>1#=Zib=wr_V569E;02y|7f+q#42dl6 zrx$kUT6ayFXggXj{iQ)^!qcnZfBaJ^mP`(^tJ5 zV_0-1RDbl9D>CBum={{7B#8MgBEe(($%Pqh_qJWEUL(30OGqG%MQ$YVgnN?P<7v5W zBeKHwOGgP;-(TlR%8=x%B9Jc1Nzd^ocrKHcvUn|K&jgv-%5Q}rApJR6Vc;qHzy)pf ze{`*^R5)4VmI6WCEdJDJW1PdH5K7batK3BdxDD z(j7ZbM(6l)>($Hqk6!)#_TyWx3*3@)e{VWE#kELU;+9-w_5&^j{Gp^M6;3X1!98X_ zExG-Y*W2K7vC+jZva<|CaoA|o=(isVe!RaLO()}@xYLc7FdeR%^WV@AzPyf=z+1F| z&0yT}^HaFah#Ml#^=4Dw4C1(Tav$Qlo$emC&^<{IpCmip+TCs+`dItxVEy8!!v_za3D)%< zq(jSXxv$|Sxb62My}gTYy@N2_KiCpP_bBUTS?*reZGkO`Wbd+Wb-ri?MK=j@m-9;Y zKkH_}f6JE3J{{|8(ed&9w@;sfe~x$e4jb%+`=PAm&PU7nK)I%5m8XQv6Eb4Wa}OD- zl!~TEso0%4zBM^L?N0t4PjFIp^l80!T)wXNk4rKhA4VMefw6N@Hg(5}UB(I<@Ordz z#dT=)`nNlepF9Gs?!S1khgi6{;fIn}ML6CXec;+ zuLKppTMSxM3V&m$6rYiTaolD-Yel_}T&EUB^WMIC`S3o_d0ymKV_sTeC;qs8E8?Hm-@?2+)!<*_DIE0<6!p&#rKgBg z=|}a}`4;hl204)!>2GdUxj6Dx(fpUw=jw$yB97|5WB`^ZjI4f3)s+BHO!Ux-+ENFHi12 zd-DtNy_%^>%kP$G*D34#Mq~m10%4y>Uu$0R@|Sy$Uf(9NjfPcrr6l#*?0iC0h8ab@01y)43tP4r70M@c6;Qw_xnyo4pf8uvX@a zMIkHEdxi4qe;x!UntNwKFVfq~;{8+UHAlS*jN<)E?WIZ^S2@LLWS>6eb?9^V^|Sjg z9teHxPV(S3_aAEc?--m8)x`}^$GW}Lk*m@A=&Vibvq!%^xb;|Iy_#+6RoLx45jiYw z^Wt&*bdn&)RhC2_^%h-FWj}J|K8+NvWozEP{prEse@lUV-wee95-yW*tqg0YHzBwC z%}sktU4>m&R1-|MO$aUY5+G7TZ_+~`^hgJ(D$=A11ZhE^U=M=4sAa`^?$x=AL;b9*w^13%&NcYD}Z`t_mS6 zfi;WyEss&zaRH%Q)Tel@V+h3GEf92C~OY)4&PZ=6D>T4%uXSdZOhp$ugzdXzPwH`|L48bnv`1$)mlrMy-Yp**k=>3 zJcbyfcw3&55Zbg_9Isd%5gc*g0JMfO!>~~W7oZ$j-n2#X$_uIpM7YUJbmd%h{}D{= zHc8raY(H^Xt-TilRgGKe-+9Gv9GzO8 zBoG{M>r$An+&X@cZXobm9Qr3k3~l^H=EFAbgv4|$Wx&BV3bgYoG4vr>AqyC;ag5Ng zFXmCmeE;Xdub~4V=)AZCbw54^Zq?iFoPM(%kT%rrF9)f6M&9wCIqL*k2fO+IONt9O zvxtCsiKMsO|B>F!MACbP1Spfm=nZ6U4wnRXgMc#48nWPVq8}p5QxZJbyo3N#5Yeni zD3K{Ay0kVoDS;`8^NRo3r+G{jED8e2%OL+XA5nESs&W@qq=tP2e==M8y9Dk&ZJ$A11WxFT_Q z+r`_>9}Sb0R(ORm6~z`<+4#-kn27Xymvh*#fDm9AF|nES^OHN%tglzAey9+AQglD3 zVTpmc{)GMFg9TECRHxSzc`HtA=f9iT?6lRS z6f=|q02Hc2L($I)i@`ufIrb~yOtk)Y8ZAz=bUb$#SkWLU6;lTVet^3$=u!zYL^~7} z<>XRI8@pqv9(G$)Ewz;DSBr}EYLg6;#;Ipw6oFzbq}iOAWK$Rtb5UP#=|?Lzt*9wd zNP^Z3*?oq3QtG;d+q7EL-lYIK?OX;?+M8wlv<B-P7%79$BpAeH4i_rJZ_Ad(i%N zYvPVjL@PK zk!7uaA5$yTSMd+)906>b#X7v_I+NW=SZ6uqG0HV;DL}OK{5HqwetpoaN$>s4$SPEA ziGP0mS+iyWW^89DAjnhvc(gFPovsoM+|Rs78$0*v82pC*Caw4Jf-kak9p_e}F!-ws zOGYuDs>mB=MuYMX^S?La^XV^*0#1jnd`>P|tkKRY4;20|$k+XPf!D`*YqpZsparG+ z2HzO26D9WpIiaFeK7q;!`7%ycK8K8>w^qJ+`49I^Q1=(%mc@7TcF|pvUG>XmgjfL= zzD&^c_%^-1m>b6~ZnoF`=N&tssxw{+i^NE1Pu_nnYtHl-W;YHJ9Gkz@_Hu6=)&a-R z3-KU1#$gxmqge&KmH@4j-m|51i8~6`i~|YK$1l&2`tuHVSggxIy;bqnI5mEn3Q(`; zJbgze7&qIZP{VC(h*@_u;VK1nUjVx2j9l3_b<`S$=o^OS##b3}r(KA{X;&;`)U<{) zGezh}PrnyJ#$36Is&KmRsZ4<|!AT>1$Uzq(xpCfFhq4^hozAnR6!cloNAYNBrerVf z88`8881zQlKxrD?+!QwS)z{IQYl5fCkGfL=!RchXeE7>5rCl{gzyaZA{qIA=Hi^d7 zvf56zRZaOJo}I5aSFj<<42#y?*R)WDh+7ux?m~;uV@{u8trRGLmMxowUH8zMOo)S9-vM$E4CvjR9(2O2E}Ur) zofp9nl=CJ_k3saSi9lkkIQJL9FTEQW_AS$-$xRrJQF3v~6r$K=M2Xvkvc@w$Vv0@T z((EWib~Pe(~@P4 za#C*yx)p1a5(6j|L_|y&3xrKn@z&lo3F8J^=Zr1OPNuc-oRO`=RuKDrezA zc)z8DGvqAZRrz`VBRSh+B5w0{#4j)gY$8KPgZiR)itKw_7cWQo~ zCqH-%2kf~GEzEa@K?cnJQwK;w%rS0zfO)ZS!~-g+JMJMaF!_Ibz=2TMp-mX4G3;thA3udFDrHRr)Nz1^I(sC$8gp8sbN*gYR)Yg)glhf8w zP|#6>{r@g>#6bVeV`Svu|64~Wm^t8+aS!@=|K20k*kvf^90&10=eLn&HusFf@rvbCk?F$f`d}R0C zbNj`^{N2JIVXyu0Qqh=WMP)xuj zh){g#7Hvw_3O3c-Bbg!3)t08OW%p7YIxQmcVxCdT%_yR(W%ZI6dei{$5}#qurS9fy zDnQq`OOhLo&Wt@mOOg>rmBuW_Z0RhD0*Y%(MMr- z_(K28Vs`_t5PoYUdkyxBsAJw3?u~n^L7>D89Xv3-24pkV#Of8m@jazHctEF)ds`z< z<4!KS2LFrzd(L2n1~ZymqKKs_vY?2|Kf-vFEHKi06cqT(Zj|CAqN9js8w7GtR|L>S zX*w{_MD{umFMhBn0;T7@I7mf9Z%623*astTP#0UPsgkEuU>|&bwZv&N#5}}wW6TP} zoli1xaSQr0E%@OE)FFKwEKOG4tzW1N8jkW?QjvFiypC4X?L{&SM`Dm^Q6DcV4hF3& zZ+w+{{B{ZMU@9e?>v#i*u2e~j6=U)(vF(e|Tvig`q^&PeAn=*w)Yd<2F1w+GDIje8 zo!oCEjI?}zhpi>xpaX2y8~F0F9}m{yN+|)u^}o=VWv(ElKz}X}&AKuGacE8%nD8=~ zzsoVjJ+v)m%$@ed56)g>>>lw-Xc~Q;}*76x(E%ML@`)b@+te z$0%GRR-&lz<&dygR9S={Bx7DsqvJV>jBC5SiSllgn5j`8s)Wpe*NDl&n%v)6l7gd) zswgi&(2~L;lD&cbEzqM|tp z1&KTEIrYTMg6o;)3-tc38T4FofEV%Q1EO}NPt5+|_efjTW_fX@^N9PM#wXS^Yn?Ap zY?VXllQcg_@VWLJBzRprANtPflWW=^o3|A0Q+$C*?F2sg#B^ZlzBbMLqxA}QQw{Y7Te;~(}Cqum2Z4&3bEhTn@t1=tdiIvr{ zf5ZE%#78GK(-bc#2vZ6Ib60AX!P*5%Sn}HT0BQN%q$;H6Dt6Hv?sS^F$inq5RQGAa zMcdS)6ss#jokPs{^=e^LPye^X6ssRG2<^UF%q~4qzGo0$w-L?45l&~}S!aOe(A#)| zDE@UAU1-95hS%HZ%xbts^%TEmid`2}X_f!6b_iy3WiD*Z!R(JkuaC2EVns^)!{1Y81n1@O~nPL^E=pu?7<@`ODGBVWPe9Yg@1A@@7`|m-|Bp}hrLX%AJyE>MF6`Zx*Eve(NsGP(uMnS)|0-R z4)fh2E^*H`fowqHvGGmb+S=BPv6PxFjA>uOnRCnfO=zFSBm<^jxs>F$3t)n+>SHf- z$GK@2o`8TH;)04OkfbQh&!?@316DkIDp0OBFuR@KFAgv`$W6QGmZFxEF zS>iY)AFyot_w`?=e)Sx6Of0k|x@e-GV@R^pqLBZU%;A06PJJ@W@^?_$I`hHLNd^gp z^5>Az3GUGp#3of^m=BlQ9N-W^O6_We@>ZK2$MGO<>FMUshbw-UoAqoI)e)FMI#OF& zYbw{8d^|JLN3-@bAUJ<%Fj_zIrVW+=x{x>?5#S3z@I4idswpGXlmbaQ&TbmG-hsN%#%tS^0#*?(A9=amlv}Mlw4pzFK5!WKR z@oB(K#{A{3u0!;uCZN?BH)-j>5(@&?I4(?(yfKl#p%M-=38h zq)8?C*GwB{cLWPE3P7v6j?Vn~!~Qs#qUm>=nw!3-ow>2{-RHwsm|t;dv0R?li$O#& zSmRK`R*<8SL~ki=cI6H0xJGRyIq|d5*lVsFL_A(WrtWY)IUEtSbN_zweeZF zSZqbGWM8GA^sJNIzO?>@@l3aAUH0J<5C|9F-lnnyb+fZuKCk%OTp<%BRyPk*dGqxe z2_1I%yM4##tJXE49W1O5o8VHS2544OG%K>1@IIv;twHNl(4m&3Fl~L2Nn(LnBzEn0 z&|hOAnd3{aIDeh9SZ5hFTQz<p@6PC{kE zcSjc%+*o{^PC!=7?SDf`sZo4~u&X?aS|R3`LV|1?LV4CL`MHMUX?2}n2O#%g=I&no z9{Tew-a9CWZdAgCt6fZEU1iNjUh$>g04&f2bo6XnHKl76JuI?#66bFg=DVNS^y8UW zZ{yagjQTLm`sX-@GyJ4X+&Tz;z4B8ktD>`X4q^k|*R5`uXlm+5aB=eVIe0kggw)vC zCPp{tiV8*UOxgF56VZWd0FxTQv;GpRtZfWcqh7-LmH)7o&o2M27<`x!Gp#RUSsDPW zt`nwzCW?NWuS)cD%jZKZ>W(?($eWQh8|K@!XRf5|K;_DM8dpfoa4WX2*hoiiVn)mz z+jz&dnF>NNKnRlSXH(|oXb)(O`qi}w%oEmeFZI^@ls?7Qy?DyY0CoP&O&P1RYag9r za8EdyvphmugZA-eu{7oITnk`1tZIMg)c~KzY47xMml|B!cC1o-_xx3ka&~+*I>VPX z`e>ngmgk2cfjDH(TShyKyc^vTo5J^70WZ53ztt}~Kerv_3 z?$Z6=9Jkn|=aU_1fQX3GeWue=GW4XsHY*E__iN)#FV1f5a!3uD+Nf{=b^RePHd)=SET@jIy5OOx70NT$u*l97|$n)GK`4=udpbDuAPCj+kYWPQ;8{&jo%QV&KMvCzy@yz`RnYBKI@wyLB;@i9TGi} zPmi@B3`pyb7o)C@#AG)3h_~PT$q)@Dl=)4$Eba4*Yg1|wg2t0=zeI!nrwToQo+(X4 zP!}7RHJ!2*pej}2mTT|Y%Ah~2y7w%bBKu-(D1?zFOygLWULA{^Ja#f;zjjm78iI-Mey^ z`-x#X^0uu&s}B1YL8Ckn=3BN!mSsGPfhj(|%$!@WfImWi(%4ct!lltAf#y1l6}_zM zs-`~4<+J_>XE4)+``V{tq$T<%^=+w%tkLNZDm1XxJP4e?>iv8TA+i2(V#vF}z|f`t z6}JBvP)%=x%W(_x?a4iUEAGc|Q7sJ3XwSJa_9TXo+k=^y?Fx&o&D*K4vPiO&FiZ}ggR|4gKg94|J>2d$8+O!rOa>FI*P(dRF z=nZx2RE3vU2kGxfTh#5ez1-};#x}k$t2l>L1WZ+GtKifaGOIDg>BRWHszD5Yo_6Bc z=!?+uyLWJfbi%mrY85pA8Gl6kpIk#@yQ!I>RF9?~Y9=g1#B1Uoh!G(W=&5yRJ)Opm1R2|hlr*tX6#L=={}-ongM7_3|nU)_Z&&Wk7Sq}X6~{00S5D$|8K$+|PxyP3Qjxw~N)l;~w0o!t=#P>e z3_=mrTuu_D0+pjyO+aHr>M9rzs~v1GLQdBvw32gmsn!}fQj5Yf!qK_f6Krz?vJ~aG zo2NH8C+IZeuhaPq=uFY(D$t2gDl5@u-*C`Vi&LxOsBc*1s8J1ZV>0%Tn*O=dZQ!M< zkSR_K5OyQ&vEIjWLccK0Un%?o)Yde@kdm!cLCMO+HB)?g7HfFz5teJ@FIc_co~PWD zkIq$V70fdhX`CAUO=)oyu0tszDACYuVhczjxfU6e88Ep?lL6VDx8Ntda0E)TIAYSs&aNADoO6NiMcqJN!|9-68JIEfGk5S1UR zTWbf2mj*T~vcqpCmWwG?r#g`?C&v&cxs#L`ttJbe>?Vk;&&oIxn09;+`w1t8gAumtkVZ5#HP`WxGi&i`6R(ms2I_gf%yvLuiyU z1k2OQ8?jcfH&auh@g>0|a(LA#_a4HmP#BQYXMZs=gEw{~$#x5*djFdzsn3ASHgf#Jc0dT$uuV@x2gz1%}m8|W2p?SYAmgB+%m1U zcI8)U-%6{%$+K}}Crg?D!lBisOHZaum@+IlZOG-;m01W~VN<@I$KNqBZR{kqEw-pI zy_6~BTDZaeo5~6}4)_a!>az8XL5qV>Te6s^wIEzzAU=0J^h{@S5S&XczEfjSc`4~C zRPNreo2AGG401Y;D^`jk;(;%fO$$dpxsDx8TRViGyJ7u$EMXw0an#9mAr%tRc0UJ% zP)?DrE#T^4apy@iw{Q%YZf~%x42>AtPfffILSr)ZSZxJFsy43V6v@i+uSN+&W}1c+ z8Hl}iM4G`GtS8Zc)N~h~Tsyz|ybjsEPB9~E$YqUw$grg(y=Yijw@xrncJR(WQ& z#_l3RMZ^P;^B(BcCDj#eY-qtv2x9Zp_*Si_chv2ZTvKvN%1ptw zSC2YT4~)&|YjBZB!Y2BlZB^Yw(*B(*dvGR>&Z>7mFtrBuViF^>fYc(iUy{9tdr#K? z!)e9WXl?NG8%7kcpkbRk^|9^DjGLzC5THvg*mUlcUS zjKsmnO+>CIZRAH5KP-Vq;oC;IaFk-3jrVNv@5k8X4)cH=m64l)h12!ewwLns>|X(H z`;!5Pi5pxY*zQ`M4b_XYE*FEx@5T9(IJ&8tW9xqI8+LD4F5${xe3`Tnq!MRIQRRrK z9wyfgj2RJwf7^H{E3g~lu{nOb^+-*@M@?$jhWy2*IbksvONXLq?vyT5HW=y?2)?|l z?(Rditlk=jzN09kFagy*dUV+Q^<{Q^w;K*v5fY-?U=cq(gpxa5ncd$6mb93a4qNNW zwu$S`L}rTQ)dd+1xtT5ak5&QRUQJv#j7(cynQw&p|!u|{l< z4`c)sNw7#cx$06hO!H++u^JB)3r6TeOz~wyXMZO|7(eQic1jS;E>Ehmy-*}%pTEQF zc}GXFo^RIO97T7!m&LIr|ILn50c6BIwLj{(-lTar(_HiFXyE60YP0MD4)W9UA4I)* z#Zk=^`jG?;A}{NaR@bexOIp@}4MSSf-QaTSX?zZd+du^q;B^YM&va(~Q<-DSnA$Dd zI+TUI1cjVPC~~(urBARMfw6~5 z>IDxI%AnPFUo4-4#A-b%GrnZvxR2?TE^uKlGI($2KQCOt3H1ayYD6{&!eANCCsQ2sMS@qqNgD0{UYaw81 z{)WA8g?K5f`!_OYe+Ypo(3Yf;KWB%t$3Ob29Q_uU!{`Zq3)vg{{)d}sJjP=$wO3ZG z>$Wr1uAP^h)Nl>Ci!W!$P;L$3yfiA(VxaW^Gsh{e)O`#9D~peT#rdBs@5^0NLY=Gx zTv4vPDPi)y>v?%O?NV;E%R>tGQpSLWCl*H5^|4*Yj10zrzV^_+8)!q{{vC=} zdbr5mse=X}b+>Y1rns58otm$@ZuP3iV0-00%t*9jS@mVY;7FWy@f2YY??Pq+EhuC| zp`)tB^3GK_X)UaWh&nCHF)vX@W}JB?=9z%c}*Q( zp<_KSwwq^onr;SpPoV0bvHZ7(JB%?s;coAotI0$_Md zfS55}co$s#WmR?CmXyUtjX+|owRI`LdGb=rOu|K50e9u%V<3(ID3_sjoEHgiFrK?l z9nQ{J@tJJAXg9%dTX2A(goFA4zM45Bzqpwg3HSfmeaa3(`(tFP^+WHTj_c}AN}qRfikT~^4kIgVFU4p3;&}xBFOF1Kj4coiG~^*q(ig`59X zrnXQB{7%li@Q?&*CzeQo{-AQu-Ejo-n&C)!90o3Pp@ueZ^Z_+C@E-t4))KOm#1k2$ zYf-(UK84PsM*=;jdY-i{ZR7UXraHQ}E&0I#L5hCkV|yQ`IL=HJCC^BYFs3#+UiLIF zCSk-YR(8=G>tPWLSG|@eXV<4bWth(S?n;0}3+q>7(ZP6xrwnMpg)Wb4`a&3f36lPJ znNkvKkm%^)IwrZC(>uUD>QM{IH>6)M?^x#U*Ye?>yq&+gvXrv zf(O^JpkdY6elwpza6wIrAQ7nlr6X(ai@@b=!r8Rk`H4+K6?L$C0NqLj8%jtB)GsMo zV~ObyWAUj_lmwAOI76F;Un6CU3M5q$%x}73z+)wx&)0>Cr2+gw0i7EsadH-hd(T1S z2RZn<7r*2S1P5+6Soli)(@1jr@o_qI#vMQIjWY_Hx#PR}s^nPUC-XDqhM7AtnbA(`tS$@ zP0P?9dSKKe$`D|mb)XI|3)b7k7zP{Rq(Xd+Eyg1Dp4W%a1zMOfY^X17+@zck!fVVV zZ5k7(q%0{JnZR0B?-35b5ayH|?{^uPJybu$l-!GM`B?46l@Q{Bu^giI>&xS!ez1Lw zszJmq)eep;I49J26e(mxVgv6~DNakpp*xTVPCO@- zHh!CDjY^Tu*@1q}Jp6G^Mu_N&tR^jF;BYO3CY}}3k%|`^=G&I!;)15rjUkmf^oTqc zO5jA6xC4}l4nmqlXp+e#@x2}E<|HKT4e93+hNXlB@8f@q1YQu`?(PpUucAOhC=fED z8)*w)Q-atZd#!u*+xKZIa9rpHM)(7$)1hjR>68e$zvCnc#L`dDzy*a*tZz?JuLe<> zH9Qvm$@-L(%MLM}@JJ%pN1bhQf=3I=;EdoTWdQlS&8R2$mY{wNi`>^#R1Al6-H-Qr zQ225o2FzdjJ2-#BH!H8GRPhiZfEHox7C2+)M1Xg{`)`VeJ0-HKcLGT+mmfALAiB*Dl7oJD}(ly3^4;#A~U zegL7kYw2|Dl!XG&7o(D@dSEot$S?1lAx_uik+moF@?g{H>)x)=Y*QvIe8j`;+;v>Z z352!EhKynCgQ&iROm1vz@#%P6?s18Vdc)N@aic@P773cp<1H@Gk0obl8WU?sCxKO8 zSE_KzFn_luRd+fWLh#wB!#DE!yq?PL3;^HfBAW+2Xc4LyoCb5k16W;6Pe$|bAwrxi zekw|HhK1!Dhegv|@jS0*Pi|0yXo3C^Liy7nhzrFc??lpgBOeb}q{g4p<)l-{Nf&36 zW4EZqv@}0{RQ7s*R+&k12W3aHWaJS^9{JtT2gyFAP`=0D2HoJisS$7QRWE&#nE?#w znL<_>vUi!1v|_B=^2)GYqA-$UvBRh8lMZlbggI21t>2TY8;UX3*c=@J%BI_J8q2Rxc3n9f6nm{RMVL`@Kx%jpn(JU-3*e` zEc){h!R+|y4|+K>9^8;y<&0zS9lGyKwhW@X(JFO~zsTi`A)Pv`Y&P)t&1_ET;yGUjAe>L@^&Z)jXq4KuXS8pGSxBuC0x zhop%7TvZGji(5>6(T21X?=)w9Dz?TwCo_1cZM-Zl{QRZS`3Iqp!`Q|m)p1U;i%#r4d7$B(O3TL> zK}mXg_tSZ0sZF}8xu-c*or2`dzvlNK7I%NK$n5CUE{V`mH|z2)E`8M*14zb)EVk)Kg=9KIBSgQbB1+nZQZkfP@m(>&p5Vp{DNqAl2e$>2;>9GTX zz(e-FRJI(d)Cqd;DVu$o>yblNr?TwOwSQGM79wE^i_py=r8Ifb$D7ElP?N9-8GXx& zu3-R_$gk7xeI-q#dXFIZw)#bWT**Sr^Q)*oVIEq%9qJL0EJaM|kam^ciIjlZ?}XKF zi|kRomM#FbsK@8mtprm*Z9sq9{mo32fx*1&gD`z!b+~N-1_s@EyJqzpqy63C2t>|4 z-J)3U`=rPl?6zi3QHzdl-+PztH+L9EL}c-PMY1#DlPB3J8sMBpi90Uc_wKUqR94_b zse(*7L)%mg`NZmxkC!GzJQF}Usm#~-UIgV+I+jSMd_m+WsZPCJTgU$Td~u6~Iuh{` z)JJ?fEFLnEccg++Q^?h(ARs>cA??S%tOuYyvn{DU;WR(&<33Q`BZyrvQfamOIFFr| zZ0~}Fk}z9OeBe4X#}?s?O3oU!GFTePAEo5~M(0Cohqmm(z3_5IodI}$Zm#+oEUZMv zYww?+QqmM_FMX{8f8LT!Y%fD! zd39A^hSg{YikjZY3<3C8<_&V1GiY7aODvfSZlx12qVnA77*ZY1Nb_s5i4TNWWbm62 z%p_={qmwC=mJmHlhse+l1Ekne)@_LA3yj4(n*wy+PM6)2W+{rKMw1}iwoaS3z*bIM zjX8-zeF}v2bb>9CsVt{pI#DNmmnPFAl4Y-X5l{A_L*6ph7XY-z5_PEF{=_U2o1H&P zs|;Pv2GGzlcT)s;soFzj$`Gw)^b|?Cu{GHfZYyq&ziW0k*e6nz`fxlPF)?!SQ6ews zLlfLM^s1D}kG*B|rKeONexJ@Mh29>{9ubz^GNX4WI9;VrPw{l+1Gt#HTSW`O3I3u3 zcJQqpe@HBKaR3)}A7`o^MlH{j-RCMfGUfy2!&gMIi($*{u`D$|M3O6dO|2fCK^XgT z9_urR9nGXbo1A+WE&{7v)$%8lQYdB{8Q=Tn2mDmd(FOwbu8p#c zSGj)hGHHHgz07Zf!TOVNZ^2duE^-`OnQ(8VBnF9R*k`ag%C9Un^qOzle;V{ftNnuH zHs=?Y#Q-u-^LFIHU#QIDghnppI&#Mzcl>^E^rINcBlVF(B;Ua%bOute%p)P}+ovn4 z)dya!h-C}d@E=i!HXmEexqJDT;ozrKBDmm?!e>gxi$ zBI6NDO0_DvQn)88cKFWt?B6}xK8RbL{)F#%x_~S)=|`X0myDxcT0-P1&$=AnHbl`* zZ)%;iGv1MH2hA?3kv{huXum~q_g`5S@Z0hMcOuz1mmfU?^?kvUcqQ9Hyg>|bCoF!@ zt3H?6_&D&{XECFFF!Lf#>a%3BIyJ$NO0KxGTiG$u^XpYDk-uDB(w-8v%7PsyQpG7b z1^|qQJgIL1o{>9oXps-B2tMRmIVa7?kO^u>>fa<23s}08-v{j6TLg0^C`8#>L7HAK zt$Kh{_J0GQX-$}UJloHD7DKjh8J)Y8@#V@DCN!1CTmKT(PldSqMdrmltKQoGW4`KrT5#w; z_OnG;0S6x*@MD3tfHjop3qO!U{}4S6RJ3svI!@q?R6V(#bzel|4BWFmUi)gZ12jD? zafi*+{4q%ajgh!3C78l+5}Bi!B|~w-fdRwDc(VWVVjB@Q6MAIMi}1PNOeS$xeL%IF z=N^2t0Oh`2Ayl9^PWV}~-Q~%kBVE1Vh_>P?+sto1b?oT0ye|w<=%njE>H``=p_GSD z8ixeLy>gEgs|N|!U0VRO-dD6<51420XBlBrPJ+q_XW*d@1fB!G(})HLN7^-&rILFjLzRihCFJrp8x}j=z;jfdo)=Pjs{wc-49n}>`+mTa z6XV2Ocx-k1ooLIn6z=jvX9FL?duF3eJdSTnkdC7=m9c0qPg_$2oL= z4h@^68d$asLSLS?js6vIOI0A8sx#Pt9I2SXiehh4bD}G%`qL&gYo5d1mQVJTlqU}Q z?u)9-6tm5mkz{dw5QwY}i&O7Q;ArfxIMhVsu(z)qd-$>5 zKkyxIFYabZMFHrK_u$AHv*M?eW^fy z(xZ99Lrn}rJ3Alh!R8;pep63k%{6bJ8 zYcu)GQ)FvR+i!E4?*p&H#UOT!9Jf>dB+>j?znjA$)LgB|lo^!Bv$%t>A=HoB&~OAS zL$)Z871RGg1<=^=+BX2yexX}vi(u)3AWU<)$=5o;gb7{jlR!vLqz{7d%k3#_y}DYP zzCR&*_Na}@$w5*vIe{c`>K4}02(SxsTCdq~QJJNJMbj+t{_^yiNV+W&UXB4^?37%*tbta2%RAKJ^=a2DWjqCF&^AaBdde1NT?Nhoo_X(; z7#!qL0Hf5zhCNPtsfv1rFX6hv>cF>XXJuK-TSij=l)W10s_qRBF_)L}lWeZ2^Lk~- ztYE!(RdqbEQ2u1mdI43VfnPk348L4@!M_3I3LJecCA2fsng4CRM+hmE# z=#|EoyNU)BHWxU*xjhJ?4VF2&{{9M9wa9GrK3poP>)Si_e(0-5yW6UNaWAYkCC`OD zLp4x(xii3I+xQi0XL3A^$l68U>4(T4fDD8P4>U#6`?1$q07n(})xd?jShy8G%HSXy z3_mUM#~3V$n0!Z@#4}-B#TUz-2x&l8PPKTJwHKrqjaC@KEg-*-BtF8e5W*s_+hNQu znNIM9bcK;IDlR;i&z?s*J^u@qW-OoIgNur3aOnygH5Sse=KYb_3fj#`3R~k5FfI+W zhcvKlI5H#o*ZJ8vCyPudGE4?ZGH4wW;PY^X+gX!etf2T z`P0UjG07Ng3)Yv;T!E+B2jNyi4HUsv@IB!HyuoEr{cJ@;qlV8a^5RRst*;(~EX$Z102-3-7N z&JB3PpX&#&r;*p1UUzguuSGqJcn-?U^DC}d6I5P}dnD^mh37GVCI?&TyYWG9-$8S?I9%>31{vfr>|BMw zbEgqN@JugPD=jp=U<~mGEYDbPJqwHUA8z~nQe9evgXDc@e}uNu4;M2 z6^Ep|c(5)hU%woQWMT%WyzBTOYR*8V+n)!%@eguEmukE+QE^%p)<}{ zX{L(Av<%=EtlGpksRhYo6zB3HM0+K}ph!sP#>8pzf?o`P6L6cja$Dk^@d%{i>x=h? zS*l~SJ8?FGnn^v3G>Xos)*&YXyC`{USOxkT!yjwea7a7yHu4Zzk=PR_f2s2hTW=-* zmi;YjtuB{_pe)++=pucd=tiGRI`RWadWiXlE>aDT7Ft9^%6vuPtO7DoB8AhTusoht z4cU#<+Nm9YcWv!R5!HHp=#yA4o3#jGaRMvO?-7Kkw!bz^(o)3lVQsX*$gBo7zvM3U z4l4o{EW(Y_YQ-l8lPG$Zjg+2)4`2N{bkX!9WjQ#=TzU2yIykNH!shN8>Jm?{?fM!d zVthUv&J+j(@Kxv~bj7fTnB|^6T8Wz{EBmNi!dV~ScW#Ep?e4=AWK%$ajj=+Gr%%*l z-$oILLLiHuG1Z~+(o;_@uf#(Wm=7?d^?@Iuqu=3E~=r6H!=@Ui(%193drv1ow-tr{_jpCQ<67zVLP2d#F@Z&3Xd0q0*<+CRd& z1JwsWpL4^zBKnp(i3c9+Zx}<5&rCY@kIH&Q8Xd)!WvYpxFZYk>(z~@4CGgeFZn(i& ze;s#b;(-V+zxikM(j)Or88xeY_Un|7ajI(8*A@GLz^?wk#{-zstJ6)Z!8nKFhtr=I z1HX=5zkg5ZsubAy1JYFZ_3m`8p~*NubQx=PS!q z2^)3=iba>TJdU_UJX^;DyZhy7_<2I5xf+KA`8I}P<9~4-8y$Q$8v+w)ZJBV@|J~8* z|4PERJ+hIoJ_lc-ugh#`8ZZgd3(W`2uldCq-bh?k(@uLCc;J&_Z2quTWg2i)H%b)n#zq>o_S6dIkRJ8U>mp-tryM5 zT~@b;o8}>9Qw=<-Y%~kL+S520kUMJL04k6b?EMD`q|*DC<^&0h3c|*gR$vRPlmr_u z0K(3i=I9U1hRedv@gELUCSnw4CSqa#A2j$sHY*Ve2TNK(3ou$!UHpnW2p3C|5C9E? zo#THYVUM;}-F7Rw-&#)?6FJmpl}mIVq|lE2thkenAgA0*H>Q_2Hk863~x`vWL;lldU$1iSetqn!+tnp zU0=>`h>RUyXFmqH;5=4saiCK{uN1FO{ge4d0J_b_^Qk|AB6ZcfDrAE##d;MFbq%!*ds z1w5F(%<3R+QSTmKRFU0j9vLqTKku%Xf3cAN^3hOMy86E8V(h~&>e|+$Bi-Z>Ew?m8 z03Ispm?$TW>!?3X*rONcQFAL(XE_i+N=C}7BS8c5id*;vgj}x{%fo6A^PE`p@P-zb zt8}P8;FL1G%$PAm*bm>eGCkiwt%#(lOLXA3%#SJJl1O2qsZda16a?z7F05pWpF z{YPb0iRAeIOj^Z!s^WbVl@opidbI3)0xtb?Q1l1qy*)4~^%rz430eR!hd@2>Dj=IC zFR=LF`|*l6qmheD_p?N$(^a3JsxfMPw`MrB&8w+xarVM>7oNl&d`GRz3KHPutBVUb zD<`9_5tV$g3a;J>FWavbCR`JTJMYb2sCMj3+*{M$N~!F=$V~!BW={dm9{KBk0IoLy zch7o!k1O&)ESyvi>3$59pMd$Rgbmr3Dz$TmB+z61E^TN%g%;k=ffB+9H z&={AbvOiebr18uSs#x6xQ6|TzZyvV85iXrSb`mVWpMeN1`~X zn9|OA;dlMAr$Zt;ToDtB9QxI)`s0%ORYd$3#b<(D4vvF?bIkQS0OODjOAy(B&V`lT z6_KC$b7_eZ8ty&6-7O!0n<)OkbCmPc2N0$n=rPQ!)i&oWtXV-F~PZ( zk@}FjT38u0xOj6E6%xY!FAj;*q0nyKfJ%IL6;$t>3a}dqaW0AoSz=;^nxaasUqeI=BmpZ004`EW&V=g{I0Iy&()Pi9n7p`=G*xpzl8Nrk=4ntg1K9j>nRYrj ze7ejcbg{#=$$%>ny*Ur=jCheQM8(sqrqrG--ztv{79;tUvB5C9s1`&wC2|BU?k7i_ zI}tztCdTL4CZ8RS`5^2v+#zbT$?IAQ+pDe&_1N^}l!z(}KnAp{dK9V~y#GX+=*@J< z%C{5OOv^D3zp<)JUF`Ys_;oje1u#H8La{9eB)}Yxnk8FTX(BJ}KqV=kR4~ZQb2BcW z1>#~v56Nwwb-^-01hyww|2gVG83(3Wx%jK>ZhOLd06sel-Cv%T&?edcMl)V9J60@B zEJmt+p3@Wz@EmVJn9gRgDsT<=JMhPo6pF_mCsDZlqj%Xk9#(MhJnsZb+CT`=fh%Z! zGfIQe9=V-vz=;X1+@wi1{OY)3U{{T9?q@mUK#$Q0Y|33)zwrZuiLX<%cu-Wtdi}6S zaH_rr8!57_pIPlRy$QV=5lIF6FX&QYgAld{r!5v9Um_Oy8xyca|^ilE2DW4+4TyXN&a>Z0Rlhg9XS@)@X z4?DP$oVd6RZpIO|iaKC6veoph^3{Hnf{{jD&hR3}lcuGqhEWJiRO1who2K>+sNtkX zb7wb}L9r<1eTtsm+7p{IDS@6J19n0IfQ70NDmup}7H5Hv?yhiwPS~%CE+~iWr#?4X z%@V0z?dWL`q<`M8KYm*Qy{B~6~WFG;;nQ3>7anlWPbOia+_++(ETK? z{e)~G4%kbf&Jxk;{bJ-Uzic1#OU7oxXhd_#z1Su4> zR7@BUP=CL%L0Bk=<3LJf6k??0K}6m>!tD*T-GFNK%CXd8iDWP@M~DWrFkd0{STd`J zjmmAxm)S~3Y5Bis3&QI1U={0QC$B0!zl16URO|jRu!E_*aYCO0t3TGm0VazA4P})g zBaM?|aNHy?C2~BgA#9rZeW6Fux4*>K1hLZpD?MPVzuD~Z%O}WPzKo(OMYPX&B((i) zp(W*h+~#|+CQRLdh7+}398TJrGv|CHDw^JC&De^CVuMsM2#=lhm0P9Pvl2A=7508G zTa5RrFYZd34OX^YpAh7<0q_8IMSPxY(`FeSjQQZZ374|N`25i1gp zR9v}5I^O9~iT?u?W9CA`3s=pcFs}KpYKLedfUJI(6Vp6fbQviH1t1t>YE9dfe@ZsO zf}%us4utNN$_)vVvL=kFn$j*$;ryUp8f;qk^idPDmfKQ%tbAuy13;F73s>!AB#Gh3 zT?^dqVxy5;KpO%BC0ObDrsDQ@(WWY)rU|F&Pri}lZcTT77CHotQ#SN$sn&Uor15D zd@mn=gj1=r6@e6X1h6SNZ&qd#3_)A6weNVqx$I4lEJVP(b?20+KSk}}?~pLfJ+2%Q zMMD4@m$^hJZjs*i6{o4HVc@eUwtBXmR*4$UyLK>F?#r_s0_EF;BQYBd8$L+s^6>QDM?H+oA z<|}1XsEUmal!D_vlF8jymHK9 zMD31Z8;|uuFS5^);CeDUhHmUesm;RR?;X_BVON(v)W|Fp24<^h!4j*uB$U^y9P3fk zCp=)cb9Zbx6#wjhZ;mrud%3cW1QC^L8GNObV1$2Q0&E0#B9N1yRurk7xtIyEB-f0C zc{HObSn-vAw8iMu8woQ6me!`}9dNnvSezPd?1-4ufpjml<>-_d55n_J%ARE_?h{L- z_KEe^q}nD^qH9}tWMaGwVcL6lR8`4!P(~xkjk3=isAeVUU~}LE&iqA@Ky||Ki?X}2 z8#<9W0KkJ6*h8f(n%#CpLCFsO`caM(9WzqYz@0Tg+f%4;1{9R#zWOnpkw2)CYF}LH zS>mGwTz~dhiWTFEAKOP}VRQN7>$C{TL}=<#r~oVDN9ARoyinzu|=z%kzV6onj2Ml&E*=wiODY864^y>Anw|5b{<%Dny&L z?vlGPT^g^H&7b38BY8@hb0QhF%$~z$iT;SFVC6_**yBa#)+FBKu%2RIuVK?bN{vx` zKn(gb#c>=^&O^x|FF2xa@*;H7Pj}h{29Xxyy=AFlo{(@n6O;EZm=i}BzeK^LP19cH z)g*4}KM8R^tl$XVzJf3h*PZ+HGbU-6`r3_29f|o zvAs2UBVxbEoKlWesD>OpK7wmrlnWX(0GB%Wy$7qQb#|8NdSL0Gn6zABB6Ep^;4<1m zrY$6|2w3+8{pnc~eHD7Gff$U}zr?P>QB|FBK?v2rIDKd^9vdAzDSQ^^C~0Cs-* zOvzvz&0i$8LUWoFucN?f6w=>vdd$Oqj)F-(bh~;aexxrTs-^#ed^P}*hFF<3CYRG- zqvUYqEV$l}>mC^fC#vR`-L$Sw>trSM5m7joE&*r@PTp>2$B-7o-I-w%l9nVm#T9yy zdIZQ4$hfqce20nM<)XDLzIE9ZfEXFnGa$GCZk#T3h(v$vzB1gBZ2de0MKhR6y%+=5 z^Zr(zNd92gButtP8b_WNAyPsi{<`y8yR8`eP=Ux*Wk|Q#SE9d-K1XalAzO-Ok34Xs zT$aOeOolNNq#3gGFv*34S6#zQH}P{!F zfd`xl3c|vfmM{TAlU4@_OqKR31S|u>#Pq)*sM&~EnOOcW0AXcf|GyzvnK=Jnn}vv# zndyJQNK##VSekYKuvS~52=M;@!_+qhSJJg>$F^i)HBtqXTUr^6EgQ^RvJas7W!6?~5Xr@?TY$l~JP9cA6&^pCN?Qs7!99|;9d z+$N-|B9OK?HViWi3yf=2oQw<_TwXJ=3FzJ&Tlc{=oEE?mH3KR{l|_PBS<=(8giMP~ z3U&nzChJkCI#QTjVOUE+Drk$^yj*CWONCBMn2z0*KwbcS$FP*yaBm^bmKt_op-J2X zesUeh1`Y{Y7A);ybR+eZjacSHJ2y#hqMCxF9|$09HIO0+6HOco5*%EtBtz~6Ych&+ zSuha7Qf5lk^dfXJGl`y_YMdC5VOS6`gwiOsrZCXXQv_66$@l)Nob>+CS{>|2TaWWI z(X&7AcimT4&3xR*V_7cx^>_rxhxDKJSMeSpafrq5TkA5O2PIP}jNlrB=|WNrBGj>P zSpWuT=83}n7K4g{;75K$`s{>RLU@pZa?^cKlxFRg8kAo9n0D}5vpozSO_lp6%i`LY zNYo!Dcj^6jhcDM!V@qLHA{4Qrk!G=}2&$%)2!jhAq+S{rdg?9pz^1zx4mfbC{nYQ` ztd~TG_rS<8MM}n7EGVw49hczRLJHsH*0f6uX-T0z*weEMhgi{BomsQ?>ZWV-w=@0Yi< z=*VlK{!-N-Q=iB-$s@Y7k3^he1Aw72{8L3$=kgx9(<@}`nu1`zP#qNd_pW6&+^r$X zQHfpIE$jp+lb}C#1mdU}q3;%YW3sP4o>7&X!HOSb^59?*k=txnE=F!MiSBZ?vBuea z_Vhrce$MEDQgp3mm=(kz&!kQO)I?ofH|icj@Nx->OiM+WuTbLiNw>{$#uo?^@7keI($XVm;=U)hV$i93{9VnEjRM4FKy|TR?$L8&Zt;>Jj&g`;!4+&>0p^n z=EW64axt|yb}Z5a5iqBX@nBiUw)W#TJ3o%7?Kg>5d;Z>bW}2NuwfLk`H9wWjT=X6p z)hyZeI`(#_+Hb!@ILqso#!Qdd6FX4vb-6u`f?t))?(x8%O03$OI_6C*`{oqOoSYf$ zmGiyaH9ecvJ=`_FiD&gXq6#bw^Nci(~KmacA;mt99A`L9_F>Iki zD6mcuF+GU=-hDIqvk#{Ms~>%t3=egT-vNMR(5-G5uwcnKp##WrL!2dbOJkba4>V3T zx5Yl%vr8I>X~CVc2lv%Fv-10eX9#!wew(Ohd=?gVgL$sI2OYMwF?3n!)Nygcv9}1$cMmn@HuasJdvDD&Tz3Lt=^%ti z`9hLR(u-Zui!F(gB>7?^z?f- zx|6nWq8;H#TO#xEC}WmmM3+J}u^l9_w{h~VARPGn$h0W<6SWk>a{`@F#s(t*;ghms zMdMLs6-%VOxegRT>G-X_eT6ilWu&RA3;W14eg1Rq^6kRr5XNsD?bdMGtN}-wAWnz) zQ@$jahD+24;M8i<&SPO0A%&Xc{tBMmj?Y`7Os<|KG1TdDkbX{?yV zWg4a>{I~&kz`${DuW*Zm>AKwj2NNJu-*cN~E`(%JvLuGiNv9!QH;#j?{HKhRKEg&3T&E$~Rfov-o1PWqoFD$J@8&09*B;wQzulHc ze_cwxjM3G3(O(0Pk<{z}P-Y}bfU+&whv$_X$)q|oCuIkc0!lvE%w;DW+)0b4^*o_B zA#K^}cq>hB048)bO$}s@9JrVtmjlHLgT!1WqR26y1xMoz*S^fvtSldEMK(=^v<))w zMJkxciIRWJ)fUJ*=ZFg^%f7x@^4;1h9*YGnGx75NG2mQS`N)hi-K6hftlnbAnmv>^ zSe&U_f`#DZb0k(b^KqTOeEk!kZo{n9?*;7c2QxhYp;RKw}#!8GYf z;f3(K%<{)li?&{Cucd}Ke;~Qe?;#6EjkI$wkJb^1$mIHZ&74?(W%P1Y$JJUg0{%D* z#hA6aD?Wa1YR2-(Aeo-oD#o(0#;asGJgIcw(jiyGap8n!bz3K1H)jZE43kEigjQX7 zv=fjf!C`!X_+HzrxO&dBXeIqQtbuNQnD3hIim266z(AcEr(Hi)1BNJiSXI3@8RS*i zI^`<6r&^nYkR2;vqNIxeG8I=;4~b)hz;aLVvOS2(v;8L9S zy=~^wOWhX6%#Y#1_=aw_ZmJvHc1?Yq(^xP4ynp}*Bi&3gxJRth`s|dg-cT+YJk zr_NR6U=6yeSY2kd3#y5OQ!cKZS~B&kjE^tF&=?d;^8r?KUYnQp-AJ8N{V&2VF?0ZIVoUqq8}f+9pDm#43=Bx*C25KHeSN8)IL+y_+bcUm6eK)g$36h zlA=H}Q;G!d5Beeaq1`j;P!T|t;k7ROC|B)EEl__ry7hC~yEiAm{1h(+t|?%Qh2{*EHlu{>}s%xdLVTsqwmV^5EZzJlVsu4`U9=i!h+fGL2{9!sIriO5oeKct7t$) zXfRiV(>RimSmoTa91BSl46Z?H7$Jr?C5)OYRDfyloSluMWPv>vJ@Ex`DQf%OqGU-z z-i{49)VP;S)MfFZk~Hm7fQo?`S(O{X0)>d zDg8+14{^bgbiVPgrqQmv^J{PIQyZ$SCg&uCugfX><_sP0W8tff+-LvLi%Fm}WTu~7 z)0D!#C?nqo-=Nn6u51aUfbX*g(y%b4QcsxMeTS$yeV`c&wc|;n%Uu-jcvMQr>wCC5u%ko1F*vC=Q zN-eBvcu!I>mXkuvc=BHjKk!kaJ0(EDdLx+IpD1@N7&jFT0bi&4#nm&cDn!Wx*I3MBOyy#7kW>R z*R2KbR2M2{uhZ&zdwcT!Bm~&kG)j}n3@5l~xaExk6I*N_Gu8!Y&lUg4(ul zyr10D5Pa@9HJ*uB1A{AolH^@MX#Ll>sNer4!j? z@(z9-{t!=}-vTRT5EW3A^)rznKksN)@&2aV^F_61d$b$)dXn<_1l8vulApUgAM9Ql z>HEM7WvpL|p+()t>Y-$gAj9mMW~jcF;0@V6A`;CyQsJ_*=_pFKj)d=|9i>jaB@X(0 zYUtO8VIxS?6Acol6j)`sQ$X+zq9s$QkB|>G8O>#= zGb^7jvYM;AU2NcBy7MeGYd`Z=5y`e*d(!N92a(=QA#PkZlesCAfbOdkO5#M9`t6pP z)LY37y}v^BD?3UPD;R}%*lYK2__OG;vdD2Cq}|rsFdP#4otk0qpbV<6>ka}$;e$4& z$(|CN0kK0xK-Fqsan*4FkRY|6mK_apwfkfj$DKQTIGW9ACj9&=VdTCp@I`M6|A#OT z77n9T8u1bYbOF*Q&y@`E+-ph24;dN%7jDEIX;XGY26EzG$mY#H^z=f#Z!-@Iclu%y zFe)eq2m61Ldq1|FkH?Vx_UjLH3TP6?@C)|Wa2=+o$8{P{A4GE?BSOelf`DSD^&g(N z%ex_g$Q_TlsMVqQ`Cl^oa&8Lyy>?!+`+MHcZdNrNhCQ42ua}v2cld3S$hN_STP8nE z)5#~#ZaUt=(|7+o-qhFoUY_*r@Abk6y=EIGkKq0W?2j5Xap&lNKEvoObEQpzQ>)iS zGxxpUpgDZ)^u68@f7+hHD0(va3bSIe1blb<~<=iWkg4YaQwZ=`Z=Pt5oJzRb46 zyYA(@{-x`xzRxk`K=_-NPkXrk^LH)EdA420%hcRO^i5+AbdPZKPjpW@%we|CUUw|} zyT8ZsfD9kA&AW((ce{hzqm{H$)zi-P6cf#r64nLKZeP>$HlKR_I@Dqma5C`^Jwm0QPRW1X?Wa)q)Y`>FI*E3egy=1L?6{ z2RwK;*QBc(pXbiW#4_}<+6|i=>$Sp|HqIT5c2YH73$*>mT1TxjNWSU;SCM?z_B1q~ zZ4lYU8&Z>HI)r;nAp0~%dD=|@>9%;W!6(%ZYy_4px^bA=*#`(?-Hk8fZyVUlsbj|h zU|1x{ti3j8mP4vpZ_JG@UMh|R!7)kRhoDOJiBaDKZ)C4Bb(lNtiJqOzS>EKf(N-o= z@fo+PiJRta!F|gRCxxfknekJVmSxw)P>NA7Q?s4or#GqT?(x}W6}1QFHr-?oFZPSJ zs!T_WGnbGJLVjy#wur3m!W7S(o>vV50Ms;(c5(-S<>o2Oy2kZQ+AR}&kqt$`wMvyc z1s@}$Jn=<7I!=Xuwx5OGjCXFPD4`-o*`FsxExvD^uWGXfV%Q} zPb*g|v=KRjq9!e~Z`7gW59T!3PE@4Nja=ee^+!*e^p$#ipX~)p(?EKkn)n}df|>B# za8pW`AVZsJGB1p@kt_ysAeV$&Thc{o-OnZTFlN73f9cjo>1(cy5g4?jF2Wo^5lgu$^<`hW zHp_!h0Z6M$Ft$Ziuv-&+t*K#0Vr95^DWqvE9%tyW7N0+KBsblP?9DUYa&52tiHDu~ zq91ytoQ^_vUo=Dc;V45R`Ygke_siKQ(_1J!Ik4IRZm33N{$DS4ByQOefO=b!SHQLL zpdy$Jk;5u32yVJtvIc{CUB!;+r5$!Cja{JzS*Odt zo3LO2DYPcn+s_9wrPyw!4?KluY=lQOHpoMerKca;EyFu85Wj&rg4(c))iu9jw1f9xNxsT zP!997MIujz^S(Ds)rAnI9pI0n%%^UN z6*$s3@Qc;jP`JG(ivQ&7#pnJAU3ysg1Rc&PW}UX$n#j6E4IsAX;P5a}OxF-PMVD10 z#ov(<+KTSW7Oz14uH^(s$x$YuO=O~Bq_q>}Wk-(ek0>riOC*#7yPDiHUPF?F`=fPy zo(bs=|Inv9hZ1H^|rCwZYBK3T$-Gt%zUat(Sw zQ0CYP*rn`Pxjg)8Q5IqVGPiHM+m5cJt1TQNex`tj-}vTEV%i-5ufml+_7}}dq5fcN zmf@%GmlA_e?wpzHa#Ds%W~q{X}}nZ~@3!`FBxJuYZb zkDWHr08mPRib89Y)|ur+scdA+WVd)4b!M~NB+GOPpR5Xb5W=f^Ba!-!ECN)DnA$5r z)Z=s2yS4;;TQE0pE9o1W^poOh=laFv1$O5ZB6MRT&%{;psJ(FK+MlSrPC@c0W1$G} zikU~O_L;Y_uQY#_)r@343}(t`9G4$gXs{^M00>-Py?4dhkZxs?A?@`17Q-Sea!l3& zON#<Mzed4zF)M-o&5Vi+8psYY_Z>zE30ktIUo- z0+sVDFRHLwATf^KIHM?b#Dl%&FA#sSuk;b>tEz#!bgS(Fm0}jPpAvPtV2XJP z0-d!GF4|qK1yoOe<;O7VqPU2A)f`f8*^%kkmI)Kev(6W(ZJI)veLr)LHdnA^lVrmi zaK{OgfDL9B5Mp0A`fg9TPWTYX$}uN;>&X57`5YW}*{c@1^nLEI4tgPYKhGbUlY-I342>YGKuyP8_o z{1VLap`$KEstbJ;eXS?uWyh%5=}|ZpN4>7g;C66 z=v^c;h`@-mU}4z_+=odBrm}Ly5Y!HJ>hLjz8;pGaMD1pMw$o#x^Z4GHGUyNYD==d5 z$-P1WEa(!l##u5pUEw{~nu2RBfUx{{l*$bQ${$J3i|f=YHpRvj+2~vfvpMger*8@B z0Kc^*gP4`~AWnCk<`)k)?FVX<>g9iB45_;HWH*a0Nz=ek)(Oa zD4RLs5!o3!$U-YOvBG!zh$*T3%7+az$|Gq5Z7cKiS+A6=>;dwac-R9p0nt>UT?{qc zImb=bvHj6|txPbP?|&~Bmsu05T{@k`ZloP-pC#2Cz^XHjl+&c30J2pSR0ueE(q9CO zcsWnkUohnHr8M%pd5X(MfTo}(;!QrHaRV&U^vj|WccDWsO_5&;@8+ox%SX_U!9 zrq^7g0{SeU)E1Qqrq#!refQElNtKHeR94)c`6FZ7+IW<4c949pa4X7Rs%))X7E z!7(ZxkMB3<_`hMX1J1?E&nf?coG$5*MJSS`maJy|ne@d(zFsAuUn*nK_-i~v+XxYk?^(4VqR3J}LMOj%ar}He+~6%;4pK%1SNw|sYT_Oh-A(&|Gy^bSA}?hVis4)? zS?yv^buteQ^(b|pwV|8jry(uq1V0w)vwzCucgyFVY@-}U5$G#h9^Bv|pEP47yll@d z&(_~p6{{6@32>&ciC&QW@scB+D+=*0x*FHp1Pl_FUrE6`OGq(U;Jn@n~dqpV;3|7lU?FMbJvDe$&FGb|mAXJ{dF`~j*?*pF( z`(ZcpXNKJ2yy&bZ|G8@Pi%42Nn)MAI9d;y*pn3lz0-)n2xDa`s&`el|^LJEEje$$9 zF;0q!d*ebDj{Kh6@-sJFi-#?r5D_bsDD%&KC0bx-6!oy-eVe@%K>U$-y-4uXc(C-H zvBh2XDo@^HzF90l1yP;!t{&x~dCCe|Pb0zZH(u zUX>ZaUKE{o4AQ(!f?kDY3)qn!`uxA z-nx(s4CGau%rLnq->NbxVPpE};j0?55Qes-#a$7KvIqoeG>TP&=0A_PvNfzgyiDQl zr|pq{AwpV)lm4>-W#;~mOJhtPSOA#)9}5LL1?zu(65n15*8kcyn2A`~{%hA@{l@>= zHU0^fzyY@hC<3z~fUq$CGgT%C8`r7(<=fqH)*fIF90vAJ ze&xG_H2vG&fzZzF0{jB?ola?I@deHR1>s=(POG4TvatU9Q3yco5AAr;c4Ys317=2G z?|{S|$Pbo;@YyX~TcrA2Pdvd5|AUqvS#jh~9r3&Pf8TxG8W1Bv9fAx4h_(5$BpWZP z<6#^SI1gtH3_ef4Rxuq+TvY9+>E`TjUO26?9EzzrsfS#%%YV&&-kg;ebbR0AP%m3zRO?=l~AJSCa_-W|jAd3sFQ_gNc!f&%KSgAX~P z?>*C?&qB1#!z?0iLpM3#NHY(%&zRhwt|_REH!t36| z-l!~JNN@>kkny%fml5U3kLmcu%WZj=y{{yJu#1@gzS#V$(L-Ka(P4azctM`Nf`!XgHsR*zSb9LULwR@HBi4A-Z~g=7=He zpr|K`(a$_-|E}v!E7bU z4YO~?*~JwbVcdE+1dpG{JiE>>1!#z?3x+2m`$#S& z*~;R9zr@oFe8o#%E!{FYpq|C#)-)J^V#fh;;(<`~*;ntxw))o6T#2Bs&?=$Vu2jc5 zcFf;OYy`udE*2%54jOG9En=|Egil<%dnVh+b>OKOTHxrD?J!F1M;6%SFyNc*@DvSZ zd-CB-Hj2H7mb&dJyBf5zSg*3F6x%vG7aZG9Th^5h>&Fxzn6UFLjh*?Gli=cG&#C~= zVwA57Ikon@I8k=9jtizt3Wh2kb77Akc&vD1^IjE!NIXJl8$)_F;6q_oYa_mV5B3Fm5u-7L| zGv!!xc7#xxu+g~Z;YjMy!4L+(|M&uEGcb`A8Ur}LlzT{G=_Dn%JqS$4uREP(kkhuQ z*e>55sne|-?I9<%(D6q6C;#qtDa^pLXc8J%X8f3AsI{)rSgj^9ww`>G%F^|z_59U} z!Z8^DfkkWyQF;DEYGx;b`UFK%BlydwY z?~{Z&A9m?;13rF(@EI0=asdN?z)XH;>182R@-E6R22B*4i`l!`G0#|b1H{74M=f-A zTV?CcK{g9bjP7-Ze=Whga#Vox;H+l%alTwRMwII6Z~0on&pW0czit-KkBD&D=5jo{ zEisiQD%)=A?P3T-6;mLrf3y@5MU~XJbJs}LWV1_OD1Pfp7mg^6k+TI*xF0}GphA)v z#Ze5J$}639<*aB3_N#*khuirS4^1t}3PPHVu7lg1U#fWM?2qg{(o>uOxT^i-mP@*;vR>$FM}WsW zRJq&*{`BQ#+gJv+;PYBl0ON@vDsAREKxnG|IkLlH2wLLWA40T487VcR)0G({jih4;))HP>=_>YOa^0D$LE>j$qwNG9-HXNw1a6ymU9AJ-vSx6QH-b|iAai{ z&A`zEuHHn<&?x0KAvhaSKeiB$Z{CKyV|qDE7~ux5(xtlC3e;e2K9VY(AU4oK(i2TJ zHB?HR!FJ{%O?k+OKx0+$z3Fcmaa>z98!?+l;JGRt>f-SpQmsRko3-?s&VbiG$;NPf zmTJ55g9lE_*?|BOGrF+$JD|s_C`Io5Q5DXlGK&vQF**uGQIolwWlrg+eyOgw1^0#A zL*XV9=Q_V1_j23u+?6wcL8ryhhw=|KUO3eWjo!m+WWk+DTAMJz1q{!jC z)}KEm?(sOs)2>aZ&107mSD6@K1u&8}noF*if%qP}{d38Vo^O2j^|l0wK%wce zX>!OtL~NrLN5V4#R|-t0?_= z54%W(7y!}=wWf9xT+WquP|}6p8X2DI5mp=qH)wSK=ho57*?~;c(;tRz$R8(> zTiw29vxgWO&kU4*>GVadC!ZOLFj5r~UHtJm4Yps{5HK~+L`#&V69TR<&tKY1>Y-Uu zfRI}qTC{h&nZP2Fk3g%$eyc|#|N7$_KC)>FGWlW_S>y&ZElN~Lzbg+!jQFCd>T`9ltT@VG+-L`=<1iCW~0X% z73qih5U9Pa@ut@B8$7H-6z^p@`ydb=>{f2Ew{!2Fn6zK?r4>XvFSE18t99|B> z=PT_=#Xa{&&6{qLR&9X7ivD=a0(?fVF&2X`0EHa2UOHPU-95qDw&Y4X#q!&a=dH31 z*i3~D$LmDGxBGF2;hFpKzNB~ce z;rD;;wylU*B{$`JYw_fN8GnBrv38*1^2ih>C`#C8e|EH2?^p82;mySGT+OS`N<@No zBcx>5yR$ljb>EwD{j)oLBAUZviB6jhn8fzP>ew7aXZR_F;@pw}^J?6d^)PosNA%;x zhB4LET%dx}fWqY-Du4wP8r9km1Hzv2U4kyC!%>0zpg?Ly#>=Tv`QjeJy(hTt_twqSw*T#oDC9buH70;jts@|hJOu{4`Dz__{D|r|WhZ!`$qS^^$oDMSE zbS{cmm0n+3*_90G6gQO_y2b0BB4TP76Eo%D0KwA^8SR&dt=bZig@|o$3OZL;CXs04 zoCb|SbO;m~GTim=9*geW^$gtxfUwy3$k5=59^_uQRZb9AN#s?dO<%B~(DI8(uP+z~ z<<6=nc`g@0wS(vlJDAVq6YEIo4zg6@#t2%*TDUY~!vf1p@U5ielAIB93}E?#X;j!QM!iZ6;#f4x=GUyNXZw|&XUMr19W`;{>1UT`(5NS zIw!p7g(FXwM_Hn3!!Irw=cgs`7$KnY{e=flc?@S-g-0*BPWcp<_Q_)88%S^=35D$8 zCB3u=s7l48%v?_P4dNrGFCG|ok{u7RGUa1*T=Ae)^#`Yy079(KlD0~VXZ6_&fe*gO^;GF7B0kB?&9|8)IkT1 z7K>$i)0ygH>h82FUu#OZtT;8cp^}0d1MlFI>tUOCcrQZSFnmsh*|CB5jMAos=e1sdieVKhZZTU7ZWxVRG8XAB&ml3RyJt5O3_EyZT=zJbJW;jQ)zvUEt|& zTt*b$+XEc^8eN2=`YUDTt?)2l6MR3da@hK%rd~OWCPpP6D$n92ufW=? zPnLjL$tEjsUIb0fEWDvBPYT`l_GN?2eXP0Kpl4kFP}jl+v7U7kMQZ$}rupND{;zng=J#(yr1ksAK^>`Xe;<)n(3 zn>EmCPWPX(OxH?j=p0_<#+?x3)R_<;sr(c{p_s`LB|hw-3}c*B*uGHxu*N34`{I~- z!Oa<21YjywOddhIVe=kbuc#}2{<*j5dDwWkj41XXk&8+#Eshd>+NVSL3!ka?)euhp zq4P;0?o<(9Txz1MJt7wrae9jAanGl?`WK`zs{h>cDwJ^bFC^Vl+Q2`g3L5bLa4Pr# zz{(&T|FR(*@c)Ml;ruTd!pYkHO@@ep1OG1#g4M1N27Chr%+byt51b5|9^V2C2m3$$ zeVMu0xc_wt_+d>uZ*$z;)E`uqn$v`MCW@JRJLX94Zb{=%A6w<(VWA{(6-uR+Ty(v? z-53i?r)0(<0J&pE``xN2Hy{)p%#KE%jyA80fQMyjjSM=YTjny&nI%YMk^g|4Ch0(& zy?T6DkDf5VS<{?`MaZT)0FFnKB}JW3W1*H$aA5obbQ#h_s$Eejv>~wXJO)^-Zi$9K2WBVqoKwu;@gkRkd>EUWtOmh`HTRS&>SI{{SerYO$z0&epFQ3PNw1wf`{-)8 z{$suW#gA(Ble()1I|8;ldlp$0Q#;qkP<46Bfv+GB8ug}q;s`deJ7yuysQTMBzO`<5m-6VwTrXh5I66#LIIlp*9|_@}d-lh`*?=t$dD&E(tPAeHmpN z4ij}RqB0%3L#o=5dt%i$sv`%}tj+$ShGe4-VEsN9-C?D)3dw`t;`IJk{@ua`u}AD+ zoOb^BCw&6B!^Dl5Ur}E(?{_p6XrBvpuJb_7<&a)_>$-_!duxw51#fRnIz!!OJ5VGN zNuPKir-Gyym1ijLCuq5U#|cO{$m%7iN4O#QA89J0ymOh3=^F~lQGj48iw%8!iUAt5x5 zYX9s{)}yR1#&z@k+viU(S;kXd=cuPB0KD%z2d#~3T(oFmX3TOU=tJADu1+_%N{4Lx z#ynZ_As8lGwa$5K{u{ED^$>EFr&E`-DaZYwOYSu(f;Jr}p7OT(pdm3NCi`K%DI%g` z(vqjGWY;#r>-W-v=j zfmiZr#m}+H!l5S0D!Z@+NghL4fK%Im53fgnqfmU>z@h9Zvn4Ote4+UB+9US60{DGg zQ=}Hl0!2bXZDf3RxIDKMm2pQr$+s*<+3^*1KsWzVYZB9uD2sFei>Sik2JX!NZWB^R zMDDF4w*@&YJGP~bp_%Pw7TPtQkqEe!WmA*z2C=+j$&jN!DW!}Ft0`RtfOFdOROAz= z*)JNp%LGgjOtE(AB3?`3Lg`rgZx;@MP&nwe-x6;yk0PtBiDfS?mMX~ll;a%P0KoahNRDrx5OSXlmMqQ&$yJ0Yx z67e&k2w16SUNXpNxS0+COfA%za?HxitckA#s3u3GsKi)m(bN>uAoLR46fBDqV93lC zcsy(tBEJe+2Voe*%bRaJ-Q~3LN`X0(tgesW^8f^?GY14?)aZN6a5-$gmuqCY0~Mw| zF%%U`9t5w^2Dj!5H(AzLNb7kst@I=Pk2CgspLQDc%?J`jHP0mgB)mHGi=Qp%)>xt{ z0;OVW-{Ex#G?b%&^@D7?DU$=Q1R@W=F@-Q9s@4 zy2T1v8Ue$s%W~5@>Iu??66<>ofn`*JCZF)@e${!JYLhwz|GObd`}@Pjv&vq^x&RN) zmxH?adaw<|&c}Wb`OQ$6+iK8b53_x*--Em-dYA@;_}Dn#DU9XD>fI5GL)8t7!=MFi z52;G~+*=3*pl>GdQ-^0iRZqcc$*5BT$@&pR;=SwF00e`6UKn@y1XCrSKRq|>1T}1g zSc6&)yxYsgT1d@VXMd|tIgA954AlM8LW=<}k6*Eww*O}`u(^y@gDi>UP?S`K zv`Tt@*P&?VoPUe1!foE`fuB=Z$*WR>wn}D4kFh7AB<)H#;pG&HAm(aC80} zm<6zM|62&dDE%F4<^ET$`Nph7Y)t>!`t>=5tI-N)c1g2g8>wB#P!pza``WKj( z>YoJfzhcn8)t{6g|5S6ft5Snd5P>lLt1XfKGqXVZjx>nV_t&N2%)eg;>=Y$HKU_R*CLUs$+Il9nl2`ysZC8TBbS&OIw%)2CxcKPJ1HZ&5<5E;RRLN% zrIq$tQ^hB%rXfYUD^hGz2*a^XUPc!fxt7sfhsIkllLBLgaV_`LW^o;31b|2cJ1uPk z>#U7`o(s_$d?S_2zIo z-LP*Jj8IE^W)eioik38Z2w+tNltQ%voC8M6362gkB90}LplV63Fgz`w-Yybt=_hd_ z3Bi_EiI7wY42xQYxB)yI#tCDo^eqb+lhSA&`bGakc!gZxT3}`(jgSKA;&9SmxX#@)B0$^V!j#orQ=lzX zc&Z4D5?{HbfGWXda0)pirn3qJwYHrNi$_*X>%0QfNmuIwq2eg?`gLKiSfn4WTL4-E5hQs0XJWFO_T!jsM}ix zbBb}qU~AR06BY%)(2N-kbYRu2gK+JB1v#@d{!o9RH_S7$Ihs`8ZPsty9bEvT^-wOD z3x@n6EXIa$OhM9h=YdpdE~amWK&K^w;Jd&Lku@RgCzKxWMaLUQzmf=UhrJO_OobT9 z;<$D##Dx}~kaxfYpBPL+eNKa=-)r6bj$MURcr%F((*?F>xzk zrNByw?BQi(s@znX*TSh-Q2I(YRAoer_+pzNVo(}^z~NFFNSz{LYobAx#2*ifEpRK@ zL?Nn)e+ikPZ$pY~UV5jg>Vy7G)>75hVjqEfltFu-Fw+l)?O9#8U%<_xjSde!`ChRg z!%(N-kLCa>(s@FHiZH&EomBH3)7n9awFOQy@`#xAq~uDKJjoFpH&|~Y0?(~JzRul$FUH$J1*#}zDqd$Q>d$j%!7%F z6cPEDs0BY720JHkPrF40Ek0(ke@p$5D$i46K1?XV8ow==2=F_91mb$NJYE z_T~=vcayWuTqcPQU*?Qc#2$jicBsN+zGVVJ&6Mu*XHJCE8^eYTZLo%7pb9lMex`zt zW}T+cSU15WT~&P1+$=Q$$_RukZX7#p0?S$hIc4m*OHTwP*C3Ug(LEkdV{|5SJK`Gq zaL?YfU3m~#ojUkqbyhBv#-3w_WmJ-bOom7&RdVG6&aDVQ;Z^%OMe5@@=pPk~iSqz$ z=jXEvuG8k}z5kD}a|+TVXv1vVwr$(CrfqxLw!XG)OxxzPZQHhO_s;+CZp7~1>_xp% zmDh==%zVyCG+uD~$X?|T-Kb6d#aUoQ&hhcIYPWI&{>)sq_SH~6B3o}&yRH^gw*9z8 zgJ!$(p%;ww%}SekRj11-IksvjGZSy(2EuLj?kGe5arI$@_SJkg{e0od8cA)neYsKT z_2L(ly|Btq`kC2?Eb82Cw|({e8^FuPVMB& zYpW-lrNeEZ=0bRJ<7&1iE9)f~Q!<@ZlJuWTb;8p>K4)(xPn;j_Ys1x9V0dv>UsrzY zi(|O*+ZI#hH$N1fvYudc8vHVex%-#jxV)6arSVJS6T63Ltc1`t)9Bi}7XZ)H4*sdR zjahZnslCklg8`{vi>=^ldZPw^Cko;MW*PZ(8_pPSy}d(|GfZqlFQTTcY|XQ!#y!ng z#opb7@mbFs{2bWup$5Fv)g8~KN!8~um)Tg8X68}opmce*ARs`5g=E9T_ae6*t);#xW4ag_gm;&PR^fJ(2B_anxHbNjK3 zG|kZ^KWON3jX%NuRlmsfe4A?qJ{ciYgWGPlH;JGc629E+DVQ@}f8ukYS}GVBdSRJK z{5HJGfH7nnb#G?O$w_)gkRG?1Lov9ZASa;{&5+J_Wcd5+@{=Ol%M1D|I=kQQy>FZk zv%DJ1hZfBFm-j~|01!Vtm7KpTwLcdlPpUoE^z+_lHuO3m-1&OM_G-`tZSpV%-`cqB ziAkl1MBr#pSKQ=3;i4~SnMR!ijmgh}iKas#$g{l0CV`@fepR}`LqH7J5*7cpo-9V@ z|0ekWy#0Lz1~7vb7#?@vv&}f^*$-y<3UWaDu3d5d&eQS+21r0?37N#lUs9gKM@W++ zc?}-#UX{R-!In$AP;R3oEO{lK+MloIur8vE5Al-p4~ zuw^durE9^bPCI&w1AdJ)FN#f9JW@Qb_AA#2QJzJd1VgbM502XT*K~nzl!mjOsnSSE zL=?huf$3n((8ef(h*0}!`1;s!c&s~rw-_BN9tSG00s3=Rs+|b1pNDF~r1KXc9sy}7 zm*}6t+`8sWG3|H|Vy$;huN>L;+@Ao8C$r}v%nXl=2>W-!4>`Um0bYB?XVIjK8CD_( z9K7hsGu&QfI&7d&U@GN9!qf}Mv)^T z@&KL*;Ev6^SSHXfAtLwtoim;4tY8%XQW<|cQGaIzVPF>D7&c3G)q>ClmQqYUid%QK zEy(q&Z}_(C2xVBU6z|kpu!fm0yDcc-snwApM`!?!ks3B5);DJWZ4OSa#Q6Eb{eY8E zveVEE+1%w7yq&+0eP(s_TI&&AA>-60mxWCX@a{A{nx~$+b4Wq(&(|b41;UP9?}1<7%T=g;+ePS1nBQ9iFr}foLT+E%S|@I!j?;5zw*>}~ zsALfQW+qKLMW09=2E5Pf94TRBcbLzL!P8Acugk6C2_F-1nxnpQL(an*kb5ady{DlL zfRA_J4!B@&0F-#yah#6}>42N?H=y?xddFr@Te?G{1S-te5?}DO>1lnn)n`pTUCv=-u!6F;1qxP+tm<7FZDDhh~Ye z=yTwN(g_RKIqqbA^G-Kw1Hr^sX6A1_;Kifn34i8qcaW5)jM(OZ@0&yp3i6MKLATOa zTCUIA|F`gAjyWgjhX0>kZU2W<__K@8~b=(k{B(C-S=_sUC3A28gl`7wi( zPqoG#8c*uPG(f>%2%_alvYY(BqHmP=-~Q^6$=aGSaVE_g8K0zo!2=*d;d-$P8%Wx{utrdop@MW|*`! zyXBAfU?D>6CngX9D&00C6g!U{fO>)29vlz4FKPHT!kmL5J6t+w%FpFLJFN#Py}SHf z;VafBpR>v4@HUPD>(RFns!kLwt2Wrm#k)^OtzbZm+pWe~X&|59s)0un+WHmoA2|w7xKLfO&r``j_UQeR|IwX z_${D!YzD`w`h1&6?sX^AxBLi-l?K*u^$DG_xiXS(JEL^ZjxA(9om$a^jq16k(6meV zvYOO{CeR;W3Zm4uPnOl10P}CI@8-L*P7}V0Ypl5a^&FUl^Sk3F=(PnRyeXz0Jed}26Jg6-#`c` zg19m?bX5F3*G|(mBZBRyluv)<%HR`zasAn0gzXiaVcq<9irJm1&lEf^!C^(e)RTg)xBg6NF|R77)3$X%ADIoLm?N7Dth3^aG3C1 zGA!2TS?)%FLmgtw88pHI`Lj=_SkT+-&Y;si^N2J4)dm1A--jV-D1T-o)Dth%9})l~ zrSNd<(`H5fKAmIjJ!*%8q8gMm+I8} z#+9>hDXTqoC#Q0Z$c7-;q6+>V?B?7vFXs~Zm^l#+dw^8pT9!u?dzyIdE&b9}^r4>! zNw@Wqc&+U1dD@MMM_!Kheb54gN$%I@pybJjS)Aymgk6XUuD`fY%yg!l?m{isjz*ye zlf)1roz_p*1GxVv1?Bp`Cp~cCPlKhI=$JI&b`Vr7mj7%6{XfP@X14!joQ%=<&xsV; z)(36VxezOHFl3*kd3)XEXoLwjfvhWqYokB8l<`0qjEM9??F~5aJj#@519AZDF38Dd z>8J%SNnvH<*}V;m$sW~^C#PXZbU4hNFPY zt_9k9p>Vnrosdp1m^`#Q;5XsH3}u*PoG@1CDM%dXG*srt4O^xA(s#-hPX@G85r;V! z1>7{p`E?PqOqv5#^z4@=Q_*l>VG6#QTN6ENlA2!--HIDhVVmw%gEIHoFL$|f7TT8U zc~m+yJ9H!W3)+uSq%K;{6o#S|#9I9-7PXox?acRys>R;ab<@gqz|;>(Al);Lo$|VM zyV}lXzQTc6o($m5NkR&2!%`FsX^a}fFvcGRPvV;X24g`dDyWc69)zWYYF8R=Txt?D zRACH(YJ}MqUC}<#be3#owcIzioTr7+*Af=_AU=$qc80HJQkXMo<$x#}JCn0SlUGz6 zI+M21nM76^t7=tN8B+B`xb4WABHE;gW0K=i%WEG$Zj7gR2+i$I#E%G9y@j_q29HsK zT*}>f_;kJ z3;8?A3mFzpz_+6JjG7vnOh;7u6KBBZxM)cXotpPVF{Xhkq@K)3VIpMD`*5vaFqW(N zZm0KbLz@>IWcNwRb(dGgNNXkWKq6WL zyuB?&B5{W@6@n_>Z*|2n@k_fqH;N!P1aA3(qMJG!!0fCnq9r0Topq^Cb+@*gs{R$M zTC7%N^oD6N9Rw(EQ955~c87!pkQK#8iO;Gf>fUB8{9Z9}=;kKqXDLnPaP!4Dv|r|Y zFc3PU|AX)@D-Z$2@(56_w)Nj?t5NNMJmz+*Wh$?_Z_2;keS(mluPgRkFSs>LbVS_P zjDG(_Q`}?}T=goRjY#?BJvMC86}q%iIBgo+6g$Ur+QPn! z?n@nB| zlZU$02^8z{fuZNw;=|NfJ=QFC^;y5AyaDzUth3&m`D>^I)6*uaB98ogd!p)l4Vj0^ zD+)_)EFVq&6g2)tl>V3J_J3KbO5Y!EmW3dEiUz8`3dYsGx41X+)vZq)Vstc+XxG=P ztMphA3cYfF54(8QS_k>YpB~#Qrjf&#v3;)f1p!cY=w`klYxGA_0yC!xGeV)YWPgA}QUAM(@b3e)g^wK6T_~++0tBOF%nWpoFfHuk-)nCOXcTrD zB{?WM3Nsfw^Z&+VvJf#dv$LlmjX+_5v2e2f%aL4zQUO}P75}VZ(91bU-|o&^IJmpJ zL!;o~SX4Q#AkhgqxWivx_XX+vYhIBwKc1fK`Z0LYcU607Y0myTqxG@4f<`7&@z_sc zNSX(>I-HP`5fA}MM8n89H3?>DY;5-L5iZJ81m)HS{xOj(dWC9o2MVzD_(mtTLS%FC z`5S=^;OT}K53&Vfkyi&YR|jH+L}aBzL`C%viHP_R3^4R2@K4WY9R(wr0g4CS0K!M2 z7VB%adnRIa9$9hyet^njDS=>ec6OfsN!Mvfv*OtFf>7$ zoxAFhQM}9GpVDVdoiIz<*Y< zLx{onXL673DTL9iF3t_j5$0uY)ppBdypv7q4bfU}-w80rfCh80Q!Aj(F z#2mr}td7;uvB9Yx)ZYy#ud<%|;07G9xM>pr;K~4G8F=5`*MV<<$Qf(^c?HM?)Cz_Z zAcRf}`4@Jf&Q2Wc#QsQxPfh}%#@CbfLRmmG2!EEoS7KYdQ~L1(yyIziQ5Pyd(}CrB z1Ob9)-87S6HYP{U3a^BclT-yY734J{9~u(?UU9MB8X!b3bqpYWSYA~Bj0ivsZ13Cq z?$eHIGMq5r?p33vDTW&Bwqw=4yB7C}aC#aL;eTEWoe%Np&J2EA=@I}`0Z6VH7@Ri! zd>{J!sd^Dq{{Auj?il}3JNucC>K_|FM-*Jj1^yxI$!8xOzp4$_IOE~&jt$^@b;Yvz z=2QiJJ-HkQ{Us#7{Bx1w#s=`ghcv=%)R10aZDeMu|FVSSmh#sPppYSwGiG9>f{u5= zm?FP9z5=+kfTQPsh!}th`**@G&@I1XS`8lHek5v@f3r0;u>Xwof8p3Z#^oNN1LOZ3 z9YZkT{KTvR3RjqaG=w>T3ZZAu4X`IKfh6*-E}npzjlM@lA7xfA4dOw2E$$d!W#;eg zNAG3k$7k0^4|5Cf|T3Q`ZZ^`)l6zI{#N5b#gsr7Cd zJe&L4NsaEGMKe#ceIS+(_pr%Y>FJ;Ke2w#Iz^3gFj3@cQeQmr3OU+MU?N{p;Z1;8j z2jOiGnDM2l{ZB=nAjO9o|4$!ab4e(Vb{5Og4##HlW2oR<7BFEO!qg4>wG2f#DZO7! z0sAFYL(pjw_Zs}I-DIVG-f5FCsLR57`|ouem@y9V$SXugS)}bUq4ubU9WWb zIPsKKjunRsGH;m-5l$E5flQKf^URMV6P#a$4;4sZ&R>B>l|xw3Zf>GW#SBZvWd@>+ zw$4KXqkL2c1!xr{sHx=QzH$;_rD7F9-Z@Bff$3QEg;#`aLtHLF>|^rJRV@YV89;bY@yYk!87cx{N1~8&_C)bT za4L&1aD&~0irztGoO}cHe6ECdkK){z zLG`a-C%|?M-y#Jb|D^9c!id$PiKqEU1Oc`pdL1n*?LM z7=SMW|OaXY_IMQw=r$AxSUR-VtQ&hYfkXJH5&FF7M1!FaYrdb5w*HN^9xG8MTb=^ zTFhB@>D}K(=AS9vW#jjIgzthSzyIDqlK{XNu^T<>W27re4>u)h$xY@x$mO|ks0lfqE~diMLzS042@lbtJ|GTXsxX|tYILB@yQ=kJi~C4r92x<@3)to79}0u*R{JZ0bct&3JzfCT zc>y9@JL4t3&cUPn$@+iN>`h*i zMC=KPSvvWL`S!T(umw0;$$vk0(7n@YIl7K#$V^Qa?E9<62x;B+cnRXga7bXWU+oe5 zi2snDRR_F(<9$Clr0v!WUAyG+0hWMYHBp&^UBWw8xosh~mQWgHcQm(0gu z0b!O&*!Nc&?Sq|R3ECHmQPeiO?@H$7lHr?!z-A7Tko*4TI4{FO-{)Ubj#rWXd4$8C z<}upGHiEC3L*$-LF_6sRRN$)~fukO=Nb^myQn#Rx6IXj>z-t$>=J5y5o#udM04jCq zOXwJ#g9-T)N02r zwftvoVULX#B6a;h8o`7d2_t~W?f|IL&e4)ind7{&)`AD89R>z9vq;sL>A3*#_hu_J9BSJ~s$ zk+i)i9;lp;?^jwR8Ul!xON?-+^@oixu@{?QktyJWhToOp_g@iR zCab$cL?_KN3#*PEoehJYvwl&xu_m`CKqsou*$Iu}^~PhB0hm zee`h5ekZ551aL+?Eoc!4*S^g>7!DEo%bO=UY4b3(ry*{x681y4_gl$%lnTVZnET1{ zD_mR9RMeTB4y&(4w;KVYwx{xqV(Y5CUFNje!Dbz@?YiaqKJ`c?(#%X*P|6GyRU&!g zJA3$#m5P3)@}J_1%NFjM;V;dy|GZAElGJw*G-%8p%)1Qt3VDm{Uqnh}zGt`WdT?b2Bruh%htvM^@Jjvn=iJ&PG3R-o8eZN#1*!^Jz7 zaWbtVi6<>@6*B;qaS;)%sdT5-^=fkzl2&5%HV-ww)I{T|lKfT!rpU%t$rSp-)LAh z)BG%nnnAFLj|by|j9q*}%Lv;lT2DzUn*?311EGsU8VkUn^d2CGS%AhK&~@)>N}td2 z;Tju<6Pxe{v#=McgZ!O%8osANwt1hG663>!30pv66Hc$WT_E*OqM@qo4p4KT#aeU4 z=M&@XSgrd5h~ub5S0QJgj(y9trlmREC|mpd;B%Xs=-3P;|IC8)IeB*t9GtJM$M8m0 zTGZasi!gvqDK;&EqSiX{X8yc1bEvODy?$e;bu8g681>4QCkAHd7pB3aQdeyCj887> z*%gS8l>il)VsQ|%X)MxXm}Gl8M$WmGV`IU4C-F;(<6=E`64tpGKg!pu17<#G%c{0T zAjAc+GJF!RlW-?8i-PD}w4Z?AyCjxO$YRbF2LhlWAB_%~x--`MT7{D4V32GM(=jTJ ziNJD_7;{qe+bYG8@F<63b&JiEf6JGWfdfUJ|KtL1&xu}4?iobjKCNx0e%ih4*Mvul z4WPoR()WAF>h|!@n~k-}P3^b(ccFyRIQsxXC`Q+To$|iExFvd&_j+w(8KWnXaJWwe zUMyhpulklO#FWI?h5<24CQ%V$-7xf1`9f0?L&;-6+5Tun?9KgR%o<{t_)4&dV37B! z(ueNRVCY5&YR~c;-xJm0zfeOj=%GlQ@-J__7OlaueL=)EQ}ro@QU-sFfhdU)*a5t4 z*Qvk3AGJHItBeC0LdbgI0GgsO8qW&BP7%QN6XQQ&M9fZP9b=I2nn5V>+a|yz z8yK{@Gqc2N#Dw;$@3j>r`g+l?fHW7@F9+i{7-GnZtP ztuS8fR#F*bF*KbG5H>_24}bqHT>GRvnb(H#yNzebTs~fGK9p$Mz$&|p?jtZ2-;vB0 z`e*iLQwWWD zyN_N-c-vZ`k8h+ML>366oYLY`Qs)G|oHZ)aA#2jh`-A)rL z6-IS^bwqe;;kQ}#05KmGcVG)X=%2L|y_~kb#yN4ZL8zl;ya5)8Xr_0i*>SH9IOp~I z53(o9V;^b8Wi+Srw|4qqERQ4^g$nq3niig3P2nzeXYzT< z)c3AoS!%0LurkE}mB7)R$${xUQVuc8``pO_seduf(ZVQLPHHC}%I+ty0E>QG>yM!; zUb{^VsiooL@p^nXEoWQKAFH=c1BGn#!cB%Y^Ye4l0}-^)Ccx`Z8-LT@_&zo~i7pfK zfuEydxML%51pvJNQioosDw``$p7}tDuRJ%un1iKR7`9skv`*L=F%f2ivAsUfKg|nD z*Ob`(WCPthAX%>+I``h&ndzUmYXR%U#^a(`_=E}V3~t%t^g>#-^%FJW^YJ*aV~P z^sMwrp>1Z9?R9tQ@l^DCgOXRJ3#(zA-YzsH1_G?^`rO29gd>wfP$&|%pJ@{>BizNq z=Z48?p?i{EL+D)?oirf9N16CKlY=}Brxw3VpDlmb|XDZFmyph?$}Q3mli3iFX#s!u3p4962${D zp-o(}o6T&4xHMOkXK3ml!pe&S*hZTm(YB#TQ-;f&pPrxmNiA*gw^25`P-i-t8$RvG zObg|_nFlLT(9M$w@cZ{Ua5Q6|%xb2i$^f1kZuaWWg_(j{%$VESnqGgZxfHS{gtXte zQd^mMl;tEeq%nIQm-I*7GX#Pba_ch}X=*4Z=DZUWw`!m+58xh-!A);}87sGbU^~lw z#Sj187R0GHz2CM2bhj#O^7Ne5F~+dOCwqHu*<~6H&gbRsnS;4fOC*u=Dc|2O=v(Zjb)NC_gA;8xgHKyEygyN(lEFva^9zBLPe!6YDdB zvV`sYrKv54q;UD(vg!VIL;R72J4dw@c?%|aQO~PeL6T2Jn9s~2fWhyn^jnQHc1=GH zk4>RN5Pv0tO?ZQ6=mY9w1jvIp0J?P!8SSXEK zRgml!00++x?~Sj0hHW~(XakVe`UnZ3YS{|bj;mXy6lCr1hY)fb48pD@#T&Vc*G<@G zsIdsEv*5arFOqn&PkCc-OwYzg^TqvI+E!)M{5Do#nsaH5OM6#8vo|g`nrJgSDPieb zk!(7S90W+`p5wuNQE^9}spT(uT-yC%bJY@%QGgHRYy9Tw5>X>b!vpl>8)0a_93GDZTG-JA&TUDE`f8(rr9cfBFmL)Mz^-zJ!ULIQ2scn{kRoE&10!ZvNHGC6tQDc|;#} zDw{Q6uwKlYV~O4B`T)$$5~E$nvoaF^c8kX!!cgf+c_4Uu zz*G&Oti*cR`%*s!UQiV0N%Bk)l^)6^0(;L$()N`Mm?%uNE*fmdu%XA!*F+L|`RZDvP6~wsp!2YQa1|iJn zTZcp6bTSLwumqsLid(RnW#A1#a1*v7w>{3~zC3xEIWng5WNYenYatT{=9r9hvOban z>lX*6jp7GwTqLPV&1`L&7uI(}zY z*fEpEgDlL9Q{P&run^*#p5WpzwZKyD4-+ms9o|T%vI(GwSggYlU37mU&Gwi6hLJP3 z;El{Fjw7&kS|eG7~3g(3Tk%?G(r$D+()x7b=dX# z(XM|sj2f%c4r%$9@KAVfhSZMQS$KrmzG&Z8IA(*9FlK7O)v<~SQ@w$#J=Ft;R|(zu zWpwPlr#~tD4oxVyKp0F>QMBI(XWyo1v}2g>=>?E*K0&?-i{g@*x3KvD_KP*U?$oJW z8K0wmjU`YJ;1*u!E zxZ{L_Qrkq(gNO9oo?fmEo^s$EwV!1=fz}VUzvXmSt?YJ34;}rkX#DzBt6QIO zi-~$fU3+ZNeP=_YEu1jnA=@QsoSBV=IQPqk)@R>w+) zVx?jIp{jSUtial{>D1{_-EJP5O=At62pGU1Zf)BqaeOH|QI&o&@wiNw$3(lTALP)r za9IY`)F>~Qs2}e_?y2y_txrST>}p=|u#fj;yAZL7{>7SJI#Q>|$MjB9O^wel06fiC zoxz-{x?$nnOOoao%}6oFJ61-_(uwB?eB3UKNK9NCRy#Gj$dnbot=)LFHB;`;T^#TR zs?XM^J&Lz9t7vf3R8lD)YrZf9BP0tYNYvmcksY$M<0tJ}bLNkZ!o{^x?V$=MVp;};jPjuL;<j1*=SinJ_qzv%xw-lgT2E>OR9 znP$i8L#IL7|H}R@8Yf9`mCy^9*WDg+i**t7R(xO7`6E0;R#dzfE?fE-eYG0Lsu5V2 zV>Heox})(+(F>oo)m55+3Ih-|SmmF7Su#Ix9r|jb zVW)QJS}HMe8q7;_`WMEPJiJ_XnWFXbT;tE(l8sYzY6fqv&)UXC5PN1r!Mh*cnvnjr z>~0=y0ESKG9P|sp?*uIs-j&a`xfe5(3{ID)+w_i;fL2Yp3&U~2aRz|t(z^(B&1
zflGBko%KL-_=Q;C($p1WAYQR+!95>C;r7`>7j z4Nv%^|5nbgxDERL*teny6G4t@hT10hM#swVJahs(R7z2F!F+(%^77}ONs@J6n3=og zkXz89@<=4W7YSS9;$gij=NEj1TNTOSuQ<+*=ijG$zarWgS6G_I3?vf&B-%O_u}(-w zE8O1d`9E9C!sAA-0cS<7oj!E3=QuJp=xrxx9}t*D3|>p{*iTKKYa&eKg@aM13o?cu z5GT#4Dg#-cI57g$ja_lohCmRYW!OgIoU4vrHjvJ2-f?F#e8?onZ(*<_`6H)p_mE2J zzOU4_`&@eZFVLN5bwd&?of5&YqR|_CUrY6fxh(Ki1Qh_ zna-&1%p3$XM%4KYhk`kIjT67EHJG`=i?3Ln)?Lj@D9pHKzFEqHg|@V$4y6@NxACo+ ztXB@qC9Gab8MDsl{XE!ts7a&)zd^OehB#A4Q-d@+F3VN~g(K8JDV+;YP)8Dct*PLl z%BlOByea9Rrq4e-4mES+0uC1TxB zC@R;y)%Uf72;SZ{ocY!VtuuA)2wj5JsiBkEy5~v*PBm|l314|4picPZLqv$=Gfkl2QOE$K<~aPOc^-0Zpp;;}F_i`wIsM(n6Y0sdL1}W;Ixt=sd%V2`f$I z&p?3s6GM$al%Wg0<)tlqsB!dyiHCzt6{9*UqIj2p%Hku*Z{yTZ*7sQO6idGh{9nXD z`6CIk>Z4>Bz!^jhg19&l(rwl7Yv_*-wa2XqnN)M@nySPacaB~Hpg54__|Gb#qOFy; zC9w;8^EwFyL(?jz_XgD$2iV)kcL>X%Uo3!W-2llO0*m$u)x~YFw$dTm!QYtB>~biou{5;iHKEEzJXK#4c8TwpKjIHJ zX6{5Akc3V-XadMS{Kr6(^x4~CKF;K9<400SlTX|ZAO5n(7ix$rIiSGrf@-yBBU_vk zmGsCL`O#d*J1y!f51q2!A=9Z*MH9e<&oM}6OGWB$%7I5v6B;f|K&%=G-w7(2@H`zD zGtB2WymJ<-HW@Pob78}#4f7&hXnG-}F1LsB&hXu~6!r_rz+n59(h1NaC~i?6X}nDyjQi|H^lhsZo_IibkyR z9&sa_^%71P);YP4&z8)VM+@jyN+bv7Q)Y~ggF7?7x%L|z>e+Ji&sRO0$JC9||GOb7 zWXsI7a1LZyGB^JHL?V2;br0ER159XP=ay>GcPnr>fkE(@#;Nf47|M$C$O9qQQa0q5 z0CpyOo;gjz$NE{~fc@&btu>@_y|?l(N6JwweMrzE*prWVAVjNH-V;EF8v>@^xheN( zgHT(+7hHD-YaccLb@NG>8eqpU93cVvbe1fh8=yK4!V!v?KQTsP*hXSfF$(C%WnP@^ zV5|Z|+10;c;T`rIRa|1H+QTI{4clP@zu(KuRv}>2%H_;lIXTr)FYK|-N%Fzyin|*DlGLgT?$fB9d>fVIid2s)3`nVF|)jjk(GlpYsEa?lh?x#cXtdPi^PEW($4QZxmAdnU;&Iy^p6!7-*`_>!Ip(7{AKDsZ}`;bCUZ zyax|E%ZFzh&YpB#I<7~(lWSFgNoo7gh51L&iOiq}i2*dlv<-O#4RsGf(M{cx(b(lg zmL0IbK^XXPFd==^z(MCUop~9`t*vNhbS5a&Ia*zj&1uS~o*m$&;DpWrY^aPed!->Z zu`(pa_Z>sF_6%etKzOBJ3V)|Xs!J}K!8^Pw>?+aa$}r*OI_@mpsqlCdFm(U1{wT-D z8fG<6gan95Ze>KzWa7e^eUGjn+rbD4YIlL~g{dPLFjHGsvGybsSH?Qp`ZlRrrsZRD zi)iO80UoYAc}j)N%>dkdTXzM0T6Ynz; z=hk8M9g%EG={3nD;%YM=u62pWJfKFuU1rvTLiI zNEOqJeyaLR@AQwT+EJdufqHSwF)jXFbMlz(Hg|-8(iPO{k@=ZaX#_SQCs23)CB3xC z)C*7%ip2%OHO=#SRzv21CFD8pSeGk4%*ZjO~chFEG2-Ft!!BQ zG`l}RvbJc6IYVTZC}5whT(kk5Buh;U^E5#(J=b(rj~AfDNW;A+OxQT)2o2*+gd}M5 zr+5yX6ZbgWV{(o~@(Yh~EJ{X_=`P*@5didZoRtR8WQTFamTD#I+dMCk-p&8`$J^hv zs;gadhf%}^E$i4(*xpF6n9 zC-Y(Xmf+-aM+bYG?u=!z$lnqhaAJX@(9l6CLyujKq_=x`@-R>CIB^yr{jG3jD+8d+ z66`98C>>@U$-`H{qa$FrSAvy8Ix_R+5 zsb_txp}|Hdf|)P_z5>g}Ry98n@)+OvqqT=+&E3xPbYYldl35mPOiN$7MX+vf1do`U z&IkIbab#YRyf9!({`QK+eTiw^*A0-o4P7mMh-paXAzEL+Z207d=TVl|xJJ6irz#

&r|dH^VF9lY?ij)k7zKxr0+6I#dl{0Q{|KsLWQ+lF07O*dPMC)?IM_Xj~eIyVxWqj*10;HJk|?iq_MlWHOzNa(w*qdt|uC z(I{9k{+=@Y)#{h{xQprZxkpbtap(4dLwW`q?_NI5HwgtTkVTKtycs}yN+?Pk<*=OI zSr$>mKp}t7YCzJZJsBqYGkLaC9W014N|MNieDV?h0|J?xR^(gsY%y;HyU-z_`q_itS};f$ zH^vz8S-S5kqXWx7TL!?O`&=|)>HgyXCcsj@ue<<_sQ(X^bw$kOY529dGvq_-JU8Ri z;L;6gDnaJC>a$dgC^maqEN`Z%`1Y3fXQtLHHxjf>34s!7tX;BuzjJ$fgemf6bp4MD zJ-O7-{Mxu2(p0|KC-6pA)JtL^=e@~qw1vA!a7N^S=`JPyDR-s z7tRneC4}L8}hR5`IlI0w)e+qsQ^nbefSm|NZ>l@cW=lG zM%)}hsVN{CzbV(&kQ#WwQy{!mq(8N0^U+2@|ITNP?H4QklbU8Y=yOn#?ni}3?sQSZ zR!0vpC~@SaA4cp&7qOHya0XxzjZRm|rkmRfk$a^sD0$S<%OT}EGS{wt(+ zg0Or3`czoM#x4?mKZecYLN^k5;9PqugMuBz*=S>IUu6i@%DQ7r5AKv!^tYF&CoM_? z9T#t7tHIwp3FqQnKjty6Uooghqu54p)?`vLy49r?T?@ao5d2G60AV95v7K~6v#ADY zkA8rI0g=M>3Xt(Lc;KKzb*dCE_K~7dF<8z*9m_24y`@pJ z7ukba*zwW#4{@@lY(iw!rniGNd$Q{gZb>%>H~@VoScKO7b%>IyvzXwl2icIgDEo0` zst27=SCvj-8*)2)UAr5YfcQWzWbulffh%Bhds9_DF}-q2xZj)*%CUd4o!oSZzEba) z5Ub8E81YuZWaTIo?Cy@)R##2k5?lRIX&LH8&vROJ~4Tjmjn^9qvOU!}U z&qMp*Lg?GvN&yfhTnLs%O@rHX=~dfGP~!L~JJ~Sv&QlS3cYrF3(nwkkwzcG~xd(78 zB)j*ncDz^^bHCJ4CO!X4mo%D#SnH^spjK()ttm#S*%;Dr0gBR_pt~?yp$SF{f^k=n zX-+AF9jb}9{|8gw6kS=cwHw>$*tXTN?T+oFV>>&xwPV}1?R3;JI=0Q5bN;KhsxfO- zjWtHS)ckbv)Sm8YE7XMfHBq$1tiA6!Yn4nkEBmIqglqbf`KG^v_m=*(&KhZsE)_>X z5+jEPS8ZG2)TEycHE9G@;{fFcfJ?uNJKiVg1&l8x!ReT^4ThgOZMnz5MPvWsJQR_lQ@ zee*WIUi20GD7pQ#9h(*hL=3nAtr9{B*k`5fr;RYScdIoqC<^_mss%wxK;k$JwzJq} zCt3>>J|#tr?w}}$2b{JGeTB^iC-P1cn~y-G3I#MPUcUXa$TRHc`itQgBW>v|MTxZ` zo3r1#Z$rzhPge)H2x2c=a71WM-0QoA3=3m;Zk!oEYq)*v{vXR;R8ldm*3~ghK$~fR zBE{9#k01P{r#<)noagyZ03dhmQjU(57|_{p=u~%Z75CNuJVON+fF@UX1b+CvetR zHQT(&g$jb)=7O~2!OU2#YP!^+a~*+M04_8G!hm~VmU0~qb>q$gfaXi(Fcu@`p_MU* z!;to{bQ#L@49bSF%v%^Nch#9IP<4IS-nS8w{yAwfyeGx^E}a>F9yoyOGU8nc%=q{g zPo;oKPx|w+G@s6uWhG|~NeErEZupQ6R?mDYkElLeeX?EQUtw?IL#8OAkcsky6~KDW zOG9fasEOCW{%wE<1p7H0kIpVut=ZB{#Y?hH>_TW3yrgA8JKh^ajonXsH0)hpY7v)) z_Ei}T9o)W}ZAu}l%fX{RK&95@V|D<*tzTF=gGr6B;;Udg>&-L5t300SuKlj7NmDGn z6Mce;8LQc#z+0_i5lnm%*&40N?-1<+1mq8c(FUF4n&b0{$u7;c*oIRiW&M;CL!kq= z(F^)04KT4SK5qdat|m*#|5c&@8_mFfv`SopV~_wVLcWI%$PgLc-{~L!k)dFs@UZ?r zjR_|Y8w+qQ1RMi`i--Mx%$8zYpns`;+bVWTW1%FYp`eHnv(wCbGpUG0q9lR@#0#PQ zl)@^)D#lV)%D}KvDdGYZ_PP=liZ7X;jCswz%f5R5(tfI`YHmDAw=kdWx|^CDw6dfx zgAiU5UP^$8&I?CTg32o>Eu}yN0UHkX7gB~JBt*xCLi&I&*43f~vN}9+jL@a~>ll+N=8j%QE7RYCa;-7;egcspoK;!W% zzGjO$Er<;2DIL%a_T2Xlf@) zvoF7yrV)nW+x^7M*eAPiA|8V%NFd(W_;Ach%a{WmMwcK$Fn}>8QkYnY-CbcqVr+cV zFtYa=d;bzMa32>KFOYjbub*7SJHXBz<%K?s9+ftTLYQ#(K_Ly5E|mb=UoP(p z1t_SXLWYJ2eMqMuqDt@C1kj=vUn_0@3Jvu^WQi5LNkMjMw|lp%9B%`OQ5~*&PkN8& zY|5%?3kt}uOZNb8$tel?AO^z9R{nWq5Gcf8$;lv6y8!+7uk;ZVh__0Fo$Y#U7;kdM zUB&h`=U0t!LI92bRqe$N^gCO+02-AJwf~iUiVi4+zu|`H{#Vt*Hp!P@#Mi`aZ~1#K zyo!6+AfWTb@9m3-YYY=gOK<$^A+UVV)P*oC9QK3;N8qjyg^A# zg75JA2LJ;V3oP=h-^kK64GY5#_Y5Ij>7^-qX5h@zr(!~zk?7uv#To7}%vb+20Wn_-`sx>qhm4gjru1JYIQ2z&!*8 z(v*xO?0DxZT#4YtO{Ca)z||m)pO*p(^KGDm?@lLlqiTA&9cCWZ=G27e&;|*GD9O%WrysTk07Hd?VIM*E>Y4K%b3g zh-?Yb1>OB)H6^;|ex~XP`m$-P5(Ru`k#y z4ZgjCRo|od7w!qgln`JJd9)!^%bMz!G`6xw_0nxLVf>RcMsvrC2mR@A{?`eoHh^>w z5jCq-*JZo1k*=s{b>rV*%-?nQniqd zs3|E;-XIPDp-ZgfkjndMDj|qw*aRnA4*O3m=V6fguhz>k!tV>T0)9d9f34)q6hw00 zw5JG@;;6Qd*Ak;4aW>Wjwha7JrUQs4WdbHCv6$a_*^qE~CLzH@ zZ8JLPAxT}4FparIpXEgkUdEaYzb{Q1646KQ2|B1u3|oEvJw5DvRf#^r7a!&F{gKdJ zuQ#L*9s3NOQI6MNaFKjoKU}q}w4%8V->O6K%1dOyZ@iE^z?*7TD}k_r_#ARX35Q`x z=cZLBG*yky!bFdo8CYJr&j5~}nebikmj{K&Md0Gyo%q`mrq>_`bs=7Zp-)kqV?Ko( z3~vMv;O(9uQ&vtUhn_xsiF-21=z7EMezBq0c4ruv0dw`a-sCTJ&4ID3QxwL?baD1# zD;_?p3OF=v9mv{GZC>3^xP!pS`x${h&CavndEncPRNcA8Z-^vRv4C92w$vN48mCi1 zz2Hcy>283=oxCI8T3fHNAejQa_>B3v!h_IoB=l&Q>`xG3t-M&wJ%V)nmae5;1w@Cb zUEl?uiCV%!RVXcd{R9apIAO>KrP`$A*&QP|?eg(2uMqWC356wS7I=A3cm82}pu@VE z7sfrFCBIy%qf{cY31H2NiX&s9ton)2ioaBxUF^5ilK_KOr;#7qm1LV*mI$k)`^mbC zE#F_xhVIdlRt6xh{*b_RJ+3A#M%9o^TPsg50p!r)i;AAZMUksP1!DV@<(n#(R)ZO7 zQv>#zle<@}^v@$SmdIhE1WB8@h`3|#(#wPDzyvzzW=Ia&_H)#=&}bMg=vH@Bi|@PDSU+Cu{coe?FRR~3?;xw<@0 z8c8;5!ny{7ih#acc|;>1i-_HYtMk7($_xtIoVHHDj(l+{Bn2~k425KlclX_PWTjPe zdfX=eG%L>Ai(`MLS8#}Cb&g_n*qEhiaddZp` zMB6v~`}+EzWN#pP=R(Wid6stXU6+qrO=m%;hz(ee+OG&?rx2k!@zS$`pN~|=RdQ`o z-J+A+o`AmKL`!ed$Lg0p_;Iu^rOakaOO)y1o;Om(FcF{MENSjtH_D1f1X~(6&vgt} z2ckbmy2B!UcAX56s%Z`)x@gVRN~-leXZCI6o; z#XUeDmBK0~uh4!*`Oll3x5!8j*lp5iyw|_FD$E?Won4KWzW8bAameSa&JF}#{XPQQY&-9vNN_?qVA-1-y zLA-RZhg)?AkXpBd1>?-bVq#8g7J-!&<+J{L=-uz!ret zr#2la&dnPM&YFIz0OIw#Ln-EB>xdC_r- zGI>rMcXHDN>(k7er*o|~^z5P2-%+f?0yUF@&Vdapxrsn_jZ1i39%cf&*(PtBB=&Is8SY(~(@T+yfygmd8 z3AK{DS%n-FG(nOyGuL>@)=8@*XjeMXI|8zb^=n*g_`X(zX6UR>pmD;!!#(4BrE47B z0kZhib;i9f4k%v0Q5vyzuHbHwq)LcrL%B_IDj_8?{7AECVOl<*X0ujLG{HDFa z9e6#a(ydLi-~?$1IWP$aN?Y$NBRfFHdD)Jw6I6|-0tJBurlxJraOPuI6mt`m2+epe z<%x~_HqNQkI5~ryXqTxQ=R18H;D9kCuS`MuWn#&h^pyhh(}#fK$f@myJ}uyuRoJln z0ZbcDr*fC@N`_3-?B*Vi2>|fk#Msjdh_K&21FW6y+bBb?P4+m_vN4@bAS2l|IW+csu7nR*e@2Vp3>WzQ7-{hA zFyYEhLx*CXjP-?bUGn=mxr@pj@uZ;l!2H2A>~WMh>EXCb@NtmU?7hm#xz8M>F&Y*d zCyA+Hs<83oH)JW^WDKd)M>^EWd{0k6GC8m!*LE~+fNl!l` z8sqpljgI_|T=_S7rtyPeugp;VZbiwMYN6ku`rpPr3~$fJ+S-)I-p5D*QZonoK5Hrq znNXcuKe=d2%fql^&Cj#JwW*Tx+PUzlJHQ}9+7Jx+#vYexH>Ksd2qz-US?Q<*VuhY_RYgG>~JBKKZ(LIkFqhw|HveC{$Wy7#kQz#?Gc7RD`IDAPA0o3BM6Qu_~iDx zeXD14f)wy^IB0K;==*a%!iu99i9^F283`rogTfD&M1l0pT@GiIT+XF0&=6a)AMj!c zC+K94ZnPdj5U+>GW_y%|=B+$Wu^+5*d_G)% zvevuSaK_C~a@ZbCk3Ut0sOjvn7SOHIQn3Gm=%%ju3$$sc6s5WtOxV-iwUrtb#e z9tnHi$4J{A?G4Ur(U_&6Me6W27!hhu(llVSCZGw1+hJ6ODG3`~rp(tbGncZ88*5P> zB!bgmY)4RTR_)V;&uiMvDCU>IIq?O@PWk@H8CBxyiul(6l5BiF*=>8X8xAEBWd37( zR!^?u@b7is5tpvpS@^9e?Q8{s8=Nj6nOB@(d7-8@wUT`Gb|LuP(pi5&=H~14F9#Oi zN3&SKuqXNQW)7y|c(Un3A*q360h+S7ykhJ@l?OMmv|`q@KnWu5aImje$@RiBi`*mT zV;OjFO~Ry-4zaTC%^2g>eQ^GMw6X+%B8Oem&$k4IU8~5qUG=Rt`yCj0a$}jJQ@> zxB!Vb2L92OajTB$)&Yy>!dxJr#&#aou06L_0H!)lxk9PyPZGycH(|1zPsrsPOgZyW zX0o&7Uxi1fHtnyy5FJ%vQ?NI2{qjZkSc9mz(s=0ai~ri+UOtZ|sSPsc{{bY*6?zXx zN#1>xH!GE6%9YQ%w0xc{O2Nx9AJoK|Xt+Z`{-6j!x7UuAS>Ei3a(M&f#Ae9CX}|Xc zmTj!*R|3l;H=4q|Kjw^YCSZczP5o!2G!L_JTbL`>HEEZB+f)6DE;ANVjGatbq9#a+ zfdR`GbHoVBq5aYuO2A6EG+idXnL0V*HxRY?DRfc6b5RBr;VPLCa`Ct*bdDLSN>jjJ z>a%-!onH&;4z>3JCo2eGtVF!!99y5=M&rHe2BysC%zBD%=BZo9YJbulG#lwHQ%$0W zJ9GNEe^c(DJBf?Zo@s`uL}u7vPc9rrTTChrFv;L{*>kX8_CRZmZ0D5FH&5F?8k|g5 z&WBp2R0BH(G`erYX!%O3dGt~V808HvdzQ?Vb8*`$URG7~R&4?2aPxt?{i`gGhI}nu zuF(!UGDE!){80@q2P!QkO82)6`RY3Dlz$P?rELg4&YN-f&UnlTD;Eu=INc4LWhmoN z6zjb6S)x;m)AmB^as9q#t0I*0=rxmwnLY!}G0R;pqmbme#c}v>t z(dI;eu&%hghv&_P337x?&7@&1@;X0pp^uk9is3~VarAK&@|zV$20rxQTytF0=6UgE zxzzNfNP;R=Sf|w|Fs~flRl5u{;;6$FKFNZ8*i1C11`+Hpj)_I&C2gUw=>7AON`UAyJA zYo7zbKPjvLHC%{<2_qxs=tCQKsYf|25t-ueTGy12;tyz{ zR{csh{b{*-HS@}nnBV<`&Z27X-Pc#HjSCE2f|#-Z)aiu{5>avUJ+)#i44>q|?H6E0 zw}M^&@;2n`X6hGr>brulebW0uV@GJa;p&>S)o5!UxOxQn3IC?Fjpx6V;Vz<8wI?;w zF}p$OeWj!Nm*+07lNDiP)mk<%d1yuhQ6Rd{d+p!_Ig0`jb`ztS7QH!!WX3^7I06hs zb;39xPFk0d>6WSM`t_SFt`qDkRx+8ns2tIHjSh`U&-xfl)CjLRP&;+XD5lHVGF#-p z&e^ij)r@w#_gar53)R`EBpKk~-jS}vR+ifH*M4^ke}z9d?t#!(Bd^~&d)!V9*eT#w zg;qKi<`zv`A^LmtwZcRx5$KO+Wv>bD|eLwH(heJfw zy>4%226ZyYJx$%wrp>QOO)c9-!AN+Te_ganH8VhQor%)YX*D?>krjr~RrI?dX>iBf*UsKfJu-D$#i3xA+DiRr zbp&a)w7jUQPSxu4NafQLvY5R{+&Khbex?f^+bIj;H&kv+4MwM4@k8;P88W~dVC-J) z78SkD~ZlUGH>X6%m5T3St zmy@$zpm&iCC+N^fL&%veBzYPy|45W>K{gpWoxra*-xo~mWbqn$ipzhyNc zwt>4Dk=tzLH@){o0yT$PmK+QrGtxoY(2-Y7ptS2IWGO_j~cn&t?p|-=BY`AKd*`mIE>SF(FJKMKo1G3 zq%o5(zlggsG2YFv4lt-2ZR)Q81X<{XH?dRYw4W9Di{I{<_Y4tLCnfIvCgFqB#kf=j zwHNOT^#Y1wi%Ix9JJ9;Whl7ZEditDKUhkrAZ1Mq!mq<%do4$x)8h+DL15`~V?bG-p z3>-B_;{3UR2qve%^Xk8IF<5{1F%(6)zyh#lvP{ScvOR`U(&=M3jd^7OGaJKOa+>B~ zfma<@fl};rWU5}Vh2=8z-Z2y+>8WeBEhe*#F=Y%!gvaq! z-v2bu*0MiZz=TAAihO)jtaC;uDdhdeY*=b#q4N$72ee$wQ*1a>*q%Yh(0JKM6F{ki z{kP$NlTWJ*wY&Ehx3U{^a`6uLQBilJEZ|~}Tq|K&49=NJnxOB1is*^#1#}VE3-`*s zyd~!4_SdV%b=p(;pj=rvm3I5ScgJYY zWl?tIeCjVdlD=~TemK=hFP2NU`~1Nn9L$(5k9wiUO)ceRFR+J|8Y9m9I4r!%li^-p zfVHpoVsy1%%nQsYRCM-v80wd-zO-_Lm-@<@o(Hj&r_l%)uy|cTz~!*D_2Q^i8D6)V zQ5I)4b@-%gH>v60!T=z9UCU*8}rLCYG{zm)INE#MSiGG0yra=tood$ZVlyXp_dl zBoW!r@jB*+y#>E?k!Kz&yplMc6|G zsE@xZV?W(=23U)+tqL!om`I0{U6MEup9ePVwdn1Z&lGM~?w@Fm(9Euel#9sSHS2I;1=Ih* zxMSRD3v2`(HD_$VGNe~TZBz(tREmc3K6$hN&dI$yhFvZGBD>K4&E~h)!9iX7br5zM zk{X{IO2ai)Ka}UgepBlg>yvHepnh_36FL!{LiJd%*vy(D_~N4S{v!-|+%sEeS>(1x zNb`O=$YSST!pvK-$!(5R*MXgFk3^uYm;S}F3g-8X|L5BFzYQzS0;{b$@W8nKy9q|CJQwt!El`FH>PIVOF*GLr|M8ww>W99v1*W`0{XqE+tNQ;ie^wS? z05Sw71Q!bz%YVZ!V_YDAC0{)<(nv{3FDrq`AZ*4Y7rC%aAuuyf^-cXD2`6UAmT?aY zmI;HAsw|`4B-tbeL_5!XxnJs?ZE;xp-gVFVJ^DTQZ6Cd~t=uENq!3Gam*JuX{QFiM z%7FiJL6|NE5y}e(L7|o#8a9I%?8^rjkusuF!y$^27JXp>gdrgzjcQobP!2OngTz3Z z_;x@@Q2pcEn&N}mQD9J@z=}V`#R`f*D~J`bA@ayT{)mv~S&}kh7G)yC9Ucd_Zyg%c zf)Fw@2}#Mw1YWuS6dK2j3=9y4@kfH$gK+KMLPZD!M@TTl#@M@yL*r{uz>dyr%F6=+ zfgod;4E|yOo{76#&@Mx4sy2z+oL zSPdvQ0)lv6EgK&S3N@TVH&36EDqh|Zl=wS|^%u!C=w}Y1kd)Bz_Q9w57cGS7y8=`| zzmx_S$^dkz0|a;9j*@>)MhnOY-XJsx32|>YF*h#2GaLC9qJM|@&3fQ!Mi_*H0vzPn ztFL!2zn?k{4~oG*v>y;d{aiWqK(j3=^v!XoQG$SyzpoTj!-j+F5pW~w7vwq%6!Ic@ z|765Bgb5$4B_SrX3mkz->-?s2JTM*LlD;Az!VH59`K#>i8Vi9e5`a8F@xp%6+VAf{ zeGmXJULg$>QFe|(UBJ0@kwFptd4y`e!r%W8D^i05-TFg4e_ZtmYKV!5fWRO`gRY1Y z7{L^L{pt^Wt`w(Z@sbI z^qNdeNZI=n_10nd8_39^5tEbCBE`qYgFpb@8>0na9$OK208SjJNK*cbxpuX2Ck6!G zpO_%C2F}-DpIWk5p4Hma{zrnOtzeO${maPtU-so+3NO7h?;0szz_G7R5)NeUclWd- z_xCTDAYCJe+}>d8IWNN+#Jr%c7SP?^@{&J-B})M!hJ-a=1WQByH8FX?4qZRkdgB3Q zAWHYhNTQYfUHK@!*ih&9lB@uSJp=WiK@=rla@@zWb*F4dh)<34H3LK7*%kDwHn3Op zkM-NT^Y$AIGurK>#enfH2nrnJEeV+Zq_n0S!VZ;-trX1STWca{AQ%k!;ZH~heiM*B z&pIEc`Je55pk`d&cYn8EWb#pl3RJ<|$!NXO&0y{9hML+yioqiFt? z67w8TQL1hV{iuVAfr>*WTK`>El*{~ZjdXX)dvKW9oznR|s|l8Madb1~f;hz`sQfio zc5p=Cg>CD@k@ehm>4neZr#Km2kZZ5IuxFTH>z7OK0_p^qFe`VgRFMR|k_})Kk2z^o zYgfyOD#Nb>Cgm7St6ZD0%|^6VQ_2=Z7gStw--aeS4ZFCi1u=`-nRA9`%D)}clb&LN zb$U>0!DAnd3hsw@)dre8;Iz>YCsW}19DoJCPM=4gi2>|FD=8XXu2ehfdJn)iyqZ)2)fz4I2cis|4jG76mArn?(>DunAlxVWdzV&?nmZ?%opBo-Vn&=tY#Dow$U}L67s1< zQ|C?_3Zn$xLRiDjuqxQp-BmWlJ|^z60pd#`l|~;5u366at?Tt4w~WxXub=EX97%hM z41!$v_O8*)N&v4uGk0qjGr~}dUuRIBJhM!o?G1`Hu!IrRi#j6>rgL?0 zb;?BBN(k3`Xia2AUCIft77hAS`v@j`%O4`^AzG|%ccre^z>4NwQBhe2jyp9N4B+{A z|7eZuxUPyMD3Ozco;Bn)E<)%*571U!tC>sdxV{Y~3!vh~pW|WhaRO34>f3`<9qXfc1U>G$d*>B>QkvP zKMdVsOVP+XE99q8_`X2 zq1m^%0=s8uq76~ijE2CVk-6>M&!%yP#mvx!5>R2IkZVAI|1i)7t8N^&Os+mE5wS1^ z{FA@;_Sc#|X4yu3 zvd^S+V~XQ!v1qomdptl#j(&=TM?r0v@+fKPNo4KcZ=|tN$&|l87|b`jWbxBtepQBG zX#wIpHSTZ7EPV%Miz>Ry*g#rQ5Cq|=8FuTE-+UtX{nX`&w*JIeWj8-vc!A8By``6% z(BCZ{cXBM8Xu37P^XIlRYk(6R9}cl%-`sN~I(z$uKS=05&jTr*4%Y6o$_56>E3@}B zY^EVWLWy|nD4l1)AGr2!k8mfZ|Z)_*t7jZ>IxAIWm}iu?y0MR15* zEyIeUwDY>CjF~8(vX9(vJh&aFaVf} zLu5XB9N}r(*QVy4D4e16bR&QuKA*3|qG1)kw#m$&^{PVL{!|^y^h(a1Q`gDA20q%P z&l z{$!zyW2*h;b*h^`6*#PP;>1a4Ft)&3XXP4?T7IJGP5PxFi{GU%3$VB>gq%60MgjzU zaT{0%X&~99soMdxi%ip)rLx^DkfTePEd0tr*~(qGt%&k7^pd>K&s*rs{sK~x`TSaS zK!f){n)8}Q(lJin@^HwCSA=Q+11sszb#jySgCme{pkB4H;l6{^D4%9bWP3X&t!$D| zmtI;|J92Oe{O4WqkCfH5bXtjBC^6FJsQq%gtSk)a$&9CO`%}lAi?yycbLOU^waMyD z&$dIB&*0&-6Y>@!!7OeCxqv)9t9AXyv39l;mxf3q;pPb8Us8TMQ%1U1r4By%;Aoh! z>rv;kq#v)_CT^S@1Q)PM3FEC!GLp2qvj!q>8Fkr%?QdP3nbAFQm9S@U{l*;gGIIm{kXd({zp>6JsKJ#(7HPJqUrLMTvTM^B2s zmpda9U9B4EQ>8^2c(vzw7OBM$TElPrqQg0}SMy}ZjdcPytPWQm1i^hU{E*zcQfVbY zIwOY2&$KBmoBx%Mp>2Phs?5g1dJ&e#=@S2#Ky07AV1!*s_Bc#A z)Tg|&EE~N@L8nM7RNUA!O?pLWQ@*QuUteM-(R9>we?r4sRLlalNS7cc|Fz7nW~eN()*9`{tiR(4 z?}tpf@Y1nR+W{CwJE0`@n!h8Y>I^nroLXsCt=uFTG&SmFct#Z}4l?b&rxdWQ%zQR_ zmLyAwX;4b+h6WA3z5e7`cae|3g7Aw?K_h#c^oUQxydrEi@#LR0Qlh$PXQ}^nwO$PQGyIII|oq zP*H{|js`Ws1U9DRBsjHyYb8<9T>W^OIVSyla*dJ)qjKy-e?d*l3G(GfzRB^QI|7CR z6$d&4sG9ytp$@1bJhqNB{aKPUzLR76t+x7>v&lgG>GONSXpI(F#xr(Z!;v5G*OqhC z)@7ianhx;!8Mk|CMoTJ!25F4Zbl1tNU?~Y5$WqLexMcUH?xx(L;b-1{By{{E8$$GH zhJ>GgyS27B6$klF9BIL#@FsAT_)>2+X66Vh^WkDaHYcZFKCWjEJ>u7mpI}`Qf2QP&iCqT;2PJYE!4@Ubq!VTD56Blk!C5KTz-JHF@`I+ zoEQYu6Lv^jy-ThA;J9s2OXQ(rhfn}ojCgVKmtGII|1k6X)ducu(i5TZySK7#^Y6KR zMX_Z^U-BJAZ5MN#F2A~2Sd~96RsCpMigb3h!B&!{Tg<@h=Y_0Wr{6`24iukfY1WLX z`i2y7@^zIJI4Li4+N3<)ToJ*Sn2k>-VU`%edQX2K=IfgGZL~)

ffl18p@GPG^qzigM#6x&8L&E%FU+55d@}_Hn2HD`)(A}rrr6Vu(0rY z>Gj{;2$&Y~PZq~5Xc&wop}47-{RL!zg=hv*X0LJ#YiHV4i~7e)i19Czlk@9UMPY^k zyh-%(m$7(mL@`QbE`xK6&!R%T| zyJ~1Z@Cd6F#-GhQy|ZLINDdgnv-5dYUDDGt&JVy10#t}iT`0NMr#I*i(h{NoqWjk_ zq2;(w6XIrwQDX>Ot2LprEI-QQH7%XiLsD zg*fTrPJFlgU>q5l7(E3;9@RfKe+sY6Z#`EahEEJ$EPx$2osNvI^OU22HR|5AYgNEW zGt&$Qcu$Na=|_)@6WeN4B!A07v^LKDVX`&SQ{v>GMN=&p*jPwN9i~6kJcHfu`P^

K`VE)OF_B)rTQ$`&(-_Ol$F)I2Wr`ng&+?q;6&cHx)+cDvH)YGE;d z>g{F{gY(I-;K=1mn@vZV#lt2iqnLGZ%opcHQflX5*r$fWH17WZC zIB&*azsb!mtms8kNIb{@ux#`%@bB@iM{v23=Gk8BwV&ipv9X6U z$|#!F^I}r+rKzX@X|jWY$1PAZ5Pi!iu-GBlkb8VBsMeR+CIo3Te8n7gWdoy)!`L>u zGpu0?St?$q(B!o64NAPISll5Ft*2b{r}9x1#pC+72(35TEW9JY@anpr-qzyIOOYcQ zuLJLf3I7f$`<<4XO1b`YTp4$yf|YB+us3q4(nN{G%3nf8EE80uf%RYPv00%-eyIuu zGy*+NHy0?L^sYI`QJ>mMew2Wsv)gIvVwO9IBlp}n-3Ks7ahj!&Hax`L%l!{~nw#!zZ;$2< z!6yj&4nJDhBjs0N=6W}$R~7xFuDM+`kx#vdg&`_p-$}8+#$2xaY7)KgCV1$U{`!Ky z%kMR-y*L?@cSh^Ys#|zTZ}FlGsQ2N@*-^#*V}ca1ax)J=kJ84ZXU!0N&VH)l<_{S? zyrl6TDH;-;1@@laD5G`R45bbgM8UqdFy+KV)6a-zQ>wHy_>SnR2T}6KjSis_aRj-= zD9Xx46zIEdo@COIcw7`2W<)*{(`cVLJQyLxJ-<1y%w;4|dCRic9<`nW`*OS%&e5_VPk8jopTC zB@e8`VspXs6oSe-DCyZY2K4680#$w+dEKor4YWT$QTZz5yw#^|yeHNw!|KJB4&HP7 z;qfdZZ!3+{+Kw0Uw~3lW;~-p6q*jLKW*|2BySj7+-W6Je{Pv=zOvm|yTs(w&#p zwS3|>^Yd(Ck>E((pU*^%u`axipoFXZNF5;F$Q?v30}A#rd6>6@dcO-gMud0S_SO|EV_5 z_PSmR_-5a4F`Z=)l5eV@u~+EFJT^ct@3pDs%j~eKVENE-kL>J`u4#_rC&)fXA^};< zm`r^(b5a(ZixZf->bs;T7Gd^Yn7?`|2ybbFlj=_{WU&HMGLAFD;vz*z6ccUZJeqK< z`cGDzVKjOKQEHccZy=B)Mr8SM|63ch?4{LI`ng)*Z&Q&XPD2KqoO z?mB}ZH!zJ$<}%{IIThI%o8nx+w`ATb7J!HembJ02K5mD1x4cc}ymtyIP(*iO-xUpG zw3I#e!wJkE8F0-?5z&&hun^ZqSGPeW|g^dru zz&bR-YE*i1v`oJCtfeIxEIgeqp zqTNC^jQoS!LI2K<`Bbv@If6sL>oFrK3X$lzrqGh7eMggSlbyaXMAa+_LKnXpA|3b1 zSFLd@9T{W3lS$&^34Iv|lM)8Ebqo!t|0QMH*2>RGbu`gv7n$eeQ$*by+r=DLqbS#-Ukg+Cdp7l22Gy$G2j(Rdp1Hk5B4pPid;dM#RPy@CF=7+|OU4j? z$~*k7&U(U z_zX*!`GmzJXq3qhMbT86{fI?EV|NNMfRh$SgiCTHq1Zp#%l9;lVN-c;Olo%QSca^! zW+0RkYhi_ICCu42(@q#~Bt-(ilN3-~B-$x#lz^whiUEpDnzWo2K-V&3N$U|%_}uu$ z|BHLrK=FQU;e?g-rwx>U$PWV-9X5`2oI%3O;huj;4_;6#T}1Z*`T;U*8Uo zmDogTPLEGi&!Vx@2H3nm<{`427dBnKy@lodnd69(8Saj}Y%cJhA7s$KUwJS*t?%w{ zOk?)dR0!xodwah(;###e;d!hIdqR)f`0OUWWUQyUvvd*xPl4d>EZkA7iV3fH5}aN% zuhg)*IMwH`ZrjC3@}I6xe}8=)X~a3Fo*kLJTF&ehhpl?$n%uR??N-nv(Iu|6XQgjt zrlYwIqx0w726hwt5Vtu1L_}(?lc2o^;Qc-S=pO;B3AU{V6p2enRm0o z{n{tGCbpg13wOkf(yUAV(Z2v0*J{hYoY7ieWA~2DCFfEdYNuTD=ZC5A7)~rb%Rxef zq2{JcqDMGGWa}<7Kn~BFy#Zy(5&E6)T194F)2kcx>Vg6`!A+r;M3>TKuG_RSzr`+2 zt|-;4fOBqPT#=Q)Jbykhoc|y0p!mzbn(z-}h)-&86zO1>S`$b}g# zA)Mykh|z~ktDdU12PiqLa6+HNFM5=J;j!5dNVk=GMu!{}t1BP)@AntO3$GC?nW*1= ziz$NIXNSjCTk>6wA!F`iq8PyoKupNr)amgWrqg+|U9(+I@{XiLtR{!fi7SL{*J!-# zf;X@xacz%3fEQZN0Ms`&eghGJ7xBj6lnYlr{CA0Eriy0)?1StxVe7&3e%;>pQrp z-qy`Tpq@0qR3_O`Fud3|G=G1 z;IQPhz+w6CV7zk%SW<9sb{@7?R2En^a3H(ie|ohHu+6QA9I&=v-xZD|VE}q-JQpkt z81Q)H`;%!N*yH~cxPbkFuo$gVe6Y7*z-;jE;hG5S{eLPvt#jhA(BQx|23QOru^g;; zD~mMj8W<3n1s1dQdrk-r6k~_QY8AADB?1R}bA7KOl7nq*rEvN_64=4>Jy-C7?PxW0 z{m(8feBWDpd%*UB1AG5-;CHDX95^cSKdahXH~e8?z=5;k-=*guSlaKuVPyphE5cGC zv;SX(H8Cd(Gv|LX)(c%7=k>Nkzo*)rJ)UOF01!}=4Vml;?g+LSUt{lG@)T(865H$q zia2t`#FvLRS1z%Dj+KVJ#!h5aOi|($OZHbJ6e-jR`q{{x23q%L237)k+8VJCZP60C z|5eym1=SfW>mtG3-GeOLS!i%~*FbQ0w;y*F4#6!*a0?DWgS!NGcY*}xWbb?KeZ5aT z-7{75&|TG2^L6*P%Xu!uXaSMW%rrF6WYk2Q3?Dx-v|2eV*KH~8lBU({iNx0cplmGd zA&`d|I;Nwtapo)HQ&L827up|CrS9+&dYX(C5948dmtSF6g0Jw7w1|X!Fjalr;q;{P zG~im@6w+YQMDeomh>uWA*tnqyM3_)v?KJJ{&4Md?2noq>4eU4#puQl98XTdzAC?#J zVBq8R?Bo5|Dr7F-;mi;`+2=Y0kW|zx$$I(?hWv=x zQuz~uFI2HD?HWgYD}pZDCXo==rW_z*%!ClzrF#6LnT^7-P+fO84qek+j;$uC4tO^) zG`Y!CXfEh5cmq{fPou=dYMF$-P!|(g+d>$t0-W~pKdwW*szJx%@Z=8wyHDM^3tF)O zTIOnEN)A&6Rz6SIxvTA1o24jMWjqznHDP~S$esaKDrI-{v`^o)=@GEe7SGqt}a zlUuTL0Wmj*M3Htd$*}{L`fg!L(A1xV+IqTaKTv2CXA1Wy1{~&WiZx$SL z3T5jiaqwX>GiRViPRxy9xS?DdUlGd|I)^>nvQZpE)rr!!w@+wM(I(hwcWGg@j zkvVyVBQArzyJ+CAU0BW(ASrL8VtnYmJoe_bFZHHide=Tod76@5cspiGCR z&w45%tSFjv!dDi$U8j9GvDZMJn}=Ff@$VV`UfS>LTbL)Lb9tpgUiw={2d6-L;_q8s zw| zps~4tahG8WNS6)dA0z{;H8O{Hdrz*q##8h5>_QIn$52mvWQqn~)IX}uRkm-gsgRXf zH8K>=L>D{)1qC;Hf$NXIf4=oc=wj5eQi4w0CUupG6li*gE96xkvN<-O|5cXM*4Eha zq*Afgei1l%7Bxbf-!nzME9teo`_T@8{rsgzYqD48>*vH za5$}7lFDUn37|&#EYy>{ilBR2q@2DkQXk%;Vt#xs;ZC?CDZ5>C3C`!{>5R*o9F|&0 zSuP+4^2Et5ERL)syra~FJMwzM71>Tk@TC9oY;gBq^VKZSn_1sFujHfFTv*LkdfrqA z`>I%_5W77!9#8$kA@q_sxlZ||SpV=YT`dOKRZcaq!BKnX1q{z$mrW7u7($vtZ*kk<0<% z)V+4#QoBz3FDhq2wMlHJ4t1^E7p^;8HCI>BtS~$%FCXJ66JDAsl!xobzss?qM9d(Z zR6$!Op)8983M<f#lH2ba=NEW)kfO&x>Uw6ayo7bSjo@gQglM{Kz7TY9Spl_7D5H?IcIHO8=O zc|~75LrD1Z$~3yP4ma}Q6|zF6qMVHc6wezkZwn2={u!;6`~bc#nHOgEW1g<0yo8$o4QIPE?1y4$-pBNr7ucKhR601-Kx+x)b(U0eEIJ^yIp=@Ky8 zK_2H<%)#uzDNK}K*`1n`9#c)cyDaH#9xkFZjXNkd%AULxp{8go(e6x<|S%=XmE|5hPuC%^< zYbgwRsyFdr92$`^e3P#X7wcGvK2Z(gIEI8!fltqINMoTqJBRi$g$z4@=S_nhLxcD+ zLrcng7Qfc#TUaefXO>?M#R)!nZKZ7ZvYF*vqMDjNNp(Iaj4f{JNZVA+FUH5o^=ZlS zW`INCCFm`Dl>H;MEuBXs-|5L}DzTu4LraI}2C~P?@(UeV$8+i2kJSkCOb^i7+IN%i zuYZGod~mK~mLrZ(_->XDxR*35eRm7$cxVY|VM%nY@AeScD58b_EabPy4*DHJys~d0 zTTjC(VRGOya?E~On+nsb!DIhLolbJpn)VFU^IYh9&U;>Auf0y8KT!Y^|3NX%jAcWB zwQEc4WJdxHF+SG-hJ&9fXxWT2jSvUNR^cR-%SROIPn;R{fXApRV855AHMr!NOfnW` zvGhVeX8<}}WBV$lo~7+!IObg;=OgbiU0{By3J=)PxJ0miHnL zik~9zg?u(?onJ}4Uy~#q{X*$4{ zfhDrzI3#-(^n+jryO@Drq@FKINuPQ-UWMPQd}$mm++=Fkhz zs}npzi~mT;?KkcirY*HQb zf3<;&u34G@NE*16HZpQ}^qHgcI$_dwbgt9bT9P8S)RoV}%Wm;0#;vLMJ`PjzCs?Mr zMYsw@8etU3;7e@U@mTS8Zv}Vl_0sN(?s$J8#tRsyP13?usj6M7BC;EyiE7)oZ%&J! zuKCu93g&-PC82asKG1Wi6CH+HhHUtn5jGyS%wsD8TZ(APD;lcTXVU4WAOgA(21D$k zmPf7@bD9B=EQNyg3Q;fv4-yTb%Pcc{t4iznmfxVi1GKXGyro5BD(Fd+IMfYR_-#$e z+gurw@qAB}_48od*H$_Izn#cfliJKkJBY2Ltw+_DVPYTx!+O&4+(M{Ol_zGKU&B=E z4LR2U(N}q773LvvRGfQe=cSQq7=$`@uj$}44mPlCY~$kmew&y|5_~XvB89Uh!LMdXE5!yNQ)3SMdUkT>5X2 zJsuv3hEyjL!?w#B&hyeoG;~&hLkbuqd(Dr4C)w%J-)&twEO^XuC?Z`umuGRf3=8F1 z%+UaKyMzp+&T|kPwGRU)j?f3`LiVReaJI>;7#ReM;uqCuhkJ3D#wXpR@^3 zGO$91+Cpj|7x^%4y~H0@1}uE(Gh!sO(6t~MG>E^s#nQWE-p&W)StZlFrm1VUf=GJc zkBeb&Lh8rFbgGg#b`be81^YI%8 z9c|&NC@Rs$VZjd}aWDMf&C>g2%WRt($(ykCFihRohGVKYNmRnTg6gG20MNoBz4oo9 zavh(MI>bxQtA(hyNXgvKW}So`vNHiJ)YVH|!32aPFtz=1T#|vu2HRQlCFTMQnHXm( zt*MgHeQjSRFlKJ#gq2Dine{d;Ht;TCFfUygW0@sm^T+r(@z3CVTA(d4GIer#G;@Y+ zFPJ=Ouc(}T`_oCm6$0qz%HJkP>$&lj|JFTI%6gj@F&$dOa0$6J+bgp$e4hqx=7DpJ zfC=Y%MbDYQ3PqQ6ys605xpuQ^R*sZpGltQSml$ROh0MlcSsd4~4a!5Xc>R>i*$@%? zee>B5NhiHXg#kg~%SwLJbXq$EnUk0u!n;|+7E(L@r=85R_zkg7U_fDvb+anzsJGqb zruFOtB!q_83a)u)pmNlQ83*_QK|gm<*j@=>Af+{*kP`Y1rr6Q7CL1zU$tg<32+ZFH z=_ObQxEw(|h6ZpKR7b~%BV%vavj!h(@w?F=tx?VOcs_ za_~y0AoI~RO|cyV1r1+@K}8z{ytzLyqVhJ;j3AqAMS<3GOLK44LBOzC9lJc>R<<_XCX6i$CcP%ncstT(7@H%B)$tifBgD5%JIXXt2H%#b_WG5;N1cdkLO z$()69UznmOEO4c*e1(q_`n$|h_^+4Oc(YX}0$M9|ekZTJJsd_dNn7*lbe256LEf}I zl1k5ssmjEVox)TY4dD6-Jyv+*q4*}LvnEx{9F7dxm4rhj{*F^zH>}bJgh+7dOvu~o zj#Idrie;rfD+~G{oYB4Ipv0HYSWDl zUY(_hZU$vaEs4`RT7eah&MD=}Qnw*C$MJT!q9svjjmOSR|BkiUXztb0x5j&-j|5FC z-SLDCe}{9tC|$fHRy*S?=B6X8^O^Nvjy8btTk98P6yL`vs<|D!9Ok**A$GlP7* z6d%)Fn1i$3XJ85cKu&J!s+y*sCXoIt&ga*Q+MlUG$qYLr6hggJX?2xvL)B!NS#S^r zc!E9k1eSEZ z{kRkrIO!dxHKfC87?tY!geGk`1GZWfALH2PA9a878U#j(+PZ^wyv7S_N2q88ny#yo zLSuKMNq4*rjCUq3hKgT*Q@o!ZXRn*(?he+84C#S_m-aQ(6P!CQlC~?+^ZuaQXP47! z2yCKyAppZ=cMEL(?l_9qV87Yn4wyWJ^(i-HsK2rS9a1(M^|1ian%*DJ2s zfAto_DSJIA^8FfpxbQ1_%x6J3M=c;*(La$C1msfCzb$)P#(fU9-7R967K%S!`8Vv|+IfzYKfmY&w&4Vxl=Aj0PY97Q zd*WOxaGiO;N)an~eLeAZ=ylmMbVxD*KU7SqUSt&0HV1Gi(k4~>ygfO z1?s0k1jp1c`STlOIr){)lcGD^B9me&wGnp-gZrp~9jyhu2`NwR)3uT1yFJ1X_?znJ zrFw_2+nU-LIv1>+0&nNb*)6rbm9PE%Xybk1r84eD6MpTr`#8CC?Zs}lOPphba~6v< zuFQR`#t!Dc$(2KXTJ1mKVz><1eXLTx24Ftq9ze+H@AzHMgeufC==XCwr)GM~4k>z4 z_@1+G5%Qx3FkOz%@FkcXIS0Z(CCkI9Wp!OGV8qQl{EVBOSMP(s6+XOO3!OQ65F;eGMT@ z+dSRr4h`O`HoD)uov%nV$eXil%la+p``s3NylT8<4@VPF##|mPX^}>5+cnx!V*P(U z-yhG>E51aihT7$;^(*{UC3fUPwyjyd#*RL^FjUf$1Kh3b2tuA;u8$T(O8-FrYERpm zMYlVIRF=D3_}p=YjA=xA<@fl`0&QE1U32Lhwm1|-U1B6L``znAo0z)=%r2QVF*9V{ zw2|#1s%S2$?=ke$ z_dW9j&+9f^2&h*^XLT|dld7r(l1qkIKAy(J!!Q3Ly?Z<^+nwr=Oy9I620$;qyW^`u zb>5h8Xvf?|S&5K3{HnR&aCRLy!i$|~2>PbbyCl+ZBpP=A;~QAVEbtfB4>y)dk~K`L z`}5l=9~-WX@rV&2KP7j+T;&!q&DTGk@2_RX*4?Y&nYe*n($9-(hZsdDC7a)7_jA*^ zWPYA?i3G~0Iw5jMk(eUC1MJD`q$Z9XZeO2_+MRUiGrj|}C_~wO(Ol&2n#|XIfA4#N zmz*h<38N3sd;8fLwIULA2>HAC+4F%4vZEt_6_u8Ji0e1gBN3vypa%AyC zP2Q4!vw~~yD4_@voOIfoTrhWDb}ws0tiLpVe{%hcHrH-sf8ws2FFdn%)Ly)e!!fdF zt?`eCGFKSceOZXVafW6cwCMl#vbW!hHP>z9u_KCf&N#D(-Aqg&?!*^Ka)x4E0stkzNm3vo0FG$mH7HM5*H@vUUh|a3NEzf8QM8>jwLJyl9qw2G)6RB3JcCT1HV8YX?MF zEnr_4z&~-f5k6tf^M7Tl{x<8t0@<-M^S<&*>Mt|?batU0x6*np6PVNdYN+-$-fY}G zSh|_g`lH+A*+F5`0;#F$FyQdpsVi2(k)RkJ~*zwoCOyIUf@kmPmWm8+8dG$$jn`|J0ELB6-Y@DCXLQGpPi z+S7-gO%4%sWh(-Wx61*-duM;{zISnVdE87t)uvl4+f3?4#iUDhOrBcHMa-lw%Bb67TG?7`^GmVm*^G3Hm={hdsYq%gW`KclR4Zm}9i;!5SFfn9J44f= zmaxVs_eSp9bj$xGn)+waviV~4l7HVf+q9xT7q~3vvurp`6-#6x?();bgXmuL);5B^ znUr2Ak?z!urfCHy7av&8vAmEt?Ux8cj&h$Z0%`ck?r{Ut^86l7E-atak^RK&0+Q!w*osyfA{r@(4QF4H}`Tq+&r0K3E&q-qj{M8vZ;|0|n zh-gTu{a_CM*wdhwPE9o!T8B_;>oAgJ&ieS|ZuikgI(hKVPQ88re?WT~4>n%R$1oS= z@yKfwTD7{Z6?H+ujZJfj-YlZ9%x?d{XW8nVni-nx4A7v@gm$VlrX3Z zC77?2HL}=srm2XC7yu^Li6{dVv#~5)lxkv&dC$K}YHySrU#58FVBQ{-~8Vyc`w01j@F#P!{}JiAH@eCQ9<=o)aBcan3TPpVOg)Rb&7s_=R3NT&hT%vF$ZQeu=+6BvFAj${nMwPh&W+qy#775uX1|6s;oRie;Oz9w8bG|pV_peP9=^RIGiu=|$3#z22BYRH|%SCXeM9LEFbLig`YI8KCj?-uQDSMx8&Hub(&rGWC- z6Dp+R?fQ9+v2wj%!HZ$;yrl9c?!J*n%6iw7(F8V~s9KijwUPcho?xrQQ~@@1>&?u* z3GCrlEo}>_Kn(#ewLL8(a%~<%4Lc%%xk5SZb^PbbrW&J6;|{I*5w7`wv>M0lM>GTv zI6B@K7=fBj&+L9)md1B=O`xl`5iojw_wRsn#_qYk2iN-}@tum@e>ynpPaOKS;!sIE zZrrXFz+e>nSfVf{G{#5%;m)lY?eLGs=Xr>&WEq_!;a(~{AFovX+g|hP&X~sGz04-A zk6IyJ4$4@S0NdE@s+OMpsCx_S#905u35lp&Pd#w&aWb7(;P?50wkOr-*B#P zT2>;<%J{z#ku-R1=3L^LWX*JpYpoW0fydJynRGVsPb>E4BAV94#*KWtmAyV6D;ZTe zej%`AGAXCxKg%&vTR1Mel{>y7%{^e!>GL$OQu4~{{faK=cDY!ximUZxnYO-N6N)Oc zJZp(j*Mr)=-uCpTPCV3#6&U0WeAUmeX@LVf2D3-5lyp>b8<#Dt1l$8JPOF0gnN|H> zne7RGkE#7fnpc_E13=9bMr}yPz8P9fOEY@;va4zrn-YyiMs=!i zf)bH1@szzMUqNIc<`nQzorF-2j?@gNm0Kr#+Nm`fbhe+A3D$9zz3 zIEqF+FZiEIB|w_i&+V*?epj~{DYsQ*boFEv`WO-wrGzD0IeByhZzBURDm z^p2*`j3IKoDVC+u6UcLZQr2ufvL@##q+;;*ExzJW_e^J_rpRdZ%HEMsIlTd8dgl(~ z#BW-#H0XyEsQE_BRUqRemFB?_**96m|E5#+)ZM@y|BYtnTG0Hj54*R8Z467oOUkgeeE#*5`2QRGDCLg$@JqeYAbG(MDB{!LZ6?Hx5idZb1sIS-n%Bzs5+nOaG7~R=-9h(#c6eQ-{mL; z(%c5AsbOiIvw$`6BF=JHt6>((iXdO-bVETk*m6tnv^D<-o-4vOOV=AhW!U(=^rY;u zI;1)#AZLvQWZ963mqNvAY2nlrF({m&!EcuJ$TBBBsI#uZG# zyW&0WF7$p-Kpd%Hc{!%$a6;1DB7P{X4`Eq~hLT{9I^Zf$3;BWSp1djwNzqA%g%I;4 zYshIsL5Tk`x!9($*2S!exH2Bv)K?c~niNVn{32K6R>z*d)@hR_&h ze8AvGS0Ly?ZLEfeJCut zI7CbZQ3BH?d5X91uQNhL_6l~Lh4@$~T1YBbeMdHW*MAxvKM1`*1+_ew_@iErb4=QxpBhxKY<_!ZJZosLk)><1>I+N%~oeT1`Z&?>jA~a%-e2$i#?@GUNFF8e8Yk3~@rLrtE1yl7~S!X$DxQr`$GkvRWwL;U^tV@0Z8cidkWWRCm^nGlbi$s`=9n_naM=#gZ=AHIy zwc@!=_CM)|EEv5R_ES}M@RC-kGg%q7^3g5gLCEc1R8l*p2ZcZ^7al$6ups|4@HFuJ-)Nqvfob3!V0}1SymGChA;SL(uYcm{F8u!(|KDzbc40nP lSUcO7y7qbe?dsF>3&5H|Qh-E;%E8A8=0~Qbl~R#L{vTFMv@8Gs diff --git a/report/main.tex b/report/main.tex index 840e208..44dc526 100644 --- a/report/main.tex +++ b/report/main.tex @@ -47,16 +47,16 @@ \newcommand{\true}{y} \begin{document} - + \title{Neural Networks for Breast Cancer Detection on Limited Featuresets} -\date{\today} +\date{\today} \author{Lars Bogner} \affiliation{University of Oslo\\ \url{https://github.uio.no/larsbog/FYSSTK-Project2}} \newpage - + \begin{abstract} -We study an approach to classification on the Wisconsin Breast Cancer Diagnostic dataset using neural networks to expand a limited set of physical features to the full feature set. Using only the radius and area of cell nuclei, we train a feedforward neural network to predict the remaining features with a mean squared error of \qty{3.15}{\square \std}. Using out-of-fold prediction we obtain independent predictions for the entire dataset. Using these predictions as input for a logistic regression classifier, we achieve an accuracy of \qty{94.74}{\percent} and an AUC score of \num{0.99105}, comparable to results using the full dataset. + We study an approach to classification on the Wisconsin Breast Cancer Diagnostic dataset using neural networks to expand a limited set of physical features to the full feature set. Using only the radius and area of cell nuclei, we train a feedforward neural network to predict the remaining features with a mean squared error of \qty{3.15}{\square \std}. Using out-of-fold prediction we obtain independent predictions for the entire dataset. Using these predictions as input for a logistic regression classifier, we achieve an accuracy of \qty{94.74}{\percent} and an AUC score of \num{0.99105}, comparable to results using the full dataset. \end{abstract} \maketitle @@ -69,72 +69,72 @@ Sampling the cell nuclei of tissue extracted using a fine needle aspirate from t Using a very small subset of features which can be obtained using basic image recognition, we enhance them to a larger set of nuclei features using a regression model, based on which the straightforward classification can be continued using the large set of features. Starting from the features of radius and area from the Wisconsin Breast Cancer Diagnostic dataset \cite{williamwolbergBreastCancerWisconsin1993,streetNuclearFeatureExtraction1993} we obtain the full set of features as presented in the dataset using an optimized regression model. Using this regression model it will be possible in the future to use standard classifiers while only needing to measure the radius and area of the cell nuclei in diagnosis. In \cref{section:methods} we present the methods necessary to implement the corresponding neural networks and how to optimize them. The results and a discussion thereof is presented in \cref{section:results}. Finally, our findings are concluded in \cref{section:conclusion}. - + \section{Methods}\label{section:methods} \subsection{Neural Networks} \begin{figure} - \newcommand\setAngles[3]{ - \pgfmathanglebetweenpoints{\pgfpointanchor{#2}{center}}{\pgfpointanchor{#1}{center}} - \pgfmathsetmacro\angmin{\pgfmathresult} - \pgfmathanglebetweenpoints{\pgfpointanchor{#2}{center}}{\pgfpointanchor{#3}{center}} - \pgfmathsetmacro\angmax{\pgfmathresult} - \pgfmathsetmacro\dang{\angmax-\angmin} - \pgfmathsetmacro\dang{\dang<0?\dang+360:\dang} -} -\begin{tikzpicture}[x=2.2cm,y=1.4cm] - \readlist\Nnod{3,4,4,2} % array of number of nodes per layer - - \foreachitem \N \in \Nnod{ % loop over layers - \def\lay{\Ncnt} % alias of index of current layer - \pgfmathsetmacro\prev{int(\Ncnt-1)} % number of previous layer - \message{^^J Layer \lay, N=\N, prev=\prev ->} - - % NODES - \foreach \i [evaluate={\y=\N/2-\i; \x=\lay; \n=\nstyle;}] in {1,...,\N}{ % loop over nodes - \message{N\lay-\i, } - \node[node \n] (N\lay-\i) at (\x,\y) {$a_\i^{(\prev)}$}; - } - - % CONNECTIONS - \foreach \i in {1,...,\N}{ % loop over nodes - \ifnum\lay>1 % connect to previous layer - \setAngles{N\prev-1}{N\lay-\i}{N\prev-\Nnod[\prev]} % angles in current node - %\draw[red,thick] (N\lay-\i)++(\angmin:0.2) --++ (\angmin:-0.5) node[right,scale=0.5] {\dang}; - %\draw[blue,thick] (N\lay-\i)++(\angmax:0.2) --++ (\angmax:-0.5) node[right,scale=0.5] {\angmin, \angmax}; - \foreach \j [evaluate={\ang=\angmin+\dang*(\j-1)/(\Nnod[\prev]-1);}] %-180+(\angmax-\angmin)*\j/\Nnod[\prev] - in {1,...,\Nnod[\prev]}{ % loop over nodes in previous layer - \setAngles{N\lay-1}{N\prev-\j}{N\lay-\N} % angles out from previous node - \pgfmathsetmacro\angout{\angmin+(\dang-360)*(\i-1)/(\N-1)} % number of previous layer - %\draw[connect arrow,white,line width=1.1] (N\prev-\j.{\angout}) -- (N\lay-\i.{\ang}); - \draw[connect arrow] (N\prev-\j.{\angout}) -- (N\lay-\i.{\ang}); % connect arrows uniformly - } - \fi % else: nothing to connect first layer - } - } - - % LABELS - \node[above=5,align=center,mygreen!60!black] at (N1-1.90) {input\\[-0.2em]layer}; - \node[above=2,align=center,myblue!60!black] at (N2-1.90) {hidden layers}; - \node[above=8,align=center,myred!60!black] at (N\Nnodlen-1.90) {output\\[-0.2em]layer}; - -\end{tikzpicture} -\caption{Visualization of a neural network, modified work of \cite{NeuralNetworksTikZnet2024}.} -\label{fig:neural_network} + \newcommand\setAngles[3]{ + \pgfmathanglebetweenpoints{\pgfpointanchor{#2}{center}}{\pgfpointanchor{#1}{center}} + \pgfmathsetmacro\angmin{\pgfmathresult} + \pgfmathanglebetweenpoints{\pgfpointanchor{#2}{center}}{\pgfpointanchor{#3}{center}} + \pgfmathsetmacro\angmax{\pgfmathresult} + \pgfmathsetmacro\dang{\angmax-\angmin} + \pgfmathsetmacro\dang{\dang<0?\dang+360:\dang} + } + \begin{tikzpicture}[x=2.2cm,y=1.4cm] + \readlist\Nnod{3,4,4,2} % array of number of nodes per layer + + \foreachitem \N \in \Nnod{ % loop over layers + \def\lay{\Ncnt} % alias of index of current layer + \pgfmathsetmacro\prev{int(\Ncnt-1)} % number of previous layer + \message{^^J Layer \lay, N=\N, prev=\prev ->} + + % NODES + \foreach \i [evaluate={\y=\N/2-\i; \x=\lay; \n=\nstyle;}] in {1,...,\N}{ % loop over nodes + \message{N\lay-\i, } + \node[node \n] (N\lay-\i) at (\x,\y) {$a_\i^{(\prev)}$}; + } + + % CONNECTIONS + \foreach \i in {1,...,\N}{ % loop over nodes + \ifnum\lay>1 % connect to previous layer + \setAngles{N\prev-1}{N\lay-\i}{N\prev-\Nnod[\prev]} % angles in current node + %\draw[red,thick] (N\lay-\i)++(\angmin:0.2) --++ (\angmin:-0.5) node[right,scale=0.5] {\dang}; + %\draw[blue,thick] (N\lay-\i)++(\angmax:0.2) --++ (\angmax:-0.5) node[right,scale=0.5] {\angmin, \angmax}; + \foreach \j [evaluate={\ang=\angmin+\dang*(\j-1)/(\Nnod[\prev]-1);}] %-180+(\angmax-\angmin)*\j/\Nnod[\prev] + in {1,...,\Nnod[\prev]}{ % loop over nodes in previous layer + \setAngles{N\lay-1}{N\prev-\j}{N\lay-\N} % angles out from previous node + \pgfmathsetmacro\angout{\angmin+(\dang-360)*(\i-1)/(\N-1)} % number of previous layer + %\draw[connect arrow,white,line width=1.1] (N\prev-\j.{\angout}) -- (N\lay-\i.{\ang}); + \draw[connect arrow] (N\prev-\j.{\angout}) -- (N\lay-\i.{\ang}); % connect arrows uniformly + } + \fi % else: nothing to connect first layer + } + } + + % LABELS + \node[above=5,align=center,mygreen!60!black] at (N1-1.90) {input\\[-0.2em]layer}; + \node[above=2,align=center,myblue!60!black] at (N2-1.90) {hidden layers}; + \node[above=8,align=center,myred!60!black] at (N\Nnodlen-1.90) {output\\[-0.2em]layer}; + + \end{tikzpicture} + \caption{Visualization of a neural network, modified work of \cite{NeuralNetworksTikZnet2024}.} + \label{fig:neural_network} \end{figure} Neural networks are the continuation of a linear combination by introducing nonlinearity. While linear regression uses a linear combination of input features to create an output, neural networks add a nonlinear function to the output of the linear combination. This is the entire secret behind the powerful possibilities neural networks have shown over the last years \cite{elstnerLectureMachineLearning2025,goodfellowDeepLearning2016}. A feedforward neural network (FFNN) uses a layered structure as shown in \cref{fig:neural_network}, where each node represents a separate linear combination of it's inputs. A general node uses the inputs $x_i$ to create a weighted sum $z$, which is then activated using a so called activation function $g(z)$. More detail on the activation function is provided in \cref{subsec:activation}. The output of the general node can then be calculated as \begin{equation} - y = g(z) = g\left(\sum_i (w_i x_i) + b\right), + y = g(z) = g\left(\sum_i (w_i x_i) + b\right), \end{equation} where $b$ is the so-called bias of the node which in turn leads to a shift of the weighted sum, irrespective of the inputs. The linear combination $z = \sum_i (w_i x_i) + b$, can also be rewritten as a vector-vector multiplication of vectors $\vec a_{j}^{(k)} = (b, \vec{w})$ and $\vec X = (1, \vec x)$. This leads to a simple expression for the output of a node of $y = g(\vec a \cdot \vec X)$. All parameters of a single node in layer $k$ and position $j$ in the layer are contained within $\vec a_{j}^{(k)}$ \cite{elstnerLectureMachineLearning2025,bishopPatternRecognitionMachine2006}. We use the term layer, as the output of each layer will be used as the input of the following layer, thus presenting a clear hierarchy. This hierarchy is also useful for calculating the gradient of the final output of the neural network, with respect to all the weights in the layers, as we can use chained differentiation. Calculating the gradient with respect to all the weights is necessary as the optimization techniques presented in \cref{subsec:optimization} rely on the knowledge of the gradients to find the optimal set of weights. The gradient of a single node given the input $\vec{x}$ is given by \begin{equation} - \frac{\partial y}{\partial \vec a} = \vec{X} g'(\vec a \vec X), + \frac{\partial y}{\partial \vec a} = \vec{X} g'(\vec a \vec X), \end{equation} where $g'$ is the derivative of the activation function with respect to its input. Then given the input of a node in the previous layer we can calculate the gradient with respect to its weights using the chain rule of derivatives. @@ -144,85 +144,85 @@ Activation functions provide a source of nonlinearity in the otherwise linear co \paragraph{Rectified Linear Unit} For simplicity this paper will therefore focus on a set of three activation functions. A very simple non-linear function is provided by the Rectified Linear Unit (ReLU), which shall be defined as \begin{equation} - \mathrm{ReLU}(z) = \begin{cases} - 0 & z < 0\\ z & z \geq 0 - \end{cases}. + \mathrm{ReLU}(z) = \begin{cases} + 0 & z < 0 \\ z & z \geq 0 + \end{cases}. \end{equation} While this function provides all necessary properties of an activation function, we will provide an explanation, why it is not well suited for the optimization procedure of FFNN. The issue with the ReLU is that for $z < 0$ the gradient of the function vanishes. This poses a problem, as mentioned above, the optimization algorithm relies on the gradient of the output to adjust and optimize the weights of the neural network. A vanishing gradient disables a possible optimization. This problem is further amplified by the fact, that the gradient of weights in the first layers of the network depends on the product of the gradients with the previous layers, thus a vanishing gradient in one of the last layers has adverse effects on the optimization of weights in the previous layers\cite{kieslerLectureModernMethods2025}. \paragraph{Leaky Rectified Linear Unit} As the ReLU is otherwise quite optimal due to its simplicity, we will mailny use a slightly modified version of the ReLU for the majority of nodes in the studied neural networks, namely the Leaky ReLU (LReLU). The leaky ReLU modifies the standard ReLU by introducing a very small but non-vanishing gradient for $z < 0$. It is defined as \begin{equation} - \mathrm{LReLU}(z) = \begin{cases} - \epsilon z & z < 0\\ z & z \geq 0 - \end{cases}, + \mathrm{LReLU}(z) = \begin{cases} + \epsilon z & z < 0 \\ z & z \geq 0 + \end{cases}, \end{equation} with $\epsilon$ being the leak coefficient, a small number, set to $\epsilon = \num{e-5}$ for all studies. Due to the step in the derivative at $z = 0$ our function is still nonlinear, but providing a positive derivative over the whole domain\cite{elstnerLectureMachineLearning2025}. \paragraph{Softmax} The last activation to introduce for our studies will be the Softmax function. It shall be defined as \cite{kieslerLectureModernMethods2025} \begin{equation} - \mathrm{Softmax}(z_{j}^{(k)}) = \frac{\exp z_{j}^{(k)}}{\sum_{j'} \exp z_{j'}^{(k)}}. + \mathrm{Softmax}(z_{j}^{(k)}) = \frac{\exp z_{j}^{(k)}}{\sum_{j'} \exp z_{j'}^{(k)}}. \end{equation} -Additionally, to providing a source of nonlinearity, the softmax function has the special property, that the total output of a layer, i.e. the sum of its outputs is regularized to a value of 1. This property is especially useful in the case, where it is used as a classifier, where each of the layers outputs corresponds to a probability of being member in a certain class. While there is no guarantee, that the output of such a layer corresponds to a frequency in the frequentist interpretation, it is very useful in combination with the Cross Entropy introduced in the next section, as we have a nice behaving gradient, useful for optimization \cite{kieslerLectureModernMethods2025}. +Additionally, to providing a source of nonlinearity, the softmax function has the special property, that the total output of a layer, i.e. the sum of its outputs is regularized to a value of 1. This property is especially useful in the case, where it is used as a classifier, where each of the layers outputs corresponds to a probability of being member in a certain class. While there is no guarantee, that the output of such a layer corresponds to a frequency in the frequentist interpretation, it is very useful in combination with the Cross Entropy introduced in the next section, as we have a nice behaving gradient, useful for optimization \cite{kieslerLectureModernMethods2025}. \subsection{Loss Functions} The function of the loss function is to quantize the discrepancy between the prediction of a numerical model and the true values\cite{elstnerLectureMachineLearning2025,kieslerLectureModernMethods2025,goodfellowDeepLearning2016}. It is also sometimes called cost function. It must be a differentiable function which is minimal when, the prediction aligns with the true values. The choice of loss function needs to be specific to the task for which the model should be optimized for. For the regressional part of this paper we will use the mean squared error (MSE) as a loss function. The MSE is defined using the prediction $\pred$ and the true values $\true$ as \begin{equation} - \mathrm{MSE} = \frac{\sum_i (\pred_i - \true_i)^2}{N}. + \mathrm{MSE} = \frac{\sum_i (\pred_i - \true_i)^2}{N}. \end{equation} Its derivative is therefore defined as \begin{equation} - \frac{\partial \mathrm{MSE}}{\partial \pred_i} = \frac{2 (\pred_i - \true_i)}{N}. + \frac{\partial \mathrm{MSE}}{\partial \pred_i} = \frac{2 (\pred_i - \true_i)}{N}. \end{equation} The advantage of the MSE as a loss function is, that the loss value can be directly interpreted in the context of the metric on which we optimize\cite{elstnerLectureMachineLearning2025}. Furthermore, it gives the advantage, that compared to the mean absolute error, which uses the absolute deviation between the target and the prediction, that it is continuous differentiable and that large deviations are penalized more than small deviations. Especially in the context of small measurement uncertainties this is important. The second loss function used is the cross entropy loss. It is used for the case of classification tasks and is defined as \cite{elstnerLectureMachineLearning2025} \begin{equation} - \mathrm{CE} = \sum_i \true_i \log \pred_i. + \mathrm{CE} = \sum_i \true_i \log \pred_i. \end{equation} The cross entropy loss is inspired by the definition of the physical entropy of stochastic systems and is useful, as it is minimal when the probabilities $\pred$ and $\true$ align. Combined with the softmax activation function from \cref{subsec:activation}, the derivative simplifies to a proportional relation with respect to $\pred - \true$. \subsubsection{Other Evaluation Metrics} Additional to the metrics on which we optimize the model, there are other scores than can be important in the context of evaluating the quality of a models performance. The relative amount of true predictions of a classification model can be judged using the accuracy. The accuracy is defined as \begin{equation} - \mathrm{Accuracy} = \frac{\sum_{c \in C} N(\true = c, \pred = c)}{N_\mathrm{total}}, + \mathrm{Accuracy} = \frac{\sum_{c \in C} N(\true = c, \pred = c)}{N_\mathrm{total}}, \end{equation} where $C$ contains all the possible classes of the dataset and $N$ is the number of predictions with certain conditions. A second metric for evaluating the performance of a classifier is the area-under-curve (AUC) score of the receiver-operating-characteristic (ROC). It is defined as \begin{equation} - \mathrm{AUC} = \int_0^1 \mathcal{T}(\mathcal{F}) \mathrm d \mathcal F, + \mathrm{AUC} = \int_0^1 \mathcal{T}(\mathcal{F}) \mathrm d \mathcal F, \end{equation} where $\mathcal T$ is the true-positive-rate and $\mathcal F$ is the false-positive-rate. A score of 1 indicates perfect classification, while a score of \num{0.5} is equivalent to a random classifier, as in that case $\mathcal T(\mathcal F) \equiv \mathcal F$. \subsection{Regularization} -As the number of parameters in a neural network grows very quickly, special attention needs to be paid to avoid overfitting the training dataset. Overfitting means, that the available degrees of freedom of the model will be used to represent all the datapoints in the training dataset perfectly. This will lead to very poor generalization ability of the model. This phenomenon is part of the bias-variance trade-off, which is ubiquitous in machine learning. With increasing numbers of parameters in the model, i.e. due to an increase in layer size and/or increase in the number of layers, the number of degrees of freedom also increases. With a larger number of degrees of freedom the variance of the model increases. If there are too few parameters, the bias of the model will increase. The goal is therefore to find the middle point, where neither the bias nor variance are increased. +As the number of parameters in a neural network grows very quickly, special attention needs to be paid to avoid overfitting the training dataset. Overfitting means, that the available degrees of freedom of the model will be used to represent all the datapoints in the training dataset perfectly. This will lead to very poor generalization ability of the model. This phenomenon is part of the bias-variance trade-off, which is ubiquitous in machine learning. With increasing numbers of parameters in the model, i.e. due to an increase in layer size and/or increase in the number of layers, the number of degrees of freedom also increases. With a larger number of degrees of freedom the variance of the model increases. If there are too few parameters, the bias of the model will increase. The goal is therefore to find the middle point, where neither the bias nor variance are increased. A useful tool to control the bias variance trade-off is regularization in neural networks. Using regularization we can decrease the effective number of degrees of freedom, therefore decreasing the tendency to overfit the data \cite{elstnerLectureMachineLearning2025}. The decrease is thereby controlled by the regularization coefficient $\lambda$. In the limit of $\lambda \to 0$, the solution approaches the unregularized model, while there is a steady decrease of the number of effective degrees of freedom with an increase of $\lambda$. Regularization works by adding a term to the cost function of the form \begin{equation} \label{eq:regularization} - C_\text{regularized} = C + \lambda \sum_k \sum_j \lVert a_j^{(k)}\rVert_n^n. + C_\text{regularized} = C + \lambda \sum_k \sum_j \lVert a_j^{(k)}\rVert_n^n. \end{equation} -\Cref{eq:regularization} is the generalized expression for a L$n$-regularization, where $n$ is the degree of the norm in \cref{eq:regularization}. The most common types of regularization are L1- and L2-regularization. Using a 1-norm for regularization leads to the benefit that it has the tendency to force an ever-increasing number of parameters of the model towards zero. It thus effecitvely reduces the number of weights that influence the final result \cite{bishopPatternRecognitionMachine2006}. Especially in highly optimized architectures this may be advantageous, as it reduces the number of computations needed for each evaluation. On the other hand it is known to produce less stable results, as the L2-regularization \cite{kieslerLectureModernMethods2025}. +\Cref{eq:regularization} is the generalized expression for a L$n$-regularization, where $n$ is the degree of the norm in \cref{eq:regularization}. The most common types of regularization are L1- and L2-regularization. Using a 1-norm for regularization leads to the benefit that it has the tendency to force an ever-increasing number of parameters of the model towards zero. It thus effecitvely reduces the number of weights that influence the final result \cite{bishopPatternRecognitionMachine2006}. Especially in highly optimized architectures this may be advantageous, as it reduces the number of computations needed for each evaluation. On the other hand it is known to produce less stable results, as the L2-regularization \cite{kieslerLectureModernMethods2025}. \subsection{Optimization Algorithms} \label{subsec:optimization} The goal of an optimization algorithm is to tune the weights of a FFNN in a way, so that the cost function will be minimized, i.e. the predictions and the targets align. Most algorithms use a way of gradient descent to update the weights and find the optimum via iterative optimization. Gradient descent uses the derivatives of the cost function \begin{equation} - \delta_j^{(k)} \equiv \frac{\partial C}{\partial a_j^{(k)}} + \delta_j^{(k)} \equiv \frac{\partial C}{\partial a_j^{(k)}} \end{equation} and update the weights given a learning rate $\alpha$ using \cite{kieslerLectureModernMethods2025} \begin{equation} - a_j^{(k)} \to a_j^{(k)} - \alpha \cdot \delta_j^{(k)}. + a_j^{(k)} \to a_j^{(k)} - \alpha \cdot \delta_j^{(k)}. \end{equation} To decrease the computational resources needed we will use a more advanced version of this algorithm in this paper. The Adam optimizer updates the weights according to \cite{kingmaAdamMethodStochastic2017} \begin{equation} - \begin{aligned} - m_j^{(k)} &\to \beta_1 m_j^{(k)} + (1 - \beta_1) \delta_j^{(k)}, \\ - v_j^{(k)} &\to \beta_2 v_j^{(k)} + (1 - \beta_2) (\delta_j^{(k)})^2, \\ - \hat{m}_j^{(k)} & = \frac{m_j^{(k)}}{1 - \beta_1^t}, \\ - \hat{v}_j^{(k)} & = \frac{v_j^{(k)}}{1 - \beta_2^t}, \\ - a_j^{(k)} &\to a_j^{(k)} - \alpha \frac{\hat{m}_j^{(k)}}{\sqrt{\hat{v}_j^{(k)}} + \epsilon}, - \end{aligned} + \begin{aligned} + m_j^{(k)} & \to \beta_1 m_j^{(k)} + (1 - \beta_1) \delta_j^{(k)}, \\ + v_j^{(k)} & \to \beta_2 v_j^{(k)} + (1 - \beta_2) (\delta_j^{(k)})^2, \\ + \hat{m}_j^{(k)} & = \frac{m_j^{(k)}}{1 - \beta_1^t}, \\ + \hat{v}_j^{(k)} & = \frac{v_j^{(k)}}{1 - \beta_2^t}, \\ + a_j^{(k)} & \to a_j^{(k)} - \alpha \frac{\hat{m}_j^{(k)}}{\sqrt{\hat{v}_j^{(k)}} + \epsilon}, + \end{aligned} \end{equation} with the optimization parameters $\beta_{1/2}$. As shown in our previous article \cite{bognerRegularizationOptimizationAll2025}, Adam tends to converge faster towards an optimal solution. It is also thus widely used in machine learning optimizations and will be used for all training procedures throughout the studies. @@ -234,49 +234,49 @@ As we mentioned before, there is the possibility for our model to overadjust for \subsubsection{Out-of-Fold prediction} \label{subsec:outoffold} \begin{figure} - \centering - \begin{tikzpicture}[scale=0.8] - % parameters - \def\nfolds{5} - \def\squaresize{0.8} - \def\vspace{0.15} % extra space between rows (relative to squaresize) + \centering + \begin{tikzpicture}[scale=0.8] + % parameters + \def\nfolds{5} + \def\squaresize{0.8} + \def\vspace{0.15} % extra space between rows (relative to squaresize) - % loop over folds - \foreach \row in {0,...,4} { - % vertical position with extra spacing - \pgfmathsetmacro{\y}{-(\row*(\squaresize+\vspace))} - \foreach \col in {0,...,4} { - % decide color - \ifnum\row=\col - \fill[myblue!80] (\col*\squaresize, \y) - rectangle ++(\squaresize,\squaresize); - \else - \fill[myred!75] (\col*\squaresize, \y) - rectangle ++(\squaresize,\squaresize); - \fi - % draw border - \draw[black,thick] (\col*\squaresize, \y) - rectangle ++(\squaresize,\squaresize); - } - % label fold on the left - \node[left] at (-0.2, \y+0.5*\squaresize) {Fold \the\numexpr\row+1\relax}; - % MSE_i label on the right - \node[right] at (5*\squaresize+0.2, \y+0.5*\squaresize) {$\pred_{\the\numexpr\row+1\relax}$}; - } + % loop over folds + \foreach \row in {0,...,4} { + % vertical position with extra spacing + \pgfmathsetmacro{\y}{-(\row*(\squaresize+\vspace))} + \foreach \col in {0,...,4} { + % decide color + \ifnum\row=\col + \fill[myblue!80] (\col*\squaresize, \y) + rectangle ++(\squaresize,\squaresize); + \else + \fill[myred!75] (\col*\squaresize, \y) + rectangle ++(\squaresize,\squaresize); + \fi + % draw border + \draw[black,thick] (\col*\squaresize, \y) + rectangle ++(\squaresize,\squaresize); + } + % label fold on the left + \node[left] at (-0.2, \y+0.5*\squaresize) {Fold \the\numexpr\row+1\relax}; + % MSE_i label on the right + \node[right] at (5*\squaresize+0.2, \y+0.5*\squaresize) {$\pred_{\the\numexpr\row+1\relax}$}; + } - % curly brace summarizing MSEs - % compute top and bottom y positions for the brace - \pgfmathsetmacro{\ytop}{0.5*\squaresize} - \pgfmathsetmacro{\ybottom}{-(4*(\squaresize+\vspace))+0.5*\squaresize} + % curly brace summarizing MSEs + % compute top and bottom y positions for the brace + \pgfmathsetmacro{\ytop}{0.5*\squaresize} + \pgfmathsetmacro{\ybottom}{-(4*(\squaresize+\vspace))+0.5*\squaresize} - % curly brace summarizing MSEs - \draw[decorate,decoration={brace,amplitude=10pt}] - (6.8*\squaresize,\ytop) -- - (6.8*\squaresize,\ybottom) - node[midway,xshift=1.4cm]{$\{\pred_1, \dots, \pred_5\}$}; -\end{tikzpicture} - \caption{Illustration of predicting the full dataset using out-of-fold prediction to avoid contamination of training data. Modified from \cite{bognerRegularizationOptimizationAll2025}. Red represents the training dataset, while the features from the blue part are used to make the prediction. Finally, all predictions are combined into one dataset of the same size as the input dataset.} - \label{fig:outoffold} + % curly brace summarizing MSEs + \draw[decorate,decoration={brace,amplitude=10pt}] + (6.8*\squaresize,\ytop) -- + (6.8*\squaresize,\ybottom) + node[midway,xshift=1.4cm]{$\{\pred_1, \dots, \pred_5\}$}; + \end{tikzpicture} + \caption{Illustration of predicting the full dataset using out-of-fold prediction to avoid contamination of training data. Modified from \cite{bognerRegularizationOptimizationAll2025}. Red represents the training dataset, while the features from the blue part are used to make the prediction. Finally, all predictions are combined into one dataset of the same size as the input dataset.} + \label{fig:outoffold} \end{figure} A further hindrance in our approach with using the output of a regression model to make classifications is, that for an efficient workflow, we need a full but independent prediction of the dataset. To achieve this, we use a workflow inspired by cross-validation. In cross-validation you use multiple folds of the dataset, whereby for every fold, a different part of the dataset is used as testing data, and the rest of the dataset is used for training purposes. This provides independence of training and testing data while being able to test the models general performance for the entire dataset. For out-of-fold prediction, we use the same approach of changing the training data to leave out a specific part each time, but furthermore we use the model trained on each training set, to make predictions for the remaining set. Combining all the predicted sets into one set, allows for having independent predictions for the entire dataset. Having the predictions be independent is important, as otherwise no statement can be made on the generalization ability of the workflow. Using this approach we can expect the same performance on new and yet unseen data. @@ -284,9 +284,9 @@ A further hindrance in our approach with using the output of a regression model All the code necessary to use a FFNN is implemented from scratch in \texttt{Python3} as the \texttt{easynn} library. The source code therefore is located in the \texttt{/src/easynn/} folder of the accompanying git repository \footnote{\url{https://github.uio.no/larsbog/FYSSTK-Project2}}. It is separated into two main modules, called \texttt{schedulers} and \texttt{feedforward}. The scheduler modules use modified versions of our previous work on using gradient descent techniques for ordinary least squares techniques. It is also structured in a modular way, so that scheduling tasks like early stopping and dynamic learning rate adjustment are easy to implement in further versions. The base scheduler implements an update method of signature \texttt{def update(self, gradients: gradient\_\-type) -> gradient\_\-type:}, with \texttt{gradient\_\-type = List[\-Tuple[\-np.ndarray, np.ndarray]] }. This can be overwritten to implement any optimization algorithm. The method takes the gradients of the cost functions with respect to the weights and biases of each layer. Each layer is assigned a tuple of two arrays of the gradients. It should then return the updates for each of the weights and biases in the same format. The training process which fetches new updates after each evaluation of the loss and gradients will continue will the \texttt{cont} property of the base class is true. -The main structures of the neural network are implemented as classes in the \texttt{feedforward} module. Again it is structured in a way, where it is easy to be extended in the future. The FFNN can be constructed using the arguments \texttt{class FFNN: def \_\_init\_\_(self, layers: List[Layer], scheduler: Scheduler, loss\_fn: LossFunction)}. The layers are constructed using \texttt{class Layer: def \_\_init\_\_(self, input\_dim: int, num\_nodes: int, activation\_function: ActivationFunction, regularization: Optional[Regularization] = None)}. Using this structure it is easy to control the hyperparameters of each hidden layer carefully and for example choose the activation and regularization of each layer precisely. The weights and biases of each layer are initialized from a uniform random distribution in the bounds $[-s, s]$ with +The main structures of the neural network are implemented as classes in the \texttt{feedforward} module. Again it is structured in a way, where it is easy to be extended in the future. The FFNN can be constructed using the arguments \texttt{class FFNN: def \_\_init\_\_(self, layers: List[Layer], scheduler: Scheduler, loss\_fn: LossFunction)}. The layers are constructed using \texttt{class Layer: def \_\_init\_\_(self, input\_dim: int, num\_nodes: int, activation\_function: ActivationFunction, regularization: Optional[Regularization] = None)}. Using this structure it is easy to control the hyperparameters of each hidden layer carefully and for example choose the activation and regularization of each layer precisely. The weights and biases of each layer are initialized from a uniform random distribution in the bounds $[-s, s]$ with \begin{equation} - s = \frac{1}{\sqrt{I \cdot N}}, + s = \frac{1}{\sqrt{I \cdot N}}, \end{equation} where $I$ is the input dimension and $N$ is the number of nodes. This initialization is taken from the standard implementation in the widely used \texttt{pyTorch} library. @@ -296,20 +296,26 @@ To check the functionality of the developed code, in a first stage, the loss his For the comparison against ordinary least squares regression, we use the framework presented before by the same author \cite{bognerRegularizationOptimizationAll2025}. It uses the analytical solution to the ordinary least squares regression to find the optimal parameters. +As a further validation of the functionality, we use the Runge function +\begin{equation} + f(x) = \frac{1}{1 + 25 x^2}, +\end{equation} +with added Gaussian noise of standard deviation \num{1.0} to create a regression dataset. On this dataset of size $N = \num{100000}$ and \qty{80}{\percent}-\qty{20}{\percent} train-test-splitting we achieve a MSE of \numrange{0.9136}{0.9178}, using between one and two hidden layers with either 50 or 100 nodes per hidden layer. This exceeds the performance of our previous work on least squares regression using different regularization and optimization techniques \cite{bognerRegularizationOptimizationAll2025}, where a MSE of \num{0.92} was achieved using the same dataset \footnote{The $x$ values were expanded into multiple of features containing polynomial multiples $x^n$ of themselves. For optimal performance $n \in \{0, 1, \dots, \geq 10\}$ were required.} and train-test-splitting. This motivates further the use of neural networks for regression tasks, as possible non-linearities in the data can be modeled more easily. Least squares methods on the other hand provide only linear combinations of the input features. + \subsection{Use of AI tools} During the writing of this paper artificial intelligence in the form of large-language models (LLMs) has been used. Two different LLMs have been used for a faster writing of the code used. Mainly for short rewrites of code structure and brain-storming the model GPT-5 by OpenAI and the model DeepSeek by Hangzhou DeepSeek Artificial Intelligence have been used. At no point was artificial intelligence used for writing the report. All output by LLMs has be checked by the author. - -\section{Results and Discussion}\label{section:results} + +\section{Results and Discussion}\label{section:results} \subsection{Introduction to the Dataset} \begin{figure} - \centering - \includegraphics[width=\columnwidth]{../notebooks/violin_plot_features_0.pdf} - \caption{Violin plot of the distribution of all features in the dataset, separated by benign and malignant samples. A target value of 0 corresponds to malignant samples, while a target value of 1 corresponds to benign samples.} - \label{fig:violinplot} + \centering + \includegraphics[width=\columnwidth]{../notebooks/violin_plot_features_0.pdf} + \caption{Violin plot of the distribution of all features in the dataset, separated by benign and malignant samples. A target value of 0 corresponds to malignant samples, while a target value of 1 corresponds to benign samples.} + \label{fig:violinplot} \end{figure} -For the entire study we use the dataset presented by Wolberg \cite{williamwolbergBreastCancerWisconsin1993}. It includes features recorded from nuclei in breast tissue and their geometric properties. Each physical feature is present three times, where the means and standard deviations of the measured physical features are used as separate features. Furthermore, the mean of the three largest measurements of each physical feature is present in the dataset. The physical features include: the radius, the standard deviation of the images gray-scale values (called texture), the perimeter, the area and the local variation in the radius length, called smoothness. Additionally, the concavity, the number of concave points, the compactness, the symmetry and the fractal dimension are present. For normalization, all features are rescaled to be centered around 0 and have a standard deviation of 1. The normalizaiton is achieved using the \texttt{StandardScaler} class of the \texttt{scikit-learn} library. A quantitative overview on the average of each of the physical features is presented in \cref{fig:violinplot}. The violinplot shows the feature distribution for both classes of benign and malignant cell nuclei. +For the entire study we use the dataset presented by Wolberg \cite{williamwolbergBreastCancerWisconsin1993}. It includes features recorded from nuclei in breast tissue and their geometric properties. Each physical feature is present three times, where the means and standard deviations of the measured physical features are used as separate features. Furthermore, the mean of the three largest measurements of each physical feature is present in the dataset. The physical features include: the radius, the standard deviation of the images gray-scale values (called texture), the perimeter, the area and the local variation in the radius length, called smoothness. Additionally, the concavity, the number of concave points, the compactness, the symmetry and the fractal dimension are present. For normalization, all features are rescaled to be centered around 0 and have a standard deviation of 1. The normalizaiton is achieved using the \texttt{StandardScaler} class of the \texttt{scikit-learn} library. A quantitative overview on the average of each of the physical features is presented in \cref{fig:violinplot}. The violinplot shows the feature distribution for both classes of benign and malignant cell nuclei. The separability of the two classes is easily visible given all features in the class. On the other hand, there is no single feature which allows for very good separation. Thus reproducing separability using only a limited set of features is the challenge of this study. @@ -320,10 +326,10 @@ For all studies we use train-test-splitting with \qty{20}{\percent} of the data \subsubsection{Hyperparameter Tuning} \begin{figure} - \centering - \includegraphics[width=\columnwidth]{../notebooks/nodenumber_tuning_regression.pdf} - \caption{MSE for different numbers of hidden layers in the regression network and nodes per hidden layer.} - \label{fig:regression_hyperparameter} + \centering + \includegraphics[width=\columnwidth]{../notebooks/nodenumber_tuning_regression.pdf} + \caption{MSE for different numbers of hidden layers in the regression network and nodes per hidden layer.} + \label{fig:regression_hyperparameter} \end{figure} As mentioned in \cref{subsec:activation}, the Leaky ReLU combines the disadvantages of the ReLU while conserving non-vanishing gradients on the entire domain. Thus, it is not surprising, that they perform similar in our studies with the LReLU providing a more stable training process. Due to this finding we will use the LReLU for all the hidden layers. \Cref{fig:regression_hyperparameter} shows the resulting MSE for different numbers of hidden layers and numbers of nodes in the layers. We can observe a general trend of decrease in MSE with a decrease in the number of hidden layers and an increase in the number of neurons per hidden layer. This general trend is to be expected as with an increase in the number of hidden layers the gradients in the last layers can converge towards zero with deep networks due to the multiplicic nature of the gradients. Additionally, can an increase number of nodes per layer better capture even complicated relationships in the data. Due to the adverse effects of an increase in node number, we chose a conservative set of parameters with 32 neurons per hidden layer and two hidden layers. @@ -331,10 +337,10 @@ As mentioned in \cref{subsec:activation}, the Leaky ReLU combines the disadvanta In a second sub-study we observe the influence of the regularization constant with and L2-regularization on all layers. Further studies on variable regularization per layer were not carried out. The studies show a steady MSE for a range of \numrange{0}{e-3} with a sharp increase by a factor of \num{10} with the step to $\lambda = \num{e-2}$. Thus, we choose to use $\lambda = \num{e-3}$ for the regressional model. This provides a stable insurance against overfitting of the training data while at the same time not negatively effecting the models' generalization ability. \begin{figure} - \centering - \includegraphics[width=\columnwidth]{../notebooks/regression_final_model_performance.pdf} - \caption{Average MSE of the predicted features using the regression model by feature label.} - \label{fig:regression_errors} + \centering + \includegraphics[width=\columnwidth]{../notebooks/regression_final_model_performance.pdf} + \caption{Average MSE of the predicted features using the regression model by feature label.} + \label{fig:regression_errors} \end{figure} The final set of hyperparameters is then again tested for MSE on all the features of the dataset it is optimized to predict. The results of the per target MSE are displayed in \cref{fig:regression_errors}, while we see an increase in the upper bound of the per target MSE, to \qty{e1}{\square \std}, in comparison to the unoptimized model, the general trend for the MSE shows a lower baseline. The lowest accuracy show the predictions for the texture and perimeter. Both show an MSE in excess of \qty{5}{\square \std} in the mean and worst values. Especially the perimeter values are surprising, as one would expect the perimeter to follow from a simple relation with the nuclei's radius. All the other physical features behave similar with respect to the prediction error of \qtyrange{e-3}{e-4}{\square \std}. In general the prediction uncertainties are smaller on the error values of the physical features. The total MSE over all features amounts to \qty{3.15}{\square \std}. While the performance is not perfect, it is sufficient to create a classification model based on the regression output, as not all features are equally important for the classification task. @@ -344,26 +350,26 @@ Based on these hyperparameters we trained models using 5-fold out-of-fold predic \subsubsection{Comparison against Linear Regression} \begin{figure} - \centering - \includegraphics[width=\columnwidth]{../notebooks/regression_nn_vs_ols.pdf} - \caption{Comparison of the MSE per target feature between the neural network regression and ordinary least squares regression.} - \label{fig:regression_comparison} + \centering + \includegraphics[width=\columnwidth]{../notebooks/regression_nn_vs_ols.pdf} + \caption{Comparison of the MSE per target feature between the neural network regression and ordinary least squares regression.} + \label{fig:regression_comparison} \end{figure} \Cref{fig:regression_comparison} shows the performance comparison in the MSE between an ordinary least squares (OLS) regression and the optimized neural network. Both show very similar performance with only slightest differences in the MSE. In general ordinary least squares underperform, compared to the FFNN by a small margin. The very good results of the ordinary least squares approach is surprising. -Ordinary least squares regression is only able to capture linear relationships between the input features and the target features. Thus, one would expect that the more complex FFNN would be able to outperform the OLS regression by a larger margin. The small difference in performance indicates, that the relations between the input features and the target features are mostly linear in nature. Using OLS for describing linear relations is also advantageous, as it provides a closed form solution for the optimal weights. Thus, no iterative optimization procedure is necessary, which saves computational resources and time. The closed form solution also guarantees that the found solution is the global optimum of the cost function, while iterative optimization may get stuck in local minima. +Ordinary least squares regression is only able to capture linear relationships between the input features and the target features. Thus, one would expect that the more complex FFNN would be able to outperform the OLS regression by a larger margin. The small difference in performance indicates, that the relations between the input features and the target features are mostly linear in nature. Using OLS for describing linear relations is also advantageous, as it provides a closed form solution for the optimal weights. Thus, no iterative optimization procedure is necessary, which saves computational resources and time. The closed form solution also guarantees that the found solution is the global optimum of the cost function, while iterative optimization may get stuck in local minima. Even if using regularized version of linear regression, that do not possess closed form solution, the optimization of such models as the gradients are simpler to compute and the models contain fewer degrees of freedom, leading to faster convergence. The excellent performance of the linear regression suggests that the additional challenges posed by using the FFNN are not worth the extra development costs. We thus recommend using this approach only for datasets with complex relations. \subsection{Classification on Regression Output} \begin{figure} - \centering - \includegraphics[width=\columnwidth]{../notebooks/classification_confusion_matrix.pdf} - \caption{Confusion Matrix of the classifier for the regression output. A target value of 0 indicates malignant nuclei, while 1 is indicative of benign samples.} - \label{fig:confusion_matrix} + \centering + \includegraphics[width=\columnwidth]{../notebooks/classification_confusion_matrix.pdf} + \caption{Confusion Matrix of the classifier for the regression output. A target value of 0 indicates malignant nuclei, while 1 is indicative of benign samples.} + \label{fig:confusion_matrix} \end{figure} -Similar to the structure of the regression model, we use a FFNN using a variable number of nodes per hidden layer in configurations with different numbers of hidden layers to create a classifier. To use the model as a classifier, the loss function of the model is changed to the cross-entropy loss introduced in \cref{section:methods}. Additionally, we change the activation of the output layer to the softmax function. The leaky rectified linear unit is again used for the purpose in all other neurons, following the results from the previous study. The parameters are optimized using loss minimization using the Adam optimization algorithm. +Similar to the structure of the regression model, we use a FFNN using a variable number of nodes per hidden layer in configurations with different numbers of hidden layers to create a classifier. To use the model as a classifier, the loss function of the model is changed to the cross-entropy loss introduced in \cref{section:methods}. Additionally, we change the activation of the output layer to the softmax function. The leaky rectified linear unit is again used for the purpose in all other neurons, following the results from the previous study. The parameters are optimized using loss minimization using the Adam optimization algorithm. As above, a parameter scan is used to find the best parameter combination for accurate prediction, whether the nuclei are part of benign or malignant tissue. The results of the parameter scan are displayed in \cref{fig:classification_hyperparameters}. For the number of hidden layers and neurons, there is a clear maximum in the area-under-curve (AUC) score for the case of 64 neurons in either a single or two hidden layers. Using the full dataset in the form without regression AUC scores of \num{0.9962} have been shown using logistic regression \cite{mostafaBreastCancerPrediction}. To account for stochastic variations in the training success the AUC is averaged over a total of \num{50} runs for each entry in \cref{fig:classification_hyperparameters}. The regularization constant is chosen to be $\lambda = \num{e-6}$ according to the figure. During \num{50} runs this combination of hyperparameters is able to achieve an average accuracy of \qty{92.70}{\percent} and an AUC score of \num{0.9781}. While this performance is not as good as the performance using logistic regression on the full dataset \cite{mostafaBreastCancerPrediction}, it is to be expected, that the performance decreases when using only a limited set of features to create the regression output and following that the classification. @@ -372,14 +378,14 @@ The performance is further evaluated using the confusion matrix in \cref{fig:con \subsubsection{Using Logistic Regression for Classification} \begin{figure} - \centering - \includegraphics[width=\columnwidth]{../notebooks/logistic_classification_confusion_matrix.pdf} - \caption{Confusion Matrix of the logistic regression for the FFNN regression output. A target value of 0 indicates malignant nuclei, while 1 is indicative of benign samples.} - \label{fig:confusion_logistic} + \centering + \includegraphics[width=\columnwidth]{../notebooks/logistic_classification_confusion_matrix.pdf} + \caption{Confusion Matrix of the logistic regression for the FFNN regression output. A target value of 0 indicates malignant nuclei, while 1 is indicative of benign samples.} + \label{fig:confusion_logistic} \end{figure} -Thus, we also study the use of logistic regression to create a classification based on the outputs of the regression model. The logistic regression model uses the same framework code as the FFNN for comparability. For creating a logistic regression classification, a singular layer with two neurons and a softmax activation is used. The cross-entropy-loss is used as an optimization metric, which is then minimized using the Adam optimizer. +Thus, we also study the use of logistic regression to create a classification based on the outputs of the regression model. The logistic regression model uses the same framework code as the FFNN for comparability. For creating a logistic regression classification, a singular layer with two neurons and a softmax activation is used. The cross-entropy-loss is used as an optimization metric, which is then minimized using the Adam optimizer. -After the minimization a test split is evaluated for accuracy and receiver-operating-characteristic leading to an accuracy value of \qty{94.74}{\percent} and an AUC score of \num{0.99105}. The logistic regression thus outperforms the classification models by far. It is especially noteworthy, that nearly the same performance as for full original dataset data can be reached. While it might seem counterintuitive, that the more complex models are not able to match the performance of the arguably very simple logistic regression is due to the fact, that even with regularization, the more complex models have a higher tendency to overfit simple relations between the features and the classification. +After the minimization a test split is evaluated for accuracy and receiver-operating-characteristic leading to an accuracy value of \qty{94.74}{\percent} and an AUC score of \num{0.99105}. The logistic regression thus outperforms the classification models by far. It is especially noteworthy, that nearly the same performance as for full original dataset data can be reached. While it might seem counterintuitive, that the more complex models are not able to match the performance of the arguably very simple logistic regression is due to the fact, that even with regularization, the more complex models have a higher tendency to overfit simple relations between the features and the classification. The good performance of the linear regression also helps fur future diagnostic purposes, as we have shown, that with the correct combination of neural networks and logistic regression it is possible to use a very small, easy to extract subset of the physical features to achieve nuclei identification at the highest level. @@ -387,7 +393,7 @@ Additionally, we compare the confusion matrix in \cref{fig:confusion_logistic} a -\section{Conclusion}\label{section:conclusion} +\section{Conclusion}\label{section:conclusion} It is evident, that a classification performance close to the one using the full dataset can be achieved using only a limited set of physical features. Using a FFNN based regression model followed by a logistic regression classifier, an accuracy of \qty{94.74}{\percent} and an AUC score of \num{0.99105} is achieved. This performance is comparable to the performance of logistic regression on the full dataset \cite{mostafaBreastCancerPrediction}. Using only six features derived from two physical measurements, the radius and area of cell nuclei, this performance is possible. It is therefore possible to reduce the data acquisition for future diagnostic purposes to a small set of physical features, while at the same time conserving a high classification accuracy. This is especially important in the context of fast and cheap diagnostic procedures. @@ -403,10 +409,10 @@ Furthermore, the use of a large set of features in training of a FFNN to make pr \onecolumngrid \appendix \begin{figure} - \centering - \includegraphics[width=.8\textwidth]{../notebooks/classification_hyperparameter_scan.pdf} - \caption{Area-under-Curve score dependance on the number of hidden layers, the number of nodes per hidden layer and the regularization strength $\lambda$ for a L2 regularization. The off-diagonal plots show the average AUC per combination of two parameter values. The diagonal values show the average score per single parameter value. Optimization using Adam with $\alpha = \num{0.05}$ and 500 epochs. LReLU activation in the hidden layers, softmax activation in the output layer. Optimized on training cross-entropy loss.} - \label{fig:classification_hyperparameters} + \centering + \includegraphics[width=.8\textwidth]{../notebooks/classification_hyperparameter_scan.pdf} + \caption{Area-under-Curve score dependance on the number of hidden layers, the number of nodes per hidden layer and the regularization strength $\lambda$ for a L2 regularization. The off-diagonal plots show the average AUC per combination of two parameter values. The diagonal values show the average score per single parameter value. Optimization using Adam with $\alpha = \num{0.05}$ and 500 epochs. LReLU activation in the hidden layers, softmax activation in the output layer. Optimized on training cross-entropy loss.} + \label{fig:classification_hyperparameters} \end{figure} \bibliography{biblio} diff --git a/src/easynn/feedforward.py b/src/easynn/feedforward.py index 2fd4ee4..4c8d4c5 100644 --- a/src/easynn/feedforward.py +++ b/src/easynn/feedforward.py @@ -55,6 +55,13 @@ class Softmax(ActivationFunction): s = self.forward(values) return s * (1 - s) +class Sigmoid(ActivationFunction): + def forward(self, values: np.ndarray) -> np.ndarray: + return 1 / (1 + np.exp(-values)) + + def backward(self, values: np.ndarray) -> np.ndarray: + sig = self.forward(values) + return sig * (1 - sig) # === Loss Functions ===