From 45287f9ed7837d9be925db19045955807b5b210c Mon Sep 17 00:00:00 2001
From: Robin Rombach <robin.rombach@gmail.com>
Date: Sat, 14 Jan 2023 13:48:28 +0100
Subject: [PATCH] stable unclip finetune

---
 README.md                                     |  36 +
 assets/stable-samples/stable-unclip/panda.jpg | Bin 0 -> 175412 bytes
 configs/karlo/decoder_900M_vit_l.yaml         |  37 +
 configs/karlo/improved_sr_64_256_1.4B.yaml    |  27 +
 configs/karlo/prior_1B_vit_l.yaml             |  21 +
 .../v2-1-stable-karlo-inference.yaml          |  74 ++
 ldm/models/diffusion/ddpm.py                  |  55 ++
 ldm/modules/karlo/__init__.py                 |   0
 ldm/modules/karlo/diffusers_pipeline.py       | 512 +++++++++++
 ldm/modules/karlo/kakao/__init__.py           |   0
 ldm/modules/karlo/kakao/models/__init__.py    |   0
 ldm/modules/karlo/kakao/models/clip.py        | 182 ++++
 .../karlo/kakao/models/decoder_model.py       | 193 ++++
 ldm/modules/karlo/kakao/models/prior_model.py | 138 +++
 ldm/modules/karlo/kakao/models/sr_256_1k.py   |  10 +
 ldm/modules/karlo/kakao/models/sr_64_256.py   |  88 ++
 ldm/modules/karlo/kakao/modules/__init__.py   |  49 ++
 .../modules/diffusion/gaussian_diffusion.py   | 828 ++++++++++++++++++
 .../karlo/kakao/modules/diffusion/respace.py  | 112 +++
 ldm/modules/karlo/kakao/modules/nn.py         | 114 +++
 ldm/modules/karlo/kakao/modules/resample.py   |  68 ++
 ldm/modules/karlo/kakao/modules/unet.py       | 792 +++++++++++++++++
 ldm/modules/karlo/kakao/modules/xf.py         | 231 +++++
 ldm/modules/karlo/kakao/sampler.py            | 272 ++++++
 ldm/modules/karlo/kakao/template.py           | 141 +++
 scripts/streamlit/stablekarlo.py              | 381 ++++++++
 26 files changed, 4361 insertions(+)
 create mode 100644 assets/stable-samples/stable-unclip/panda.jpg
 create mode 100644 configs/karlo/decoder_900M_vit_l.yaml
 create mode 100644 configs/karlo/improved_sr_64_256_1.4B.yaml
 create mode 100644 configs/karlo/prior_1B_vit_l.yaml
 create mode 100644 configs/stable-diffusion/v2-1-stable-karlo-inference.yaml
 create mode 100644 ldm/modules/karlo/__init__.py
 create mode 100644 ldm/modules/karlo/diffusers_pipeline.py
 create mode 100644 ldm/modules/karlo/kakao/__init__.py
 create mode 100644 ldm/modules/karlo/kakao/models/__init__.py
 create mode 100644 ldm/modules/karlo/kakao/models/clip.py
 create mode 100644 ldm/modules/karlo/kakao/models/decoder_model.py
 create mode 100644 ldm/modules/karlo/kakao/models/prior_model.py
 create mode 100644 ldm/modules/karlo/kakao/models/sr_256_1k.py
 create mode 100644 ldm/modules/karlo/kakao/models/sr_64_256.py
 create mode 100644 ldm/modules/karlo/kakao/modules/__init__.py
 create mode 100644 ldm/modules/karlo/kakao/modules/diffusion/gaussian_diffusion.py
 create mode 100644 ldm/modules/karlo/kakao/modules/diffusion/respace.py
 create mode 100644 ldm/modules/karlo/kakao/modules/nn.py
 create mode 100644 ldm/modules/karlo/kakao/modules/resample.py
 create mode 100644 ldm/modules/karlo/kakao/modules/unet.py
 create mode 100644 ldm/modules/karlo/kakao/modules/xf.py
 create mode 100644 ldm/modules/karlo/kakao/sampler.py
 create mode 100644 ldm/modules/karlo/kakao/template.py
 create mode 100644 scripts/streamlit/stablekarlo.py

diff --git a/README.md b/README.md
index f413068..9d72b85 100644
--- a/README.md
+++ b/README.md
@@ -137,6 +137,42 @@ Note: The inference config for all model versions is designed to be used with EM
 For this reason `use_ema=False` is set in the configuration, otherwise the code will try to switch from
 non-EMA to EMA weights. 
 
+### Stable Diffusion Meets Karlo
+![upscaling-x4](assets/stable-samples/stable-unclip/panda.jpg)
+_++++++ NOTE: preliminary checkpoint for internal testing ++++++_ 
+
+Recently, [KakaoBrain](https://kakaobrain.com/) openly released [Karlo](https://github.com/kakaobrain/karlo), a pretrained, large-scale replication of [unCLIP](https://arxiv.org/abs/2204.06125) (also known as DALL·E 2).
+We introduce _Stable Karlo_, a combination of the Karlo CLIP image embedding prior, and Stable Diffusion v2.1.
+More precisely, we finetuned SD 2.1 to accept a CLIP ViT-L/14 image embedding in addition to the text encodings.
+This means that the model can be used to produce image variations in the style of unCLIP, but can also be combined with the 
+embedding prior of KARLO and directly decodes to 768x768 pixel resolution.
+
+To run the model, first download the KARLO checkpoints
+```shell
+mkdir -p checkpoints/karlo_models
+cd checkpoints/karlo_models
+wget https://arena.kakaocdn.net/brainrepo/models/karlo-public/v1.0.0.alpha/096db1af569b284eb76b3881534822d9/ViT-L-14.pt
+wget https://arena.kakaocdn.net/brainrepo/models/karlo-public/v1.0.0.alpha/0b62380a75e56f073e2844ab5199153d/ViT-L-14_stats.th
+wget https://arena.kakaocdn.net/brainrepo/models/karlo-public/v1.0.0.alpha/85626483eaca9f581e2a78d31ff905ca/prior-ckpt-step%3D01000000-of-01000000.ckpt
+cd ../../
+```
+and the finetuned SD2.1 checkpoint [+++prelim private upload on HF+++] from [https://huggingface.co/stabilityai/stable-unclip-preview](https://huggingface.co/stabilityai/stable-unclip-preview), and put the ckpt into the `checkpoints folder` 
+
+The, run
+
+```
+streamlit run scripts/streamlit/stablekarlo.py
+```
+
+The script optionally supports sampling from the full Karlo model. To do so, you need to download the 64x64 decoder and 64->256 upscaler 
+via 
+```shell
+cd checkpoints/karlo_models
+wget https://arena.kakaocdn.net/brainrepo/models/karlo-public/v1.0.0.alpha/efdf6206d8ed593961593dc029a8affa/decoder-ckpt-step%3D01000000-of-01000000.ckpt
+wget https://arena.kakaocdn.net/brainrepo/models/karlo-public/v1.0.0.alpha/4226b831ae0279020d134281f3c31590/improved-sr-ckpt-step%3D1.2M.ckpt
+cd ../../
+```
+
 ### Image Modification with Stable Diffusion
 
 ![depth2img-stable2](assets/stable-samples/depth2img/merged-0000.png)
diff --git a/assets/stable-samples/stable-unclip/panda.jpg b/assets/stable-samples/stable-unclip/panda.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..49aa1ba4d5a818df96988f3dc2ade49200c767ab
GIT binary patch
literal 175412
zcmbSyWmFtN*XE#uTY%sYhJ>KO2KNMacM<}G;O-6q5`tUM!QEw$Ai)RM!6C?CgS$hP
z_xtwjp0mGp_jYyv=&tTl&#CIV^*s01zxjV_03t<M1z7+J3IKred;tC}0p0`9P*ML=
zo(=uEU|?Zjprd2p0D+iTcsO`?xH!1D_yi<G_yoiRxVS{*M8q#h$;im?2wzf=lTwh7
zl9B%BBPeLk@1SE~V_;yD;^X3z{y*ElP5?0$$~Hy=8VVx-l^6w$80B9tfd09i7%2Z`
zfd6SIsLyT01Y%+1;67J?5&=+A(9lrP(f-rw=i2_y&jIMf7$mQ_r7&M;m;o7`NqK_f
zbFo0uwOwSIlP65PpIt(*amZg%P*S~q!~B+ol@H7>ASfjK{)3FHoV<dfmbQ+rp1y&h
zxrOBytFP8Ju5Rugo?hNQq2I#7BfkHLOh`;hPWhFZmY$biP*_x4Qd(A5-vDiFYHs=6
z-P7CG4;vU9nwo~s%+Ad(EUs@LHn+BSc9DChXXh7}SJyYUcmKhK0zmt3SkKr02KIm8
zB7Vk&ijIzk4*U-;6jaaW6O9-h;}th1iIfJ=%=rZ)PcRm#bbM}Y7dD7j^MvfP%Onmt
z6W{vl)BiyGFJ%9Bz(W4Nko_-U{~OmL01pl2x%1G70TO^)d(s#KL68vg)<<WcGDs7K
z7x>#>V%+A__hV2p#dy0><42B0GkFK^s_F1vS#}H?DRQIWvz_LRDlBZ5Vo?aIhT~?4
z^Lf)*j;yz2^!Vs_Kr&64bLY;!-XeyJf2=FSrY-(x4JCezg<uOm!#OiF-CJG#(Z{Wv
zknV^};ZTjM1U&^fVIIkeVMdH57U8TIpP=wWSCl1k_7YF-ERT_g%OglQxHAv&;$Z8~
z&pp@K*w;yC!WTm28j_<h!E$K=Amz9ZdZY#{ytDGHIV(5ZN86STt5d=dO3_96V56Jw
z6D!r;UaNQ-B)DB1<=olNGL}QocBl2ZBg`GC$-Ii0G2yC8R-kNJ(Zc2zDWl8Ff1op;
zf1YzqhSpUZ42m@@lE43~gqL17z|!dC*E3bA+ALV~Mh^#2T!JX1cH&ku{v2G*WlGI=
z#?uhGugSU-iQYYhnV{}$B;z4-qs_1G{mH_;Jj@f1K#)PZQKzA$c)gX|p5}K$8LspO
zwG2q4c<tyK>yH4u(y&Q3-2m6hN!NqjMPbjycV?FnJI~sS*~SQ4@~}hi^GHBsnFNZ4
zL9SX8TAVI9Oso!gS!7{l#WY2S|BBh0uEC1=Z@h0-b$6De!A(K#eG`A8(Gz=HAnQVk
zsHEiK6S&*U<pQ}l6FvTQ8~r{ed$t|(6kTil<)_}z%VVwC4c^z1vN9SB<U(YHZy)~(
z&w77^)Rk1Z$;=1!;>H0<ci{y=h^UPsKx8`Ej#z8tma{PFaATwF8ETM4H7If{r7yX*
zNf6v#u$;;)%N&zh{(CGHlri?*uX4<mO@C5x2aJ)J!oVU_c_^ZgHuTUcS1QKBL97$L
z9)D=^Cc~qKjFGzw0RofAwjp0wKaum)d?9@GbYLEOGyB^bb(%W<6~&P!7~~c1NiC2%
zS0fdy{_~e>^XA>T!P!s9J)xVrCqUtn52g_#h?&$Pil~(Pv(qOQ^|yeh*TTQ&#z*G{
zw44J+w+Io|K3PMF!Qf}3TD0&jG-+EqHx0FmFR*^akiZ3~7OSzZptw8$o>%qAwi)~y
zjzc*5!Q(~;%moNU-c`z#3{|0P&?mWzHyk%?;S`YpwWh&1B}<7xgo=}tyrR9Gblk}F
za(9V@Fu7H;Ab$))_`8G^=NXvP6F7pY5<~)H_ni&Wjf<YamV&3JI8e-ZI40e><<S0M
zc+bA1ZUG6&DjmNPjQMMT`En)pMyF)@p1)_*&im{n|30r}X*fLnv%+%i1zqv%Z!=s6
zGZr@<u+9QS6dho@hRESd(giJstu5j9pM*ar{Ip>fCPSc~xZN=rM(CG<fsw~MUbwwX
z{{Z8e*+U)tG0XLLRQJpq`plhFH+w1t&-WBVCOWu;-7%Z@MR@;q$({xj>a#ap9`}7+
zGCWi#<9*mWtPDSL4mQ*<#L{x6))VB2N!}UDJ?u^%xjE#oR^GWNcQ#0nvEc~kL49JZ
zTAaU&Ev=x{%HuN`(&YyUfgsRQQgoJE=cdwf7J#znI3;RDg~(7@mHrbH#+o^D9LASY
zq+Cm5<=9eIZ{%H`c<#yAbY!N+p0E_>c4;#Rku0CSEdRC;mEV%c!)_JY?9r#=q@PUh
zby=PeQUJjgb6(6*Ev6OwJ!fY!#{Yw(ayXL{<+Acne83&|@)E2%L{7nEE>=7B+k_u1
zYaBhgzsn{J719k)TFa;{HZ$KU%XE#JO;juk0~&uWRVaaxkV;d=YD$FXq*&;;Eg0vB
zR+rP8nX<POvq949bY*-e^u$;8`7O>-gcz>C4&G5^gsw=WPJkVvggNKfgvfyj>O4`T
zk8<Z${?kEKB-Sv7OkK`(mrZkaVaqt`lRI<HgN+rhzVtRnU&_X^6=X)ygeks5ZH4)H
z@JN{$C%$+1x%K9@7G8z8Ay|$6Vv{n55}Vqg^!o^BhlhCo@J@r0|LbmM<wD4OJgK+C
zQBQVv9p{_=1Bk!K>};+rEAX83gI?7c%0oFPJ;C1Z29o6k<h1}1D2bPqwjekGMNIxU
z@;E9V!(;G@<Kb<8b1H;4GFQNGq2=T%AH|V>%xgeIEj7&h!B*^I@K=R}o^&||JR+i`
zNrkeiHCoSW4nlxoLoLvGt9U+}e)4_(%iBSTzk}oJMyx*$2HeuJNG#|C>5{IZ)?9CK
zYW8rs8Q#&maV~yK+z&P^O*<o7A@DxdDUjxch(rBjP^fDmH{0a{x#0ljeWmn~F%6y~
zGR2Q6PzMl5h@xU)Zo1M<RsrHK-=GzDSawD`e5^I_jo#5PcuqpfDD!Wj9_1+LAudP+
zUXgi;ptY&E#gjrVX3WPFV*j$Wg2Opk;d$kMC`WB+p@Q2`#w8Um*HJ}bU+qjM{I6;8
zQu#aa8g-u=aL|-(18qDx`@IcOb?OQEjv0J4x8L-sKpG;#5rqa{Z}&BLHj<CXDb@7T
zU}$vP1>q7&wJ6^Y)vzqauWjNL8GG!|zb`C%;w4VC<M!*?TA6<yp7KhL5g|4%{HEcb
z{sF|%#E@->Y*TW-8b{x4lITsKSUuZlyytiYE*+GR=GC#-gpaNGhiSX0vf(dCWm1jJ
z^e^N)nzG&q+LTs=*yjawD+i!!+UyD&htD5HQ0_Xbw3|#6H<6~;pVi%?-D*$6o}#ng
zYo>gY-*IpygR8AFJjwo65HZg@H@+%KjS-;zp=ERu|GEHwFZ5P-aE0i1v=}R8_+`J1
zo?yvpx~(hKpB@WqBAB|(OdBQG1UMSy4lcH)AJz57)hJpC*&z)V6%bD@U<B?*M#?v?
zonBOY85>ftvr?ci(<z7uj`6ou*<f1|pD{qJWV%;*R5|AHwpC<mwN>D>ZsO;}G)$9k
zVA>=ZWF#=?N5;pvYJK;&8Bq^sK+&Z_X^OG6NE>J471PSpOzJbd7aEmg_a!uj%TJ|Y
zj!=1hbg*9c8Wq!~LDQ$DCgKU;V}0nEDsm1N+Ci$w4CCFw+i@dX&~uIup^%OlCUe}F
zOVM~#Z3JmaKJr+f--mnRn`|%5B@(I1O$J!oiXgL)rB2>`YstQeV8z7HR^f2A&&BVg
zTPPo<BP!jbE3PfEj|*dFfLvPC%M*A>RJt*2v3`&2@Fn2YP4xaA1cO)kOe=dic#SJ&
zy`%&s?qYE2wX{*~@e>V&(s<nYPN@wr;1|=Xft^2_Q!x0g*jPf(lM?<2<j}GyHJ85}
zD9p{UwwvYxC1$Qg?lt(6rpQ|A8vAvH`}_k4|CE`q6kyf1r<FaL*-KZ8pfuBc)iy%l
zJbkAJ&hW{8AC{A(Wd&&p`&JttGDJXTR5LDN2x~c+rlbgCh>CB@eS#h;<d!y_OA9)A
z3vS+<H_g_9v#5_nZK2Mcw1M<KekB(>5t=^`o<6vY++RC^M+w5dvy}BtBUJ>G@Ob!K
zx#>JsFgp2R%*@ZnAGE!7j?_NMEBjBrmnECq6X&ez^%Li_b#E6a8U6u03mvTE-z74N
z!v)8V4i6)j6MKe4p-IUOshxe2i=Ubj?Oeh+#(Z%ChG6gNoIQDus?z~^M7TUrCHu(8
zf?s+s32-wX=5czOoQ*V*_AD}I-8JYA{{Z_l(ah4TZZ(5pOqpbq4|@nTt%N-LxTR2G
z2JVz|3O2jJrE#^Vc59m_@_T~xU#_CIq!BLPyAU-4hY#hR@+X_={sYu#IU?a(gJ#Qt
zVFJ1-aw&mHmGvxP)6Q(6H%>k-V))v*Lr_}5mhoiJXilT|R-!X|WG^NyVY{_AHVu&3
zLuj%2;z?&A7Kh>hmkcKZEnfgPv7bi*XB&!@@Ryt&^*3jCp{fhvEyHyMIA%+`o9SA}
z>NB?41Y(OZ^^N*wld>PIrTkZ}JiSCIgWH|1MTWsV0s;inz|wcmhI*^^eE1Kp#gG&7
zl&6_jU|qps7o!m8uF@%@*k%B=grafp{j0U(PG)wsePw3a$I?I6T^baNtAGCi2o)jd
z1XUwB7Mm#3NZqt?R!fD=#VjAsl2peojZozAzFr|zrA(f(>>A5_U6#;NtR2Y@!{SJd
z?EI8CU0o>IXPFcvNBg+9Ct1e=ICK!c5{Kd-?>!Ff8eJ5~I2ykueZFM7?IaODg;5|P
z>nP&K4J{_~%)f5ZSXBx+H_=C1zkpOidi1TxaH|DNgZ}|=q+7r(U7tIasiMOb$_cov
zGh8JrblnPwPqiAnN!i^ZK4->Xc9tx7mnXa(ETNPXV~MKLgEVNJ$^iUXNk!-gxUYY7
zmB$r)WZMySzATYWGibuI%KHb1Hf&Uxq=|=_FSe@jWlGbMs~7s0EJ0&b)$FwLm{`zW
z8A3XV<VCN=c8OJ;L!0q-nhLc8J63|nO)otxUvu9^mQ2Qnr~(j-K@{?6F$JDr^eLYB
zP*~2sR&5tkJnXD2fkX;M_WD)fFwv46P^dItT{|=IvMe>J{PC!mmUUU{h@O+|=O;Az
zRx18>Z+cGWkU+)X>LlC)qNjZtKSI`aoXRJ%I<1;7k;?JGuH*Yk%fx+@JGAZAt_`1v
zVU;S8s(n9QA%u!q`QuD*B6r|y)!PB2?spG19t-o!oHv4f2E^Bo)_?WEd9v)aMM3no
zZhZIOxq8jpWdPWW+3yWs$`ep{7O~|)>kb$zl{13GY}(Azn_pgk4q`nppG$Yiy(nj)
zgpsi>Z6_a4C#T?@M<y(iSUlLw$9@$oeowc6$?8DP&JWT-EubMYTS^ZP$kE<KV<3&(
zbxWy$y`Q!mu)vv(`umMUHV~6kdN(2KAHaPw?g{iDS?ndCQ`eN@|4!SPJR>m3iY>cd
zUi_f>g7Q_+UcsH-`;9;0!mzLkM!1RiOmT;R3idLHIhV;@m!_Nm>q1ngC@m^)wwR>H
zE_2#BrFd{`%4a)06Uv?ErG2%(*iJ4I+hpD78MOJ^+a<6~;!&q_sfw-gBy4D)-a3IF
zJ>T!cmGcvrJ#QZ`O6NU+VI`4vTn~XD=ytjhvx{cCK3*aDgFemlnrc3okhB=AP=VB3
zOz5i0n=1G8&CCEL<8LL$i$@W&Ex0aKwWEfi&xdTQOco(6{ejtcnz2b{F5IE%Q_#e%
z?c&)3(H2oS4Y9!qZE>Zg+gH1SuEz8x(yoX%wTO(!G%L2QHEmuQbN$iK%u|iq7Qe_G
zhX9Q+|5A)GIb&_U;fd+<=6)aei861H>t%&wgH?x(yhD>Y4ZZQy43J~2Y)T{ZN}|$M
z|8Hv1zyNKZtA{2DDT(*7ku#t)F>0rjbehLm_}+HQYo?AnCbgsrO2~nQ?r<QYCxL;C
zmaER)jx)xlj!~0V0~hrhH@;5M0;MoToo5@`3chs`xg3>tagB$Wr8{ZnTZ<C3tm&36
zhgXmgjmT+fEf~sCL086yNnFa*wA7$b-Jgdf4=F>u*~jU2b<_*wvc|H1T(}=b-TK*F
zA(}06njHxUza#D-Y36GgizP$e1>l$Y1<sKbHp4Ua-p-R=i_-~RKbRXT|1FV=|GwK~
zqVmr1EBVPR{X=GvjB-7WlHs%v4_ezm!w#G!IkpG6Mqn}0IC03*!yFE=Hbg}3nLkKt
zybA&6&M!7UvTY06b1z98#6T}rqd$rh&cHAD#~;dfzMFN8{{=N4u#NO}=*E@~Ax5!%
zo<a`}(NKcN%?u|O^VvZ7Yf)rkdu-Yi1N$rO;>rF5g!FG|@8`zHJgwl<#ktb63l;hw
z77|we%m^Jw{I)rL#T~Vgd^Jd^DdD;7`M`p`##=LfQE@F$_hIs3d57|FUo{>1No$(+
zQLUzoXC<sVv`I91)Y9Z{ihm1o6)HGwpEl&d`Pas>ONe3@Xah?)sS!Pbxjv8W)D$76
zD=`d>^vx6oy5wE0JQvl%QDBDmJay<CnZpd3G4JEiu>Pb<IQofiV6>4^d*<B*u(Yy}
zyg^pMcy}YsvqTqIa}BoJ`$?4*1#EW-dHl7_O!>$XTWW8d_>Cs<5yMs3L3xUUL$!`W
z<MSIiA+))ci}J?LMpK+sMi%T&Ww320;WZBpW0T6WR@@)`Yl0z!T6?zA__{`Ey$Q)K
zLO20PnA8Pbc9s>Z?j)^x`kse?Yj<1}=YY{s35B=VW0I~9FRv3%&vZ`3dcYTRrUdU8
zHQn;7U;B0n>V@4YTCV_juJMrKsb{(d<x75VIy)=!;oM>VEy(FR4KHD!l;++Y+)$?3
z_EskEM0Z7xhf7TS<74Wa(aHYW_0Nu182%((DI2<|CVHBh9F=ir^1$Oixiosyt2h$o
zf3Uo8*1;+!g{Pk1g*bdmU8%VDYsJJ{kjB&MF%?7l&6>I^**p)zH%f@(?*nfp-8@g8
zk~&}bd+`sxI>J1Ps22ZYczsL|(C_m)EF@Y>udNM^{f_-4E7C}4%^d;X?norx$3c&l
z&RLg^0*IB5Jqh?(pHz2NV;SFs_N7;Ns25n~Ht1q*AZvS_nKcy<Vi?{tR=W1m2{RXR
ze6{rf=N@wS{{VQcC+AXnExs6*Ka_?!Hq+$cXbB_B7?jL!+FJYQ8^#YFPoA*9i8F=v
z3e4iYQo_d}&+1~n!A9ZY9>nZ;d2@Yl;<b7e4(ALnE>X$3F-iUDNPw(waJD<#;o8;(
zRe4>IZA`4faIg;Z$WLQBU2o@F<iZOLS@nmaI|eh=1baDHh6WyA!vpP(gVQfiUi(Wj
zRg`J&D}VlsC&ZlCc^s-g_e6CQcnVT`nmv*DUcrf8qQ5Ab%Y;rRrQ1KAdsRLD4UTg%
z+9{@6g-qZ2`@{9BT<9AQ7i%Z}Aj5^_6T9y<(&Xdb3ajFcPlDRGwhx8hRfAnwm4(0g
z#&LvUb~y9bYIckFS!J7kRcb25xm6^GuYX4IpmeEo*%Uu}dmAy$)uCy>4?8;37IBn^
zRRuhJJru?w{wu|h6}Cy}DuL17BG=4Cn&};kGJdl&;SETH>EpUt{0*`3<yYM<5uN)x
zdW~=hJx^#d(3&&imnt^X<<+QWTDa@Av!QSZpRz0_8~IiYzFW5+EJHK#=T|!83o!Ao
zWy1d41WhR+q=smi(Kfm-*uJHY`%GSe!5lbSCNG=`uYBb&Uy{~R`J#64ZL%TPviT~Q
zTDs*DR%u|#?q7-*7;R6<o37r<BC6Oky;x=<mBdWj+-F_)X(K0%r$T~tLB3mkqC1_c
z3^`cooeMKfmbPG+J@phUrcUPH_*GLJ-$2KbHfn30k^Cd(0#)TLI{YWy1`8FFTsUnQ
zI76&#Bv%qdrqj<t(EDx<{KEgZyU214&dy~M&w-B*EpwM(fvtc|ll6<y+BMOwk!RVw
zKnPo|ZM^fET!ctEZlT}>W=qDK-k!fuw~c;$&FQk#hY@;ny_Y$V?Jr!><50;N(TJ~X
zmtccf)ZAC|(l(6+0`r0w4Dij%BIOSI`f8V=kl(yKGqsHOA*}nZ=%LL!WZ7t8Y>B}T
zs8ekZU0SK)63rAh6jRHQk-j}n#s*2aZLLnXmi$moetTUbhN)uhW!*GClr|`bRI$#<
zyW|{#V$;n0GQKD=u$kR<4~$^s`<Uvh`;EkoVTOz*y@hCOv3@`{)5SSm{eB5psP}$0
zMS$s%jlaqsQ-R5<gc3}cxGOx9`s7KMY-UDts>+QGxBc)XRg%V8BlU+_pS1!xR*CW_
znn)@*h%Jv#@+B?4VtM>j0(hs`u&&0c&?M?_MKxoPFwJFI{i{mdP#T1&XPJxCIBIRm
zjxt&DkWHprPA&kaG44}nCusunh*yQ3-GbT|HK~}3%~(9a(#G?t0(qF>kSu(n-?%JN
zi>Oa+Yg0Ya2clN^8DTyV(;ELF^SrYZPY?}_`QcO(Mt&yZ&pISE4z)$Mi;L^VDM5og
zZC7%nqD^cRhOqOiY3NdLqe$M}&l!IvN%)b${vi{Q{9yBxR;U!{R8C|08&w2vHYK@0
z!JR?9fSktakC=?lYOo*P5mB3kMO>#_^?Tf42&0}aJ!O{jYrHZDa$u{0KKy(Y3*M_L
z*^lWv=UCteY=dm+m^-<6_sieSwcb#7&w8Zn8ovAcVgUD{9Nxe|wK}3+#>m|97rLuG
zXxCB|VAH0oUY5(kv~d0uaRO%N2wj$?Qr^klG1T(HGRh04MuBh-JoBizX}PXEBPYB%
z0m@x670W$wyr5S&!6ukSv}!kViF&w$wRhqg{X^xZOJDrPopU+91!$KTiZA*2I~$LG
zZxej2o2-Xccgg(l1icW4uFJ0&Q-u5a&ZY0A;~E(!`FZ=-IpKR(%b8y-6L3N}aokpp
zyf6z32?e%)iiZ{j5?qFHW{9$5X_k%wO@;NKdjT|GkL07!1`p#6>b<O0*HBxSLJXWR
zW%y!MwY{NM3Cp{G?aPAEtOUnZ{cYOBDcP7Dw;%-oZwbRo`}aK^rkVX>dGbra%`f1n
zf#W`vwuK*D3@i*@c*xN_iv1ZkoMduQBc%N5#N6Gxlyc_bU{`9M;IBr@l5$72koALz
z@DHo^ZN%L4>aAYW&AV>$U$L&<>{k<gHZdfN*UsmZL1;F?1GL_(D6n?%<F1KigL?T|
z&=C&ar_A8Ez;XIQoiWEUS^L6?b20u!Z(2}vV`m$0OpNfFkL$y}EUd?A*u{Z4I+gsj
z*;@j|yKHr!n$ND0lDtC*mrj!b0a;LgAS7@c3fdL$Noa^u!3!G-bvXMTYO&_t@4O%9
z>sGYN&4z!FJVtL4!xtkPM81fkqs)_gcjY+Mme_|kb@W3$*D=G|Jm}BU?<)?qo^vDl
z{zG))+%GySJ?5F8=zL-_an2u&6u;mzRnj=<I~6lq-{*@xK`vn*KRD|YziAL%l;Jdr
zLYcpMrfe9^?wXjB(-V3+&g*L$Eo)qq3}{aQP+doYl^(l#iZQ`WC!Z6wLhpy++-(<V
z%Xo&nER3!NcKfi0D!><=2nlU>tGNl-JMxG&XM*s^avL^g<3nFc`Zc<vF%N6OYB#>E
zN-ffk5p#4&xDk1ElI5Z+VAC@Xj}h4B1>y}_f-os`m|Yc;@2L7)pjk{Z48SfC0P@Cn
zre9B}Y|`3<lru<~*KJ%)&1Wn@`s{H!R!woq!?OPZ4`_UGr^OWW_T$?Y5!Zn=e>UA|
zq0gDRC8c&bCzu3cKHBw5@$>+<J>ST5p^eXEPE0T}HMR~>BEWH?3W&0C;V0`>YJjMF
zNeB_Om<^&2I>sYHgA+DE4x&9zw@wtX`#|V7J<vIBsK4<_ULs4ihM-;dH#rQw!BPPI
z9{|!&mS&m?{-WeOM*&<+W?Tfe>wIr9eh<k@W3&g76__t3%UbTsRA4U`Y?S}roM}?5
zLSV+r_-Uwu(pgY04ssP3P>?Xi1ZHCHW$XdUam4@;bNqRn&pE4y^FDLVJ1_8IuU*Ie
zW|B;Lgs%Sq#AkHHdonL*F3H>h3wvTR=*b0iW!{K=9nR0Js=$*?pV_ECte1m83Nq=<
zO`^8*OoDHo1)ZZwS<kL$LdP0Axe-4<(?_Q@2#rn+n4bUk_WR%tq-}}=uTk@;D*|vr
zRs0jg>Vze@kx$PUMIbf!1Qk`Jvh2KMCMEr9kz~r74t-&-w^RJdCeFS~N|1iG;(hcO
zS3tYOqc(O5StyIB`@vgHHDChE%OeEEoqCQHAIY%#i#C*TN^rode*jdZ!3if4<uG#c
zyG3wNM_z+0M*`lbd(9aC(t<eap<CprA(Iy(ez75+&c4H+jn|Xz&sf97_Hrs;FOOiS
z7_Tn5XyofobIm0dOHg&862;BNRk^~;fc(X5JV~6aeBvTD7V=2WY_Krj<65b*w@{%B
zHL?I*0F8ikgw?|hjLYJ!1~vW>pz$7+t3_)}tbF2$*3Z2CL?>qoOd$c~Bggo_iPpw&
zi9?9tuYHP)TvL32lXevZUA7yorz9fSUX**WK2b_Zh$YpTsb8zBy$xIl?3~NkGvv?B
zHHB8?TZ5DKA6~pNeiWhaF1I#}KKkL)Waz#;tEco2!12n?YCh(hFU?$?7cYV~rIHc(
z#0w=cIDlo?;tW$~{sWLI6K)syvuR)O*70<Z`MrccfuvraW+E3qXktB0Z&A|eN%T-%
z!(6<`$`SOzrVJ-K{{VL{ILq-IP?x28A`Woqqy!J-!N*S|kg-6Z>GBP(W)lRHarSRN
zXzRU=651=)6()v6Wamap_76JYKR5pVR#r}>6{08dMY39TM1r^a79pLezGzc*D$7k|
z=3U#(CbA2Tr1rmFSxG&xi#$n;r8#nefcrOE4#%^EFKmEcmTc1mzV)NXp#(p;5ZJI~
zc-FH)n?3{}?tQ4<r1NU(fi-k~YnEZ`6t`<ZTe{C7f?#_GvY^gB7bO+dlKDhIJ71k8
z0aSH0so8pit^lkD*6vZLq*mr-#a5}27(vNEQ;MMAYp*KrTGDS;WaereXM-5cj)s-~
zjFf0-fAji((ba_^`2N`7`0VU!7bJS2yB8fZ0rb+tBj06ocop0$`K0GrEEY52geYCc
zf;{Qq6~>+|<vZ&wrR!PRQJMZiRrXL#-r%re4kUnPyRff5z#`@hrXTZ0n{v$}K63e|
z&P%N3&S&bF<sA)OXBVDRDp-9H3;`;`p`T5QZE1>cY7}<R;ltd5ok&;_8{^+i9-w2=
zHSH6$wZK|WxE7!2tTKYb=^wz&U+k-@$^y;de3U{8ipz*R+Sfs5q~Qc|*dfGsdx`!;
zvN)v-p;tHUDO1$nq=gpImh=#GPxlWX)hA;s)?e<j!8on-pvAT3541TdR!cU?OlEcc
z`cr^JA<!u}Nate<&V|OF|K5d<D`QB{_P$=jK9)T^NJ(KO+IkSRLNZFXa@AZgKlXQo
zjl3G|m_OT(8Fzrv_a(H<KhLo!&u13+kyCW*ckrW%cR-W#Ikl5wOdY+<#ocVYtc%b5
zSo?2Y76<78r!wES$2YUyAP=A<tHEoYS{E+vppMX*vP~;+Y2GBo6L_avxaP7caLW`(
zw1|lbpe20fWiP8AG=D<^cUJx+dvPw)t0|sxA5ED?%v$^d^!%#cH4LjkeVp)T?QZTG
zq@W3GH&`DKL_&%?SY=z7n^x9*^u8lJU6Dt|2|DW70~d<L9OGi%mt|)wF9lboOE1dw
z1iUfifGw3E<sEn03+9$ql%PF1Ww(k2%V9ql1WG#RFB^DwDQ!JC07Pr|>`o9W6#q8&
z(p$5{iBR|WcILHTBlegV#pQKLVWQK17=5L*gA0G~z@Ci9Ym3j<@6sCm$~uFuGvIfG
zHdObX)EzPp=LVW9aokxs7uO}TJDv;(GSOcw@^iK$K?;jN*Tu5N+-(UoP~H8kU*Z<d
z^?}6rEonTr#^7!1y{2=m{roOR+K%4qx}|_GW2n`0HP_|pD<MlD^qq3LnaE#q2iqLI
za#_9aUXS14T)YeWh;{i75czvqcUBz?Nv%vfrtvG`Mn0$<8lTtceTnygdj>!7Ry;_~
z9X3c1E<z74dQtCPYGaP(ck@@LIuJJk_Y60lUE-?68$W>Gp?#8hl$iEwVn#Oq052ta
zecY;g)}Ff{i)}qN+`hpcb%qf~VyZeamVf;ppzUx@ymL*jz-Vdk)nYd5CA;rjqGox@
z_w8~+&B?FPZ1PjoZXpy?f6>=LQD&0mpAb>b8eMi8K#J0bn*8lYwjY^|@`I=EzF+(U
z0B;<1=AhoQb4t@31+6WqnfB2yndu%QTB^I$VpnP0KgK@Kd+xE+`>#Vxl(wv|EbAIm
z_Itkq9fZUS(jOK60Yb99Zrf<@c~^F8<R@2dAa`s4QUo{iu?6MDE6M8qk6gy`ia8&c
z8O6&a?q)7q#3`l4XO(olq1#(*ZZh!J1l_-#_G05cNn;Ab1S@-G3YjkkY3QXca<>LS
z%rQQ$AGK^Nl^LOYBuFRMOgc;;Gg8yPJYX&0T2%~KX#PXo0qfyTYj(W}M1X1~bWzdh
zDSYe3pJ@AzS(G@`bmyMyK}T315*PtxWSR!c5Y?%=ru3sefy8h7{I1DP)U*HEtwlm`
z5suEJL#v_=G=zcGmsNHq6dP}cNaQ?#WTNDZ;QYxa;Oxg2ZCp>}bDN=NL?d{!2fqDk
z^xt=F$Bv|cj)lJ4squbaZ{9h2f7ur=H>1-bR`|3h;Z<-)JaAc!2U$tGLL_!7yV$OO
zRp|Qdvhp<`KBlM)Y)GwQLMao;uNQupZkH>)-U1HT>ltDQ!OOy_mbto{HTi4yS-1)6
z!<BU<iW8RvPGF;ycaYB?!*hn%6cj`y-lj`Z3o?{SWVtWQpDnex2_Q5}fW=_LHyoYF
zeXWuK7$2k5w1?MbrnAoL_pLF>?)guQ&M>e+J;YXfq^N;af&9dl^Xt=RLMfv#*EkVf
z_Yt{3F}A!6!GfqCoj)jRAv36NF3~S9%FkfUG2$A^_kvac|Gm}pMVf~uc%&e97nl>3
zprus$A{Q~L^u<zlJBKa(w|NOutUX_6(QwyAiHWoBLcn{NVHJYxYl>91p_4<>?7qD4
z#6+6Q#KUu(z;3qy-)!O^VEAj6c$+TApCMb-sgG=$4RIb$gd5I!wAj4!JCIwUhdmYy
zgSLo)_U738I<7|>@|;N`VtH|sg<V-zJBt~kK|GHsft-zCG2%FwHB8lJ%DB8&2t72i
z_c!MeS8vNsq<!1A;0pG<nx>3+F||*^O3+BJ3(7c|XV0gC#G(LQ_0>wVgCsZRJ$%#2
zs?q39kIz!;5l|-LWu);ZIJ;0edUzL*;X-_dk{*oftaC%rr|Ml!#60A4S(4;ly0#kt
z&D}KJP4&wyHQP^@1p@!3h_f0feIAjc+#wVGBM#N&U79VVH1*ZVXL%vCqCqyUx(v^8
zV@{W)(Xg#iU5tq#Glu}m-{$9lQ1c{pq{RGLC6`ZZD(7cZcs$g&)k?||I<+$|%HwM(
z9$E@oOpx)Kh?(?$Pg$#9=TwP<=)UT^lV?)Jub5*Me=;6(q@hXk$O^01Zd8hz(`~1|
zq|qZibp4$GVv{sBkIXD7;*>+9eA4kNwxOj_rq7IQBI=hqoXx>exM0veB@o;Z8?~k=
zaAVxmh&?H9ybAXY2~PZ5X3(Urs^TV25i}}fHb!AJ_ptqY7Y4CmKRXv5#}7}InLmEO
zSpOFIjyvGAhG%uQBl_fHsIigb8D%m+WvGo%G61@rO_&qwW)P)#;M|lu?n3DBXuE{?
zLky$gA7^r+Enss_5cA)^VCSHi4%u3M0&heORr(yN)ax$x-1B{3WUOe<B#vx#EFZVx
z;-+c78nxn^bF~;_`b4lCr|_}LCVS}}JF`GaVOLX^HIda9A1m<k0iyIay+_p@4t-R)
zKJ|xZf=9w)Wtc1NtW7y0ZMA3;D6L#;JaPM7M0TLw4*45dLaECmCO=j>D1vlWV#@q^
z7hC+<`u^C#Co2oTe1mwLxW@1ve*~wN7!KM5On0Y>DjSuu2+;HH8rW-TG@hGd)Wat3
z^w2gQu?tkX6od|h1i>qQye+?UR5HZX5SoGCJq3^0=GuhTeku+3lXORT4O_T_ZuNTf
zrmQZR@yxfr<=dps-!iVQCqI?4sZjA?459XZqHDjQ4Wx?IxbgNx9ke-ZefzvTccq{q
z5su@%p$}j?&L9L>r$nMImD3T5H<z=maWK$i=$}6w(;N0=am4=q2hewTq9qK@xltv-
zGS^owoWG3@k*gB66042!6e?x9Q4#53f@s$DcfMMb16J|g`0D+w^8zWG^;NuwaAmQ!
zLC~igi8!Xkgk5A^XX~%h<LX;m=Wk4&s7z>u;FwNRJcppKxG~-cArM`UgrTn&oh>cy
zcvSteXet*QeUine6qpeD_1fo-**`b7MXN&=DNO8=G@Qb2YkFiG)xq4Ra(s;f1c{AK
zlrhWkF_I60SLIsD7hiYy3Z@_ty%5ku{`8{Qy&+TA9#AR@EhcSzYt&`&=P+&5q@mIG
zGvS)f6$qy9lc<m01Rd86c!O4q4rFKPK(HghwL=MdxTt!ee`7CJm|4xFJY8?)mWG>o
zclGQ~GpE{Oi;GRebi5RJduO}tr9Bm-a0iBRJH`Bkj+c;{dmqs_3s2_fDxtViuIS}d
z$f`}%&GR-+A&m)0n_gKq&LA0FeQ^62x8*$nkKh9UxqN{%ZC_svn=rhQxDV8B4Swga
z)|d3|3<Ea2O>+T^AuJs?911wA=joNs{L}(DR^vrPi>x0c<e4xj+11vPz*t@v*!hbe
zD-%|_8UE^StUI|$W%ag`H03!=oOGRAm6&Rule+PDb}uaqEsY#n{_@#-*&kRc-X5WN
zMTqam_ETrhe*1Tm`l}u(B0v~g3mx?0#<&$9hi<CsSnJm4$4xPhL2K96@N#HF+rrte
zp1=i!;mkXZs)B}lQR)m#nL!YRAjPi#{xaX6g&{V1^HusRM$U<tHu|2aj+&{5*Ve_-
zWacukl}@B#s9aODpcG^VTRrspvx^SI`p0actv6^*BJxjt=I^lf-fpouJl?omRp%Js
z`_AfkanOTUc6XDq)wjG4`apHX#9hK_KY!aR<~~yzrSC>Qs0E%E)hQi9e(XQBr#R&n
z7}I$Q&b;5xpDsOYH=pXWx-;%+=~|U0I*T&{zw3*dZ^uG*wo^*^i2G|9D4e+T;ua)!
z(g#|+89~lWRZhZsxnAd;uDdUf@orz~r_K{<855LKM#ns~{e&mN1GCT5Zv%Xll8*w$
zF>o4Y6djPMdam(|#An=~qpPa$A*6%-oT%Xa!exd2XUsLqmre}kv(U9hb%9p^x*$Dc
zCP+lKX$Ev~JIHEU?9rwmp42>u!1r?0=1>E>i#oz^JhSQmWn8VlNuzd$|B8*>+(Rw^
zYWIPE=d5f$PLNF0g|^*`Y=&ANk+RMd$mu)d_qnQi4BK%k*KH8kgw+1>VJAvQBy6Kf
zzQA=tM?rxh=5JCeqD7-m&waa+3z1vGWH>TkA}5}tmFoH&f0tqk?hzVl#8h)t(YuWS
zFyWcg+JN2~#CS{e6TVk<(ou>t5Ftptn{(bK|3%`0DZl>bZgpMv)$K2u7g6pT<+A)$
zLU@vEG08TJVa}6XgBSE+0I__|3K#<krmgUrxf8{Mt;Y7_Qi#-;I8+ru*@=K_*1WMW
zeoJ{g4pu=hSnP;*Bp36PewG8#KI?s@t+c$$ckp!5$fLzVSruhL1{8>xt7<YJw#8h5
zBoi@L3`Jws{hMR{x6&vdEktnB>cQWJLftPX6sFkxGZ?lbStS@9w2>m7KO52l<lCKV
z5@R75_86epTtnG6<d%bafdmndkaB|@R-m%a%%*D3*@3M&CO2aFS^DLAGfS8345ENJ
zzL&SZd1~l=UoSr6(RnCHrbmfftTYJiNKLcRX8#8;Z2FoRguzuL*=O7!Nul{!xhDRb
z=(*zFv&f?EDW(Xztv66|{hGYuQQR&WRG+DX)MJxrpw{1X)opy$u2|bC$6VcRfMydJ
zm$RK{#+ch0jSr(N#ahPTrIb+g1g-v!>%##&=WHCv_Y*2Q_zOKHxp(fdj%-H%XNlc-
z|9-|SlGfCpS^0Oo=HvQuXHRx&G9W=#5EP$&r`DBC^dNF7@g8o+xp=HoJ4ItmjZ7SB
zMw4LKZ+#17dR+g=xXdlT(AAH^5IX(9wmZI1zCPRdTeZd2{}X#9m|`Qc>nN_o1Eu4W
z@ZoLiiS+{5N#*aLasH{*`s=xG0k8@q;nm%#yxt~-9+bf-LFBdmThLUqk<m)o=VVXW
zm1zIqQ{L_~n`}I$363{WCKP7jLSHs`G{O}wkRNH(fD2V$!qW8n4>BgUjDVDM@VwiU
zg3-6_xL;OsMRG+Dq6gz3<o-pD(zmO!Nn>a$Ai(0{qI^<`Y?dq)6yL8(xU-5h7<%39
z!7}VN9F88s3Wgig9KC9nD%bPzt1W#c+7%TNdK)>!20f9JQ0(g6RpC|S(6kg<$*$C=
z3LT9?oK*B`I_u_|dRJH$_5akoTNcN7m6cH*e^GMY$D4TxPJ>1$ghnOaq?Zc8-rqzv
z=(z+(Rx(-8rIJok<>9Semy%0&67)3hE792V1%q^xV=XAbt;ILeXYM@4d>-njDM?#@
z-aWOOWXuaqEe|!t?S$A?ois+5?5jM5*-fGJD5}oGJ%=F@tOZRSG!sm3dCOJ^f*y9U
z{sAnUZ>MZliNQv@$#}?O!x-Y17_|pLw9EVjex-`|UFwT>Sbfwso*x%~9bVfVhU<Zf
zTm99@ew9B+wan^Ei7)wn3Q!TAi@iVmQ>=L2%I3Rbs&aB1X8bVK#4~s^XP^5OyXbdI
z4~_qU>&^Y#2kR1>F&4Os4=s6c!ScQc6*4TV%tgaDR?X*#!<1TIG{b_ZvyFV$u%%rn
z?GuSg>c)Y@_b0M?YfJaCx?k_{JH&N%^?fIq{iBY}U6rIf`hFb|?6jM<#hXVVg1pE~
zDfE;ZMZc+wN3zA70%K5LhAOKyn$X2yntyy^Ln?e|ozF%18?GvwMJahQyb+WOyo$|v
za^HHXl@aj?=32XaxM!<m+!D_mUzj@XE5IYK8pzrg-~9%NK8qxQdQu$W5(#TL%L~HV
zMNe3xqhcvyZwlSj?$%a^+DPTXR_3ps9o=z1{TZ80;APjd6d`*eejM;DgRg&~r|~B&
zAN&(0@o}P(fvpulB=p%=M=uou1VIg2{D+7GHnQ(`?`CjQxUD4oo|1d8T)+Cdq#NNu
z)*dHZM6!Jfon7M_=riEcR&j?H%(G=8=26#sw!cYZ^f*H;f7J*Hb<#l%HU)RG0}Yi*
zdq}vW+dc)d7Bh|XP>#u=P5stnL}I`ZBM5de3?1)C3r`@#x?X7$@94HQ`#C9GRzZn|
zPAt|CN<LDsz?^O{=JR5Ds=z=1B}j26NVxW*#9WKDR^6|Ce&7l8xLT>w#woqlB;Av9
zryK1O!?~F7l69HLa#!Ci%7;T7%qj^&#uVj9=PynDmcYD`&^{e&`2FRV!}N&WNc`Yp
zJcDTci_WV{L9(c;R<L2aYW-SMck6hhI1~9GV^^_ZFvl`Xf^`S*$>&aQ_qxY1g1m!z
z%I@Hx@@dqPB&dYoS>tR_Gu(Ls_)#>RFo)aaEPOW~`Bu~A;h-e(2Wh&BSBiy8@w`TQ
zSv}R^Nvt!b`vG1O4^H_-TUHhXpd})}_}UGbFH>--VK5);=Y~Ko=HSVg0;oQh`^HaW
zj;=pu4s05}H{;!kXCc?mfX0x_&n*`4Bj)CJBnM$~?6OMPl44@}vJ)STN9t6#L)cw-
z0C(Bvkx=Q)dlka5@f-G!kvoxr5yI$JM5#xP**gQASUN*jg3OS$x$XE{S)hXi&i(Nd
z&Z?_z;Hg|!qTvX0OG8b<l;1m?ZWUy<AC3892S(=8WnF+>X7ESvkO&UGYzwD%-i@N-
zI+ANRlvgp{Pf(IZ!gdh9n&@2qZ+_?{KFqUSEO2qlhC4y7v>5X?+y}*g>K%`!$3THO
zz@yYGk*A1oC_vLDUjhEMvoaR<>)C#KdwZ08odSDgj%WtT)p)zPt_^W3dgj(Jt&#ZZ
zW2jr2dj{?<`QFA~A@yu-=Z3XdtsdV`n^fW=eoQj~^j;wd^^k=x(++7rAXwesm=jQA
zTlY;7_o#k?PXi~7CGPmbwg$KrT$LkCJV8A20a4iFs+py`Y`e##equFqcAR`mOYSX^
zTUeiZLSY{(1N0*M6d^qB<`qV@cOLX)s#$N>Sud0}@3RxGNk|=69M=4I4i63LKQyx|
zz%={YeG$i9Vsh&*ztags7fcnoUU{EpyAT$7%VgSuAEfQP6+QXo&~dJ<Re2*b=^`*0
zIn+`<P93QfBe}anHY`>|0?WTgh0!(ak5YbtCK#i;@4SH?ZulbWq`LOxd>p?YGfq`R
zy|-PK=PE~D7E>j-Dqv3*Q#okPBgo%iNJl}6A>8{Hr4y$?fx=Z8+0|%5LQj9zwU|{~
zF5k4XEuRR#u)>9j?m#XLPI%K(NxX{M0>f<5Ud#toEng3mSjD=$h|=lKek;}o{`J=X
zZnnmX&5$DDgQtI>V$@qn$^GTKnb!DlOZp#|W%o|5G09E*#Uc}*nL%KXtPo=&8km75
zefmAqFU{6o87aQN<d0@BpGwwEu8ZwB5mIHpPcsJ7cuuVMF}ylk#?`dX5=kT`qsm)J
zai~O+Jaf&S4G+#~WS%~!ug<I>SCf;}qfHJ?Boq{I=zOE4xE3$omf`+0blPL?p-&wY
zdFc0kZAI0<?Y`dTRL0YqCelQHWhrvJ<Ag5X%yvhSOgk#-x?oXxMq0%Ll*SfYXfWrd
zVQz*ie%v6RCl~wV9XjD}JsV$P7}k;*u&EdpjBIimwYlloVum{k4w8_e%}qS%aCBJ2
zxm4*cY&YpfQ#(3*yC`BW6hF{-)VG*Dyh;YcEqi@mKITV8&b^OH++JY<%~iU+c5@n*
zwZuL$CLv*He%d8nIe57<uU1pZHp+wUU%SFS98>UVM&lE85Q@Y>x+=#jHgdPphndAc
zPCi#)x2P5}C1t4BYp2db(Q`xAmLxdGfy&4A!6@t~_{2LKlgy2RXy-UKB|2uVvt~mw
zuTsS=4Avq{R)ahr{3w7;r=YqQEVroQO(t>3#m3L@EGbh+6}N`Bm4u;=n%YpQ2yRmL
z$mJNzZYO=EWQlGj#nCmHjUEQcUL2R2grQH|W)xz=XU2YWP3vczi*QVGzMd|tut=RX
zqn<HcUMfc;iB}(^+8<!PcKl#sLw2D{v%+qy(qi{=udFB)I_0>SuXY!<7`JAg3tOa;
z5pPw7a^+_Hu(JkF5De|Kyq<YPMc1MWMxSMu5}^)9{4F$9I82YKnqZlXoTQRTLAd05
zQldUHYRuuRNqfR=Gyn6<IXbS|DD(YeBU8%ud+-Gw_Ex{Mi8fQd4!^f|<;(BgW0{_{
zmvZ55p_X*nF2)M@1B1kmHx&MB2)Kpo#CH38{aPc8F!R=!ee5>Xd@MQRrs87ox-fl|
zUp3>dYT04x+=w8a#NAFUOs&wJj`?Qp^<{f|@0r|o4*|Uy+ACF*YYF4@d%=wslDYS{
z=G+4*ZyKtBqpnkq9j{a#e_|xVotsy_wVGOtI~nCZ*fQFC(5g7_Eo|{epsMj^b;h6|
zO2n|hYl}_2HyRF{;;T5)wYT>uLm`4)2XZw#Q}u2iX;dm#k$OIMr-5I8`p4K0rU(&1
zh}%zZVt~`w+oWqPpM5CA2`j$`%;rNI#3Xc^+G{v0Zc^0RP{i}2i02p^M9J1~RW>J=
z98$Jjcw=wBi0ssY;Wt{H>D4iuxz>#L><+-RCegrb)60w4pn#Xi0J8c>!F7DDehbHs
zHmBVB9iFZiQJDHQ4M{|+VRR&kvbA`pLA?%4?$x#@a9dnTvF1^8gB}`<Fm<*sLmI>s
zcc&jkG|pSI4W!BxnB1L>QIJ{{WpWxfvotTMTWDDywA{KzUBLE)rerNY7e-(0Oy%o|
zH*{E4aU1Y?u%LYGkB9qcJ#;OGi`4hMi5PVWoBieJOU+(FEf($koZ8=mFOKXzU6M5+
zxYQIkS@}V`{|7`G_6rEsNR+)OhvyxW`_&>Gu?t=(Z5zTIJ#yACkdUm?@ylxN*$y5f
zFKXFeQnP%B@==%hy2cnuoWXfipEt7la|;qgew4Nw(In0duSaYQS)`9xB`=P|%1@WB
z(q=?e*6CxWGWNeNF3#dNPJK8qcXw>tyxR#84mRP-D}Fs%dkJ<e)&i_!J)PP=&-tzo
zY?&)dQfvwx2C5s9-2-Y)!iedU>by3ABF;1$+D=}9*FQG(cK!hlhEJ#1%I7ODjS3+8
z-QNeggXZsxr*mO`&&lAEqK@{ZhA@Uq`LUcZVB$vujq^SiyFjq&jh~KMg)RTmv*7Q;
zQ8-}vx>d(_#nL_&9U!LS8rnN?&|PyOC_2M-sEV89n5J-dW}CuyH0Jl^OX;*<0vw**
zS7Y3i$apA=c+;Hy!g_M7nlgreS&(A@ZMI}sNGU|XQPWF2>9)#mzLD2S_=Rcg%q`jE
zZ9)A|YFQ?fz6}y}8;E$<{nJEVQ5R2fv#zds<vG%xeCiqF_tH6Mb(FeSlJ<FLTwrBD
z-hQLg-z8-_uKj_{(|>^3yE)j$2Ey=zp7!3ilFY7F(W~!QW6u{mc){Ig3OYX?$vRt>
z%evNCTdN{qhuF|~lNpcriR<!z0G2Qce$q{I8u;DlH;olOZ}+94@;jfv>?bJrjdrxT
zXbIX^T9k`wR>;oYOb7XbzDfh<bXS=ReP)U%Kmi8!@Rg={C}1vjVogN-7Mh@@tXD|t
z_fSyvp*n)0c60LuQqz0MdrZ~kts8R=C0gW{<;{!BN_U}E7U2;+G^Rkq)DAi2jWX_j
zk$Ct_2~ix7NvC&6w}zB6dTFY3AR~2lt!d?X*I<c)&3h7{&tz{uj~<T+cl>Nla-i^R
z%IaL9)kv^0%4+&}(o$-U_ElbmJ#aQMgJ5Tp<Q3&7?=>s1nUogXeHd>?9RBmghrZfB
zXikV*?Hn@w6+Lu_GxCXS-n(G*T47m=8BL-`H9BTxev?S}44+4<h4{X9*9R^&wl<uM
z27$__4CB8s7kWCq^$Bo)nJPcHKaH7QrDQiCBs3`YcF=OuoVGL)K_#{fX&HUbWLM_#
z<Zi7-G5&rrMJ)qBIF($bnVgmOi?xPU)#s7rdvS~e=Zyc_+BF^gr$c)%&E2dowd)%u
zP!4RWviDn+3eQ^wP$TPsGoto^<$LVVfT(8Qb3R6vHsMo~;@6$v{e8vWMq{W+^V*AJ
zL#%=(8lA77F{Efdd6Z9-+WG1hy=00M8}cRI*TND+s>N{R#~%)sa~EL}Dgs6l@TgK2
z*{M2a?_*V2H1P}6QAjSzYh=BO<Os#0je>>N)X@@12;24bZ+0pLp93PSCe0;Z;@R4a
zFE+maN&)j$lVw!crgYFxfBD_wrrrinGxVHL6|hX;zMXQ|0c_3<J%Jx{-2KImz8*LP
zn5=Tj364405aK_<?~-2LUrKEoB^V@=S`ZAe@sH{VnD71dv?^3C@#-A)cI^2Fh&7~-
z-jt`E9XcFjd~+azcMaYnGKRT)>)mGnQfg%q&!&A#G>Z8IG>+1u@E0tlNDxzDjCx#G
z;nw$0b?a#QnscvQwlur?<sX2ndgLYTnIy$^^~5_YW0PUg=s-u>Bqe*?I(IggdCs)s
zrY*Fo4&}!TH;kDMHUkskmk~-$>9@a12Be>ViCC_O^1l}nFFtzJ4ecD_T}_Nu6cEc6
zZ=b<2Egw_~{7a82B*GQyfe2U>=f$Y>ft)kD@5Vh@9aEbNFK_%F0ja-Qq2KwR78qtb
z5aAzF_Qam8o{nikx3g21rh3U>B?Mbx3jZGft3Xu0JoU|RmO4J8e`7uEy^X~7aDT0y
zIEh61p4F;WM?1Td8XIuwOeuyWepS5(fc!x{pt@Zq<zU;qGp^mk+dXR<M0liV<VKCu
z?PXzt51|x2MwN?p8jRN94cpiIt<7H-HaD32L?KE3B^j)E#~%Hu+4;EZ?N~Lbv%f=b
zJp$I{b8mCEJqB}+&ZS=q-X{qpau0AtVuH#zIH|^?c_ph`!!s=F^Tb&2-l%@Pt2a$D
z)Z~pr;k`g{<-||>)T`n9v)9R7b<5Wv-YdL2!!{Tk6(a-@*S%S?*Z$DzC*voE0*_#~
z2U<r{{{RU3AO?L(7)PRD(hD6T01Fv!Tj)ssf{PmoB~^7;lpo?4=CnqUFgWuf{2H}X
zpt>1KtopK>>N49d+Z$E482ijWpss6H8io9<F0LezdoJuK{41|r7HJN{F(2+PU&fMc
zFHX}=Zise06t-LPtL0r0%<5ZKdHA<@)W;U=^dcdF_?nj1!IEvvH%Zs#<o^JlrFM53
zvb-kV!%yqq80NDf@LsE~T?Lj!g-=MKT%X8sTPjk02zBGJE{ozDT`hquXOJ8f2Oxem
z-S~G{mr`}v7n?c%09T%Xd-~N)E8%{hpv1FkmoZunp{`#pf89Y|h2ZAWE$(A`D{aVx
zf2Cb6H-EaH<5N0ONVrsTXH(%{0MBrVeSU^U##Knb6>r2^oxX+)V11%)f++e5*VKGl
zB=*4}{oG-3xSHT~3zK&<wm|Mruf1{dyPb07wuU~dCDyh}-_3{@m9>oF2iMlQ9dlWV
z&L@W4EK7_MI}d8=HF<*CJ8#;6WVfw(Hm`gyE*T>~BA%kONY2f&tdsfjv$4rOwOxs3
zIp&m~%8+%cM%=ukmyyjWKP@*jv<%1bqiyh>s(TA|{#En;0NO$6+;d*}<HNJTddc**
zf8}2_M|gU5&q12<v6T7bkEP1xYEoW*p|EaajRUtW{3FxevTh|TOB0qn(zLts4_@_F
zc}9P_J7T$`6pvYv%_vey<n|Rp;^P@ma{i;RrMb8uWPR`DS<cwV{cEid$qN=39Mmr-
zK3aUvdXfAzZeYJD+Q)&8xTQ#jLJZ_{$68>O%KD9q*{IwD%icb_KELgIcB&p$-`>dN
zamXB#{{Ytaty{E7V~Q{~q(qWI`I!2jf9YB8Zd8^4fKT2%t9a~0mZJsBGXR6jP%(u(
zDD)ow^t*ms1CL6UJ7kSQfKS)fi6b$LbUg(U)Ja^8SOcDV;A5_9^bhtu_%NOq@Vt8F
zrD`M9Z@zX8cQPkc$UF~Rdy4rV_J8<YC&UkkHlJuk{>`yvmf_o+l72<Up<!R85HxxY
zmv3haA`pQTfsQfB`q!Obnt!{A6)P$+yE^{>?KL}Lh~gyRfz#<$<Gd(v0mn@D`~`4v
zc#huI11WR6q-)TSI=8iIXg(orOtWqB;$9m*zUS7w*GG}+d2_1gvpV&2AuIDS13Q1l
zmOE#XL7q-I%{s`8;BZbmbNHIGdw&j<0?5NMzY(GCeSVbf(Gkkttg{+HrbXvmI94No
zPbcYIe~o@2t+Xozyz2y$WfjT>WaFxX!N*ch0<!#1;$dkF_cv1yF;{jN{Hi_w0K>g}
z*W+z>S@G56&{>&bvTP06DDwL;A9a7(AC7C+!Qm98&oc`3ow;Is8T(ZDi&oV>Ax@gh
zo5K|HY`sE2xI)JgFX8_H*5bStM-p5}!-3D`U!Wfnz6fcaHf4`eo>6yhe|d2L3I`b4
z3F<zT;<~r&yQA7cC)xFj$r-o~(XQ0zraRZrCa-0&RHdRv%?PaOTd)}?(wgtU<Bwlj
z_dQ4894HnQ)J#_KjNxP4?t7L$(zu@!d@Z=sB`s?yn%O|h$TpQZ^sO~}9Z`B(=LWI`
z+Qfs-Hva%xv2ha!<9sNMf!fM3y9bg*Hsa6hmron8aT(f7e1r6)Q1h}zst+4R!Po2S
zSv#kxNLyf7x`@@-t^sb|_3mG@M}qaM?-JSH!Kd45x>S+~r?b1rihLY}XCLa@saEwE
z?_O7+=#$TR3{k)&az;To{d-s3zp{^nMxS|Xw=WDE-x4aw8C-S*V{85(73vt*lbBRd
zr6{hayMV)5YODJv=tqIA{=?zx1XzWtEKB==j48(>j<wa@TZ!RW2i^xDisLkE_^nxC
zEJ}<3&T@FJ>efV-50eC9zdZWCB5`{jjcNPb=<=_Pp9(ELDELD6Oprwk@~6o>k(ct%
zt_L;wmGMX6cD3-c#1m<{oF~aHL>F6dv#9O|QS`6aSA|#{it+D_{{RZ~zZq!~#d%{5
z!nynAFj)>dsqB5}(XT8k)qAY;u+<}7dkMRCeUHl+bvdm;5jg-3aoV$P_go(`RxU~V
z$CF(zkGwQ|CE^>sG%U&?l&D*PL*udc`qwcf+ZM_J<`3fRqu5u}(rQUgD|3SorA|(q
z-oy@@2n+(a{tlr>zr8joL2kTz)E5$gkfe}0libt`E)PtK?2LMzY^0OW(pLG@=O(Pk
zy*R+dXDs00jtDiNUvdLffxDtGxZ|}po(~3)4mwrEm``Gy_MzJ#;QC^zOKxLsYin*Q
z8KDH{6{^*o?IyG_{JL;oUX^w2fd`6$NH&AkwMfFnVsX<oUR<8FVUUV~PCA2CU`VLv
zhNO<xNl{#akrR!_zpg4xsvMWeLDhe~QQlq4boq0}&^E6C0nh>1R*O<>M|CQ;7BveR
zbRMVHs+26mb;U?>!#Mf@Px1n;>0Cy>r1Yy%6D(=kEcf|k8zUWAyZtNFd_CbUCrz@p
zo_K`wu*u)a`~_s_8V04Uh<?dsGutYFg@(ZX{PnMK@Q>{K;(cn^Tu6}GL;wmOC9&&@
z#&hNBYZVlSd{npd{1xF#=v!d&e3BG>&@0Tf0xvX;lJ0g@BQ44NE9=Y8+1mG3@Ft(8
z+~3J7Ld&u^&<s}ZgnwbSzlt`$Sghd3`unW3gnLFmsjeEBig8zyE;CV$kIo%KO#2f(
zz)(<jJb*pwZw_fUelNY24Kql*)h^k2HgL3qKIa$}`lz3=_rsqJ#pewHJWZ&m8)eps
zmK8qHE(Siis=B|0yh&wfY^-3j)3o*BZFK08+Au%eMHLG%r6pu?OP>De`T3*%%o?rq
zUuxHxt>hyj&Nt;gr`EA`U)g_Fj^V|*QwDxr%rjqZYTgFbIWsJZeYvh`<H6JJ&T@XF
z*Hm%M?9t2Y9_^nTkJ-OfhlykW`vJ{m_}k#z{u<J)ZtbFy;^A1V;vtid?_$2jxbW@D
z6PUw~_K!7*f8mjEq&>R=JF8&U(Z#-@>~+zf2}vS`+T&vn;bDXN(_<l};P$VwwO;~h
z_lkbkWRs4E?J+BUxUVMo!SL@y@b8B|v}`P6wz~7uGSHF+2sz{Mtsz#Nosg@>UdmTI
zS8jg_SjJkMIL$o<E3EnAP(DFQ21XACoU5EviZDs3Wv8I1paI5Dy-AKKpj4ZFT0<+y
z2q)0cR|DRqkx6X!rtS6XP|c#jK_GMa(}Hp^DO7C&oZYGuXx~PQDF)=<j%v2zSnx+R
zJTf{C4;ZAG1Gp!F+=}UqRP<*}I&qS-7-7$`so{Y49Xr(GJZ&S9j<seG3t*4Nw|cD(
zH5Ce+v)F`1yc`Z_*|n;7Cf&|?2e_fgZrHCfaeT+n#x&e77DFNfPy}7a+uE$ZL+MZf
z`qd%Y+l-zsln^R7kcJ(qb|azbO&Cx~IXK`}mMBKxk;fe5Rjo2&x0c~P_nL%f*CMEI
z;L^n@jznSANF(b^H{8(i55!GNLe%w}NIppSu~Zy{ZZgV0rEnJ>KGk&@8HLM-ll^32
zxOVJ7_4f9y9~GG=)-Gga+Rod^9D$MB+PPaDD(3H8hTh61lIlhRC5#-FC$&`S&AVLr
z`j}b~oT_r8r|^!6;vb5BA&*J6^Nez`iEs|UM@;+ha1DNjd<p%YY<v@>+1i8;I8f6_
zeqzns`&S+NEPMnW2GfS4sS_@>d;>UU`C}R7N584>Us~x_k=eq#Mh~#96nRzN=Y@#G
zDZWT)6m%XOu(i@{(kK;1(fliq^YpHs8^w$Q#Cri;d^ch~PhP-MzOVo|ZckI|S-9Nf
z=ds?ys|=5rWPX*gWvF?)91rVWV`r-tZcaV3T`rrcZs2~EoS-FF%=B9+v5s&(t6CXB
z#yeLVrr(UO%mSZU=xwfpAm=s5D6`d}Rao7PgA7%ekyvDbky!S(By~JilrwSw70)GO
z)~7Wh&_GGe8+MOsMQrC3ySr9R9fKKjRi#X4<>%{G{{YTuj=80<or*qVC(^!J{i?rY
z>ko?-GirKRSg)5Z@F>RPkG!X`?~e8LW5=F5RZF{2>UiMx6|^dHyO}pA>lbEzA$XhN
zpNRep-g*04;bOSl;bX?!^v^tG_4Kb{_*45t+-ebd0`5_A6#oFNVYegs4!`G``+wt~
zz>7Z|Slv$yAk4_#WC{QhfA#+Woqlupzu;Bh!(SCkb!P1ix|U=~4=8`S57R%&zLO1E
zNhFVm&9IQfDN3sMx-;ur4~o()m}g^<S36jq52kwao|V#Re-UETzyeZO{G)62{JK}e
z`p?FBAc$H|5MVw)@tk8g2RQ!#KU(fS5`0CH_C<(I=EP3W+;1Fn^~bMz_b4bzRzDff
za=cYMJFK*^_9d33=VJ$VJ@_@HB!w*<%WYsdt_#BY%DCLe_<A0-yML<gwtjwImFUis
zlTk<Hl<|?E<vH$4EP=E_&Ot21lbZQQ_LY(=9Xt^#v35Io0Floa{CTf;*Zfp8{Vc06
z3Oa&tE90NqH^tW))Y03o&KE7pV;@j@AA0dKyew&9`Q(1h;g(~G$7Vc=I(IxW=1YS-
zo7J;{w;rG2Q$cF7D(_4l{PX!#woHyywzmf#F|By*V{CFj91m*vI#7#uKdUiVnMNFj
zT$Zega(Jx!ix5a1NI#u)vqTX;+B|d6R(!ExiS66zUY9D_9s;P_yo^BsUJ2_+ZT66H
z_~SLG&;oJS`cUxZE_x5=Rmu)#SCBh_inVPR7~-dn)RE9s%V^%11Kyg6GM&*XMjZg_
z{c0JY`4|C@YSD(>f+@#dxTvpV*fKu?E31~#IF$Y4p1bG;X37+HEPC-;w{f2^lh@X$
zG<?@HJ1=x+nDS0d6!JrV7IE)IBTKXCsG!bA@M9*NY>$-ej05t-dw-2cqi*b#&O6kN
zItYPsaf|`jirnEd18%q*au*mi6Di5bARfSb)RPv9Q24@=oB_~OZa!=ijyhEmHA@z4
zqi`LC2wQ@9TysJapqwAAJ!2&ZJ#oRsRe)4(NWrDfPBYS@P<IUR$3Q93$F;NEpT?vS
zDH+8IpdU~=ns`oF^~X4&P#(EF@kyfgBv1y^o+?uN0l>%Bp9CxcAQD091s+}$k_SCK
zKD4YCBl7`_WMkVsDri%40oNx0RC`LQoL~dn{3^wmEg@cq+o_-pJ9z$8P<IN@i2~#f
zY8z3KFbU(4S5XH8ILEb7n^zvAf%;Ual*C+t#yyQq81L4BsMwhhas4UVGC`$5$3s$W
z;)1L?H5lTpK_APSi4%&q3ZE>Tbft1yu&K$Zw&Z)$ZP1c(D%eig_N?Swb~7(-;jp)P
z=8<DX8Ot13$A1`pGT-W2{oQ~ONnsmd8aDgO@7lfN_LbG{v@ZxNLZP92tgF!b5NqLi
zWP6F-$t{ve&owRC8q-|O+_$J-E&S#3Q7}6C_o@m3+!)~Yr=)w7e~m^kcLS5@ip{&7
zYfTCC{Yq^%<z|JM9Zo?974&z(pN-yay`GbKBEhl3uRnMkl1Z<b*mXGR=~LL<%V#`t
z#>}jTa3JD^D{f4pq>s{@9ZC!M)#7m5fH()St82z_$*-EfXD=6PULn^e`&^7{Qaz&}
z^MSyxrJ_c0@{UbxaLID+eFOU_X)|4~izkhe+?Zg1T)#Uq%)|lo8TYS3)x7@z3|dQJ
zq{BVbdR45~63-wD_M4d(YK_N_lrdwDm=)#T5%FZ69`Lq-Vd3!wjF&DZyO<$SaL9<x
zWMiD*Ip?MYcK$8+g$IT_DF&D$gHhI|w^uAdXyQ_LDR9Rni02=lrBY29-8EyB@m{+M
z-9u%m*hh5GY0yFCCBr063W7`EWsf-rpeHp=z97$T*0AX>W25NyvFwjc2xW|u_l%ht
zeEn;Ryu6<N>Pw4>8d)ScEXO}VP;NU&2NkC>GfADsy{riAXPOK7gf~hRRs*=HQM)-~
zkOypFj)J)@M%vwNW18mvYt2$@JFy%p;}~*zEPcJI3w!x3LZC?Hi+2A2DplO1`r@kJ
zT}gLv^2&Ee?YN&`^Zh8guvv{)+71Q<LmcGG?V1U|BBG0j%|*>GV)M<PY}93sT;~+W
z`G~>IAjlY?8@U#%r5#EpySBl(OnI@W<^9LM^Q?PWfnnTx*FkSHK`3-$jJ};J8?8L*
zH(qV3gq^<nDEV9cq3CL&V&8MG@Mfb75czVCIwCn%Jg?HbLjo<Zn$u%N+?Cu%*0{e7
z_-^XhEVHO7{o{mR%-3^osA;w~u9mTxr*0S~cK$4XpsgHCZ+#3~i>v#xVusYa9))rV
z{5sVcw5V(VI<4ikhDgL)i%Q=`87K6uThps*WQapS3FFJPJ4qhrukx(>e+)gFG_7Mg
zu{qqJ7C-PK&1$C@9PXT_qb9YyvWyR~==0sTX|JEjL+`L;pT?Pe;yrLh<)z$K5r5As
z2!wz5_a?1~B?@*~-DjsUe_YfTS{|o<c12|<7~qkEQ0mG?ii(QR!m_(>w3d;J$lvnO
z$t-@P^IaB~u1{|u0gexUN=sjcI*Udozm+T?9Pe@m>T1w$4rnYlTsEgV{{RYdKR%V4
zk#UPU-6L6=_A!@Yr@v~KR@UtFi*nZ!0Qmm^Ryk(;D}dB|do0nI?W0>Lx&@EsY<+8#
z)irr;@^gZEwsOX>Zu=c-Hb&2h{9!h*#w5%(;PDvik8@n?a&6#MXtKcL^{If`C`PGR
zn>3XWd>WN`T-AemjtfpvOjgZ5Bmqc92gE)+53PUKtN#Eh`CVi=IXEA!eNpig5BNcK
z%#8m4qQCO5lbTj$$vLh{%IzOVmj3__UcaHFmm6a_<oeW7zTy2Un$3gOrSkz&dC%wt
zaq_d-(!>({)lru`)zUW)bI?<g-+9=Jxw~hkef_INq9(;CEsT;WaN$o1sy=5vgZi5F
zZ-t+;--*5_Y6)*^2(h=5t;FhZ2R#1(yYE)M3VbR)F!2mxNSf-(CqW=i4DR1@A?OcN
z=sMTw2Y@~w+ju`lh5(AQx{o(!-Twe^f3h)(@$uN#L2}dF?}aGB9Mj$Gzf-{cEB^ol
z`tWsxHt(zIm%5~(L6S9Ku{r)B>$LuL)hFyH@U<s_*G#x%jY|1%a?Ff-78&Oi^dLA;
z015}E6nUKFi~-c<zIkJD5`Po9gYM4)llCF_VX&-vGcGfO=P}1lIX>KaX1EWCKe7J+
zhP*wX&2glN?e4ADB1u~){CwF{(2@Q{ePO6-%WV^c!?q7Wn)x^OhWO^T@JXWEKbdzq
zFoC=)D;oFkF~wGryLLRxMkX}XmaND42cTJaT4mF1Y;EAXd74|9Bjm6EW1hfzbOf69
z?Qc+Po7cd`6aX{RCcN+9E{`>bhTm1V08MZ_xKkJbd!Bi&#^-bS(`~}%u73*pWlctH
zpE+K%suhL6mhM?x5gKwp_TUQJ@SV#=B==Lu!rl-<N1v3xrE@DBgeph(-Np$0O>EgW
zp0HV_L^0+`an+4wDsj1mIZD@Op;`FEXts@YdQX;rc<Y|szY|=Q=a25SjZSN}ZKRE;
zq<pZa)7H5!6Kls()C>zJ`aB1mba(Ce4wUFr$EKMV1<4!aZiqeo=#2<RTOCfLYS$=h
z`kea9xw?^svc5!PudzO@P}}Mjs64o!D~{N~{C>3@_fbzA(v|{0;$z&7m4??(6YkC;
zPrP%<K9%1Dtz(tly-oYwBF9%%M~E{4lF9c^u&z_Ye+qQ@V%!`VTcVDO?OQhbz%k^w
zSn=PCRM6@oT+Fh72l2N*kgTNT7opV!WplzjPv9%9CQcQ01AxJD4SDXf;gy$pctnMV
zJC8s8YWkY?+6`xM$Z|i3_dde8EqCF>7b?OWMI!UN9sLb&3f&}a8Cdw+S@<twc+Tq|
zlI+E|Ly)WOisv-{0EE-&lDn4N;~^D#pL6M7Q)@m9y^47vWGqKS1B35ez1M{A6taa_
zP#hT#U&^$U?zB5;$u@ca0D^o1EjCMag$#eyuU0==`fuT<gk?wz+q^P1(TGv7>+SAq
z%Jhv2NDAeQgmJsIdPjwBZZv}%!sHRpUMtG2hLuF!^*zXDQJ}d}J@>-?A=PvZvCge5
zLj*-~Fu;2^YWBYfc*er<LdflJBMw$>hv;j?ejoVsYRv_W^`c$Gen4dBKTPzmXM@0+
zb^2#cx%(~3j%}pne=}b<iH9TPtgYs0^@??4&jYs$xpDv?@-S+arD&Hh>?3g?@%}Z<
zS<R|=YGi`p3Puh*w>d-fHPYE@cDJCTFOu#3?m_w*@#^8)S6ZW3O{+;C9{$%~1OEWw
zDbfzBWFPxV--K9zK4<$l$4q<I!(J8e2EF6Y635}~6Hb!*R(2r3FgAd2#hapmj!ELb
zV7zCfLdqs|%nAr6InUu=pWhTbSz+MMihd@8!j}TpJxgSIl#l?7%%hl*$10~kFda>A
zl~t(fJT<S?c^<|s!nPe^Hj>f#{{X=`Nq=F_h!>t8)Nizl?HgFtbmq;*i)S-5n{kp7
zTx@3ps*3sAXs#@-=e4`Hmd^4AF->hc?2;!PRe|V5exv@#TKdW0Y46DSBE6C}4cBiO
z{{R~Kqy7pb@Z#S`@%ESE?J~_$+g1&2d3c9|aDZ$_9RqFWC#`z1#(O+M!`0K9NWVVc
z<b3xlZM5`9%1HRFXe7)+vEcpOfm`xD>q0p&$L8X_8?@LW;KfRbk^ttbL`fqY`jJ{z
zae0J+$?9uV%;l(yY+9A9vJcDn)}$7>0JjU&3b*!X)O63aX6Dlxh0bg21%*L1fgN#O
zB$n*#5;QxHMNyJ!NxU^YNE+dSN_i0uSPy?n+7o6{f;fANGQGNb)^zdU;=5bDG1X2~
z5#F)}i9VxzEp092p5gddqg=8701E3<?sLi()iafm7@Tkisie2GYsr=icxJnqjs#F4
zW+&(?-25N=I$Y`ulWX&vNHQ^L58iKL-{D_I_$%PXk>Pt0EOyXo7JUn<{`N!ee(|hl
zUNQJgTaoflg+FJ16nK%5uOk{P4ngwD`L`@SgN6J}d!K>-0A&3yNuDWhAQIdhZEK_E
z%1^TN74#mPrK}K?w1{c;-F*2J8C&QN3-vtLU*P`$3F<oI5v0H^;3xOGjmadpt_kJ8
z4l9bLF~w@^jM{H=i}2^d-6O-T9FoEWzyOi3-y?koL0-Y(Z->{~v}p4u_I0K>RvDFt
z)3tW~75HhV_<aS-kF;HNYrC8teTP9^IGz|3x`QC@?Z7|dUNu}~rKC!{Vy`Aw(L6b2
zqFcrxj22OV@ADtdwC24Lsrfen>ToNZ)qG0ng@wZNoZ_>!{Y0d8t-YP4L}R$%43A36
zG-X>>&px(3H1EsM+q~AI&OEq37Octdt>h$>gP-oWu5wGfsRh<29Y!*LAy75gYy;;?
zdt(SIb%u;jHRh9gn%8F5MJ1vtzuw!C`c`bZMxCevL{VG;*UBR!@~EKuMa965-UGb!
z3g0y@p<^A3Cz~q$yFWU^(FglXI+}jU+9OzI4eHU$_=Dlah4;-IlH0eVjl=UC*Ac7y
zJiW0jVU`Po>$TS$SE{Yy2R>_lCYx}v+{XKW>4T1^>F-vhIB3rLc$GJ?^Y!JBg@iYD
zOADTj_mtO#{?%GNwWq^jJW=d_w?wBOMz5wkM;vfpPj19J1|tCR$3tH_e%2Uoh1p!~
zTlh!ghp$5n?^ZsiRw4f8UWdk=O-KiNMOt`T`cvgekmD5~sL*carc}*Idu|9SU*V>K
zfr@rGphFx|0h*MrG}5G0GiU%A6z#MHIH@s{nucze<sT(0FHzQ%4mhPFCZl7nFG71}
zI|7r89Q|th+WeuA{DfpyWPB6b(zI;uaE3F`^11AP`t{RJcF~wtk3OYo<@pG3NE`ww
zZDjMt!f<*6QBAvhgOD4qrB;Q!smBLza4W7ei_x6bDNV=RX0ltHl_TcsP`nvqk;QCS
zGG9a$^Sd7-j<lD0bH?fz0;i@cl4>efN3VscQNq)AFx~>qk;k<`G$lqmp4Gi^rZUDD
zE<qetRLixw;EoBP^dmUSW@Z&2;+>p(9@Q+Xmr?Ib3%R)K?@|lMvvp?${AnzkRH}ge
z`2)DGas8M+6G5u$4Xmw^e<vuCV}YKeSA~cfiUd|*&9thXx%Bt1)PLE};Z?VQJ_TvB
z%M?RXyFnpm<f&tk`eL?)mD@glHhhs$gWR7=(*~NV&I2=GDD?VPm)mfsB!D=;<Bv*i
zsr!da<#vsuKd<Xp*LO$m=iAz?yXt(Mp5^JaQZtj(^vx09Jmg^Juo%3OD)YLwEkK4K
z26Oybri^9GjjKzIolY@Zx^AoU3=`J4<8VuWy)o9cG`U(?0AH6KPp8(06wxU&(zN|Y
z%-DM4wRU=a!ICk~O?mf+t^DjUZLyB$y>_<OQAUALwy#gEa8yYil^D%*H_@150ASXw
zyOmJKipRIOMpYz_rCe*oR6j5k!z;FX6y@yqHlv&{{{ZV#jG!L1gqP>2=}7vX+-@|v
zE_6IXNaz-Dc+WKEVcVr*+TVGHtzQ#v9fd@q)3K~8DHE3z*G@R*q2J9|QWy%0SlTa`
z%e=SV8zYXj@jvaypnqgoE%HJ>%P-9AqXZuR08^Y|zUZDH>TAQlX>S9kjl2ner@R6}
zMpWY@A5mS{TAZ~N&o42h&q=<=;@65==S#nPh@Z@X&&-TH{{WZruE*i0jx2382z<@a
zzVDch-{bH4R`0~`gj%MFCBrliBrDh{<KDbr*vn&akfj`S7{UJl_1Dlwqo*!tpF3L(
zOC3Yl)IUHx6Y&BWt&na~Hi8&7a(yeapTyEj0VS}$mGDP^{7<e}NeplqW(w;2GEXC>
zeGPpHW8$rP`$@&ZeVg2lmGl&|Dlv`fy^q3kZUDqf5qu)LlQw=Vc+gtdJW;-7JPZ+E
zGHWwjUth^Q&iOqQoa6AXRM(W=yUQaz!;Yl?054kNwG9Zop^`r@Et~<~zF#!PJjPas
z!jIXwLnDU`;q5=-J^YRqEivVlasWL6`c>F%+AbI#_{VzQeJH~tBgcm90OqAOpC;X_
z>zrb|IMU_N{RXZhnO^5PbCt#q<Ly)?Zl@Uc^sTG=LmLyF-MOrLn22JbNF%ScY@=fZ
zG|KU@mG}b#u4?>Ns^bH$dLQdmVu6(B82*)`X_W&YeKFFug1a)QJ1q!W-voofs}Wjf
z0|N%6WZkrk59?R0ArFnbpVJl4qhp#?bR!m{W(Nd%RSVmEDFe4^=)8cc`@KK?RhfMW
z+D<X_tu3L=3F>k8&}|v+dHgFXX#P|DtE0b#K;UJw*U)-aHQM~74xYHhay4VtgD_e#
zxcr&=QCAuwtUwu}u{th&9$S{@C#Gs!wA+O_1qX5Yiip5SR>@thk4*DYT!!;l5(y{2
zu4|;?E3|450&*1Ro|PPy#ODX6tymzJB%GdleQC=gj#vUS&*R#f5l~xzibhHGJ?cb{
zbZkg$bI288@UB2O7|t=?q=_;DoMVoA)m%n)!cnuw)|FJRI5-$Pv-(x0kFh&fC!pX8
zYbMNX9eVRfXg1}5AY-tpgnWa8&uVmWsg4&6H)Q3%0aG+YF+dLoG>W-m(MgQ%9ew-K
zJ_sjhz{eh*^&~0*10)`~H2{<4l^o>m1$pWRwMYywslW}82N~p4*74$eFCDn=PK|e4
zpa+66$o%SPK-{?*&tGHfLlLKKzyW<a5&bIUV+4-Apw$Q$2d+UiqXuvX-uhFx$x=Rm
zj%sB}XgNG}=~Eylw?jq1kr(r+Ko|oZX@~_#IjJaej+J`VH$>G<7e89lu!xlWX_Cgk
zxCD33Ijgb7zk)fWf&yENRru{V=}{u1vGVurg{sYMZEI?>Y*vSG8Q_fIj917bIqQIH
z?4R2U!xx&gx@53hx0aFP&Duh@IV+yW*w@WEcf+k>>R@28mEA`CtYnT4raDwL*xHjx
zJDiLd^9Ps<CV30T6!o%Z2N9G074#3o{{Y!@#6BRmx4-)wu@`m>lA%(#IP~wD`WL`o
zv7O5V8p0&81qxyjfLG{8U&fkCsZOF>AD@Y(&m#<5$W(H04@&C1EAX>b@m<QTj1JKj
z19J{XvG=djEkE`{(4i`()p9U89AnbEZx8$d(ex%k6`HJyKJtuYQE2xi)O`{1*TKJm
zekJgxgDk?@8H{|oME?N2VT^(lfKMKoHS9kQE%fR9Ut*f#ZLx`rlSLpQTq=?j`w|9y
zvMb*7-w7-Vu{1zzJCTR*k<gD%t!?;6;qHy9Y8U#ptE53YS*GK4=4_Fh_*IA9a&ers
zUD7$7<j6iH_}9aF16$fhro}X^YLZ)>G^9ll!m}Ph2pebs1<nO{p0lgAxvJi34;sYR
zEL@_Tg&bgKz5o^Kz9RS){{RURU+R_`v)*b9o0shoEIE{%g?th@7*V$aBD{L)#_vVE
zlFt6xTT7V9R*rNk-kHIwO<O}Y-SrqgY8h!G9QxCUCXqxB!Owbu!6cfqAgG+}9MxER
z5CM`&Jt-ngzpZR&o*C5qOK`C1)>2$M90RwH&b>m%;2eG(vl@qsbxUnB6UdHfJiPk=
zK4w3a7O_7UL&w!J!BJV<@D)Z)YT0;0SJYQ>*|cf>>3GF=nwO3|HQ{*8zlQX+zLbB<
z)yjEsXVfoN`jcEfv*NvP#M_PK<CuDgA^{^&{_62nWJ6-fqUpLzfq!}?n01B^n1lFn
z`qXXWFCk2NeYAG#&^wH`<<hej5Tm9k3X}ZvPnz2i%YDh;RKAu?^4!TO^mfYrU8>sZ
z(=O5&kN6QI^{n2cW3?+u$R5=WMv5Iyn%~A(Ht1CX?i~(s`c{Rn#VuaY46HH?e-e>g
zG8~=_O@ctF<|~yRjUUE)@VNUWxIf;;f2~%$__eD}ghz5=2dIG+KU&~GIjKOPENEZr
zmwJntE+vw7Jmyy{e_Em?X$~2w7-3JYDj+tT(nTapTYx)MGMt)LdVRR_BowSPhr&F#
z$;C%2oQzeQh6<!{RLR}afs#gZ)~7~2J5_b+D#Bu^2$47@qyQRd8CsPGwJ<r~i31)3
z*MIBL{{WSIizI+#SJyuh&O8CHpG!GEm3(^4Se}RLT(y_HkD$w6?&a(H67Q*&anSz&
z`s$f(gQYzUWYxAki=y;xzO}65+!X)?n~e4qgB*k(YTwY@Nh)tVbUb$N#U`{gjhnIe
zC%`=`MfhLv=FeHy?k%ikmM8K|jj%HiSg_CWmEip=jMcy3qgqkCy|uBjmOFT4^43la
zqZw1*9kLHYn(^=3%fxq<p9KCI+1W`v78mNq<7q6e$OgrB!2yOi!3VEVUkkzFE7)z4
zF(~;s3`R~pDl)t^D$sMQ8@Au(aAtKSD(*3pNgt`2XYHZ=s_~>c>ay;}FF$bSJv~6h
zZruLT(yPoa?^V-=Q|Bv;WBbZ+UzFCKI@GQ4<_M-82q397r4Pg{NX?ir0+F@iEx8zd
zKdpPeQ=xmPCGtEudql0w&(S{~{?zt*b*av#(A`direwhb(~dFtX1-$ZkB+AJk@0+9
zUoy>UD}A01F?VM;&OJ%T2e)eQJyTqhSY(oMw;fk@N8w(#`!Z=?+Mg2iSkz^0nXY66
zXBg;z1Kx*CLP;b<z1TT3^h-y(5NU}F1IrOaBSjq-0A{VOr5S)*q1*|}6X{tnOz~;b
z%6emQWgP`g6|}J1<dcap&~_s@`d1}IriL`_)RFH$(hrxPFpY{a*!4e^J@p2H3xch&
zB3!hf@=rhoR+`oACJJ%9#3T*tk&#^fyQr1Znh<hiaKD9N7`F8)O<5#EadNT!nag0w
znFt=4;*$Hs6H9KkYEBuv$vXp&x+;y^3H1T6VmMG|*a7}ET)2WMXK><YUyybO-|()R
zD-&`}LsC0Z-Zv_!H%P;|_N<Goawu67sbAm+kH)E7+!(Ff%VJCC>x$;xB$7f;E1m%D
zRTdUz<>t2~)Zy8PQg~H8D@Gkz=Z#@j!gXRRkT(M*RP)`veT7`HozP)^NYwuTlUhAZ
z7KeYN-z;pPu_vHm(JP>V;&Zp}5;+5${{Z#t&+N4jfB}J@?{qb}VRY`p=PQHvg>udb
zyOf(cZA(tKxVPL?Gl9Syl^&+KyUznyyqo1h#EvoEw{3MLwv6sztLS>QSD#q@>`85z
zWY0n}N3g7KHKH_&?s9j261TNSXzk?Bxa(9b^mde;ipd)Cc<O(<UXG(ux&%pXV>s@(
z6zH^MNXjq(oRUXxdeu~Ed&Ie2oK?F|9IB8U73z8bE7E)y@gDm}xPsn&M__paNp*5d
zY*E>J;;^REB7!ZE0NfrcB0EtG=6Rb1O<`7*JVab%jj7g^C(RapmEpe=+IX8sigZ&f
z5|RL!df?+fja$3X;*>FkiVS_h_iN@K3;5>#)@yiTnU>L6a~^OA2lTI|^sPm1?NP%k
zGazHM<Z>(Haac*<r*!mZ(P679>B=h2ZgN`2ttO9eG*|mt)Ep^M&VL`}U!4B{w6>w3
z&+%4EodWJD<0!s#ae^6GjH!Qryghiquev{JOCRmehWh=IJ4Y)*LOPPTU^gpdaxh0)
z{Hn9Kx|>*#+|1V(6G~cHWNpmnxg`77r<CD2>O4wzvE7QqsL}Q|-}?K{&`;R%UoYWi
zrqY43OULATZ6DI3_>tos2g926j*;VAIJGYdc#_-f_sevf#1lq#r1(9ykO>^{4RW6Y
zyi@&+@Y_w&RkqyRO6)-6=aGP1`vIEbe{MZ=+xW*p(uexXYgd)nebQuMkI+{pX_wQ?
zDb9skr1ecZH2O69H1#vV&8n-h*!ZLNEcjb?qT<W@Q%Hi~2ZH|qPP_B%_MV4pXOG6d
zZ`VEnTKGrCy7z<iFA}f(BnuK<=yDi>T51rLBnEPLkdd$`?_PuZGkkl!@P@l-;f-2<
zJHpzO!rfL=pR!8-0Ckr?z)uBw4CGg7`)+(QO;5pEcZ%)eZ?t%3=n2sd3AJ|h^jE<j
zTKuak%(HBZJ%&@7*L5jX(h_>>(d(*r_hhctsjDARTMa_3A8iPG(ciP_@_&KxwXCK_
zE4OTb;ID8;1b%g~WUL&2cyMd%U)ek0&%;lPe-GdO5^oGzO@Co_@JV#@I!!A+Sy-|W
zfsEtnn)h3e*>B+PjSalko)25t!Zu}!_D4_v&&*46o|&)HtLJoZ^(977cTU|8n3XPB
za;AQE=~6<)W1o*n4@&6c@WRf4;(f?@+`QM@x^KcSg?G_gnZ6x1og<b4-szY!{M}fC
zj(U!IS5F`8;qa%+AMs6D*u)1iCDTSz+@0K4BhTpbTarueG?gb!d&j~Ncz%0mF=J{X
z=M5p{kK@{#AA~MuC7w1fo>hDPHTE8<@W0_@hN705XNzxRxMmCXVISLL9OJGA2jg9)
zk?>E#nhderY8Pv&768WrL*=Re0DGU(y+x1aOO+UNMQ)$me0Sjg0NL+S@sSrYC-!!f
zrxIRFydPvA#C`f#s_35vw7(7~*t9J-GO_vY=H=u*tlpl&zN<|?N4khdBw#4Qq)`!o
zKe_;`H=YQviVcv4k8sI|9)A-<#z`yBW1fEV<vw%M{2_0lIbj-V8j_A@lZave06x`l
zX|A_FXt)n`bmRA6j1%-9!oKzFH@*XM2bQYM{39PCd;6NJ7lF-=`56q7pS<nPO+>M<
zwuUZ~Igbf=5w(lh<iBAcjBhcI^>L4XU&g&M4Q9_sMYgg*aV&fT{0shls|v%+(oFGO
zMI+l>CdnEjhEdpx)zb8+UNCH>hU5YFSnY5<2<co?sHE>EaMo27)DeAkX{ReIgFL{T
zd4mzJ(P<}^Ymt_e6X;GcOly0D{{UHUBK;R4smj;~Dp%Ja(o#p2RnFRx&!t{Zov{~O
zV}deyskMt6m9w)iZ96g`Qm~*c$NO30nH`(XH)m>&?N_JLzSU?^B9NSASZ5X0PZtVr
zksOZ)Dv)xqfc6uWR$V{K8P5Be$K-hwl0$D82I}T6Kjod8?mtY{iugkKp(Z%tPNk$c
zAFWrHPqTn+n^aj`e+8_0Kc7nJc$ECiW0%xxR8Y9p<Vegmaho_#QVwKqss8{9%NDm6
zjff(V&s^h=#MPZk#F|EmrHoOl0r#$?&;B|q&iq&LrV$fHE|p;&6tIpz$a7jn&RWG-
z>Zh04-c0K+)ij{(cXb-F{{YK4&i49O2d{YRRh&FFJ2dDI9%CKyW7Hni=Q`iT%L$7Y
zzdL&*7z#fgwa0276RpU^5pU0SC9A45o82CrIuM4)?(VN*iJZxECEV;uVw~jqlU^tL
zQQTZzd^bmtL~yO_TVqp^(YZeLrE&2B{#PO-Xq5T@GhT7=j{fsl@Uo=xNdExo2!3On
zaBH@OCaNaNuP@xo(C`qxz3Np$0BmFf^A7l_WU^@60Azo7Q&u(}$2IS5sq-Q$2^b=i
zl{qcYA8N3x#~95B3FjRuWnJC5snP5WzN0oDK~C5~1GQR;J<2jR)Bem{5(Ocra+3;8
zr0zY1UN)S4-22ict1BOrkb7d0tX+zs3l1qlmgcpR!xlbp2XQ+<$2CY>$TBt(54C3E
zZ*nYPQjt~?&Sn@0PvKI3XHEUgRJCo*RZDa#guv$_lpUw3<I=7TGG@TX<C>8)nJ{ty
zG$`Lux~Q!aqkpXDHK7)nYjXj8#1-RjueC(BAO{5e>NAX0%S8zbFCY`mY_%oSwUF3C
z*|P|UDo)@(yX{(YYAY-eDNs*RK&<}&By<Nfz`J)3N`|#PDh}(F^*3y_C|3&TuC3H`
z0;+g}!}hNl+*!)5BwU123n}gxcC2}KXXacGKLJg&x|$1wl^<*8tz2Qu?|U9&p<0D*
zdNZ=@`u_ltk0wNn2r9#j9@Ip;pZ@@@S2bbgPrDdq``v*x4ZJ2bBo2KCHBLON_o+#=
zues@;v**L(@gL&0gQNrxZe|ZY<=i$7PIK$iJ*)M<RML{m!*N8Bs@vNpL~6YbFcc45
zepT^5{1da`tKE2-=fo1l^2HfdIN)Z#LG|kZ_8v&uqxqkDQkAqv&1NmssQizcd`k@R
zo>YPd%a2d<-@i4@G!g;GJF#Ay;(L2TshkqwPTiznkUF2CHOgN@7?s}u5zn<F!?BZZ
zY_O*0>U;WAUR{rnbBb-%&eA~V9+h4;^GPgly~#DSNN~A!-*oMT;2utT3f#1rW7`^V
zy-D@1O7C4uWZH*n1A>3OQ)!y?cd_Cp&NwvL(F`=VvFa9j;>T>LhZ}g{)gKagqf@bg
zWwZ)VADH(dx!YUTj3~}?T31&F;uVY>FMRi|I@H|Nk?-JnDNQ-6qr32bi0-wBZCTWz
zBjz5J+)s1mfIjmcaa<R}`6GtcVIa$sk6c%5pi6KG>Cj`fanp>Vz2h9lP?c*_*tu-%
zBj>F{a$k~iPx7r5gJ>LJ0a<#zz*QkmJu%+2hOa}<tehjO3q4RlCK<*@=Uo+z_ne@U
zUSoH1(Y|ACq;)5)T+)0*%HlVWkM~dVHLU2lE2GY(j!>5@k5MxzPnVK9p7ml0dt_HN
zr{BjS4Y(QoYi8O_$Lm}&i(8*TM!!C&({YiS#@96UyS8Ovk&4<zCOTx%28@qt$~rT?
zYo0bd3&tM?v>kaKTdQ2i%y`?^@UN8qIR4Du57DFhb(M$uQaLWWvaY{RQS0B3Ywm4z
zW=YIjLZ(j#p4sF4`&Z8&9PFpQ@{%anD>uv(XYj8|5|s40;Z{+l?<J}7mXV@IrQ#?c
ziU`Nb0w@G~R@l+r@pA0U6m=sf@~*lM2s@{k=yAVpLiP3?vs3E69+OW^(32uC8DI#<
zxBmcMuh=H_Ii-wgU2b_+x8bLQd#eBe4TGM49@UE@TiJ#40F|?i!=UyR=sMIUR3X)h
zjD{nx<Lz7?xoRM0EsctP?mB*5tD-PbS{$C$)Oxd*)I<|ZNKkzSI3N9b%3z8xFmc8`
zO&ZPU7YEA69dXjB30aU6jDJdrE1rZNfpusl2ON>q;<4`RhTLt*{v6iiQi2!;;8ZsN
zF;m;th{0}Cs^aW(uWVEt4!to>js|WpaC+jkrneqgARL+?fF8j1<E3jDS&=H&MR;vU
zKzie$ts5vrM>zM%6zE~kI2az4eh4uiC*HT^IVDrM2t~-}C$?)pQjC-%p*_WEEPi8l
z8K`v#{z5q0-!$6N!i=pXad+|Ijt+h69^xO}viI-xu8Q(JpPLx#S(Cy!9W&Orr5#UZ
z6%~vF7Ef%^SK8dOG3k@Q9MM?vv!~i4=peY-4=PS@KBlT%;La6$bBeJGI~B>`j+y?n
z$3x~EkQuqjALsI_6CuHDkUmk(J~&jyz!Jo=oa2u9{*`Q6cIrEk+N4`>5xM#E>BU)_
zD3u{M$U64LShk6oQASDXb5|Mw>Z;>;Cj&im)~&^6AqAA~$mgCYC6o=g0!ADHxW``H
z@m2oJC5-?ak%Bw_0QIY&I#QSv1A^a&`TA6!V~W`a2YCk!J?VkZtV<QVYqRGEjQwfc
z7DoqfJu~`M{{SF)<C1cC+6MxreK}d(JhEE^1p|sfb3X%ad2ZcmFo2RsJPhNdYTa1<
zoxll#SCUBd{#97pkcF7%p5E0Zl?GH2dIQjNQpDkkw=J5IRr3HR8*|SX&*R#n2ogXC
z%spre8aCZMf6rRbiGa_4T#l7O(C$z&IUnTJ#9RPJ=b8#HN6-()j@4#Alw%!gq#15H
z{c7509S77>7gA*%Y3ezpRmtg6h2y0wmMlqIT550wWx$8}R?eLt%n$2Sos8uA7wxS^
z<~Xew;4JEqsdgj<3Nla9nQXr!dg8u}{hvM$2Qm1MShvYIHjv1;$NN~UCo2~iU8Ihu
z;h(`xA|Hl{YvN5J*HV`7dGf(`3b+{@1DfHJ_ImK{pLuXJ{VLwlIV3L8tbZrqJAMYf
zoRAS2Kmk1f^c9!=n;2rufU)-CxSZY3L)oiX@!ccgU6z#+D%&p6(+h#^TF1hR7zHCj
zeTVa}N0P-y0ZICFq+LCFkQdsHc1N>3*m!zHay+mJ^r@f2o?2%doOQ23n%+EffPFx!
zS1lYWNyB5*Q&M-(PMx)9CI0{jRFT5W#x|II!WEM(jCxlu;zl}UqS;2?OK$UJ+ug|H
zC(sTnvb(=qRePx$2^q>KV%@*`_2*jtqxQLEnTQ~a<aagOPVaMqtd_`h_t#IRX|{I9
zb==2udt~5#7#XH`lJi9PkY}*hZN;>LM*AM682RArJABG<oF7`rj`Z4CHOOXLn~+i;
zSmNl+JqhSP8s;O`V~bQia+3=kxl@7$J6A_4XFoef8RCx#UwAuJdmH5hfx0Q9A0)_n
zGM`Gx3P`}|UZLV$W({LUT`8mcPRzBe;N_xkPjGvRx#3TQF+*m$pNM2gErRY?*|ve`
zSpNV@hM9cH>Uqb9JQu3?vh8B9fQa{Dr;~(V@%jq&75GD<crMK@wd<&12sb^=$IQ+C
z%y9#g^*O7vd{43P+)ZnuM{jBuBkgY<_x}KZFaH3lynkKRFE#B>d+Y0|uO^WH0IZ#|
zK=d7l&{MsKd!y8EK0PjhVETW9^mr`3U8|_wplMIIPyio|c`mQ2>KgU)-s-pW>Nk#Z
zW}V9s><2(ShAI#+M>LJe6$z`B+iJuyl}m`^bP|PN#T%nwgM;jAcg6bshlq7gwpr;*
zrsXy?X!%lmHUJpMr?p~Ztu?sfoykazx@WyH`gG=|E%NOOpyxaB&(rBlaq{${<whhK
z!K5Q`r;$MGNU@PGHE(VfnG*cfI}Gto!WgNZO*kMF<royZX@!P3$fzHRutpS`tsu!G
z+pQ!7FEXzlwLPSwAY0nAj4ds}^aivNFssclZo{RQ%+G3Ko+%F`(7k92+?@wXvl8>t
zt0B!=i>RQIP65S8IOdvHy*)-&ft=68{{Z<Xz9IhruQ@-Jd~v)Ft$mU4K!3%%@eloZ
z$^5J1KBls%x$WQ`p)X(1m0xkkTAh1hp+v_xsgkj1_2T=Ipu*>&&tFQ`(xoxCE1rXz
z%tZtY)tF>4oP)(nOJjKHx$B=BtRwiP`ybDB5R&Iq@J^u|vZ6TvTTDWk8G12X4o^;N
z;DT9O2a5Yw;m5@bkB1u6mtxXAS4q9Mdp%Q1xEYG_*~S_+9=?aMuPptg{{Uxe9|l=l
z`1`<9KCAHKSA0gz%-US%lWjky4*;IDu#@_;aOn2=^f@tf<vBX5so(+e)~!Q6LB}W3
zvr*1DJdUEIM*)ToGu&6X8j1Bh%O;J8=ir9t+P<p&i8PmeQ(M*~-TSzf)&rls?0Fv8
z74lqUvH5;sj00bB{=|B2--mjo;A9gE&8zBIXODg=<jxw&rDN)SIhEtJfr7}BtdNu6
zaTx&qBCOcV*H(7u=MwJNKi(Mtewh4gBT~3Fg4SewV{0irm~)@ckH)NA-zrTck+uZD
zPan2=kHM>kO6C!EAE;^@e#0{z!YMqL2d)p@{R#Z5nblJ_K3gtTT%YhYui^=9rqS0>
zicj7=#7P!=0n~eEKb>G-UnR@iN*o@MzCCfltkG>2(%lFfMjAo`XSnpPb6maMD<J8(
zfuBy*w>F@$#n%TMA4=f$Jwz#y$;m8mJ%v`{Tb3?!BwsE7=v&gW=77ojsK`_!=40NY
zOW`Cw<B+GJs<P^ARgoJAe)NaGu%RZ7n<Ol@ajw80<<Bin7a4H9ao@FPTwb(_2i+a9
zQ5(gDNMK3ytx0D|9nf9@9QCbhT|~kZD-(~;@~#>kPckm^xSoMe6{#Me6Ye%w9FyDe
zq|hU;`)kTV?l|j`d94jL>N`0BX&WcGt`&8JZjN!Ez3W<bx~Nb|Ap18@&a-hH&r-M6
z{y`ckZ1?R;CZJqw$0E2pT|t0X^G?&)bgR+pd19tRl0NY~RxX{1xzgNRh(TlX6Z|08
zIe&7}+y-na4vHzmRaaf7Xg%{(udyne5KpdaNXEpn)rI(mWo&Neki_D>qrpECr;gQa
z9^(;`5;6b;A4Bb5KQz)H+&gyltxG*j#-aBzY#7FS`_?#`F{>L%T<yZnQIh3nzWig=
zV!ZIf+_NsngvTUm-^8HgAM?_`EvDGMCc6-fkF>Z<cN>Wqude<md{MZw@Gb0eOPNGl
zA|o8OFgo|Ik~}FF_Lqt+)={}co@|8mAaZLA4KJ(0N3B|JZ699z4zjbl@t1~l`<5Fa
zD)NTwwFXD?uOa=kJW-{1kK=Zrx?GPav4t$|CS&qRBKJ>V4nDQsd^^_T(zTsOQ<;E^
zOtbQt^T3du?fEFhe3#-aY5Ya6-rqa!lJ4S3y}o87eweQx7tE%5I5w}$&HWnIX_D%A
zBq(hoTa_SWWn5S2hm37u@o#{&T{d)SlT^{(Hc&H{F=cr5-HQD8@adA%OuCFXR=Z?~
z1KCF>^{;!;zBy}JRi(TSdK{cb6Hd7bpkRKr`Ck@f_`KFwYPef-({1YQ+3)I)xy!0h
zeDIsHXw?0SF0{LU4(ob-?x}mF+<m6rJF;*9Axg)OK)mPQ9c$I~uZl8w8&v-Qi9PT3
zg`f-~wx8u8+wYuz7C`8D$vksk7wB5G?eC3mwM$k=FD@=#X%Zk-G?2F=Xzj*vUq<*}
zLAmjz-Mz1wYL_wr=3GqvSVr%XdwX@S(kfu-RFAi;z3C-&=VW}`q}53%yTxd4P4SZe
z<Ukw0#2T7NEvzl=nm`8UmB>Nr5S^f&`4ze0-2=p)Jdn+GcRbdq^GIiGMG*eD_O7wK
zG4R*n6@p!BS8L*(Y1A0(VPMgN>yJ-TPvKl{x$zUk9}wqTXktw#NCRX#b&B8>^v|Vo
zQ^C{ToD%aec!woqj>5|K!`==8EgmG-wP=pko==wABkVtj`wH!RFRJSvDY;nQ!_K}O
zhqoIB5BI_7J?jhMpTe1S7EnrC!nhJn-*tVL)xS#h-A3xi!!}Q(XhX`33|w5Fzajqs
zd%bl^l{BrpJj%F^S7oaf(8J;R8DTNqY76quKkz~7tLlB~z0ZxKNj!*ueKvk#Ps)eX
zp7o0bo6eygUoIojK{-CvHSL9^t**;-X1FIFTKw6rTGZl|<Yh)~Oy9Hd6fzO>s}sR(
z`24FtY7nykCIFGpn&5xo3%Y%?&mj%L5vk#RwCkH0?c;5X{BeMKSI<h5l<f(gw<2p}
zx_M)i;p0{wtHmPN*-apSC@T&@91-}6@>_2f>UL9X^O@Hlbc~wqJTc;1?P=0AiCK?P
zv!BwnmJb&dc(nd@Jo(guc4vKSD?|WU-IR83oAs>-rY;8MjlQXqURUF9k3JCie{kA;
z#qibqK%<>2O=XeeKlIRpg#3B-uN3k3?IoyQMD2CEOVgVrZ7j(fhx>AMKS7GTHeZjZ
z?<*v7&ZR3KQ0qSsv>y+-ibjs=>~y!aK=1gTi};%2HLs1Vi6MKNZ59RT99WuSJ=oz_
z@XdTR;y>Cx?UC)Yz^V>#Vg<iJ*YU0r<HX)D@om5f=AysxN!RdE(!RQeQsS>s{LV>K
zPvJ+{zBK)(w5>fG=4-|u^3w8B$3O1yE5&?K`&8>1h$&;3qv$sN?Q!(4I?=uqYC4Sj
zukK};k0)kbhx8TOL-4Cg(jW%xBy0WYUoLCVjvo(6?RG`nju!XgZ;CuiaL4VEEP2Zq
zfdG4Z)hP5o5_p2+d5+<b`h(BCdhU?=4V~<EmX@mI5B0JD5pn$L&Hn(3rg>d00z+;N
z0af{r{{ULFb5ZzJicq=d79R<<i2}5dtYG!?B>t7ZZ}7Qo7dKbY#p)KdCcn6ppE7A5
zKCEj#S)N%|G{ZZb4D<QdE3GMLjg>Sw_+!JC0GY0(xM9>UO5{EwS!nvVhs~|DY^Tmr
z8=>GD>96l1w~3&K2LtDSQ(O;<4*fi@A$-7k{cEca&Fy1%)aS-UEHy^E9&M<29NVdA
zU(OzcS1oy`c}&2eQV&8a(_orq`9TsW{uZb&^xJ(>EUN94bqA=gq;W^VskOP~qR(q!
z<w@qJg4FKCZ{KL<Cm^mGsGD+_2dM8^$~M&6H93f&vMnG^=B<lgEd0P4v1GgMV>qi`
zPV0<gjP$IxHRiEdwERc7jAU_9n|wGGt3AwVGQ`JjSB%tCVi4~5RUDscLrz){NeB8A
zJd^VhaZ$x(yo1OUcG~7St)eP2flC+cIq6YfOR_Z}<D6tt=xr%G*s~pwjRDR{;|8Vs
z9!DT^?N*g{EMo(JD$H<h0-7;$YpGFeQBZuNH1@UHoG9nouLw>BS(?q`jWfpX1zv`3
zO?5fD>jX%HkT<d7v)Tci{og@~>91`Jvjo8^fPQsCOEwNy9X%^m1z`5?nJ&g6!L_rD
zlh%>#PIoWft-)e>1PqRNsYaWSGBA4mYX==qVini3(du&2#_}DtM1va;!!_B<q~%A;
zg*`oLt9W-eNYXANC!Q+Wwy2sok7-uYdYqNz$J*US&>RdF^{t-@=yGdXgixp52Pe}N
zt)%!-PNi~=n~{|5&r0=Ah5iPP3rMHRvr9NeAaVdTH*EKvSjH|)``7j=@LYNyfegYR
zo64v-=Kxo__^5eS#rG8o3CQ|ZKf$dg9S7m=nj!}e2$i_x<Wv43T{y5HWD*Zfx%yWw
z^FCsvl#`cp!+b-0WYr!<Ve=A3M?vaqmejBC?UDEz-PV^kb2lVsBiqur{ZW{?>`inf
zbBVyGHNlb24MTf-<;EjZ^`ww_3Awv-n#Nhp&E^5f&uZBl(4}_d#@)vx@dWBi9y40r
z7}48Gw|Q8$<~#$|w4l;aHaiW&8OAF6pc_Hs9V<6x&gj(Tk18jEW?UQ-n%c8gmj#X%
zvn|vTGtbhs?J`P%$o1_`MsK;=cqT=T_(ld<m#?6&M$+xxdyU8Lj)J(q0BNO+w@kPS
zFuZj&>N-`ImVjfE&3W~uV;^ScG&)ERmw-E0FRI(NUk&)4)z|8G5n0<3?5Hu+*AJ>|
za6ubk!BgB+QMt>EX=vHV>UV06V;O95ScdVVn;?L5(v4eCS#KSqY=hl#SyN6!<l$6y
z&3m-tq42e8+DD@Jb6sM}tlrh>I!(Zt6|2H@JHPaVAm9$U>t3~@>ZuGdFC~cQi~;G2
z;Kb5h4}UGiYwt6-fwD1Du=!4Em5jb}h6got<f+eY)#AMw_fV3QPA9~+a<u+uKsxuw
z9qYyXd1fS-VSL6?diJlNwd-(TV~m`4?ewn?@yl$J0AL)hPu&OWU9?@#7aFXY<?m*I
ze8{Ln5*yHZ<NDQG*kim&p?*USq#WRTR;Hk;vB|)99RaL+_~fy+Ww=QpnUFT-7|8zs
z4|?>XwLWfAx#u1wwTdNZQy5&Z$MJjpYsmazb>s$%Zdv~Thzw%AN5pp$TRf5gB#e?3
zvCVlWjU?X}%y9V0&#!OBwzh(%{o~CwAC^W?bI)8;VndC%{3=;dq?i`~k8XmNC>$}`
zG6#Oul2>Q1%%X<0ViCaW*!QL!%HJr?JD&A%qYg9E>M4Odl{{o0QCcOTmp%PV`E$T*
z^HwatVnYFdc*Ql>&MbRl=~gV`*}xeizxnT4MQUd$XjzIMkmL_sRBbb0pgVEf(ySNq
zl#ZvSYLXrMqyXdYob~ssc1BV6wPnkHPzWZg>P{68InS+26v`CgvyWQIzPaA*jw)Nw
z))Bmn%ZR+j^vz{nLSjP0k_Y8i{O!zo4@!dIZ;iJM<ae%CdP!K!cvmAQZv%|tigPAU
z-cknxJ!q&jvF{dE=2M-d@I8e~<?f#t=hP0>cG4}exC0sEJ#pHZ6iQ}pr|DRo%tLMl
z6j2OWaHBheZXEvrI?#BEtZYF5^V9I85Qy#p9sBh-tqj<S0<ludfI!Ai=la##SOT$7
zop`|JqFuXNcO)Kj{&=hKAVpFE9ddC<RtPN!3^wuy9;eo<wW>6zq%J_{Pj26>Pi(Qd
zD5?Xf=XZL&_C_({ObLnLf!32_v57JjLn@|NoP3~QdvR8@YZ#(}NoNEwI9wd$Rrz$8
znab_m5ddU#&)3$phlk~wAYsIYB@8k;{b^j2kG`|}WT%Wfa0%&*<NpA!?@(XZ#|aRi
zBqj-L@<H$IU5jYT8Ak)=!QGHf(eG3*EMto3GX-tHaz`!O^Q1G0HpLFWxF8&p*ix;o
z^#B3rc+GEL+aEcZlzjOkxfPxw+Svn@I3)i7Dq~s^#Q{<dGoJNkILIVt81(H@M!7%$
z?UPcfj!#e1AB_mT!Zbdnt;7NA*RD-R3|w=Nc<+zKtw4t(J%6nxXtpdsn}`@Z)a5Wh
z=}n0+K<;YF$G49D)g-a@@#C*b)w5CZ=kAWA^dDNK0vC}ZMawe<#(I<5y<_0t!ngR5
z;uzo|lSs5}*He%AXh(19Lv~jt+}a)I?D_Cx{t{^;*R_}k(-7v~7XJX2<bM$!g1)JW
zB(^|^p%q3ISAYdar|Gb1HZWRR#v!+bLLx61091~{Jl6!}bI^noP<ffpIH+XWyqpS(
zS;BkODeqTqOAp4fG$)=>)DcwVyudkND?Z}<79cfeTwUZ5oc$`)Zf7Z2zjLVs-OmHk
zx$Q?+-zidY@BFFv7h4eLu+KHkPdh!tlX6lqo_Zfo*191BlBGt+_=<Fv-C`1Oj0qp@
z9<{@346|IpXd7j)%gpQ7Xc;*7>?=m{R+-B&EJ^HZm()w0Uf3L!^HeAw#mM9N*JL8m
z!6rd*sWtV?vo;h1wNP|b0|0&l70BzXW4qm*pp0Yb#dO#I0A_~v@hpq|iDh9CkbK2+
z$Mvp9#S?sivPaAhC)fV~)mvK3&6Tvz9>0Tih;F2@Tj*5$s9SE<`u_l0+3^R&T~Ef_
zC%BL;yN$_hANsie0L1&6;58RXm6l8}VbIax+v60H*vdgFe|Pk$n$tp{`Ww=2bMq41
z_ca8ZE<06>j8qeQS28-4ga8l?HJ6;z)AXl>-ATI)tAkA{YRg5>6w@6S9g5{Cr43jy
z(+H{+hzk0OX3Vu%*^@wznqVYgikBHQDsxh7#ZV0kOMy}d@+mgrfFs+LBB{<iwa;_K
zXbAI@Rh}<CObU=IT0nM}H6)x>_JBhO0983hH0~y@SD`)W#Aby~b4+$5PSs*I#wxms
znRxu@6^p;ZJt^_2G|0~2D#Sya(r&{@<^Cw49}c`j{{UW6e=7LQ4r}c{ifR7<7T3fl
z{+y)#Rq<fXc&>Rnp56so57d|fN3Bg7iiR+HRfxzupTd#rCQO5{1kiFml%8i8G>7H{
zsm?q0qQ#CAV0Ae3t^WWB_}5YJieKB>tQYz=rlu`NM7>`&+VL^bl=NZQ*92EDAs_+k
z+K;uBQjwf-^ENt<LGABPl|9N*Wls_Q&l*RL<eK;4z0Z-TADR3qsZ0dk{{Uu~;7UK)
zIB%_ZXNbHxsCZvfdks@hy3((rBuQ;`5sk<E1dRSQ?#nS(S65_oAfq8WhC72(YW^+N
z^(g+ytHWvH?In)yv1@sfNKgGSpF7ZdXCBnlr%&P4ekVNW)0%eC{s+!*Tr5RSS%5hm
zwwnD7{e@lcJ`>$N!AG--`Jz=S10-Z;KH#_GUmj}y3efN5d91B2w35Rj13#Xqd-IHu
z+P=sAh%N4PKZiPvw22&uw<K+g1W1QD$^P)^{{ZT)%5^1CYIzu{Qlz!B^gZXr7R<H`
z6?Y-|gL)qS0FQsAVB6|N?Y6Rmux=doJ%2jsb%{mZojt5hWwplWr;naTO!^Vi+*g`i
zz1@zHCDp~O$q&s5n0?i3{{R=Jed|>g(dLle>!t}WY-A@WO!oLccptvM)~RWFr<gDs
zBq<pFMzD3Kd$^&zjFB{4`{JqCY7Z3I&fb4RRmiMW^g8)<cb3%dZb-+{xf{!;xw?J7
zGO5@X+r3|&V3zJqd2@^(YIlb0GpHz;a(eU6S`<~7CN>L6E=QGvxQ<3X?OKb&R}-R$
zVR`0HzR{^|pHa|#YuWr4@Y)%Tt>yGfBo6ypEa0C)Yu)r84%*q;i+Ev*+9TNz08jHj
zjd|58s5>NdrF47`sp<EYQWf%KBRDd;5I^7}`kKO!u--Cu^d~>e*XVAu@cP5Vax85f
zu#I?-m{hZUG2iQ6W8;6>b56clCbG&%0Ra#V+3YrP*XvhL6$hlWCP|~@NH}ldDmtE+
z{A$zArZ)a1&q2m(vGGU4OTP`>yfXl^N5erU%2VG2WA!!5wY#oW(C#Ca2RN@mD9co1
z&#6hST)H-S$0|K4<+b~;0eODFbROoiE^X$C&Jj>`<?~5tD@LKZ{W=<MC0^%Jnv8&g
z7~|A>)w8Oyrry};ST^@wa2#g?wOh$!(~8BSO^bW>c@&e!erriJ4nmAz`qwcG%@||&
z3Jvi_+FKZ+#+K07pI6&*SMKBGE9h#Dp|8EXOwqs20r@k*tgTa0Q0ps<ZXY-wp{`d~
z*3<0_1f^8+PB{E(T0KlW6|P}?U-5O!+8v~BOltEMa0tV3z&^&gZvo0Q+bQ)ahUYj`
zUs6tX{{SlH{CTXO5O{%Lk8=pZpu=Z1)96vH#=WQ601c*?hUViOhI7KN@~q^QxlJ3q
zj?>3_^u7@IXC<OaPNAu}j#J&Ya_8_ISBh9Vi-P0~4hPb=ek^#Ewa*d7YLB`b81`^@
zG9Fv}E1J_FH+Lw2e5CVVEsB>k?$1t?xg&G4)9)3q*b5R453ka$PpzAagfc!t{{RT>
z=~+6W8#Guz!vV-%m6vaIwa{!?GQ$e_!~0h(AgVrOdKgO5jJc)M_Adn5J;A@!UNl)Q
zWZqeT3X6<nW3k3-&@_EQMPDw_m^eA!+~E3w>+N1)V}E%v+s7)%W1X<E#F#+K)l~P#
zdiFmGXpE6Yp$)wF%eQEUpwCx7-srzmTc>5I^fe;|9SyAq!}97^s;;Jcw>fRx{${=Z
z0O9Y$$YL?cCPZ!bqyy)W{{VG;2fcHi71Q)@25AvqYUPXE5Fob$GF&!&I)wW5>t5U8
zu`Z=&9o>`pc>zRZW6KA&D;kw-v_59NU0Ozs!@ikjEiD9aC{UyY%Emna?$xcJ_yjen
zxxYpLjmZl(S-r{rmAT=c3dN@=y1YJp!R1koLVxw^r69^dZk>RxbvWAh)t)TncVv0V
z@O$~J$b<~zjsPCjsipW{&c-G$k28`V{{XLDgn>p>44$O=RY_w|dBzV;Kaj3^wKTN$
z5|g*Pp^0^3zE<0KOOiQu?^t@4f@QaDtF}NhitLPp5x`PE99J>pzZmGA0`T?Tx8fZ^
zEOh%&3ryUS$L^wm%Q5Vun&ISAl8hFoZ91~2te%IGYF-w5B)PfNB#zF))l9S9PU1M@
zLCI!g&<goO<2UVbs%p(YhCUeH$)o5BOjo`mRoQcTGlS%0W;ptd-+8+6R{sEK4~*LH
z#)(a|kk6-hbV-{{5+=*e-JWcH=I8GH)z3`VncMs?n%dZ2>mb|1A1efFll0^7A8PwN
z77;=!b>6*?HxXIQ+9=1;b$v=yTR#t5JPW%rPcO|Rt=-wW<oc6a@%UzKW#e6ASG(3P
z3HfZSc>%o(^!$OZL+}^D{YS*fzDS!((~llXf2?qSyhl=f2TJ-M!oLS}p9g6xJkYJI
zBw<ECng0ObNA8O5tBI#eNUcm^2{(BkG-)3Nt^`7R@3UxDha5=3oc%|0_57==he^`3
z4L(^cG|RnRBm4KaDq@5mQb-y6E8KN`N5dW&vunG^;xps-HpsiAZ=pHI)Yr~m5q=};
z9yPrv@(48e&+hGF`-*)|I-2E;N!EJoZOLK{OX4;BpSrdQVF>>Ko(ePg8t0z==TVTB
zW^DClIb-<MDB)?FrfXu#&@MCiR+8n8#)({(&eH&K@_%-(uGQcX$4pa62$&TU%YnyQ
z<EYz4(N-OCAOQ1$RW9T@ZMk#B%){je_*cD5j*MPef7C>BbNolrx$Q>%?yeqJY?k3y
z&{WfpQtm}P!H^XncRc&n7l_CFOZNBPD`bx?GJ4hzh-A%}{{Vb{TJOWJ-pAD9#=rJz
zo`*OF+*3$x!PMc$Lz;GTMgctZucbde8FdtQko?09{cD?=&*o%h<R8kqMmt%5y4`Ap
z>f&9%eQFrijOP*B(&=-O0mwD3n0d<qf!3;P5n-c##+_j{;xow|J?jRwI#jNOylt1c
z+q>lCk&#WfCQBI{>?qv|b^K{HETsNsH~hRS90O0ez|tLnP&mq{#bVB+^g@dj2+>AJ
z0yH~`{A#T5Wv!ymSw3LLCmm|t*#bEeC^tnO1cw8qA-41H?UZ6LfU^##7{?S}LiJAU
zw*pA$4nb2dA^76A)-b5N`&K=nQ4Dbst0T(6aHWAG`Bv19B8@VvSpC&rNvlQ@+_h^L
znIeKa^sa#ECS8of3ZrrQRZk6C3yD1E8&Dj4+*f}NkY&=|OnlOj8vZ7(D(vQ~Q<WZl
z7L@V_aJ;*4aZX<iEK1EM40;T8uTPi5DPsWn2g)&u$&*YZmLnhz0PS6NJlNE`I5@No
z&Pym&>x^(IWYO|Ms$_cOt#+T;;ll1-eFZX2IX02N&lnx4*wNK}8OeWR0T>_#lhBH<
zailcFZC%gqIIh(!HtpOGrtDRS^bO-7NbkG3rsHyBS;Ab$4{4&z@`cU=efGyA*jKFl
zJ<{0UTfwxfL6>mHrBTtmF!vIO)icg=2Lio+!k!Qse6Ta950FQx`qc8Ig<l<3qUB8c
zOTiN$@NS#2l<a{=;stXb5pN4xMymN@P8T@j5A)4g_)BtM!!p3<bH^bWJ+oPV5iH8-
z?&OdnBxn1;_aDl<#H}WKG^XWF9*2)=(+Q?x2Ymh=s{-cKF&ut%r!=N%U%q+Iy+lwm
zjP>tZtqp$B$9R|<OLKwGs1?m=ZMsd_!6!dYdhL8k3e4X(DnEv~sBFT@$r}Cd;RB^n
z=u>g!vDny3w-)|mFj$msXY^1xCcYu@*X`S^Y2Oib1k)_7ZhSLsapcDfE+vrx!Eqq&
zc{u6PzNYxK<IOYS&%*6;TdSpp;p}Bp3d43;c7Q&)ugQ7m`)m0y!pzS5xaGcrxhhYW
zV(C(C-J|vA;YY<CE8>TQZ1vqHVDU^a=zjceI-?H6SFm_@PWx<1z;I4JwfWKi00jN<
zVn_YE;mypmnUd@&JQX8h!1g%L*1t~t7YuJTy}Wo~67%kBRW{*s8q1OLo`s<4i)CXB
zNx&E_^{%Qp;jpz5Ou<RuZasa4V_Fbh5k^jVuY&&oZ|~YCP|$pP7lw3AT5D4T3%uLS
zw`&q|G7krie+u*CA8R`{LzAZmC@Zt>ol5TCRhB_5!V}ZWRtF!2dDo8gN%V_zc^erZ
z&mnjV=yC5~lR7uXe-n6rUK*9Iv3RpU>nh3Tzup|+^V8SXzfS%Rd|!*>Pr<>f%NQ52
zskgdPK2sn7?osG-kzMMKvbCMj^LV~#ry7dR+88rycyPesPh&~_q2@Tj7(E-(x!pwO
zIWkVwVmEpUhfBV1G$S_U9maZliuUQjSoj*%tnANA(Pi3Hj=gtRr)e{ZB1ZL4e;V?Q
z3Adu0k4`Jn^fo_dQOU(}SB=k80asfdnV?7{fopMb`d2gIS$C<y;<|OmTKSrhw2#$r
zlIEtFpR2+owthJyADw*T@jZ+)BKI;kL+RSSmA^me?VfnAnm#NjI*q<T`L{{OY@7<|
z!)|%`o{%^Px_elfO#RcieQTB1b*r1?xsqv1Z$C7Sa!LGuO6YY58+IXF5tF!eu3J`t
zO35?13Ksy8+@7EPe@gY1nx84HtZ;hm%fP`AjEUEt!}|XKpGxz86^ON!Rfue5VoHPC
zxBmcIy%WTdjL6DH`~kq^*PCCodF4V(!{%f>fBL_lwQU_%q{m00#Y<vRMu?2L=Re_3
zNR|lpj9}ytc?aoQ`h+U-22R`-BdGrXKaFR+AUDcE$8-LF{Zv^UyzWS*V+WPs`}eBS
z$GDX_9CYtjS0L`jJ-bv6qCt>-ezeL?#Vx<UaC80M)ig>}<P2k}6d3S97{{$V#&&U#
zai6VJQ8InTPbLwVXLmJ7%d!E;13Z8|f2CNQl!s`+CzFo#k#Q+iEBs^p{uHECH7qs8
zz%Y0ma%z-~h8vGxT5P1UgOi?}IjC|&kK)1q0646f+XOanqD=AcRqj4m%N596oMWl0
zQz!vg4!)wQ+-2QV<O9LykytabB@(d42bv2r4cT7wQcahYh>AD$EyWcEhtqbpWl-=H
zal07ip!+a|(;4J#Bp%h$8zn?-gba-F$gKN_O2`U=P<O{l#XFs>t0rTPFr*xJAPm!P
zBW=ifka~O7tAt;W;I;-hqBxYOC3kHkRa{mWAuRE+!j{Ku=j&Gx5T_C3G09>(deo6b
z3zYe{1sEUasI7f6%rS2~Zi)U;o>&?#A{!Ws{Ky2D>*x>p?^|}3zH5~n$Z>^UaB3TS
zact5Nh1-$2xCf2~D*lL{ylJ+Ru;V=kPxH+~U~K5Rc3Fg&ask4z$ie=#(P|oeQRwk`
zC1oJ|`$kT2zz41|R4tgKVswwomQVpsfC|t_RAMH|s2@1(`1Yw|IV;^g{?VP}{Fuh>
zap{`pZ*<+W{_~TJ1wBb8*Bx<QjN%(K18IHW;C~fz`u&NS)sc59;G7H}^VY304nA!?
z-tT}82_G;6wmQ~a*4|qM1QjQodiz%N@J$z(-eYXR!Nz&VuX@c$m7!*QGiR<l=M_V_
zn;LG(z#!FxOrZ+DHh3iU_Nl~0<W<@a<J2GX)_|LL5TnzMw9#DF0=WQ;4{vI-40t)`
z&~r&^0U!XOM@$pn*WRwA3XG^Y<06KxLXjkBHxF)bqO2rkNbjFc^!U)JU<Md*j-%;Q
z1q5Ur#Bq^R${g>Zf5N&v`tG}Jr`rY-PTLfY!1X_tYw!O6hMo+u@GpgJH0wyh;Q^8R
zx%V~lkL=%~$1jVdznE@V5D3Bk^B?}TeVuOpTnW7Mj(zKvtgWHjMn7qj$s-ZQF;p%u
z6#>ZvdQ&fNh*Ae1PW8`TUS+UOG1ON&jm@Pi7cR9Tau<PFH=2-*e)m4LO7~M(fdJ*Z
zD6W3u`^_Ym$@D#|qBPDq*0H5`tOOlPXQfj9*DaHik<zkm14)uwgWDBfafsjz+<Ml^
z2S2oI)qPu$xraS-SvtnDBepO_gjQ|j4(6kDXZH5=syDi6SlQXKK<`>4keh83CB9SF
z<|B|Rme%i{(&{p<Qe%hQoCQDQS{jx5+Y+GVkLO&Lxp31uF=LQFTIh+)%XhaZFuBO;
zO>y2WzuRRYIXJ-`2Sfh=*R5-d;6#iGBLk&z9y7lWWgsVU!RPT6X=+K$j}^x9EMp<Y
zK_m=g99AccJ|IEhh-1~RS-iM1SOe60Ry58<!Y~Al-80ubR~!39Lh|?_IoKsd1pRSN
zRyBf6q-N^Bw%xs)j+WL^U2i*x4TWBxDCg_VcU}Vht2Dn9czW+r)F4eY((P?vw(~O1
z3godQ{cGSU0a+M()Ryy@!3Bu)1J<}XwbdT0r53D@)sF-C<4*C$l$W*>31f0c2cA8@
zFV4Er8QYGAzc&69{7}_wFSN}P(&lShIcG$a!f?fR0zf^oYxIjvDr{ofq-8-FIV5(;
zH8ZkzIk{SPV}xTo^Gs;KIID{v9V$1ECZz)^G<;^DWvlHT)dLx*8a$k0s9Kc>sbfs7
zIink~Q!IN@sR@vxb5r?zbf?7@Sk&bLtyRGl!enBPC}8o;NLfeLpf^gabHyuw$y_30
zmdQcO^{nfwr7yQV>g~Lv$>Y+2YFL<u1PY@nki7M)K;s#xWL?Df6vk31oSKl1Ddjev
zlx)bR2x^X%Dh>rIt~*r89RN!jahkA)YSeM3%4$#H#b+YYIp2zn{{R-t#9#I0C-Sd}
zk=f(`9Csx21Ju{rUlyNd!!HtZ{{SvBe=7K0uAV{@c?+I#$3C@^S3R5_k$$H~7%y+V
zRP5c2$Eo$lu>N!`m?Is|Dx6nMBncY)h=VxvAK*Vz&0f<iq}6ZXvxw#0x#N&MYV3N;
zQ1b~VZVoC%oGM_bW7i$AUiIP626%r@(jyP03%49eJdMtNmFK?|d@E<AOKEwhMA1P$
zP^Vlt?M)<InmqDpH@@OX$j{+W$sumUCtOs`5EulUpW&!V5OdO{k|A%`s2!;#!(b7{
zDZz1{L({!Twr?^s76^djC#i4lk5lW|dgqF9V{1Z++eviG6frcCxeBPGB$Lw>>tD04
zj8B7Zb*meRg}>WXJhVc0Gply(oceXIHi$HNXvtDkqZK&Hk%BYmM+erlG|Se~?(J?7
z6^wF+8^-{Shv+M+n@U$_C3;-&aX#empNg#R@1adY&WuSLtk*W{hEhiar$^_vTFuw|
zKd*RiN0#c}I+l?pV>*&Yz83|~A&?K`!-LVQkno1Ao-FWZ+O6l1-rH@sRFMj+ob$Bw
z$J3@O-hLVB_u9UbExw&SmFze6tqs+t<K<3=uJz7Q$2fJ)c+GUwe4hF~axt^zXM);k
zD{^It;}CS)8@4#~2fxtMU*XT%<6#3BWanZ(89tv%_pNu}1?HTN(IVg541Ve(iIdde
z{qN^qXQ|ohUKc6^hD$_zDiy$9dJuUAj&)O5GOezsm)zS;XK6GyF_`WlKQf-V92)9;
zI%6}#akLd=2PfARcf|Jbc#Z*iYK3mLFl8f|G53f2!;iwajSIxF*{pEN!Bq>zA#OI6
z<R8e?t#=cN^pAqPQ2+}AxDIig@&-Mt)nxIem!#WE_O~c)qVpsQMqlylE9BijTk<YY
zqhv|SOA+VW@UCC>kNEbU2>5F<L3N#4>O%xkV>_5HAfCWv6ZH44dEuhn*`jjT_CFE-
z0Kr2172&;Z=T8#J=2}Q&Xzim-jmK>A{^GrV;dku?;ZKU%XqL)*cDo=6ZMW@_Sd4+N
zoSr%3IOuEeZs${;;^t}IR(WP(?B1lDlk9U_eirzFt@uX%E6p!axv-uuyDhtwbqB5p
z`=t8RI&^0nTE_m(t(p7%t7!UniR{h2{kmU8Ajg$*o3{NA-XDc{w~YS)X9yXcv^Oxh
z1pSuid4PYq!;$|00<VjH0e;)R5*B!1)cj8JTks%)OLS%04<(q4u{|?iqMr%=H`)9}
z(qsE&ybW=IzFSaKXX)CaSbF$qs*6~&o=!5KES9IrJ}UT6tZ7U{`U^<Fk>^3V@;;Tv
z&EV^l3KGKZTa4rejFLY9Kb?Mx-ssjB%zUe4c+N=$w-fG9TA_2`q@Ky*YuK#qT?mpG
z2uE?~Nam$udNc}~N9T)aS2ixn$W{B77~Av}Y8z=nV4*B~o(?^0>faT7EYTK8Be1o%
z3K!<QynV4D_2-r~<htjDA=2IBx7~EdA+Ac`H_&=lR4YcH+*&-E{x`R`0Y(maIIMZD
zMAAu)qpf-tv!llylL=L$Byb4*Ys@2Zmy<@>J9iJQZ(~@iiFt90X;q&950v`Xl=#cU
zg3i`f7%p@2XRmIR-gu7nbn9~4fo{Nc=DZ8Wda1qDq*Wu!UZ5IH9-Sp<mHa;bQ@!M!
zVkd4+a(bHYr@lT6@T7CfLh4$V`b)2HKY$;@n&Q46XzSuH6<b;{0>U5P+_3wx#=N&a
zhoxHahlodud{I2$ZL)=m#G{~5!2VUqR#t;!OWv`Z5y=lJEi9*d6W5;gy{FqPyej-|
zepF|WJDSIU?~t+NvCcgObhc2(JZ~(=7{^cH9f#7sR-{wao{SekmophM3lccw_o%Gc
z$8#Gm-i?C-+zu(B{&TeK9EHz%=ro@U+-kP#8YzZJapp|jB9Hufn&_oUG@yDkxvEO*
zq1JpkLw%|pZ&kRRbo)7eNbR2jX&Z%xFnHWZ{{VoDn);W*q8KgX)pcl=6p44GqJ70~
z$~}Ic#nji${uGkV<5hxHMu$k#NlOSu0cQRwkNZq}`eMDK!RGVEI<zrctH}gTJg~<K
zTl>U%0sO0)q;Br?N2gYNl$&o;)I1TOUiizzGHUv+<Vf+cJ6pthAV1zCub{7~{0*f2
zn$9bRk7O#lV2>RXcRya$!hAjOba6*(VvD*2Uov++T%JikTJ~#bf|Up`iZ}zH9ff>e
zZ<)G(yQjJH6I{}}EycOm0!bZCf0bD>mH9{BJxJ|PsKSs51J<Kr*uWSa!S7h$F-}nT
zo~8~w%#_M8o{B%tYOMD=n>|VBO+_T8UBDhcO0P8Q5|$(n_mAd&g1f3>6jismqNi<H
ziR0fK=^qb#EvW1I?c!KzRuLJeoyhrD*mq)h0Cgl+=0C*G+Am%B@9_@bR+<-s!}d~?
z(sX1004%}aO(6a$SYzqXcdZ}WGvn3I$6p<<hV(18{{Vz5Mx5<<nC-rtaMK^}D)Oj0
zbv3Q~HvAoJPf~+i)F3UVTmJ6e9CI`F$Kl_#eU4jz`mH~2U3>3aK8WG1SDD!!sqn|(
z;EPR7dsU0~^6;U$p8HeSe-Z6ooum939x2pgX#%Rpf^o!fkM585jc)01T==I|hR)p)
zVu%E7yd+9H4`KS6+4$G-Z$kZ?{ufWGT*N25bM}2dPX`j;TrlB(_GUeWD!7Tte6Zb?
zpraeT%v<~Y6T$v0(|j#t7WY`XLn2#7z&MP3=Wp(y?t1jB&)OSRxA5dze8W(VDa=?l
zalZE5*LhyvgFU-f$QpIFyW(w1Zxr~7?rC-1N>4WC=W{D2c~~L?7)2d@X}%=W?(O5b
z)CBWG6d);fkg~DHdCqv^yJ_KF)!o@reNBH5_=4kBc`=lU&RcV3o9ch1bFK3qudQ?5
z8}a_1;t3^yOrqf*X#N#^cNNh2fMt(gO6-(*9gNevEI>p<H9y+;!2IhW?>RoSsc=sx
ztzgl-4U=|A1347dxg>?)XN=ZbUE`8zliuY{4>g;O&08(%`l1NLai83$3J+4=wT}e5
z$#?v!h27zrUOwpQQ(L}38S77$$Pqk}nOt*$+ZB)E76R89{vpz}ZlB9gsxrr*?^r$}
zm_ux^ADTi#J9>Lo_-$&()Z)i!g{kPwP~e(>m<?19#XU`ZEsxJ`O(FR`>Ye8zt+5*b
z?oCv<QgijFEg7270>xXt;43>#pA0$YIIFt+fLbJZ!Dg(RGDORcIjkP$(fCcrhc{kv
zZP-u(9)te?*Hq`SYinDa&m_i12W*)6ewCyF+7GoudYi5Ad$1LZk<ke~iqq*)>Q`*i
zyDlL`AOLa)HET$=4|gQQlez$<aqm_kk26eh(Lbe4qs4_j=(Lffr1?!FX&5%!TVW0k
z+=0K2p7iVJBAP--W_ZTc3fRa!@!qsj#blm0BP^NQp4G3SczP?#i-(a*v3w~fl^EzM
z?qSVkcFXz=lycf>kbK5T51qJ<Fgr2zuTr|USmV>}*P(sK=mm3H9+Hri3y`SFh`-$e
zy4e|Jy*VTIaezM>-W#56X<b;gbgL{YoRS7JSu<Os<PbW2(N|JqAYs=Wb5i}3b?9-_
z*G11cE>j_Nlw|ZH(xH7AW8_l9*B$F|E!sy4qX6^(^{qFySR)YuxPS>H@t;qnR^aB*
z&RghcVh<wVd-bhNKSXPigBw>hr3RYwI+K(2=lqJ^v$F^Sg*hE~>rI(TSJab7(PFgn
z!l!&6TJH4QF(mNd3@<0%u&f|4DRYjUf6r>_bV4?+bGVH0*j6$!PVDu6g_lAri|cIT
zdWjd_wEiVVAH&iSlY@}Ihxpec;Tz+p>sGQ11#gu7E6{vhYA<w|QA&`?NGH^CD}s7T
z=hI;3ojZ0uWr>SjGV#=!cA+SM6VbX+s={==QG)~Ww@h}d-A?DsB;?__9<{PPH5<E`
z7xyguk^%me<-ZjEF=<~4^x3Dlo;SSPm581}?Z>}r!~LT^Jw@<~NK1QES+$$E*`ZfB
z2)V}ZbDo3ouYi1I<DFy2nwh@U?%w7?PSotaxcARfSk$POy0<-AG@~DWr^$brv+-x+
z7P0YDSe9G+i9W_0vP8$Pp+Di;x$9kGO(RaUYZQW4M)M_aCoUU2Zf>~t7!`!^<Qx&j
zF%n5-M2s^tkbzWk$FLadTvAQy&b3k9wtL^~X{${y?E_-yf&{sOe>EW(MH_R|(;!#u
zpTf%*mKkmiHuO?YYW$P<G4M0uzr%qKiu4OjGV1<S2`+<n>*vW24T4Srlg9yrSK2-S
z{jvN>;q5*<Eob0wju%l;%yuyOv6Jc;u5tCQl^I5zn$+@f51whQY<-#HJ$^q3cw1c8
zXOTi_b~D|QP5{`2znB&HO5P#6@z07cb$w=cm`yX>O({E=w2WH=J;@{RuhJh8_`CLp
z{ibx?Tf-VJ!(s5#!gtp4TixnjB5yYCXqi6ZCm1c$AaS0x`Hg?2Uicqh)cilL*vV;e
zZ7GW4-MDz=jFrL9<@r}4jH0j9?ZWa!r6Whgo=u#IKAw^+A^f=_nS(H3Gs6$Oq<62;
zU)g8II!2%H4^Oqza*|xBnq0P6%HMdO%cpEt=Dqf=Z>-~k0vJ$mn|U8!!?k<2?2++C
zJs(k<!uOJoE@m@Y7d!%f`j38uel_S)jF&5!;^#4qXi3rCe@y#NP`JgTGkx>V*Xdch
zMZ*cg5)R|RC;anT_C6cA*X^+bqpw}S<b8drW$<10mmph+PnP&)W#l*G+P(P7DQ<p0
zTAb%Cj@!bv&mP$~0|W3|w(n1^dY+4IBU!xa0f`6<z>W{nyywE2MZ~EZ6lWft>(ump
zFr}n%xW?1fdY}Hid9|t|+`!@G7keFNgO2FsInFt*t2|dT;h6r;@~O!z2PV2kVT^H~
zYWe#5%R}`1z1o7gpQ>CeFu5&*UK!)fJ#TJgla0L;e-?XUy??})3d+fmv<^u;SCeaq
zW|_9WLHV#oap_$Y6_w93Fqbxmli%nwPa>paYJx^z?*4U!;w=K=EhQ~hEJ+&{Wd}RS
z?gn~+#yu<6uWc2r{$JhAoT~xQ{{V$#+-UMf8f(N-vL@v+gO>VxSE8kJ=A^8&K3wr6
zQ*Cp%D2HfLJLlH8tBY}WBZrF)eL($vYv|7w_<1g_nrPKj?#T+EcKtr0xD98)5j;L@
zY30g<soX~5dh#o#CnBFNjPeanOkXMYZBhq7c_;k$u3q9MD#||Yc{Sf_mddXnlZ=zl
zaryrMjd6GAt05WN#&|#H@}|!EQn`;LrAY*UKczBOU4VhX?kKoBW9L5E?^O-brzG)?
z0Y23?I?}l(%_|mON4;2b8XN)FjE~l;K`EKnAo4Tytu(h*T;~LLtzoH+M`vR?<Pf+x
zA6mzr4a}#fUU6HOaR{SQa{Y2gTFQ|=U8(_Y2`qn4^VXD&DYBHG>j+;#)}Wb4<c>P=
zR~Qy%WAfwis@DhQ<ALexSXk_fB32B9V?6MGrBslUF97Z%1D>9>J2n*Kk4mIwS!MGH
zA#=d%$E9N=cEyw@e9GAb^T)}aD5+g!GC5JZo=Xn&RM^(%*O5Mb4o4)^H!G9bha_>+
zy*lC-u~Qig<yUS<z&-Q-0N1FcktJA%IqQM%$6Dh%+ADxvNbj6={{TIz)Dhu!G04ZK
zQ&Odi8BnT?i~|mOeQN&zlv%h_f$jPJHCF^$(e$>vh|HTqAUR%yj<s?oTXqG|s<Flt
zayi9CVCx;J86ksYdh_YcK1m_~Z(Oi$Ty!3}?@HiVc(%bJ5UO`zgVfdW8rfSz7*!1C
zInP|<1Eps}Fu0j;3lciWa1S-CB{Co0AabLi7{_nrK`h&{iwuR>70DYzWc2U;eJfQh
zgpnwa1=;g%>Fvk8Woh#dw5)iLfzx8+uS)4{ZT{BYQzqes19J=tK#pTjdE|Yhe-;qu
zpvP*>c&12*V6Tz)ladFiuD4LLM3ksNWF&);j0|uw{(iNaac*qX70;8m1Z~EBMN%cB
zkGt{~9oS}!w^cvSio=#9GI?>ReZz6sI`jU2O6awVRcWMUQjxgZw0+g%{{XLCtgPv}
zEUydlupNH2qSTFsI-f9>=L^7YbN+k&G}*`un?W(2OB0?ASccpS7j8F_K>50@I@$-G
z)#GD=ry0gL?M-M8YZ82x&SG`_*Z_uJyz$bkvTXpA+Ekp9a!C}#gk4#yV3v&SC)>C6
z;<j`R5bL_lou1O7M$56}Dva(wkZyjJS2Js&L7!w^ca&j^12Q%dM{Hz}Dy^@HG-)mt
z-St~Ju9y-chF#}q<(QGc99PHRAATijTKB|VUKYBUG`MDHZEeFJEh1hNW2ZrqYl`r`
zv#DO{@afhQ$!3#~lE{bVVV>vGp~*L6rBP{fAFp4srOl?e*Eg1~lECf?kb41vUs5&e
z1;#Q403GY$f7mYaOBnTwBw%fUUnpZ3JsY2VSK8?}mu9TTF2e<@hNAb5nr_h(QoX|G
zIPKQCYwbX(ILY^`+P%>P;EZCqE9>wNTHX*j>Q3XXZw^~8g5RA{jl;MEfscBC%jdB<
z2tCbaYkJ+jjcn3SD2+hmk}Iw-IV8=SjX_lTas1ESsV&gUAl$>A)#2VD{h_oE4K%Mk
zz*yr_?8}02?Od*%`*?gaYx{d|?U|XRaG6_X!sO#5^Z<KPlxz{awMhEf&q`a2sRQN*
zZyoD9#Fj?eO-CacB}w<Lr^0Nyo||oV5#FZ-Tz%Hg2qWqZL-88G+v)pJcxl1>2(Dgh
zO2=0xyUcv;;>Tz%zFxhtT+FjGtL`N6-naa3a-&j|Daa4CPwD+DnEB;aB#;2fHQ0_8
ziZk4KcFsv6xK9vXy~K|matO)Iby}s)+g2IDCm8mwBJgkA-3uSmv_Wf}#pHN#jQZla
z-`Wb)%kb_t#u`1H4{QqRFHjaa<O;<7qET(U5hMr{;6^@SoZ&`jdX*PukAo)|mH5ZZ
zX@GN5-3(6S1MeP{5C=8qJ?wjHV_!hMn#)wQn%(@lW06X?q0d_VQ20mYP2epqNp4N$
z!^Rp}<bBTV@9FG4EAunqXN2bQkBoHPIyfYqCY4o^KBQ-+dwpy5W5YfiweY5aXQo-f
z9lTH=U6Fm<<D7RtQ$&t=+U4*##wi4O?NR}gPXJUIvb2$sDgzD%YWo4wtH`}awKO7=
zN0!YyX9A>UAYz;s7|&XvF=HD)8f=ZmGg1-))pz^U!sKn=2C{F2rAbp-lF5vU%AK;r
zdeXJQOUTQL8K$_}PEBt^14S4IrAKLI5Qgq+L1Q?{s!G7*F?q)n;~q^;kc@j%%wP<T
zw8<#EM35=V2N~N}77)jglT)fHucb}KP}p&dQ<6MYvZFL888o?yA_=u=4?I)Yoa41L
zj8wK!$b4LH__lr}?~55fm3$=3Nj(>wpK)Jxd|hmOGx0W0A7PU}m3%UIJ-xeu{{Yvb
z2d#woS$@gbNT|exz<hTn9B1k2T}Og-h0}GGmO!ByF_YAOHOEA7tV0#YL!6UZHt(~c
z<nzsHdeLXlmYyur(#%aJp%iGNB@1$Y3iBTr_`Y8g-y~>Ec2E-0CUME?#<?W<&PnUv
zwODLJAWUSO9C~_Gx1pnVCXzKE0J+b0>S`y5t^wPhL&hr9QU33L7(=-3??ON%knZ`3
z>RY$9WXiD}wSvg5@~fEFzIN`%UV9H#?e(kiX_8!8dG5Yc2OdmeaL&HC{uCX*z3yvY
zOVSI*vRyBoZ~!q|AGm-X44&Qg_v&k+v(bFr(n#+l(>7$291?vyepP9tQcb%fgG)lW
z80*Pi^sQ{d06_)F;~l*#&|koQc~^~gu{Z=|b6qW;g7nV?*+-~qMBZw$F6&)88Q4F+
zAoM+l<63iAP7cM-fZimN!gr8q_tA+hXFHlZk&ze6)C0iJy?vkXf?FLgO}moH+T6<^
z^W&B^`O!(i13fT$4wd=MtZJ9btTpApSsHgPkuE|m52tVs;hOsA_Ga<@&cA5ODz4)Z
zP^aW$+i5+>#!Yp}ds=!QIf<w2s$CyU>fSEWGz(q&=CfyEZln;qf^_8WWA~fsUom`G
z{h&M*uWRq7+q`!QcL8?;eCwV@<UFv)anIpjJO2Q|Kz=Ujx_5|lj||5xl(2thyw{Ke
zzEqK)DnaWT1t5Yt<mSF1I@Y9@izHX_Nh~SEo@O$?Q(01sDf`To4rO+Jh*@}^-%ZvA
zo^BWH*LZtNq<o-n;e7fYhuf_~<L?8uwWM9&Y3i`u%l@`7aHYpzp+2?pw}3t;PvR{U
zXg3a#-#}V;2;<Ce;tQVSbzbL@Us3!H@uPUZ!V$rJ<~`&~<jPrbzix0b>x>HQx7&%7
z<I1jhJ)i9XfAFS8)Afxl^$1jySVQc1&#^tL!agc|SHAe2to^E6c<e0`i0xq{kN23K
zr{2E&@jt;0569bNywae)NdtZSf`_2P4ng<zuPyk4{{RI2p7%|b@4^wATkr{n&f-#1
zW&@b_AG*i3cqfX=vU1ieYN}Sb@tv24ZZ8=XvL<o4V(51DE>A^K>T53MF&)H+vI!AR
z1CEXFT~Cg5&l31L`s2ge){S%F`^eZO#kA0}Nh=QQTb3VAD~|cIgOOZqt(L~=n%X9@
zKAU+H#c=VexGX;KARg<`{SAJKe#ZX*5H&q+SmU?4lI{#)o<>#3_6yVLU!1x`3ofB<
zvO1!d0GAmbjee4T&;A-u0r+;pZ93rv%&M+Q3YjEy$gf6)-8m<!IV)P}pM2Si+gl`>
z*etmm#HSBgy1ub>Dx&?3$m$W|3IlrX=xdtrkHjXrup6Px@&Fhk;5WDjir!sbdD=-0
zwZu}KH=Py{ewaB2@~;+j+*`Udi_qn@Z9ekiXOi+%Rb#tgeZ>AfE1A_Gw6eCiytw<S
zHpHZG&Da55kBGIsFHN-{X_YSS8FE%80o?xpwT=aN$A~;jE&RphU6DtXMF$9g^Z<Wa
zDMC&~+~s^yMa8Y*askI6clNI$)}l5yLR0eZkEf~huSoG0q6tVz&+#)4<z6r2?-e4K
zh2#tTOMza*Rggh8&p!Bv;y*5F03&QFPCvX)O7PuA{jM&UE?ASu9Z4p){B5p@rI8j@
zf0(oFIUd59;CSHCqm#tiiV3`yBF+XM>u1LdM{vK+vy<j_N2K2>vDaz}zYV+=nkry;
zms(Yt{fFMRxBdvT*q)}jXyP{TBx9C-TBAuWwVUbnOXo7&TsdZDJxxoU^Ro}T&{v&`
zc8!r#msV(5V<m|dvXbTz67Ub+Ca@uV90C{I`c=I;<lRWpDBEr_sK?X0SBk0MQ_z9d
z_+LbiRJduTBw9pAZ<){cNAu66c6yECMF_CBK?#$W+nvRD$tUt2!;baL=mD%W-?q!R
zL~^mII4ZusR^ZiJ?N(dID-dS7xNU&}$IuVveQMlX-MSv#2O4zlb-o;$eP>mQ+T4VU
zA1)MCz-G^XVb|;Que-hrd^m^0S{0+rgij=5SCP8x_4={>E9Y<7r{Rounv^$JfeaSo
zbI1n_2UcI?KN|a&!a9HUozxH%3mal-030l1g;Dx(UKT$RqMTOCbC(f45@&JY9W2>t
zt0RrGT%E*v4#V=Urrsi;_o(fo-r$A#K-xX3`os}Jiu~53C5Wo(&F*;9pFtKefIoy(
zQxFv7Rf!JK(}7r8jnhu%6=I-|n2$<zD}7metcr1RV<fvrEKq=p-}+a`AGbg44KIXr
zw(u5-bYRwXD}OUnxrw&5t)pZJl=gW$p39TmSJ7V+{v&B$3A_!f_@7oGY&7VykSXSQ
zPZ7qxrJEJ`O{l)H@n6PThP$lDZf3o^H+PbC4Cfd{^$IY6dJ6jf6~ap&RpYCC5!=oG
z0Kh8~J1$6_7r{@4Q+Uf(ib*0VZgB4EFcCWdJxzW6;m-xdp=frN*3pQik}#1(qXH!V
z02AETpL_%GJ->&vcwTixRbM)1?4z*!E7`Q2LfcZmR$M%;H!NGQRXQS{_qvnmUul?P
zoSugz8h3om>ij99+4z4=k>+%SM;V^#X5cCTlb=ooeuBOU__5=6{h_`nUig5SPJ^QZ
z4uK3wyKH;8os{)y53fbVdw=b(@uEKo_*&QDzL$1?w)L<i)mq)K@Y@rDx&9sJ0DrO%
zt#~hmd_RAq-C60YZE+3N#7G2jE*r}|3l7~o_7xZ$Jim=C*`s%NXQ%u-_-8MR<h6?O
z7TQUFJ_$DNC4Y%qi~>&`>(2iGYb!C~9WKkknlM&EnYy@8c7RWIJeEDd{4gu%-78zN
z@Gg|L_G%`&Xhe~~;Duw;a2-M`U}HR*`2+TW@hfYdCyP$DxSHlcA$dS=`DlOExgMnO
z2c|LJy}Bv7m*{iq&kGi|z8JIfZZ1qt*ck_tpQU=emXoa5cvc;4<+s*!CXm~WLe2LW
zoB(#Z^&E4bYVw_a+SgyzWfwA~rir8v^vi`inn@4*vWvFn=rhn{`d5AMZ^t$k4|$<q
zS-bde#A)_>yGv&!t-x$WBxk!PT=yBS+R;*t#{|<hq5a_A4pj0EdsN$o-nGx)>Dq_G
z4<BE6Csm*AcIu*iGWQq~#KUrd?Tmdn6`?iO7lY0#&Ye3mt}P@>HOF9hs#D&TQn@)a
z$!7B7$?9sD%#V{<sFiz@!;*frd461W#a4ndo-0<`B(=BuWxygK{;EFmAMmW+x)zsK
zRl2%1QEzYz<b11+gVwnnQb%NFY@D`9Cbu<9hnjYg{`ntw(0W!5u0(c$Pv2$)*G?Wb
zsS|@7zuBsK8S3MYQ&OvZ!l<M3R$>h!1zz>^N9U!u&h{<Yc%_yt<Z)Iaghqiu&(qeM
zs<6|n%&te8N6Ir<%Ro<|$asQLCE$&I>|-B|05L(%zCotKX1P!Y8>x~hcPvgxIVQIJ
zKV!*#x!ONdT>I*FY^-TDutfvXs$Ir*AnBJCv2utxBmf9g^{RJ^61-n9DkT7rdV5w(
z(g;M44|VLn(ywT?;Igg<-CBg~BJ`;l1Ci-k`W?0r-1Z6w9@HVt*&4R+Z;)VPam9AN
z9Mgg`V0)Uu((P2v3FtZeYq;?BxCmH}yl6P4%hp!nDI;IvB(+zD_s+PS^xW910k$>N
z{gn!e%F%}98TBVM)dPm1zZ5Pux`W4FYi8OnIxHNM*dFy)Nr29zk+kC^`&QyxDn8ZZ
z{{SOeqmi+ewnbls+{g5-B(PnXanDWOm0}n-AUMJ3N3CCu0{A3*_32b%*$suN#y1h4
z(-olvKt4t~a52EA#c7n_bMM-u5$$cDV;w>5^`s)@s8BflYg*FrEaG{j!p9nxV#k07
zB>gKHFnW+S@^iqiF8!>2Ag{yEg%Ier^0@I2in(M)!`s*?@9E|@=~+ri7e<t;PSHK5
zPxz&vTpu#uR<e-B1Ct%XoDWma1Fd@>fj%Q>KNGwr_u2-pX{y5mG%Y2&?p2Y0g!jn*
z0QFbk&ZFWv7g2bULNZSGLxI<)9YuSu?790!YF`8VZnrR8G`daTRlc-8GRUBeF!vpe
zb4t_L`ieLy&`wvq&(Uv*ntH`MOrIi?l6&)BYc2Y+I^=Fak;ZfFUX9`p7im8b{38#D
zG+VKEWh&*0kalEpqz>TquNJ%1RcFTUP<sRZ6qV9FtSvV^8a^ETwKUCQ)8gA--^4$%
z?=6s6IzB_9jjgz6<!#x>`d81)12|sZ=LWw^{wRDl(mpI`AK7}Dk{RM=drR1q7lnwy
zBp<wox}T`82Y>h{kBc=Yk@W8p=n}v|WoL?LoPq7MkJh+bvv)nY;c6%DtE1%oJ59XS
zwHP$*KGsY7+lSc{Feztt>;N5W=`Vv{v(4v(b-4Av6vZ0qHs^JwoTN$RziA{RjE=x}
z>(afu;h(@=ANF*yn&?Qnzlr9E$2O%RZAcJ~Ru$|Ic07^Nv|2$CSKc1xUU>ICy{mdK
za+Tz6@Jf{@Qm%)i{4Uc(&n?WXStB@Ho~(QO_5ADUuLRmmuNsGCRRA!_>s}S`(@p;X
zN3yn5%QJ3j`nSV6BI&j<DNrOhB-eF$bUq%Wy`<cdJs(ZAM6-`-fOd{+^TYlMhx;yC
z>i!;|;=hNj)-6`VB5M#vpn34;3orn3>yQts{VViZ$N_4DrxoU(88rC(Mc~g9=#s`I
z7y5khM2t8xY$ZqHKo#Yr&c`)Rm(2cp<<viSw2_19MxU)*>DryO^_0Fv%nl@t!NF21
zJ5q*QE7@(QMvhq6qejC82t1F#@%U0(LnN<=)UiE)&2`3n>c);1cRsc7SM6!7TUuIe
zSuL*98`W5t*!z2bvU886eaZVdd}o8>3<poXw^Y1D1!h#jOYBExACa%huLNnc-(1Ec
zP0|7e(mM1N`fd9Pd^eXzx1UzHf#MFKQ+UYh{(9GJ<fBUG$k)Ns_Un;a@8o@TrFd#9
zXuw1n$5C65=^;SE;}v?wHC0^xwWNxmab5)K$+POHV5!DC6m0B_EuNoBwJVUM3Xzxr
zR5c5c6b*xej<u3}%<XD4qhxX(D8B$mLBSmbaDrKGuA2yCJutmb(zX0qd{^ZM3|pQv
z#bHB!ZRKYi75*SP*QZVzXnd9~t>iNGfC3`LTW?7?8UCKt&T4v`%@lUh67fcJv^EO*
z{{WwA-PQF`b-3Ux+^S>0{{ULK`R$DM<Q5nkl(6n`)A6pDuIG^{HuO0SRz$O$h6<|x
z06u+3wQ`;#)U9t6v@w9e_NtyoPk#OKYj?!3%(n{wkX$%k6b_^Nqx|-+3s{Mj2zc9f
zBxj%UE21Yfl5F$O7Z{|Klgk27z-3NJ_pS$2g<P`q!O0_xeSgNi3stxCXOQg*Q>Z;p
zwQ*YAlQfuKPI6CkT3G6)?V-ZzP=t9*Y*=%&xAp%38pLZ!8fN5-o=$t$TXkyfD;5Ee
zboI?slFcFyCdC6Czt0t&u6j|CTS$S%=5_f%+A`fw`R&rRr*RrLAmC(Tpt6mBv(W8w
zTRi8cf0a!(RhBGpdC2cu!$fjb_oZ@$%86kslgBku+Stnx<B`w`*AU~6FC>g(y;!!>
zBnn9bZ>}nuxRm)(&dp}a7!0{Sp0%BIW<Vq;^uYD5-b)b6z{x$0VBcHpjFXX$oPM<u
ziP28A&NgU?EIM&iBLX>390SE}-a;dbz6s>|p7o0%3eI{4ILEzl$!dELcZnR09`Bft
z-TB281ICvQ%*PzybJ~i7OC#-X5X6#9lPMxR0x&k`k4nUz$&x&v$>W~*uAWF0l33&m
zw}QFp^{N*z?*{`LdS<v0>e#Jeaps+}F!=xgk%BQ=4G8kXDu4*f9P|hO0IylfgLHer
z!km`xR?{+U$t((!xryWP{{THIg<)ktB0{Wzy>p)Yf1i3?tW}HTh5(EXywf5>vw@sv
z02;XjcnVlEF~9>jphcTFvZOBWn2a3OrMA;P(8qsFu;(54s``3{jT`_yx)MEV?wf15
znnoc`NZtHE_NGfO>GPGiA;Ar}<gwslx~mD6FsdZn*cb{k&#iN@Tg&!<AJZU{k@(iV
zn-$pD0ATJp#WGPmmkArYy9AIDa>uDXfc~`{_Km43AwF8;c*xG<jP$3-i8adYRgx84
zjolBa{V`U4$Yz=|mMZKz8~`!Wf|0}OF-zwZWC22*ss;mlpY#01a(C9{?g;Wq6$MEL
zI6eOW&#ijKxuyBhx~>fDe|OIzJ!_7*(Z#H+&Z=Cm-6Uk5l+&=9IM^F{M;_9lO9sg4
z@AT+vLr=7I<gR33ah^~4#Yu0bs@+WZ^436d#eTo%{{YonT6UQC5<?B8y~X7G4=Ob-
z<eq?FXWp&58QSJ_aW%A(M%xsOE1vj2%l`naUYGE9!B;wiYWnGohM%WGQKWI5?&BLr
z&=P+d-1v3-GWe&(T8O#UpnV5TW(^}u*`7~ey}qWtrjNsZ9`Lt?qVVKt6}6PgVMv=9
zm5Ah%+?vW$>}d*3>d(b5+9TmyI!B5%ohn6S)U4dk1epH-kC-3kQa*s!l>94LY&=k8
zxsg%%MG}q)$N1OY-?yj3zwG(6-D1HH+bls?n;gp@U8MeF@~?q>M55mM9a_kkqm=F;
zc-Z~Hr{;7;-B|qx{f6~b{@2uCJ9mTkLFf)YsOw*R>@Q{p-q<Jl*XAeeL*lU?k6l*^
zw|40=jCDEweQWgM*5p2!$qn-Lu6l~nOmx+gM=Puys6Zs1xb0jnuYR#PQgU<Do|WF~
z(SvU8dH_$$Un=}(_~EAb9>((T?eG<ZU`32EJqXTv`&V@(ba~ZYwnt~;{{V>EcZBRF
zp5pcx$^zdx$^3^k@bAZu+Y;wfX>9blBl1+>vPw?xz<_JU{xW{kde6jbi;G0t1+jmW
z#K)2U0P3#{yJ;PZNgymnN6nnplB%PkHm8SjcahO}v*J#*;+b4edntgQxISIq?_;f0
z@HCd^#X4xZ1j5s*y6(Wi<F9JKwTMqTL`XpF7khMGde?QRd^gnmB^9Q$&uysbqC{gY
zvz#wE8-W0f`qv*bZtU-IN-ik=f&Ky5tbYvjEU|Je<58dXR0>awy4)H?gkgwh=1<F~
z>t0*?6aLVi9r%;*GS2wiEuHq5rW?C?8v}4Zc?WQMRZkyyf&6){TuCI)Yo{w=S3kU9
zI|15=)0C<<anzcpNh8b5GkAwm+JKb|bC2(4rtuws(riA?gUp@II2@Dd_*YvE_L<=A
zK3m((u}R892sy{MxjoH!z4wWv*RIXPuD0<xz)*6|JplCjS9O`^t233>@`K9AcLV0f
zYUOnXDu5mz7y`P@b-v5yOCS-a@I~DI6~@fXGnQ<Udr@|c6^zjrn|B0#LpTTXt}phJ
zye*;K;E+^<!N-4E^cxFPCaD{Lb0-J#{<X^dK=6(C#5r{hD^7?{b8vQ|wlF#BeuA#{
zS7TOkN5Pj7w6`u47h>2L<eKh2AAA_`505oCZLPHV)+ZzPg>VPwUu*m;{fzWK3`ZTt
zw!-%J87jR8wR`@D;Qc4V@S<4R$8WS_D7Yu_uQkp$?s`t6O33pcfd2rq&WrGiPD%dA
zvukd@yt|Zw$`A0U94P&3*S2o$fGTy2DuAj0_86d{sB$tYlmd_-s}e3TQISQ*Vi#OB
zSCG|-bDk<PHw^ZwM3ACqnq;Tt!0T2vlobrCnzSZEyNZrZwK!fq>Ir!~RT4oc&pE8-
zAH3w&gr{ng?Nnk@EQi0=nnapVfK+#@3cRnSK@#l16=gxdsbDycK&PtnigF;%dYV>J
zPDXQ3*=Yb!4k^sJ6yiCjV*-<95U~DK&c>!;nu)lnPQ$SlaA~>fX-Vdq%mUGUN;uz(
zr#}q*P5%I1GJh)gaa43U=Dy<iyhr>iFB5;)jGxNB81ptRGnU4A>?y?dFiP-c+a)i)
zP^#$<&O>kqp=0&vJDQRy3o&v_u|AyUsC}kHjN7YW3vXl{3w9^z=~ndkP0hkZg5;bC
z)2nfxzz(GJ0<}k`yBd){ma8PgCmD?g^CQx(+arJv&eBTR2nJgu^{mr~r*&uhKvVNe
z8Ne;*r?0kpk@#0brO3Ks30XnEnRcebk0fKD^kP5nmG`JMNFs4>AQG{R08E+2G2cGk
zwdx-W{{Uy-h<_B48;=cmPU=l9+bOGR3@t6T`l~<8<Neh&OZIs9Qr{UqBc+wprq{yy
zI9vT@=y0nNoUB0lxfsW$YwXV%e$p^mT)-`zth_%9EH>8AuGVz}0i5<u1_0+3!BVY9
zQ`u3C-{g7ttXy#GQ;YDY>;b9HV><W53ma)&aBVb;LO?yhmPX|M73&ZC8hk(S{)U#<
zdUgBAO8)@F`Yl1t;9=8{HY%U)g~e_BCHSSL`0CCXq5=uvVv<15zXTqm)E>w1HSRXj
z#j4pzj=SR^D-T-v{AGMq3G&o@zJJ$43K-9BuYC^!U)a;aca}{pjGDfKF#+M49YLjX
zar|i@E7!2ac&Eh=*_+}ggQ32>vD59mb7!-En`FY?^B@vSLdOUG07>)}_D-3rNvG@A
z8YZE`qWQ49HWFBmMjuXucKmC3=Sco&(8{7g!5PQX(zIC0*m$pN9XEIVENdJ?a%vMl
z8Fk36WxEVy`%KQ8nSzn!^NoW)bPhq!IqhAqz+V`@guHKW1dh)I+1o7LHg;d9^~HNv
z?aT0XKM45F)5cm{CJzqyhDJ-rR>XGK$EHRwt0?c&c=xU!;IHh%t$b63HF*TuKZjtH
zOxnf5?pcDKL<%`}^l!VLTKh^h;VNpRUX2*e)gauD0sXw}r_{bP*~@7XTs`HQn{Ts+
zdF4>sM&-tM#{qIN$gUgV@9h5o@pDYRgIBWnYpq{e%M-`C>qmEtEQnWj8yO%HN4<Wx
zUuk|0_<1?;4BB9})J!m}Qdw!j<5|1|<pGIOV^e|hpcu)?6=?i@@Ya{7i>-S^*8C+5
z?xHJ?5ya?|Hh5Jb$=nA$1#-_2l&obHK1-!^em`m7v*(K~ZMFGtG^?)<>RunVT~_Yv
zS5~!1Vd#<Ij%8*5uF`YTzJ>S@O}2p!rz!F!ygzCPPDbWjF&@Q%KGpQ!$FGfkANa|v
zc(dU)o+G-0R?>`GUan>HErzB+k~p~;=%;Jv8;&?N@?XJSK>q;33#3CjGTvG=)>EN8
ztH&#G_y(^>2RAuNGC8p+s9wj}e-GZ(%u2`3L(j-MdaZp8p!jP|xwMotYa+{y?nVjp
z{Hx^ehMJp0bpvy^AUl65`m?~Ax0?(}CJxiI4@&c~(3Bn1IcUv2B#hsPzq8+kJ}BvC
z$Hf{grnPSv#1~S<{*iC~<Byqtm3$fT)BXu@C5gGzJSTGZc7PIt1A2M`*pGVsM1}mr
zo|&l<)V5bVAISR=US&LGN>0+VMb4F<xY|DnEPrRn{ut^_Ka6Z{Ci0{OL`X_;-yQ4d
zuLk(fK-BD_OD$^Hx+wnuMY!D)udx0f*1rD$_^a?gL-ECh#B$lix<w$S0Lk?phQBzz
zIR4HW=faI;?)0m9ZmeWt$vTb74tZ7|-uC)e(&4cb>cuE%WeG{kmagZuXudPO(wh-p
z2-~P=oHl)`_3y<!Q&K9oTBX22)mM@FSI3%X#r;dew<!hJn{HI$k)hxU>@=^7qfm1w
zmLb9WppJfrt#nG9X6|+<w0qUph~t?VlIL`XjzI+eb(wXm%F(Z$<<lGwO7pSeqjefc
zyEo(?-G3L<cKj+mSH=$<AG_g{VbeW3<kWH{u}iR?Ht|F0yQD`yC^-b+<FNXg`JczW
zH8+z;$iy)vMo&Ll=>8^n_3oy)h-Bbr2dN&Q_Z8x=G}@iRVapadcmC-A02+xmvO6io
zL`y4KHCvKcpCNwvKgd^2dAj~vXo*y|fkbh{g9@YD`qp;jvVmbigV9HSsjIpXd#PEE
z`nB)xQ0)bDZy4R`Yuts31cd+~E)EAZQ*Iam&%ImQ=F71;9m((SS{K?3x?Pe_G;Su9
zccgoIFJgVqrFhk!GqKgk+R(tAn&Lre8Np$h7<O9h{3D?KvO@O>w#F3<;P5v80H4g#
z_&dR@@<T1`TjrgEBpA-`O#5fGbUr@6u<)#7Otp|rEDO5sMf<HW=!4V)&2u>~f+t-l
zZ6;9GHIX#$8Zk>-bloB!W5Ro$=j-iJ_#(;tLE>0h0c@~bxA3lc{VRvGd9QAk1o^|Z
z_sPl6a(<Qc7r}oG?bIU?sChT8{yLvb(yaZBCeEnT_L9?NeTndwM2^EpwY82tp~&3b
z_9*H2o-64ehSo*kmguj`Gh`F}!N3FByl23gX}E?9cm^e$t-5<C{x8cWzKzkXqOrAt
z)(j%5h5rD6f!4n{%qm@bKCI`fE4I#t)D57HovOiXWc4*t2?orMyV{h_0p#&s9Gi>f
zJjWK;w{rwWIL&g?-G1@Ub~wkSUe+}+6l?+9-G`|)=RXmAeGkLm19*ePw;#MZWt+~C
zj(0vjLHM>S=CKNR%v3P&d$ZG_Nx}|n_^bX32l0+i5%}uQz<OQBnWtz#OM0l;5?(mr
zPy4I-n&dti=+Ic{)_!OrD=;2I{{THDJ%_Jdquke$M7Ns%0Ep+*^#(UumF2_TTr-^S
zozCO<4C1}3!@9M!S24vASuC^|zSBIJ`DVH&<zH|ZllWKkW?h4Z2MG*4J(B!-v&@?2
zmZ!LQgU0Ql=(hSrs|C2Y!po20I0e4}z^=#P=Zs<SpM`ag7i!WOVX%d!D#sh!7(^eZ
z;a@dqx{Uh7cY4ImPUbN{3CSez$LrLesHy(|Y0rpf!@mivJU4G3Q!y788xPHs_j-B^
z{&f|w>ZFtlR*JGbdduR@tKzSV8pn<GJ-VI5cOvDcxJ77D$;2wa;4mC2u6ozjo(l1X
zoeqI<t?G>=y4~v;uP^6A=V6;G0&d3ANx?mbO8CR!Hk|e<FJ)mR)RG^vS=|u!;zFdO
zvgZ$upmYRe*Qr@}*)=Z+>bkAf(2qvb&&x7=&vz>}HYyHs?00(rYtwL1N-JZMTC>(X
zZ}B;<ZX}VQiW}Xf=9=BkWMFWSDE{Ju_?(*YUlD0{Q)#klAzWSGv5L@v^CKP^{++Yi
zn)B}pYnQ$<@se5E-6Fyz0^)MV9P(E_!-4N!*YO`-(Y!p(ror;b98tot#@{3edvbHR
zlx}ZANbg(5DOn<0wut$w#P%RhZKPWl`y3f&W5=BN`Q`T{jC}=mp9yq{bw3^H5ZjH&
zvm{{$Ks(lko#46rPk);dk{KIhZz%b(p1AGw`d6}eAK|QD0^d>9yg<nJ`dzzCG-yFY
zW6Gv`w<AAXcBNKnCmyFoGtR$lKNQ*8{6_H2wV_|_JvJ3Hls|Pi*r(qd)eB36yMjB6
z{EUJAq}P(`w^u$p@vg0|>M!QaJk2YDda=$I@abKDh4qiLY0GBJ46%HrLF>;Il}dJ}
z7_NqGyR)N~Q#^$8^r@}m$<0@_R%R-|h8!I6in>2{T&>$fqBd&mu@tEktGpJCPFY7&
zP1Q9Cu4Xq=e2Tv;5RQsHY6p;Q&1cOpEJ+y)Sw$K)D@)`do=#l<06OPy?poqw+<H~J
z>;C{bm}djhs%h4be|fQR4hi|Ix{9>ZJlu2t0B5P_M+Li*H~=47)3mqCDj+WBu&oGm
zC}JD#u2mFv=}_MIvr5sB#b!RyF#hLH#MiNV9|31AjU={%OM$K~nL{YgJr=l4ds;fZ
z;bh9dydXc$t7_Wi&bKz}hyB<;EJN;!s~9;X@mR^++ME*A?n`YCpKyuSapt=l8&$O)
zTJ<M6u2W9p{uA%kyS+XWadH3|sj*Vrku>2VF6@F%?24?~1aU4^-;9xyp0(A-U<$kR
z3&5(kR`^mqsu~<-&PF>k4Yk6S+|J=kV;QZzEHW{Teqi$j!8qUwt0W_H3?60`tEE6q
z^Z;}$&H7SExjUn((_;C69ZAJ@)=-;x<59-Zn&!MaWxnb#!y!2}*FuJA$Q{lzNp5mC
z)TgO`qiE8T{AU2w4IOZxXPj1_ix5R=YXIt7HAc=e8?FJ*wRFK~aa3nrqVNuR_N}{R
z+(#WN8pV9IQQH-}Vid3kfO=-NIUc69l4X#d++fzsR{N9;V06wu&*fCDqRM~<39V~<
zfC0@WLnVsXzsd<!1Fw39NH#aO^Q&<NDxi;`?N>DIAnRJCy^or(M)`RsZWcJyZg)O_
z0Dnr&%G!~v+KhK*9p0y<cyC0%*Y)Xqod-<;?lIPRKZ4yoUCssvZfo<;<LAXKL*g%r
zwI2}bvh2UqpJluVepn)208!aheo^XjYu!FE{6W9?#i?sroD=^5W&Aeq*7eilX4?kA
zmGj-dj^bmGiZVi+@tXLT#J(V(T-GMlW^X0{cEJoo?x^U1dU~I&b2-^_J=yh`3>PGn
zckcZA6=b&Oc~kgxsO_}|uyv9)<Q@iUw0e!p67r;BTkfdqP|JG@$jZ@>M>~f~!D*^K
zu~C!NBj_)IzqCE~!~6Lc&3la+>-Tq>acppZjdRh5W8a~#pcBUSn!6?CjnduN%y%S+
zGuPC1Bhc69u?3%fw7^N&ZsxW$FNwY*@LkHitj0Ctat=|)xI9&ROKx=0aisnDpQ8G1
zt>>v=*^GRwLFwpm>08r$P>V;gC?`aC$IF9^4_~fDd~e||+LjBeGadb@Nt|tC3CCXf
zJu&H3b+3u=>PaQU8ic4YR3HJ4?!5EU-lfVLp6wcx)Na}9e-?f$t<h=PPu_j-palTu
zjCQPVf*04mB#re8w_=lAG>8ji1RMZ826M%DPOW!&uISgdEDVH6x6NSL^v9tcI`S*&
z?}R=d8WxFV8>vmIfrdB)57VixtRm-|li40Vb&I_vO=yp={sd^E_fn2`8IB>igI`HT
zxR3$&c)+h2_)0A-^avt?WHZb+IsTRF_Ak05`S%qZoOP+Mer->IrzF|i+PC^T4{GE*
zb))^h&Sc32##NW9x2<({M4M#ioK;P7+|Q@VI0Z&>Yl<$+?x`Mlq|eO{+86c;)BYvv
z^Xqy9vFVzevJ-U*fD%%Fh;{_^{OiS|{eV1GqcC=b_5w$nA0+#Xp2zg}uhG3@T}0Jp
zh-_&ABo)B_06*ugc77YwVxBcg6;$vy6P%ByZ>B4D`8BZICemET#2*DeVvB;%$*Mv{
z${d*jkC<oq`t~*VFM|9ZZ{Tev-p1Vm1p_!3t;N(WVIyNKWRftT0#D)IvSiod^JKs;
z=%csRv7J{_L!a0vK6km*L3_1xyC8P0OAEm;;MWxfptlhf54hu#{cCGNm&$-;iR;aB
z$|~1qW*)PY`BOITuEbg7S2e5aF|uJ_c6yQ8sQ8y%5u5{!#D)5L*CTbSw4-kCl#Y8=
zl<9RxpN+@H4aG7K5L~q9<jz6mx@NL(E#_T6?Tm;m$H+h+;fMP@vryk^A2aPc$o)Q*
z4xcp9#E?wvwl{JG^BD#M4&e32Vr$c<7i%9WR}h@F>QS2YTQ3$+@y;>;{{TH{R{dkO
zxK~!oaHnriuhy6qM2s*14snj3TvLUjtcNS|e=6y+v4Wevrd7?fGb6i4`{rGW#AglP
zIra6gKh|^?^IZ(V*@0Z=k<EJD)uC8fkwK3DVNMA5_deg9b9$|hotWZZnB?u-$^QU8
zl#?!69yO_WL99dE7FQgo@0#;nZ%U9kIVD-Mynj#ducI|z4^8%KnA~^VNjY9Se~GUv
z*7Wd8e5Dmh$;UtE-ntUH5~*f+74DZKOO3xF9W%inUwWrLg41Id$EoS~{VUY%yf&9I
zhrMG3pygdx=5gO@_x@E&RnaYsq}!y3fc&i3-R+!zPAfGBbEV=ERy@l0M{8K4i)qU8
zau4N0TA7#<3Un3PU0DW$Y=}LZwP#DF`D{8c?ni3qwYD>Ln_8Hn$w^V)*0indk)ql#
z0R3quK^Yv3dsCp0De4D&R*4xoHdmg@YPLTdS2L<Xn46EA-n&T{GIY&z`h*~o4{FXx
z>7wk;F6z+|6!}BAIT_=h^Z8dFsRGQ*6e@y$I`1_|8dI?sB>MjVD&?<qrk3DD6#T7&
z&T`%Bj!NwJD9tU|lWT4SAT%vEqbSY?r4_Wck;b7!M5r8RpXWtIk+-xy#nd7&Oxus$
zW0BgjnoZ2{pOFCi*8c#Bq4F;#9D>eIJ%w{K+#fk440;pkUP5QlF@Pag0e2jC>N%xn
z4J=&lP;<1O^ZL@Z6actZkv9Mb2mb)ANer8v*fO`wFb}7#Q4GCok@PE-2b0`pufr5!
zh5_0K1$oc8s<!3Y;BG<BpxvHxl0KEG_IZ_;0GPuAj^uiudUp_8A-8?Su%(E}!O!7Z
zx+G%aMJ~axdLLZz)~Qcx=EH=IN!|P><_EW0xuGP2SNZVqE>&1#lh&D=)a%~ae3r@#
z?Hl<WJ*n2VYjbo}K@sF<jwx-fLt3iK7>&UB{r;79*Bi3{AxHpw`%^+C)NB&c?HUzt
zDU^JuF^-jbD`9RPRUuric9Hl~Z6PwOL+3Io{KE$e&-oR%HMo0nj;pu;0|n{z`qZ{u
zQOfFo$vfqW<#5^RI&q)N^sZ}C(;Hm42?f3qJ-Hy@dJau>y2ZI=m>}FR0opr%oc^@`
z00HPJuWM;N?ZC9sH2(nJ$jmS^*k|kC(x&zuuFcPfzq4)Mh?X1eX6)(OV8aN|=4nPf
z{^<AQ*Qs6ji$?fkX+5`wGzcVUhF(J=tGV^b$u;8t02zKXOQTs`FNf|Wx1L4v(&gJ|
zjeS)+dw1r(AJ_gPc%MksrMrV$)LK%-o@r$uKhz$lo|SQrHmq`zinLF>bzj@EUkYFO
zTHl5wNP`eePlJs0<YKoxWAU@X{{SBT5Ng+65!ErQS*W>47=@lG6OqSU9+|I%G+&K&
zzB#phAH?YxyYk9MF<dN!XBa(?ze@5rek0%f6ZnFfu9-ZB$=k}fw+H2T$>cUa&*xjH
zK|7iAO{+8Q4QJvUKOVjkOqV5pw=9O))P3cUbYb<zYv3Ol=;qT?ytR|}aLVnK0~=%q
zu=F+UAMmh;gFImmh&0yoG%aOv@hSO!VUzfb=DgqHMa&xBuPww18|^q~_8f8!e+qL*
zbel}l{gnJ}d@1oSOSQO{V)>04<%sG=;rVm_00H-}*zXN$pKZ{hmmr2g)6%~P?5-{}
zjdnds&v23lj%6r5pOGDLU#*|A-^D0=bMW%QW{^DQFp{QuMit8_-CWjBk@}v^;s-CH
z-L~E3P@EHtkEpN5Z`%vROrABowX?gLJh53V=8bsZbYY#{y({)6=5zL00bsu=9Q`Zd
z&)OIERndMeYx7@4JepRgvk36U9b3?C_Vw#s6t3fBa!IJoN=N0)?n9_m5wIMU<24&w
zvo6vIZcYg1zSq&e;Fz8=@dl$JroCZrY^{$nRC3Hca4YH$f?x1S?*sUHNOcbz>UUly
zxLgxvgUGl1am#*n5?s5T%Cug`<!{5!h8`>Um9Igicw0lh)-@ShYR72{KHqXj0DGGK
z6!<~^00f=!F0Y`&Kg16izKYW*&XcCXCNYmkb~w+l2ERxA3Gl1KUk&uMu<);k^zRSZ
zD(%Ir3XFcFa(@$3c#B`M(4liQ(p}6y<(eE1(y_aL2<1B~$oOmH7r=dg_CWYCBwq(?
zqt%KtPpn_wjrWCd!=jIu01ktvTKSj3U$#??GAQoKe``AdadmX~EZxZT759(E--;J9
z=+nosA%XZ20iL5B>+{>=Z^F+9+2}X=)9Hc@QPt-kWq&U{@>Fs9SEC5cM(3B$N^|9A
z&bRwY+UuH<iEqs8Kjo&<Q3s#{si-xt+Lui5hLqCjc5=^l%Y$=tL45lTfLF=d4~&;Z
zg~CAupz?g&tNsL4FA{j>=fp}P5T&dg(Fr;FlTzxWyv~J&_q>m$ejfhQE#r+{>@{mf
zj_PL2GQ@WS?B1VIUe9fECx|&tNXOE?9QZS7B;GNxWy+G+U86ZY{{R~MdqA_2OKpgx
z9PPoWi*i?HRp6m3M8BtevNWtP2O#}9tzU$ert8IX&fqr4+mq8Kvh^ai)~>3!3N~l?
z*I)2oO<@+Hc6x3RgZi3P9iYWcU7nh%Kmw%M(=;IDdS<5D4{Gz)i$k&M$iSwunwxDK
zay=<r3XHj^*>P6dI)&GgJ64Daa*Ak-?G*VqdR63sR6yOC6%?p5SG=RQH3X6Mrk<l{
zqT?Cjpkb3&3iYb8rU^Bn1skq}dsJ3YZ8c&@iyYL}@a+bG06uD&38n=!<l{8j6^n6?
zmV?hqjwZz@e1llB(I`10oWi0F@{d}0ZNjCLhB?JR&&^3AVAW@CX%q~3sM)i{A#Brg
z#XAAV{9Gsg6_1Ir*X%NN`d7n}I8v&hkPbn|EZ)CQO8bxE%I^4U;zjkcllfP}Z@A02
zo=;z3Itp<;oEx+ssba>)O-AY)XjDA1oC4=)&vH*-I+I&77%U=dODR0YI9aZgf36;i
zJ0^M#{ZFN0he#nayL{vji~a80`)9X$yj0mjn}Zv<fGpiic^~=__TfcndN)ayV391G
zZRS<L2szpB?tQBHWwy4O=3|Bk&JR=SaqfM6@kfSqE6sOOv9qwXmiJJ*kV9{20`HPX
zJrzf(=O?9odGLew4c3k2wyp6pdu<y}f!1Zwp_of^afZ&>Ht?&QWCBUWbJeXHQEiRQ
zEoAIBz#41(Gx!-{q%@JisA{;?wf2q2nC1C~)QshJCJE22a}h$mFVrJ@i9DzTfW)DJ
z+s_1w`j^CivfqR}cdRY-&5od1fk|$yi$eDV3=H7mvCwCwK0nzb!&-Hx*>p`GP||KO
zk$Cd?nPJdK81g^eu6zbzLk$mlH=EVJU-CXz9YUree9)Jn$NV}L-XB8Nn=QP&`Ry15
zoe9auudRI(;7=AT(M1Uo6;OVB92oxqgr7rSFKc>Uukf<{;MFY}MTmnT0|RO4AV1x&
zMDW+dGpgy&r^2gj^QV;^%9FfjxBmcMy*ynCIh8JW>*v_{ni!d2s`Av+^*!^)ejRTU
zXuf5O{{Y0^!#5U^Dds_rQ|QBwt_^dZ5%II?wjmgp=PSyShEvlA->>9o_&;^4&85UI
zBZ|VrrYyvc#&%$E7z5D=A9VB9pz&vgWzsxJ{imnF8robleX&6SVQ!_5`z_b&^{)dB
zC|X!L?)KUKhp&oNCZUYOt$xz-@A5jU-4jRgo{e##YgRV8#+P>(S?sP8%!QZ`c?aAN
z%C4fdlq(Iej@(AO0z@{VsNe-1usth=wDHZZpLQVyA{h*Y-;O%-?_J3;U5M^sB1oqh
z1x`Rc>*RAxRx>7{FRMqHH=^x-_$Rr6sY@2^DQbuk*4o^e7T?X7soZw_&&RJ*UIFn}
z!uB5@bsc-bGU+qM+8cSW>gCxCcdWbw45bWexZ{$;BLLU59~)UuV7iu^(caj{If_e`
za}059IA+Tam;=WhPfF!{N8+7JThecBye;Hez^&%Ku_7#S0(zC*fuHt<Sb9|DHSu#^
z-ybVU?cIF8Q`l+S_e*2rzlZ+-vpwg9elF;G#<1|sYo^>R@mspdZSzEbbX8bpIL1p5
z+3GnsuO-wr_<!S1#QieOvukbQyRESSj0pzvN49fcLf?4H!d?dPESLIrvoDGKOLXfU
z%#pi&ksQ9=?KGb~!5BXv<AN*W{{Vr$BfIhU#6KEpdcD`$Zm+HrQ=V2h+{Q;-<F^D1
zb`|<wTZs1cxvzU@dD!Z6rv~JGcks^IONCbXhT)t6*V4X}@C@cD)6L;r4D-k};2sdx
zEv)RJRNDs7Hh_PvdTxQ^i+f1|xlxX+dwpxW8BSFbQAf^3R(C#>v0=4w^9*M`wWSsg
zFgULpPmP)nhisN@LsE~*P%#o=<R{Pp*1mf9-~RvwBGP;ouWe^)@FNnav5_UrxICS`
z5q`Dvbg&qjx4N=hwt7mU>ACiHwW-H%!niDWuQBn~pC^cPol3$O?WDZ9wvbyJwA~WV
zf@3ZRY>XdcUJ3Az_NlY@v!kW;rOcAWxWy_-8Age~$yObXImqDltv`?79wGQq@XuHA
z1W)Gbvc}qWlWGp;p6*rJfDZcu7C(orda=TNS#N(cmDH6<q}Hd$_I@z1_@Uu#R_gD<
za9ZkGZQEGRu*tsKS6t-A=bqg@oY$Lc5b5?my^066*~`Q^F`naXd)I|N6y1C~_+O^!
zGsnA9aBi>39cGi3&!*9ZKT7c5iO?;)ziIOZ@hivfcJutJ*{63McReanQf(D&ykGWI
z5~Sgy3y`mz53i+jTKB{Ytz{j;i)(cGS%c?y(APJvYEs**Eb@hlLAjN<J8O>AHEDI0
z@-4F+q7F|@)iReur6h^<ElOQ#Q*R7}K!-0Q9+V}_R$*d*0ta)|s*q6}VkrBbeNA1m
z)7ojmkG==a-IRZgLS0$3MUE*i?n<e_=uh_-(z<(eTN^-&@)9t<{`FqY)+<x7S1fuA
zWVhD33t?~J^0)eerPiU0+t{8`KSyt;VOdXW9Vw)A{uk0Ld^K#EzO@?7I}P^M2ZG15
zzxyN8y(C(%{3QC!_xgkR(Li^)fq}dDRD1h&6~$d^mpZ(UCB(#*KQza;upYG_*Omcq
zZ3z9^bwk@d+tjsrRcYH^r$i^I=}oU40?J)Wfqy8)vIE3pAA9JdrFkZ!sY!b)Pa`|P
zSdj7q`^NqzHD^@Tr`P2Xv$ETCk(c<f{44A#h3xX`w{Y7+sv*dVD3tuFJ-?Z(Y0_!k
zofDenx;k$G*^Q}iR2e6eYVRL)SCW0PoaevcUs8M&()_t1U@~Lla2#$m^OuC=j@mnm
zw9CZ|Yh!FjMg-vT{?W-J-oCZ?J7jGlha)=%<|Oi1iu1EMu1g!z_aJ>q@WJDb;%lRw
zmeGh~&#pdWUa?`Unb&f+1PtfX(!58*mxE2yypVA1^An!r*L|UDF~YI=>{VnrB}Pc}
z;=eVk8kMmZCE3LE+kIyU5Rgw{(u?g>Tf-wWWG{2jp4hG%QrFQWUn}l{0)l#z>sUTA
z@o7tIduU`J$i879^RuRYtau*4)K!B@-B}T*89SP~_lgX82*Y+x(oT8qYv<qEgU23E
z!%O`zFp=&Ej~VLSK|e!X4y)p46YerFSQNWss3V{i^2hB3;^(<&ZL-9O-dq#a*x(P#
znlenKyliVvqq*u%YBZVUHu{5D*~v6*ncVD*pUZdiHP**+d^zyJhU`muBV0GD1s^Cs
zjd|6ZKBIRW5a23pfd^8|c+GZy5$0JnT?@nTlk?9Fu~2dRK)3R*+vBy5o0IB#&YJ>h
zwpt#V=V^)v-coq^m<A-*g#OWfH~!bwLMQJedA6wNxKqI=(>3cJCGiP@=To{?RU<J+
z+zcpDUncxT)LQ=l#O}E-<)}<#{qFURHScp1C!x~l*E*J|0{y1t-r&miwy>nj95I4_
zd7r2t1Nzk861;4;HWB!1N=02eO>?*Gje$MM02e%URT~HQw&t=MT8mE5?b~a{(L&KJ
zyDk;j^}!sd=Og{v=d>$Sm6}_NmWJK~<~u|;Gst=Yp8furt=6woot@Xiy>+}Mi!4_$
zDQuFGU}pn!k)EfH53OfwUNF1y4yERaU8T5U>`z^p`B&>)?alMaY8x0?nE7vkhE+c?
z$5G8c!`iHRMeNtpz%FiCaRNEO40^Hlu7pZQ**~*=?Y5Pu$E@mbq_OGnHORLkuF;5r
z^=FWS^ix<r9{hK4s=cp=E$$lQQPd<icdR{6O#MkefEDIH40XG>&EBJTdn(w)``JKj
z24Fbj&}RU9*H`d+L_Rk0zOUkaT%WVr2Dp*iq9P1{Kb2$blAk*_Qqc2H5m~Q>o;ygu
zTsGOW$qs+5ZulEUT}Jh8d$SHyamPQ>x$Dhwd{^RKQt9F!>6Z;6CtjzFeFb`_gS6Gr
z^x-6*EX8;S`>WNe>bpk7MPF7;;wzYsiR6YN3#=&5kM9x9Qj%Z|bC!CeY91_}a>V($
z#(lb0^t>|`^{xpdnY=V(427EmsjSUX=gLxpuX9=#GmAJqWoqZ{37nF1xYVRm)s+$1
zoad%$EBpB@qs%fiPIxS*wODW6tz+ua2_nJdZWYr@<42v1{tVZVM(0#AH}51Y$ENls
zpciqm`K2D1QA~<7DY=jqJ<Uw9$AEA?n6FO9#q%iWPYlT91a0~6R}cUK=K$7E%9F;*
zG0xNeHLnI_Ad$2W)|G+xFX@)yu44oo6Ob#o@bAkrvHlw4^yrA4NF?m}jy)@`(;TNx
z{b(Vbotp2pJPewneI6M0tGayZwxf)KG0$p>?F^Ybb*$MdMsog1X?X88Yi~q@{w7wr
zAHc?%nrkeo{eErO#cpW&aXN%gj0BG#jU<g6m5!Ig&=q+f@UFi`vEJL-v2=?qfR2Ek
z_0`#^h72zpckXGWsl`%9G2)1f`kE-|8uY4`NM2RP0G=ya{tr6-Rv$JFDs3Z1m7hC*
z8tIC%IONN49P}MH`quW8GUFp4bq1>0>G@I${nO5CXHT;V!Bud2;)Y^t*`YW+K;pLT
z;RSkg=~V1(4&R%fL5kUe*pvWLA&|!&-=}hGa$Btv!aomuRpJj5UyUnEj?Jy-x*!%v
z+p3`cCvE^9q*ON6z+jLLJ-Dx)f8d~AC2MbnJ{<7EPEl@VZARf&tFoL&pJCRrl5)`-
z;Z#*r<a}@Ox8n`Z#HNEy(yW_9@Mf4;X}T(%!3RSu&HQsWa(Z+fE5xp(3bK~VKHOw>
zuT1cTj8CE8TwN-aUFcblC;-kyd4=>u=Na0^HN`!2N9b5+dnm`ut2Eo?R_%d|Q?4Qo
zf!C<#ibA{L3<g<x)Up2ZJ*U%(xT~uk{WX0V0tK2NK3wCyMKp3El~-Uty4;F%EMuLc
zBiwt`(>61nKDC~V=9H7VmgW?(v~q4!u=d(<Q(J11LilsWat=HD`qa~^$OsjkEM-aJ
zx#qf~m*CYi+&m@XIW=F0I`)$gV{_$+h$Li`A-Ur`4Ei2xvG~>T8^juyf_x3)pA2d@
zHnuU^v=^&r*b&@@&glT_kUR0keD~om5lf(I@XoD>S%SF0B#*|s?K{M}hk!2bpo|NR
z0`<2wyYm?bw<8_<j%#{Te6BaKIJNEL8i`+*bJP9{{@R`|@SVtgO%qbEgpV|zkc|HT
z7JC!<SK7Ki{1l7f)|=tDFK?pP<(ZT;kXl<KY(9mHfPV`7tMOItpRC`6(DbjdG|ls@
zZ<TY{DC{d&;n#!gyhU<g*EK0@AdR*N*8%?kUNK#9ttyMzNnQ6oV!mHNO=1&q?v~Hw
ze$xB{{{Vu4d@=adu3GqOO}W%`t#aU%x3tx*!g=Hj;T>6&wmHUgUi0IRA88&B@Q$Oa
z>enl6WeiQ0Jy?;#f!vc{mfr<zz6gHKx3-=wi&47PJVk9DZ65AWW4Dm5H>7GgesRg~
zUfJOI{9XG~YfxTmw>Nq>gKV>LYiik9CmqQLz7O919)h{uCbef)9mb^7h4~k8{{Rg9
zO|5v#Uet9v=$vVB$V$y1Cty7@j12qdn)RD6AK1m`3u}}fNx@Z881cv9gP&T&_~+vt
z7vYbBY_u&MEv+>0xk#1zAnHjR^*@b#ZSgbqjMKb7rRpNeSHF<}+2x^*G$e8Jp8d9t
zap_7<>g@Uobm+}<P1ySfRQ;f|jUM6cCXvjbfLJ)lC$E2F-o9e^$NO6@sp5CD(=X*s
zBGvH<TFI7><C!F3)SiKP01Eil@8Vvy<2(DCtLwQ|+?=JFxiRj?e_lJ9r}n3m6CoJ`
zjCK{s>-#tEE1A^GsN$n(HlyXC^$wf+PU@Dk`L<fb@lA{cF2Mdg<E?#3`#OHvm;Nig
zJ|6J_wz!>_Y!`6Is_fp34_~OS%Z0s>8+LLy$o~Ku>by0oOBBf^yssRAe}wY9^&jVo
z?Wc(P$GcJp`D|`a94s8Lm0Vw*r|hP!;)`8gQzf*dq(nr^AlT~o$x^3~YoEH+*oMn;
zcpTTup8~!p-T0ft8e-~~rsn1wMzxCKT$Yq}Z1-`Ex%981M&ORjL*oh$t#+zvo{apx
z;_BC_D9cGj+5Z5aq3Bu$iEs8hjaFN@X0*3pMCdpqbsJB;a@yyKj=QLmkt|m5<zk1R
z1K07YHrEXL4eE=TaLt#62@4?XPtvn=OKCM3*;Yu<s+MCb1sME)m2@a+a@}2H1ib!#
z<Zs){_Jl2x%%>!dl=hH!^!z>Qj2B)-p&L2L$?7rNk6MlH*qjUuj^}|>%5|xi;=3V;
zBXK7_-qky_ZyT13I3$6;u0Fl%GftR1=awVWu7A&^T%JK3zTU@yQ^;-YZj60u+-*Bp
z?KuTcr=hPa*6hC7cP7;!soEEB2mb)AUb%P-+xFmr!Bfs_i`DI;R^C7iPDsZem2Djn
zmpSK?TC+-4LIZ$#?nPF*w;wPofIAWT{{Z!?uhedBC5-NYOpBZo(0kV_dpM7G&Bk+_
zboZ?RO6=yfD-#)F4hKwUfmrWo(d4>=lhAfGzk4?3W*iKC<KOVA4pl)Qf(XVz&032!
z=Sb!CD{_uV=g=OwsK2x3f;ePkz>pSEo^hVPO6WYLo^)ACq+7D?Jvy9!p7mP#?XRbM
zhE0H-sEg4T7*!s?RV2gOnGOcS+Z=J$vt_iB;oJ-;10|UD{{Zz@V|if?o8dcorGaF=
z)MYZp?bZuexM%+Wfw>AjmbgzAUs~xt7}W2t!<%yvjHG7_#rO(62|fL*ChnZO9q_uG
zW7bwY`rpSe>o*gNTMMxgF;6}A?I0X(9Q3Q^-5X=B(uGMlKg-<LoXKyi**=jA-lE4a
zi)>L4A8QgwBN^iyn)KKP=+RR<Na^kCT(NFR-1qC#jVUfuDU*A#C>WE}u4t{bfNdG%
z_dPR37Bl8Pv+-<MzLRe!c0nB134S&@sRJjV2e)7AUXS8yPbX7gc}`Ay*9EKEm1HCU
z0O~s9>0Sb7*D|%(#FngBsPdIT&#he3;q%p%R}3?>?Z?)s>I?H)NsfG=M&sW#wW2c1
zaDr4WuhRn<J^r<KAkovT*+AbB%6!UK{Cn2Up0UNb<v`dNM)danT+{9Ll~K4dZST!S
zNAFP-l};Fr*#7`O<kGn8I@;buljX;lK4vZ7@T~}7MW1(;!T$idIu1Jj0H6M~R#^Or
zo5~<$kau%iwk^H)eAy$g??uV61X4{917I;R;d<h=G)S47B$1Dp=cnOX5NT^GMVyxJ
zgU91qw%R({1Le73SGQ64Q|3X)CA<Z2er7!e)4|V5=@Q+Z(UlM|Vg@tY9V(P|qBr@m
zmgf!Ik80LM@-1>s4*+llXDJ#+HaU$>+1Y%kVrP-a0aNBApYm!{@O|~AtHp6JwzdGg
z%1&~740pzV3h3q5Z2TRf&#KH4d*{l+j6aw4HRl&v=Zxg=uCuCJstF}`Z?sMKEs;U_
zx!@DnQ`{|WPayc=@dC%g_f2J{>4@7D2f4iY<Zjsj@IM}F<-ZYllG|UNMZ3atozZ-|
zPq3-JDR}bpQ`Bx6;^Xa3T<&KWJG0)r&f{21myibw+_60L=lNG;ERoLQ<nC#Bi$oey
zB>HvxUc$&p45T!9>IORfeT7B4{u4bOc;zqWJbQP@xFywoQ~7oJ3SSV#FNC6jZ*CB!
z!YqifzhVKw&m*mKx85p-+S(tmC`p#%&IgQ~e-3+&ynC7(Ph!%SQ$yj`i5mAxy70E3
z#LWp~Yq)ehN{{FG(`D2og{HH)W0>4B3}v|VHOzRy+e+80we2*JE&|L7@)eFaK8LSb
z-PGqAy}IgAsaaK5dN)#xGOdqKQ|(Ik(9@Td+|=+<v7b@ZG-;Ucyojx&<PZ{3%JO*`
z`IO*f=N&8D{{Umpi^t*TiXTnBnT3S%2&G2P?wcnE+nzIDFI`+rw=l~K80AI`jt`;8
z{dxSWN*@(tS$#r0M$)hoCI&-elgDi4lr(l?s~%|Re%K!o$8)2Ln4}UTu0xK5W373F
zI+|Z=VO!*2*enlIo-5~X*~j*a{?O5egR0$RsM`#Vz+`$?r$yoghglMdHw@%t*JL7+
z(D7#{vp$&c&Znot;%hw)2^DXQCTCDb%rk>tlX0tCX!@3&X?1xNCh}h<1(|>{NsR4Z
zQ-EvZEjPyvqyE%C4e#Zdmcv=JyVG2d0{NUebL(D>`+c^jqx>V${A;6zYtISmR=TSN
z>LU%c{{WcD_pVy7_pEnOm%4HvwI}Vz<8Rrw<1o^-zX3Jh!?{v5x7A{il3S8g845St
z$R`_4I#;Q9ALIW3!;g%<0X``3-m@IAC(ChdazamSAs=!?>JO)5+PO~@c-!G$?a}a7
z^TTrMc5&R_z>>!tj-u~QxW-Y!6O)24anxqNBs_hic+>V*@%M=QBXKHQUpztvo(JC?
z(3aeir`H1+<a1F8azA+;6Q-)h(cJsr<9Ed^Gg9#c7O-6-z$E#Oc*S^E#;di|yd|t$
zD+wcOX4+Sv7~ogTnm5E9X8O}IeX9EYKawvlGR3pqi1e>=@P>)6Pw<me{{V!AWP*FO
zl2{;^VZ?iQ$m%k4UEb0&R*Ar{?u?YA&yb&~rG-m31Fc-t>>AHexVCHtM=O(8^i2-p
z=IK+)eaC<{6I>j=@$ESzuFq!uoIDsWwHPnVj1-@mGt)o9%kDp|eT5#~KFnp0npXw=
zYtH@!{4BokzK0FN+K;hZD4Aw&ypQ}!NARC&>^1FCWw>Qm%CE1rbj4X%@#;Q#m32F^
z@r6~~{{X@Va&g6bXT$v$`wzmFkTCu1nU+tfuOaY8n02oZL%W!wZR`(!O81#$TUg^v
zjs^vC*Ojz7DXVHc)6wuMnY&VWBNXv)p&6NiB7#}i{Hi@ku6uMwzGCWn#jcHEHSO%e
z3uu=t=lNr~_OF}%GJe$8noHSe-WY%;wLFXFp~Ji?SZrVS*kdEGKDC$oMtpExHrH*l
z29cb{VQtj24h-x)2S5*eRKK$Cz$dY{xAE7DBWd+}wlmsW+;UM*NgN;IJx8^7Dp02z
z%X6Jnl-<_H)H-d%*BVWf*Gc8wJP7ij7z*V<BOh8>4nd+hclGZ`$E|Xm7rH7i+(t!M
zjWS22HBj!xD#9+%Dn^Gm+L)?x4CbxKa=5B;I2;`2ng&;86&#7Xf%(+*n`=}iQ}Vf|
zY{07wqZH7&HAZD{pi)Wl0&3BSWSWNI0gZt4phZ}2<8y#{2cWA6**!7sT&TP@m)LF>
zHPA(Ju0{`0>sa4YTWNCGYyv4#O0F9x6mQzpm=a1xd8*FeHA^e*9V(;96!StI#}?$$
zZ@{RG!x^g2CUZ<@JMmUf-wr%S{{UW6e=7KsB)KZ3n4ASsl56b$0E!Y5@aM!}^Ztb=
z@~@8V6KXab#9M<Reek`)`e(W8PA9K}c8ey1NGOKlbCJT54nh;qDd^49r%s(}rPQD4
zW+>NbP$MN+uwPYE_3C|dT2kENIMETzP@za|+5TAb)C23^KDBGWz7D<dw~4IujXfG$
zm&1V?<n_Sg+?sQeQEl`+6uFv6`v>+F_(!K``eXq6Km@;E6m|W=^UZiPVCM-K18Yl<
z;$|eM9e}U4ZEUXfm=f{@CFEem%jM6wJ-T$SoBlBBnpeS(1pGDd)Ov-!pRx&HjVG8Z
zO3Hw<gzCVOMtykeTi*<SXv<h3x6+<*6ik6f+J0TEKf8|KTKT+gSz9o|#}PtTciYJD
zF<I3-9x>9d8$l&zdsmILO>bB5UDP@zqj6=a+dr5*hTA2`QIiq=>o;OM*O2&Y<7Ttr
zoj5@>k~AuXT`m`Nu{T1jPdkC@UfJQ#9og$zoJV(%`2E~~;Uy>d52^Gu^WW`t;rH-f
zsRxR5<OykX;dMJ;F}c{`X9KoR$lmxp>zgNBm42m+hVMpj<LdirPZfin+t_V~#nIyl
zpGWa*u}x(;BHHpr$+u(1*FVJ>@9EaK-CIMx_-o>Cun_O1NDP+l$3G`@Bz0xaUgH(S
zSn8JA)xPUy3^B@{qP<(gx6Sbf!BhBp=<e0@r29snFY(J7^^E$E2Oib-Sd2IJbk-t=
z!L*m^e61SGGK<64`0eh#zK`@hGr=A!k6Y4h)?1s2Zfrmh#-zAmoPbxReH*oUKaDke
z&l3DL*Dkc5FZ*uJ+1tn`Ai9jN1Fsn$TKN;f-YT2J+WeLeA&V!5-GEg6-`9`Ay?esG
zH68@<9i**1mo`%V?&)M8?v0PkN3bU~=HoC;zimlfd`so4JMdWUOVX>Qz1Q>6A2-RZ
zYB9QAz0{IQ$C+4@?gl+QtLTr}>*F=$#g3uj`-2_L`dZ00oRcwBmDNcw=iG8T^{*}X
z)$sfN5N{sd+Q#4UC)77ySV#s`#_UN0{l+_fn5-{`v1&GcHqmCabWgJ0Ex>mBqBQ1K
zW7q`GUOBI0GWxDNE0uX!S}sZUvg`L)`OF)|W!bc7!8u+@HQhJ<N2cok0JK8tnnK@d
znmXD{mVx87xN`Q9xMR7T4qu$8Q^+HpE5W>b@y}H8(^*M%dvfqQ7i&od$PrXvAczsL
z9;XE7ij!8B+r_>r)>BJC5!5bjZtY|98PIt}pOrjrDoG#h5!WKTzV_Bcxrv3Y%V&*P
zTF!AOQZf(9cyo|(>Oil}aIuAN-cNSFPwUYBlKQ+b?$M5`@EPIxVw=Qv_j4lNPc``3
zlX4p|$cvJD24&=T_pcD}mc2K_jcy%Z%4W5>gxX%k82OZ``@`tUKU((-4Sjqqd8KMu
zw$ximh<%dI>N4=!${0F^MaK3G^KB;p@=3-k%RFIlb#D>g>NeI8TCLe>rPJ(_5$|J`
zF{%5!^S2~&eQWHp+Lo2JU#au?jVaCDo`}B~^^Ft5clzCqHH$k=URzW0sV5s-KXiS6
z3gR_i+lN~gam8!lyBKd25Yt{yG{7F%k17Y_n&kA9(!4LMo22nA)`K(cSBB2l?I#Vx
zBSu&Z5ysPyX_}9LJ|B3`#>)nm;)wh^Vqgz*b*ftZmMJFB%p!<0>=^7@1mp~M`YP4x
z&dLtQ%q%=$+OaI3wkO0}=f%y=mpz^j`YF=kRKe|)0lyvDt`gHu@ju1S5xjQLNqu!8
z^Ce<-#|m~V$AwTmO?nIX7r~wtj`vT~d|-7e%lQ7!9rf0gCAGA0u?5QR%NA?^7z_cs
zo<(fwJ|FRAt*y3!rTCjh(Ch;h&4g^S%*1l!ungpm`NySmTCAhXCfe<DWfYH4_<yMQ
zZ}w#HsDo40(XVbKTdTIWIeFX+Dn%b%pl}JmB;@g4jqp0w_v1&8?fxkE^G|zy2T4;O
zh5SC}<<DQWNE@jgbNo$?$}7qK9sB{%v^&W4zYt9N)#@KGNvM3vTReQ3ApZcsje9Po
z@d^tIV;!8}EJ{Ni-+PXVJwBXOC^w_n!fRW_YR_i)o#LGkczLbu51oH>4ocvVI+iEY
z*TMe)7H+Ml)uOnS<2DwpCB)%)=L2ew_lN?$FJAHVaOrZ-G~p!*S=&D-`sclTvE$Db
zw~IA7V6r4wBkpN>Z$IJbT@m*ck*ykhh~#xEo2`3Ll1bQ&?U0|{&p-#YS`f!=8p3*x
zst@xts8Rvt#-tE%T}RFMRoy#9pHPsqMtr0@oq_p4_dNilqi1uiD->Xb<hWIhesX$|
zz#hK!(rMPmi%TUk$WB?v`B&VUf$#MDMsYRI+DYTegP``!Rl4y^l11c0`-p$J80+<_
zmWHcCqr24X{5uJT)xObsavtFFLH6{nW?T2TxGyU+vmEv7S*CZ8<A}jL=aW-I1d|ew
zDcXOHhhL?0w{q@TlJ-l7Sd*_$m!PXQ*2VI?h(|8Rt9uG;Hp{1@ZX28UusQrIw($4E
zonOR08@=%r)VBH$hi#FSbqzq3OPO)RP^jicJviWZu6Z@oxXCRJUP*kk*f5}tsu!^K
zsVx)i(aR`PyswoWI~M9k&{c4bz+@0HoOkyXyQA6?7UEsTD{r5X!%E+De=6skt?wN$
zPE90s{t(lFzSA`JQZ4O>nrwa3COUiYR=&mfYo;437!k3#u<z@;x9MLr{3P)6-)c6V
zU}jz3e2k;iaz2&y4v7B%WVBzK3`Za5>0d9HzGk3pDw;cgNc256;9cC1ctCPbZKAys
z!?W02TC&R_TgD_r&5)6<I_KApmE`(m;f4hSW#v45D{5P?$I4Pc2X71kE!5-F734;e
zr>fk;n))3!wR1j@87`c8?xO=az(3+EiPd$7oJli2T;py}PRhT4{&g0ot-Q!fJ4ko|
zoaD25<oZ@oQ96cWwNMvo;i*kZcWCK^nnnEv-Yrt{TU*%-7Nva4ne&4RuO#I0z#Z%6
zZ`yytHaZTgF0E&-+uYeoA!|!v{=P<Tioku}my@4SUhnZsRsR5lPr){td2P4a<eTm>
z_l2_j*!*~3!oD&1pRK;Bt=T{uaCUpi2dO_QkIKHUE5q5T1zL}m#jl6)Ew-Y!?wN|k
z8e$LeNZcU&&TF05^|zbiHJhxPp4HWrzuM|g=4)DANpw9T8(i&aPoFW*puh&a=S;q9
zzYsmq>ds&Duda%+ZJujquj@L)4I%|+B)MmN^*!t7o9nZ8t6WY_NsyfWHR_%w)I5tb
z^Kh2r{&nM#9FlTR_?pTQOHh+Ep6uSq7@Oyb58NbykF9M-DqCrYSwv9@HmFs}$>$!K
zJY@D2k)zuErhb4be8ithcBiQ|%PeTvhAf30i9J~R)!S1dyGiEDB#x@P&)i-B9)hjf
zFWRq_m=*^Z1GYK}%=1GKI2f(}01R7smjXw@!wsvRt=s%-b)b()vzzRiO52WYpbGOU
z{{S4P?l0%fdPcLUpNIbd0hY!lSoJRvIah4ptRt3V`0-w4Z1BWjfXFs0gCRZK{p0v^
zTb?Mo9wyao^&L;lx6rMTVS*AcU02W_T9+k#O?g{ZF}x|H!{PlhUlM9Sh}?O0ac7wi
zOyawnt$oLZFK^^KC7*4)2s(uS068`0ekSoG{x9)um8@zl?Cr23fd>(&9AtXc{dZI^
zb8l>i_b~jbGt~D#TFz;;c0VL&_;y0Nye*Ci$2HvC0EEcg<ou)Fyz@$sM{y1TEFV6?
zySta)Zjxv2*A$US*_l@|`O+Nctwdr~LipfSg^gW{cI%qVzr8Na8uaOrO5KMk-RM-j
zmw7lmRxYS^z&ny_N%GIM0#9rV)jet@8e_^_ZW1U^InHyM=%u8Q#fw*=KSM6v3q&w7
zaZ`Cr7o!hgDVlsz$e_kmdu5vxHSNSs$*`6O83&5=`HzZmBIL?C{o&TO?b~StfITZ3
z6_CiR=s?bLMQcJsV;phPrLcQg=ycZdt`&jWtT0ctZ)txtg+cjGBi9wnTDdB`j0_B#
z?6kvbkt0Y!Zsvk6j+agd@ql>8YSQu}Dbs^l+I7{U0k$~~JqK#n7f|jC<renNarn_=
zHzjdlvl&mNbQ(^UWp9*c9=u|*?XHxcmIcpZO>}x@sPSQZgad)WH0(sBp}nWr;Z$uS
zoYuUk(Z(^z9Myd~{hR_~;N$-Q)m;=8(Oj^~l|682q;XB(a~jV_^DykWEO!zsZqm!h
zZ<RPbeGPBvnlNB~Pyyz(mq}E?3J<*}D{f&UoE8CB<xWqxO1Y<4;lAkuJo?t^Y3hR{
zXVRS=zm|Ah59dcQ8xvdGY(as~tzLx-0mcP3-Y}}Y>e3E9ew18nv2PFx1$=k=e8Try
z?tz#6%G;D9(C*Lj74}pB?)$w5>t7xI+maiVM+>*gyNWeq>*co@q-}@RtWSon?`|Qo
zzte<-TlwZfqp3ZC_N<s~(pYeLZiH3u5hL5{GRH0ij*7&4)hl`YsWJy%zgopBJM=#8
zl#{ex=1jJ2X>?;C1YzH$R*B4Su#6nm-loEI2ct%J4m#FV^hBM@z;0^?csrxk!`625
zL>DQCS6-@l#au$OfTRL_>L}1p2<IogShlwh5yuR?6Pl#D_bU+=pE~MMxzqZxe5cd3
zWzP=G<Z>&eknCNqzwj?wsdH-_qaziZqHwm7IpCWB=m_p>(|#NHmdjK~^o>ty&G#aa
z@I#Kk=eYFmUUhL7mdD%Psj(<SDJsD5-3YFGxI3ewolZuY*!qqigz#zu_gZRPZaf%J
zGZE-ibUirjT_=YB0AyV|#4QB3*C$4am<DB+E)UcC=Du_Ax5r-<cuf)ruEoR(yFy$p
za6JL*UV&}<RrsFb+hx9;Zm+lRhGqlb2D+nCGU=()I=EXqtxvVS6Mn~@4e-9L6~>*W
zTzIEYi!LrT1`iZ22MxD8srJdPm-d<cqOHCHcw*PWdOY9S9}x9O14nNhWi7P0TR*)j
zc`g7UK|fmj+Sa^9<NIV9=BpO0*1DtlP-+(fSkcJs8-fQOg1s-`XTj}X#MdWR)*5N&
zmkSKn3}j^<q_?*o)J4jqxl%kz+>yNRr$lof5OseNU!?kt_PwZT7cziL+)l)|ew{k|
zSBvTzmF}r>qVnQNPV#?x+%8Ww_SfwP@DEt{bEfIjYLKad{vS3O6yy<(htQ8o`Get~
z!->8n{9*9_0EF%1+WLj`cF)PkOy$_0Wdzc%9TzLm=fYKOwKb|KXj(0*L?f2*QI9wS
zr{`KXx*4;9WJv?47_Yg1ZNG>91J!&ddvW2d8r?PB5)@H=WCD4y$An<OaQGei=Dq|M
z4kOwe?e;a7vt5rub>~k>3ml}D3^#L;>MO1AvRm9tieK!qk&K_X-(OFyak0P6Tly5j
z-r*%|6;GI{Jx8T=Qi@F;9xiig%c=G^?5*JmZv0bqsm(mF4JOnycBL~MyZ&A3dSQ4t
z9=Wfnt|wBfv=R#AA5Q+&@L$7Uj<4|2YY1db)>n+JDJPiJk<;6p5$#`SX}aoanrNQh
zM3ogs`o4~gGOhULxu{}gQxhsv+3b9FYk_!Z#<W(AYIUotTIt$V)N$M_@=BsJ1~5*2
zew~P{Jv+-s22yqtv;oP_wtLl|4ruy*r4)~Meyckajw^6>m&yCtJYaVf*?5cLmZfQ{
z+l?<zTir5H8WdK@04_?mr#a`neKM3G?zViorB<?4BCl@8H7s*SX%*0hSnXusO2B~F
z>(8&XXC|Tn#zSOa6Y1^gUgxRY*x7tJXm#62bf~<n3p8xHoN!k?^Itp0-eBHNMtI5T
zSksrZyJ*`JPwMiFQ@z#w4ci;GGUZDyPatqHT@uL>p;-tyA1@qzYmTzI1Jn`F^c8LO
zJbc;h^s1C@XHKHAsdsX)50#3J5B~tFu=NPyzeD^_lk$<;sNZU?SwJk@9MwO!0KtQh
zamTea8y2T?GXBIUar{8?d)F_l*{Y@pD~`CXvi{%=?#z92+w!hkS-XvqwvnGgdK$G5
za%kjkM%T_uV~#4Dup^$QjyqMmxd|*i@^Mu)E7X;3;EFJb3gowTbIUjK0=H2#Z$hlB
zNh9Uvrqc9>*0n2J7>SJBI7XAQFII2BWd0RHc>_K3#%lhTsJ@qZ6zRwe1c{sp5RthL
zvkni|oTH<#M@Ft|{6_NDSZ$+28Jalc5gxGGAjQ0X#fD{J>cYH};(Q=@KIF12kt-==
zZifJopTr9EO><M9#2U@!rFYENQ#7r&fHTm20Kgwg;{GOEdHxL4R0EBhxc1NZ*CgS~
zMWfQ8SDzV9T`rHF^vf?byghX(X|86!Z#G6bX2S43gVMcP)-1ycoP&>>8~{5C^KBhR
zk)V^#%|oZuSF-qPLz`UGEi5fTS?(Aj<DI^w{RbwwDWz+lO^ctrujl!iULW{v;vW=P
zo6R=N&1@XUyzmJCXCIvv_P>IB4GxLnn;RQ+F+~am3%8u(l6}eN@uIMES2%Aol<}Rh
zj_f!9<YS)s^{+kF?EY|KA3ia#`X5hP^&c5P8G$1`xUL6Ww#be78NeqQ&3Gy3ef5rN
z`q|-@-DDXaNMICm`BsLPZW-Y*t`SZ!$EO&lUppv+Az2E5cY1w#{#7-a!m<x6IAO?K
zj1OOWw8?1OTt+aVP?gR>%}Z=d^0O&s!s8eN@S+IPL|NGHA$iaERl7)j)hQ@2xz8LD
zd(&>fC)t|fV#P|}gOSG-(P>t`SOJxY?n%Z4M`vhoaxi_XZrapgdXh&<&CF^g%{;6n
zSw=@6dmVkNVj0$CAS&gF9eN7Ow@)#0G6$d)a@2^5zTAP&9AwnCRgIfy!aBQ^Ez@W}
zjY)DsL>uoPyZ|}x{!Les<SQ#<X#|gBoL61&$3lx<)+M>VJMHY$FpcqsZ-1pLm0{i)
z_+H*3NUpUTWiY57S)F(zJ(s`ZT<^xOhq3FC9WFl-T3BhwyxI?wpH{)-`g&Fm?H&6?
zYno-P)|KHPmQgA4n^Kp#$3*q7&z~D!_@`XGp5pVw_fpKvT;-1^z5qQdq7ZLXWhpI>
zZt+*_t)<;Gv0C_w#LDF5nb#*g_v`hqo&G6)&zh%$;kdbBZ+jer23vd_cJ06w#a~(Y
znh9Cq)GqD{x6IygB#ibM_pQs{iXK(1y_-RM3$$fgS=*^NCvfSG3Xft5=C;=M8dlix
zFBM07_V<@lX`Hc9=5E|@Fgx`C4xCmE%foEZyi>H%wn-ayl^Gxdp8Yy;&3dPZ?_=={
zO43{Ic}pT;HfIN~J%9Su<(j^MabiTF8R2e20n^y~)oh0seanz)EvxAgY0?rTk22eS
zd0;-j(vJxEvRw-17}_<K0OUv;sYb>@_2>xqtRZ~xNejlJDPlaZJ$Wbk;-$Yu^OEW>
zmKn2<KZU!Bo9l9#(#NOxkHaf*WcT`FX5ANjkb{;*Kg2)T{#C(V&IQ0|j@B`cFk^z<
z57)8!*Qj_aR+Y4?N#Kwz#7m43klvt<;MD#h@GXytY^A=?q5D11UH4WQ`PBaac?a<?
z)1`FE4XdM;H921_k3G;lS#_dK6!!NHl0-pLJ6G4A1im%f`0fa_?LtJ0Q$dw-*bY61
zZ*gB4U0VHy;#gZ7y#T-mzA1EVYVX6ElsA^L6nNM$BcZ5MaCTu*glfl`ZhnsZMA3CU
zOTkw9zLcg-Z^HMH>ek9SvZwmMJDjIKg?*2r`0vBN8h#T&G~z4WA6>99&kOV_>xNV4
zLC?_Fz~8f{?Fnt-cp<aCl~(>F+F6Lqi;Q%wdp{oEe$0L%yjY>Q)V>*N)0y^LjLW7<
z=PU>PGd+Fy?_CyC)bi~re(K!$llGkWOR0Pg@nmsnmva9AWm`LqH&wP`vk3QY5lgh^
za*{x9;=DUr@xHU-ooWqNUAt{^?g-KmCHwh>kvLUu<O=;4_@MqB__g8Y)HLa#)vae;
z#G`MFjD6Pbr}<aQdWXSp2qohaWvWKRljK6%m*@^Ztz|k*{3m@JLQ%BVreDL)+E)8O
zb%ymc=rr{oEv}uWq;3BIeGFsuuV1zCO#c87^e?tr#Ue#^gf8>*W09ZIxjzm3F43cP
z8gGS_n19P#_cA#53twjb%pV70)m}OED>x%am025L!k&a|0~qG1UYvdAVNu4UuLY^_
z29NtSc#p@w5@oVWNv&+Ba>DP+$_B<l6VMa+SJnOv{g$*JhMIhHYnQfA$RQ^e*9yDF
zdKEnM_U&I-{A9H7{{X`u3u+phOKGHAK^Y6Xhn`$~s*ByT*mUZ8SHV6k@lKENhf<qZ
z*RAzkGf;FyH<oA2jD;8<BaT1dBvvkIYL2xTRQR?&hMUFz01mVp6cYHlAF`xDKFO!T
z^S?|LQrJ}jyi>)07CZ}osoGg-I(7b;Y3Br+RKB|N9H~El$_fSm<bZR5USFvEG_|>t
zevjh&3D}$nH5Gey$ENN^YZgC*o+oK>cKUUtusK=$=8`|as9{-iH`LZN)LHb6r;mI&
z@OM|bwbu10Zmv;=!D0K~F-Y934?qYY^Z3^M{{Xblf-ZD;@8Pl6F6C^iK;Bdv4sbqV
zPJasd`p@CTg`k97c&|nCg6I7rIZFPY)|;+rH=^AlvGLb~ty88_*}Tuga5y!a_qPs2
zk?Q(~?PaCfOg_)x$;G^2mX_~RY#(!k0DiTj<Ij%zPNDG1%UXm>r)cxSWVnxZWtwfo
zEC_#zax=Tr>0Udcd<fE@w0{%$>%{W-gG;uI&4{0Ey8+N=KR2LJjGE^3)xQ0sz9w44
zEMC{b8e1wuVIE^a2<U);fXoNsU6kOV?=zaEq^!0&&)Lu5?eE5)7tiqz#qeW8f^1sa
zoQ0i~{HO=_6k|TPue4-Qvm5{h00WFut!yncZ9dyh)9s_Ou(ge1g5D8?MFXJ!06fx4
zc&wz{o!J+*+YqB{Be3mOFbD)yQHE2_an5Q;rXA{DsHD+uBn{0y<G{$K#NR13Fl^?r
zHd_?sIK@|yIL%$TUCKzRVoa$AwNeIG3_-!A9MVhkaw*v9S|V8198*UorPxg)M}eAm
zAVbotw|cP0<rP`Z-f5C@-X^odtf|u@H5QF_&ZQZ?Wv+_i-EM8=Qae`@JjYJEJmWp<
znp$df!`@bBRsEuox6FG|CZe)Q_349L@PD*`cm5QR+zV-dB(*}BtEQ}Ut#~$`c?PIU
zak@Zqc&ckTcgB8JsG4KA=C(QZWx;WhX{KM8^HmmJ4L8i@fVqqDDCgn#h(GJjPvu`E
zLW2PIudDtd$)5u2gVxdi0F`{TZy4h}DW1j;Sie!GA&v!mKkRp4_x}JK^r%pT+>)6a
z(lCcD_!13y3MNw@t$Ppbw<~F$De;Z$Cv5iK64fP)_Y5N7isi+o=t(ZWsp-YrQ-%Gc
z{w6oY-;3Jbs2Dz#Whv7%)P6*E@RFdY>{+)l$9z>EhJHVnPq>Fhzm=bRV%p`-Nn&`7
znEr;m+QwNf?Cs6ds=t<@ZkO)294v#^{i>4Wy|^$>8AjlPp1y=0{VV9{QB>*1oj1zT
zNi*j$G%4bs>D5=i;pl#Wcnjho)2-u&P`KF{8$^;m4n0nNI{MeCufz>M<A;IuUkcr<
zS31R*Ww&O2ZzIRdo4>q3>)eGk@ZW@W%ddzZ4ldDVGi&-H3yIL4Ww*vW;r-$;6}>Vl
z-+_K2&2N7cw$n*zaSj+Cy2`_`^shQuV#T@6PSqmxU7G9ZBjK@?tK@YcsY3Lh#Qy-9
zZMQ~_sA-pe4A$j<yUQ4D(MgPXk)OK$+2<p%1XrVYW5n;JYBoB>yX>0X!x7MW9Q$K6
zqorydANY}@YEf&kLu-9?Z7bSpD&-o|=pkU-a8g1B;Cu7lyqiP6xJ0u{rr63t$~$fy
zN9UULAwR59ojS3zO3d@-l{&QR(3RAlo>uO9-^M*ST3kV{Y00+Pg4`hkAGnb4Kj1C=
z>zcgrWZIse8?2=rICJ#sKPuAk7l@DiArd_E2Npvj+D8D%8*wCRJ;x)^t$7uuvGz;J
zAX&WIZm8qBAG+U<e|nskn~~9y(QHQ$!qpXOPR*@X>V2Q#{{RwcKNG$Q>Hh%Qx?<}$
z+C;*}Sp<eTZO-44MRCa7xCgPvHO<^<ejbZL)NRItc@3oJ=UaK%B$Yy)%I}P_4!mGj
zll%+u()f4AhfKPdogZF7r`#hkz-G^q{B+v<9Cm8!bS-06yYY0^UPqT_47j@%#Iz{N
zd5N%b)EtA`t$qW?m>k0`$2d}~)*4HlM)p@vT|IR_ckuSM2ROlfW~DXtids2!w@rF_
zjn*`q$)b+RD@B7&w(_oRo;bu;R^<tBRg`?hw%p`^a5`6wc+0}DX%~&DeX3T~<+*=4
zAfGe<jGd~v1;Ojj$O!9SQ?8%l4N~f5w~o?VE0zfH&IC}pmcYWFn~ng-HMbAKEf4J%
z7q_-?d6OxMNG;urO@V`xjFHD6SHobqj#+GZtHw?*YhB&;`JbfXeyaW^xwo=+ck}+N
zetll}cTw>g-ASj|7<BlNntP{kk=up&N1E;TMtH%=J%KgFc>6_*!Eb462BjVLrEvsv
zvO*urfuvl7C4gxeN8UIAdSKVkUmrdrUFzD4Xnr5DheOdc$z?$t5v))|c7f*2w;`PK
zfO>KKS+A6?^=%_t@oZLiy5#n7yA{3Hng*SMa!E*m2d)M>l5$TK`^HmS5s0N4vX3fi
z_TBvd0Lb_($;J}Zi>BT9Q^OjQH;Qd;b?Kr4Pq13bBKt&{&e<h7+Zpe)ZX=E>fY$FJ
z)U`{?xU8-8<eq1hEbRdiKp5mAWRtWG21)m)>ZTcQr#9@*0x?BrU8I0`Vlr?@!Nq5c
zlk;#o0bf-F(!94mZgz6H=w1l;wQ-}`$E54lF!+-}QL%0B5AUEE{{WUUoRs=~-)i(&
z?{xXt$>FQmH7g>alE+kxE5O~<?#>AHQUUZ8^Zk@RXuOU*h1|dbM-}NBAA$y_5H|O(
z1F$CyGBfY_R_d=N;GxWP9wPCliTpgo&86Hd2d>0ERzKd!70TayQr9nI#l4zD&N!A?
z02_XttD+hwi*#iF07JfMVdyr#;KTjXS(iGWi7llMJ+vl4{7(P`{*_9ymds9@T3F@g
z8t;pXCB*YwqaXP8wo|a<)wme^>Gxg?zS6B4;6?JHD+tVgASB>|Jx8r{U+|RpgnWiY
zg-=kCVyjE3UtPw+;_iEZ-A5Qdhox(McNLCUUEKYpBbpdeAi*AGi;%hG)lFfZ`gTiy
zv^AFE6iz|N>~@R+T|u{vm=Jdoa!04VD%;!TyvI}^@}P7+{?%7Zt6b!5G&h+D5QC2F
zr;5K9gx*Xv&LUq$z|D3sC4?s}aPAj{+IHks+iho0u?Ot-VrCs3#^d_dQk}F$k&9<D
zr+7HZ0E*q6I`76a_?odEm%3!Lw20kZNZS!0Qb{MU^)<D7tlw(sw*otLAxuheNHw*j
zYjfz)u+t%q&TxTV>-fO$wAVDI*5;}g*wV4^4~Bjl#U7{QWnFJvQ@Rfh!1#hQ)EkKh
zUv(dyab7s_#;fD~LOZ=qRq|j4J6kJN^6jlt`;kUCQSa2$nuHH?B&{sXBeVYitaF^P
z?^!e3It{x=2Vha#*0HtK*<7g^u-luN)mH!(=smkvY`R;tmX-++S-=uV6|<IXZ5Z_f
z`Bg6u>Btzm%#euBnYu;!rX%<l{5bcoLilH)KZ!g;J=99`yluT?d@lFu=N*Os09PGa
zmv`unxcjkdt38Y1pN7*!(p}szl0^^&Bm1~4N8|kK(V-;E6j%TOmNki`>FH|>5uiAU
z2J$=Sl7BkWw{}r1j5%|$Yvv~}X;H|Uk+M2VDgM(cjE|V@Rc6{T?t4|6jZLKiHlm-n
ztfK_3f1kZ`Mso_E>raaCqBNxa-kGbERjhI*YTVuCO{WD@jQjWet4m90=fAPEk+v(h
zn=w73UVkd&;hehV&rFk=$MGME<kB?D+4th=c>JjFIw|e=_53RZLz+iqCCrjXAMwA(
z-|&)nlT*Hh$hNmbX=uFwP<oH4HS-nq>RkA`;otpLxK+vYQP1m|^e-LjBH3Yyn+)x^
zf_me>(!5eD$GkW4E?QaT@{kUvCce7|PnIakyR+p+rm?KaeFW08<)>l*9RdUXHJ70O
z0LSv2bz%JJx}CI^Gc~-7zGO?Zs+@pvQTTpBE?0Nn1Rgurv2S&pXF=mjt+qFqJrE^%
z+zMQba5=8O#*&xQ4hi|>1pO<Hfaf5SfOA<c#l1~<bq}#xGA1(P3`pj(+?Qkh)6e5f
z^9wmFa%ssX8$5&WS|c_~7;i8x04Kj%?KGG^%LFW-1}(B7@7<5`u4lv6M)y>OF__eu
zU+6imxg93e3a1fU#t>t%@A}ejLZGf(7po4TADRnWLDUc9Zuz9~cZ;t4MKVb-@~z-d
zu-mEmR1TkwV#{-o?pql>Ju5Yt=9g?^bI8F^R7TC}Yc9t@;fFTn_fn9Sk;slFKi=<B
zTwLyvO0hdnC)Cq*6#mDvx3XScS2$k8^)*7?U*TrR^`cLq8y$w5DPWlE!6!bI-9dHd
zTEZ73Fz;S#rrhj|V>n)M>s|i<hi(>o62$K-^8@`Wo<Q~{zj1FLl!y12=CEa8GDrC}
zcII5j&(2AzXO}+JnW9TZD&p1xPx|Nm1#|l2`FeXVW|2;RI@*{LNj*(r>+$BnS3DU8
ze_E<<-$pp4`-%DyXz`;WmQDF$GHSiT1ewATyqsY4sBIn8VOXf_1x{nY@0#?|>U=}o
zeT1?*O0LY|NXAb})q(Oj=~-5YW0Fq1fm)WxLlQkHnMl^N%8uZAS68EijYBEH?!Aq2
zwwVNEbLuOi(-Uifz}?OddbD#%=ykTTZVE@s>6*N=Gd98U4EC&RHx2TPpRH}%4azZ-
z@7z_`<Zn}3Om`=4Mk}q+tt6DGBPx2=HKf}EB9cJC$i;ephb)psN87h2f$u{mQ?|!L
z;k_7nYRsUq;Cp-5Yhj=%D%(gtpW{(zI(P+0QT{d7Ll9BBX#=J!88)^#_jhG@ELIf=
ztC8(n+FyoR^}u0++#YfGQDtgW2P5WgneFtg?=$;86<C|*9I?%FN=(@$)Oz?9UoSE3
zDh5boz^rXU!>wpbOKhP(=c0~3F<Wb|-ztJ<;{fnGR_sqF*_2(NMwL9kt=wm!tkmTf
zsjn<5d?%P(ST(rbgtULCq=mkiuD@E*QYAMQ*xKBbC;3++a;p@FZ$rr*mDd-{u5isb
zyE8)FLFc&@@o(+9;tQLL?+fVC1XWmMyxOV<B<&eJ&MWMTd#$4x!907{$bYt9fiEtu
zbpHStM9n>wvI{GyK>q-bd>@sE`y`wb>slqFI<VIEsQ5d_Owo9bM#u1|TF7~D0|WAr
zUV-90DjV%gB&aq|2^sDPJRicm^77JMChkj1j&}Xm^&N=%)>4bNk@c}^$*sulZh+(-
zJLajq`l@+Pf+{Qdf=SAC^sJfW-o%cBwR7_=ndnoTugs0RYn2iZtfU@&F;^9(M)MI#
z80QtrLhZkUioBPT#4@rn;-)gXJsC~Tmd8LZ<8><{bMIATc*7lxcJ|Ftk5-w^A8TjS
zRj95@j$&V~DwDb9w%J<oLyY4jdRAS+u;eKtJ;ijMWNf&2!1onWdpF!dw&gvlB@2m4
z&r_UrW@B1f9r?ALV}0IP?hi`W)BHUZ*&<aadk}F|wQmf|rWJI3(Vpxvfl#SV9`XlP
zolkM@ef#jk#-1DTufiDo7vrth9u~cUHs8QEOQy`fc6i4)+HstA>t4~|{{RDj!T|&t
zebI{SK?-?SaFAoTjE|CZ^$ak5jz?<z_xLgK6zB-LZS(myjCRK%ISHPi`x^S!;D^M=
zzp}jW47WdOy0k~X@Z9YA-hghJ%RRvvbRMG>)lQ;>uJ(JJl_@#Xv|h0_zifXTYCbQ~
zEDiGsE$mmyL{-4)f=)V<fnF{99Q;1}Z;gCK<D0n6y|#-1hT3)cb8#+4Gycl+_}8F(
zT=4$@i9AVVpy{&dliELs;};fQe;cC6%#1&TC^<Q<>rc}3p8)ta!@=4M@3esVmv@=T
zyPdJRPsgAhg14&(%1SiWhIj~0bsXkrTt^<MEyB)#+s=kIjE*)CLH!PE;BOrGW^V-e
zzgE#D1RFWpH9o1;Pvk{@fC#YRdSi<DWA?n!;<EAHo2W)G*Y-fk>A4Qy<yBgqiKEfP
zYOc=+LcU^&7@}vAqa=HZ@3b!srRJR@nc<ZfC|MK{CUek^oZ`Iwl#_wf*JI%wWLbD`
zTHgNv@#_;zfS+s+)Yl@HDs1uV#mz2W=)&=5i2nezb(!IZK13h7kMQ9BXEpZE?Bnr3
zuYx=|CCUi^{{Tao$vFds2iy+cmHF*wA!1pxxbt6Z{=&W?@a5%|-h<*>M7y{9Ky)CH
zBxIE;%Z|Y0Hh9H()hTmDHKSUm<~)2ODjc%eA6V;_4{m&yBV!*iOklR${{RoCzaF)_
z;a?v3yGD}V>{mB00<KaSet4V?57!>mtMR_t%HDexvyo&~whiaNeo{Y&1a$y+HRoD?
zi!bc#C58);Z*F3HW{8utZXmWhf_+7OZ7DjmS29hXj$&Z&xVb1zt9Cj+9oTr9?&f_<
zRF381m6Vw!1whYCcE_bqywdJGH*KcPs!R45FVxEs!FSj{h^q1d>^s)i!V`U}c#ZU{
z`<K1aZeeSAW0}|wEU(@G@Cx8_^{sCiK8vc`eU+~vn)^*bX%+4cE+7XuTy$V~{0%>7
z)RjAJ*vZmTrt7Be_PguX@HpAXg6V`4Ey`dKBzY=9_B{P+!b(|Hi5&I%)q6c9?jBrE
z5~2oe?ARQ39;|x`(wYrAKL$&7(@MISxe4Wx6nSKI0QEWa=~~I9u@oGhjK|bu8yGGD
zUYPW(SoIC8Ltu_T$j=n(i;~10!~@T`u3u8VF<dxP!Q>p!k<}M=YHHe&lw*>(9dle(
ztE&+F9{3m^=LVqEwaG?M^yfau`ql;2*?9rZRC;t3Ngc3{b68xN2hGPlYOT)<#GGT3
z&rfQK&fsy9GfRH>Esiic`c;?9b}pS}FRp}a{!%D1O6-2>j;H)SwEqAkZeprcN%?nm
z0;^dP6!FRDpYzRLj1&MIU<`B3JA+mt@>m{(=Qv;KSbj8sP2sIVJ-N&dMmg<W8vsvU
zob%J#vvoMd)}K6NVlx{9@5f5UqP$u*qWLPqU79{>(jp0IrxzI8d;vYiI@jA@v#-OK
zxzy7BcnZm%Dl{w8`@|o~e>(Zs;pc@`=I2W$5b`zC3w34O1l>ch?x-@y@UOMCZxCC2
z5BNO#!VTKJ`U@>0-Tw129J&0<lj>{AllE}2^_4%`DsrzgoA{;sL2Dit@r~Av;d>LP
zSy%xC)2|@TGF%U!Dk!f99}#YRAO8Rn`)eVo&;GXV;~C8D%MAV^isQ7NwmmLdanZ>7
z-^EtNmQIXwoZ`IuTedB1p1Z5oJV$OyxyNrx^Q~&vC<fe({{UL?Q|Ns?j!yF27*K#%
z;2tT^L`0}|0h5AB>svRL3T?*16p_!T_|ziBwv6)KK^+BM#mY{u64L;%<n-uC?^;?@
zh}A}ZZsb;K{PBp+MhN;J@Ts7Xrya*``KNItc6Kg|n}<=wYohIqNXO&exT`z%hDFDJ
zTIj9s#H1iSMt+rx(^FK9rI;eB?FX^Rr$*ju>;tAVwzGu$vT>Yo!L6GJ4#0#B@^>Am
za?;}EE+k%Ep_RZYGI-~4?OundYF3{Nei)<7!aYLyaJe{kcI5t5#9Ym79;abzI|%J!
zb&_U000$>OfazMzrAOmIcjIpqJeJxVD$!|HC9(Uio?#!}Zq(_;%EZa?pEP*SNY(x%
z>ZaRIh2*riAH0`4Q5(4}+v#2{@gD2p-@^jwu6U6zwPhbQhMGfQ{{Xbr7wx_KRv!#R
zo(J$e>!&*cZcVSsVT|-rM^<d)lV6`#w^y2l!_RXxlS=1vJd8;{TJ#mE=#EuQ&sNW%
z{8RDo!_r(cY1%KB9(u_kc5lb0=~<F|a?<bAMA|S^$v9?GNg#Is9C4cPk!>@fWkz5+
zkUCKfyt2rv8kS;k0Ho&I(Hq7IA4q9_H1PhbrZt(jTX~U#sbCpZj|6489gh{sc)|y@
zR!L&{uo#eX7$4%zae6O;b?fLzMv54mjG3kjgWA2P!M_ZzwHqU4sQ9BoYj-*FT2|h@
zv$!u>okyZP2UD{dGRJ_x;HdBKT0R-Hi&xWb>>TZso$>n|U{|{MkM=;;F71T=9M^Qc
zT4e_|>nV)4JYb#)$F3`Zli)wZ4+!c;=5Gw%#;L~Wg2<nmlbY5W>P)40=x2Bi`Rvcx
zr8^on*vfbyitpuHfMV1vOGPqr9S#bf;<#N~N4tvTYB1Q{M>apwt=c9GG5kPd^R0~w
z#IE+=yKIIX4?OzUa%*FPac|y6zO|ugo;{4bjGAV**U!cwnEE&$bo<wt>;4YbJUbyx
zM#4zq>a(xS8xK&!j8|8vYqKMqMu<=GsK$Q2^;`ZD{aDB3EyFZN{#j=$_>PsF>8WZ}
zIZhg!kB7WLdEvcHFD&M0o+b^sPw8JxcwhFI(sf-hS?gfi%E__Ijq$kmAJdxo+gU}p
zl^x-aelkoTFXxKGXpjO!HcuD>C-bb5lr}b$V^LkB>D@Z(!0>8D`#{zkNt#Yu>=$hc
zx780AuFBKJ9v_VPmzuKNfH^NPA6$`&_?hgs$`U>DNst6-$<L?%09w7%;eYI7@h{?N
zhGC|*k)_B38Se9)r@qd$Ue-%SXDSeG_dSx&;`O(Qd{1TJKZlxd)pe`cRxdIjOi}u~
zH*mdc?N1o%Yx^jC8PN3GX_D{8+Vp`u*}iA!43d@q0B1NJnH7`p1NMFJ7wm8F>i+=6
z`ux$&c?4Oxy_EryNB$V&{BeaH`q$4MJNT!ld|LR6H;nbH#?GcaH%?~CT0l8P`jg8Z
z=D6zBayn?yih3g#_LJ2$-xBLKT7>=}H%|;J6p+m1#{uh+>73y4n&*5iFO9U_9yYX!
z9ZJ_2A~KS$Ki%t^%<)dBi6r0Ja^fjiv3EHjlb_1DLusg5$TbfO>hCOE6C5vsKA`ok
zZf4Qg;mfhw_>WxDb(?RqYgz+omzdAmZP&{xeRiA!ToiYD{{V&^q|o&Tx`FvHqu{Xg
zC{1cX@y}AVLt)|z8%<0ARIS`#upd)dy5ETWGp$0`77H$qa-fk4w)j}%r#z3stz8-D
zVd}?R)!F5;(It-PozHG!bq($~;QLlJhlygdy*hrmrrc>aOn&r=KtGu2TY7KCU0T}Q
z*roK7+OYosR3<7^_8`;lwQC;~q!DW?E#-+j3&`7~fsb(KBoETI<(A}K5nX6LAn|S8
zhlcc-=Kjo?{?PHOW)#N;Mgyl>_umLUgYbVuiWqepdp5UF&d-;5r9Ox`IrSs}E5m#N
z;-~O8i5bKq=`HetH*>Z{{vW6ve+tI&4!J(Be?83XT2^v*emVUsbv{^}?w>rokF2~U
z`%p!HbjMS;({Al|H_FO%`-K_cDf#1mGr`(Vt!?YSwYPw*V326~KBKALra&5{!o>+b
zxQ{!&mGG_2=ELSp<wtTlfBNe-^6ptP5)`+LADG-WKk?DcVSAeUIG;=DpSJF+AC2v!
z*CVn20EBjVj59%NYN*!Bo*fu+cJYvUiu!xPo;1;XX=b<DCZlhuTQ>#ZeC<ArI)nAE
zguWQX@Rw54V$u9nX*=suInJ>h?t>?w1EPREM>Xs}2)+nwUktu0*y@@dvM>B4bdDzT
z9}*QK<qV^S=NKZc+8oo2ZhdU8gkMUKNt&t&7#`I1d;wSte2tKK>s01U42qr3;rP^2
z{KEpYB(ANvW4&07cokb7QL7P12o*?7ltwYum{bn5{M0Uap%O(a3>v32nH_3b-x(D~
zXU`d?NS9LiaqxLH$?GtzgLxSuw(jQ3W74x_0%Aer8p?Ye5YXjC^pH)!AXM<)n}^Bk
znr^EIib54?t)4P-T*q%sPKZtNAYV#hkb})vf>6oy;+*eYG1%ggvnN?{dee-Y)lkPY
zJljVmv@<b&AwKT~>r0-tu0Kloq9ocY=+7Ium%=*EKI+>!Kb3sw7-Ky<pK4Y;oC^0P
z>OTV*9S=(OPuZTzS+$=WNwE1X;N4g&{{VXtVE(*Ul-XL_!vkFE4p9Q`jse3-=uh`Y
z>VI1IPuUk-gI4kHi*$LT3in<R)Oba}P!kx(VS~W-tg9>2f7esfmXU1vbHiFZm-=0^
zLZf$-prPbAz^<0p!7^V!E#y@&Dann`%%19c*Dv9%PU}jyfIv$|5C&7h;=ND8UO2J3
zDGlwW2;l7*HS5%=B&?4uLhO^_Z^Q2rct^*2HmRWMR`Il#0U=4^IJ(`zi1uLpG43m}
z_|5RzPY(P`)-BUdy|*Z-a+iKzmleEacQkCofhvMX3_f9w_2`}i7T4_;+0)EO_PU2$
z?O<}OGth!KBc*+X;lB&`XW}-MZ+GGyGArBo_e&v@b42^oVGg;$1QU>W2a(Nu%fw>v
zxTRioIpb?=lV9d`V`*UVcu%XUT)p0|?{2#z;V*@ED>7S1@fqWJ1`+uidxPzY!1#l0
zH-fc&M@>nImeOL0BaH4t^CNyLL9eH7{{UrOZ^2$8mfyoR&#UOO$QIR=)IFg*`ErIM
zC_75I<AMq1xlh`+;AQ58;jI_M_VTW%yB5P!mI(HwGAaGj0(mRW&Bu(6*{`Fem{O^T
zPn$`re9y@-IYuK3g_?|CxbJn<d3?1#RJrj5?akPh;(0C!AZ6P)`U=s}E-mh@ZtbpR
zXs%-U?Yy<&*QrC_Ki&LmJv2MUwt0*Wv4_}8uH1W^^d9xwX&(eNU0+PnFKzGbV!yo)
zGHKU}V_^XQ09P-W$jkCVk<eznj8<cV%W0|N>MC*W^gLXvF2-h+XACtrQb}r-r#Tj;
z^2U#C=1Xz%h!QjN7c2s=u1jE7+#VtLkzt|gGg<g%C@!^2_Hiz&cl*nDCTQ1mFh;-Y
zoG?5Af!x=Nco*PR=Y}Uw5O{}J)2*7-S)eL}M+}UmM1{{*T;*^$Jk<{o{8{j?fNm0f
z1{c+}H9(E7Uam@Y-5d$V&PH>N_4!tLnc;Y+6AC$cRU=OR^}XA^*V4=9<bKb>`3-E2
zg-V#q#n0^DqVI3(x$AnLk6`$1rZ%hMeN1bfD3=~hwcBB#AZ1GcG4l1>jE>&5^CyeH
zYkhl5O;}HTA~vgV*Ecp$$+2Z(HvHTjoM3EUsUVv1{{R<!NY#90eI?D!^zlbDzdh4r
z=WcRI>cA1kIW@w?Ja-bysN5g4!DqmC<(>1C0efyEk=Fp4>A~bUd^{Sf9(L^A*HzhU
zkEW@Nh54go%icWrm9F@zcG8<}mu`{*mT(P#F&M@@NIgj>-nku1!^@`1@ioME1Og1v
z9mrewg17Micq}@eoO4tqyPp2-rjfkKrI4`)ki?vv)|~CKQ6&EWIlx_uz!)8er)*c$
z)1?}5QHruWIn<Qh=I&%&s)pE#R{(sZcI-c`S7HeDAH!3*X{Ic4NjW$@sE*!Smn#~V
zU<(`$!@X-`BHxGYnr%fvT<uca{*~*#8-hz;DNJTh?=teBcKoZ${4#+x9pzZ<P;rcB
zEsFb3;KznEPY~$#g6bIOnMvMdj{thNsKs#Nt9uA@HFS1SihG`WXQoee5qp2%g&T;7
z77D(n&{9ol69-8eFhb{JAHDCI`eJ{BJ|2Tll6YddvyIq>7S}i~qo6tM&T(Et;_nZ5
zYfF`3mhSB@T)06raj*&f@0X=`R58%6J3{BB?5wpsKULF~DA>B-oRZ3aL0pva#;!sI
zLFzxPdVZ~?Xm=t3{j)KaT;K~}Ndwa-t#fmDQ%{<A7S}V$&KX(|ocd$dzK#_aE~k@O
zB+epLQVJ*y?by^qPmzg{*Jx(s4!t<9Qq%qwUg{9$({niF;yl#zd>YmgH*l)p0NC=|
zIi>ZOM@yN)H|lv*Ry&w}?>QI;x%aE_+sP*k#YY?gfH7UQ-@`pt(T4Plfz%cw`c{0t
z8NRpqWJOO-K^2?C#p`pr6t0Y0Yi}(aZpM0($I`8=SSf&5jy{LhterkYc_Vn3Of!&q
za6066{Ap4ka0p@8gT^ZtQB3IMwX`Zdvjt<b<Ux|y;PxM_Qd=xcNLDW62&@KjPC54e
zRnUNAf?jZ?NC!B{&jZ}lFj}+C6mcf`q&q^MxhEf&T>U6jbZ8wl7K+Pa`gWHjfG8|V
zMhkqzFAbk^arsx#9|b%WHiO}0X`@xRy=)RxWB_vCfjAZ9dM=M|toX9SNr9GVU@E}>
z0DTbk{7KI>_1A~=PbS_7ND3f~DkvZ4xvw`7QTs3P6=bDl(88HbsE};qkXOAU;x|P<
zFO<8Sd!Ci*3*h8m2~9J^jXl(x!=9>%#(y(h1=fcpowEo309h+;Yz_kd0QKv}P}b$o
zRjHD08Jab5zaYDxK}kKxbXd{8@yYsArnPj9q+zppzz2_%dsPcmX(V|Lagu(3RjnP4
zImH(5ZdMyrL&JUM`d65E&suxighGDn7r_4U;19yPJyPq;j#p+N1teFF_`6y<bXT{e
z?r4{6YC5x5ejXgm+LE@Xk^DvR2*2@V#1S#_?IQ9|JQnI}2f*qc{@zy^jt$oY9-Ioq
zYloTj8<`I0a_zgWNj+<0!)z^dtCRdjEuejC>FGz6Bze^(%_hoM_hxx+t{7#UlZ^VB
zheXR>!Uj%r(xAR@iL*FnJk`$*gZ-%=Ol>vV_9Z<@{8HuYVMWO$&j4n*cVLOfTIf7Q
zc=m{{qbwyj2hzD&W-<nL<0qi)Sf<<ZEx4uTJW2U_)TnX>PhN(iwzZn%?2St|@*T-N
zht{+$9y@2-5fe;C(<mP*`zSsA>Wpr5J{}9EX!jSaaWv#Z9z8pnw{LJEwzz(cFTXX9
ze|QMDw2Y0hE=V4ytu}j*3dWfmdC8(m=8%=QElED{yE)GSx*r5-v8=ZmwbG9w*>Zup
z7Oqwu)0K@mEI<HPRj%C)H%YeAU^vJRDILu;mr!iVzP@YwsV9Cn5=YXbg#6W9vE0CY
z(Z&T!Zgauyz^s_BsnKbzlBDIA&>HLX8{rZ>LBK4(b$SZnte>F&04BF|c^-SzKuam?
zI*<OfWYK*M$?gLZn}af)rxb!gp0r<Dh;EfooUq6k?eFhQ^6&<8&q3>1vuG=lNHc;u
z8pH7(R=^x*_k;S^V)nMWZHhhc-75^F1JmBQUlv5~r8Ue^n9xe29fos67ru<~EB6!h
zF*MCu*1!+7TQs?B=NYXlt$xN>K)0~NF&{QDisk-!c4czgxB|AVbPJoP*UIw2$T()f
zuIWjpe19#|FLE5Q7$>z_g^22V=RK+k^y{lcEkBY}9*ZVTC9dx@WqN-q*7rC$S~s-z
zV<&Dp1K3wdr!a7vPv$F>iTMmT<E3^!8@tjrg}<IiBqRg12^gh1v81mch8}o_%S~!@
z8f-#BppY}$*1CIp0t%7X^c^!#(Dcs?T#d`Ev|w?x!pEPbX<yvfTi>KM6Pvg;f*qG>
zKDiY%v2?1-mJa3^oCYThxvIv<y{Fr3WxyV_>7EdXA1z1(_04cPQ!KdS>GiKj@RVT4
zhkyq_E3LE0cj|g(l>m(a$2HK|!{*F`8BxwF6HJhhy|LQ2>>rXqIRd$)HMy%32<Hqx
zgk%HLG|P*IXB&n`emm6@Hs=KP<ec$NdxLOD!5*Y{tYn2LbVozsI|S7)+FAZvJAdBH
zPk%+H>RQ9z>gM3K1Dq^fcBtu^N4<+m(P7h~ZPH8o>xM(i9?P1=6EJr<2;}uWRC;ID
zvrAjES2Y-@N9J$EcNO505?Pu+L;7O4&1NAC`&aJ_sr(Ifejm3HTio8u`?UjrdjZ<H
zeP49<EKeEa{ZBQ$Gb)bhHe{=lAYOSR9QUss_|5SbPVnxdso83l66(4P&d^_oIX2~h
z1nvGTcg1@9NMW5AAG$%}ywk*Qrrb-fTHRSgDzO&~UxG<&AHe3i>2A#FG^Vb6HSu5J
zwa0_Cy%O(Flxp4_)EJu$N({0(5OAdSVd!hjym6!XjK3icgFA@F9+mA{pT&J&;OE7S
zZ@{-!Hkz-4rn)o1r(8xqcCI%A6MqTo$4t_EPw*9<v!uu3zYELu-4=24CO&u@y0Z2o
zr{S8V9%#ECSsJ{FMqPG3RJ?<#NTh+@v94e!Ngn|IKmorp`qzK2X*(q&ZaMz|BDpK8
zV+@@*>4RQvT6aAfQ@xBdlMkA|6pG7}jC<2AH2uJ`4tn5KXjsO?9-Zr*K3y5=Q>kq_
z7sP&31JrgtwWAA(6k~(w>r~)jEL%=+Iv+|DG;hXhMW&_|W6Lje(nw2|IRNIAJ0Qk*
z?^RLUj`c8&isYOD+}4LTsc|(wF9UBu+|%zhq`tE-T*f%%Wd{XM^{hLKoE#@V?NwtO
zlCI__XzplpX*3*(yRzEqZ3LK&)b`Fe{A%}x{w8Za6!8t7qiK4}J(R074xJ-)IZ{2l
zA7fn2(-bWj=m*Sdq-a|w(z#_e^g1G~sq_}H`%nB=@ou3my?^3syT-X<7_>;xsLpl;
z0}7+litD@^`&nvMN~y2t*7xy@194>LCm!WaLHsdXm+ZZECDtzV-4@!>d}niQDrx@!
z+5^qDy;3spGB*Lp`>o%neAi*|yY_h2yd~iq4;wa@Z==aMKic|5n{SHE$-=rGfm^zQ
zcAogH;W;IvBicAh$oGE(d|U9Z#1g5c>Pol7b7gh=xrfnDU&6eL_RWMW{u7Y!{$0ZK
z?Uk>aMblf#vd1e$G-q<igsUm_1atXUP2&Fm9(aqz_S$Zr;whSW^ptqxl5`?s9D|4T
zBLfGxHQ1|jE?F4Vo1AqyR$&-52)=0a-A*KId9ejP<8U=zX}7xocoj}L3G>Ey*mGR+
z(OAJjJx1AE$`PdU+ixQ{>P3Ag`x@9t-ZX7W>h3#TCM~wJv797<pOq>-vG=-m?_P7@
z9~t;l!af<kvetBit;wn&4xV?E&=omHjljybdgnPDR`2Zb@k>wmZ{tly^6$e*u3BB#
z&2JO|zj+j&dMQBwkClMjanqU{c@$G-Rolcuq!fPrnfmqd2gB)q;7eF-bSF2uWD&el
z?h0mrf?UR<uWYFu>&Sd3;7vbXj(gkvVChy?u^A<6n2UlsWMJh;^ggx8{1y1S;$IqA
z9bZ|oxwY5i(*$<<g!3DCe$9aKmFf@72;5FvaRgUC;Xm39JYBEN9I!2w#E&9esz>ET
zGZH}gPeMoFO;@$6MjqzJ#^Z2Qu{F8*bxU5~f%Gn+@iRu!d>0mlqTIHve9G3%s!47l
z*%3T}+?;o>Hk#ni8C~H>!5AN>y>TM^+g`~I#z3swL-pJ5TlUwP1mNyrfN_!jf1P_&
zqZr2Psq%H>DzaSKTSIEwQCl@b=8){z^1|RHSe@4L&&++`6NFCW9;Y3DPHPr@QSD-s
zDy}9_8!Nb}2ZB8fQMA`@nrDg=B!&=WEfD1b@>@NC>z>uR)Z}#eU6HwW;mt-nN4Pd}
z3s}Zk**1Vb8uRUPZL=3&xx1hluW9gKgnV7`9@5)VmijGILY)}NEcaP4;l5;Gc?5Oz
zHOG8T@SlXDOQ?K1;)@BjOJLr0io1Do5!Yi7L2Q2rIIR<o+}fHbRD)7*YCYNF^D@nK
zBfd7r#xhQS#<6c?5y({zK{@2{Q);t6?^HLJBXkSU{y*f_4EM#33Xs5alg&eRJvc*~
zRA|L~7D31)8on;P(zznJn`r}YlK^wZeQQR}eZ&!*bfV`&g^q&M@_6T`)~=w&?m!vo
zj&oUd$0s?)MQYo=@J2_j4N_)KAmy+RuUys^tKw}BMY)R4Sk|p{XS#|(3*GM8#0+z^
z91nBFbm<40;ZaMu;05*kYlNS{z76=?`mU*~%Xg*d9vVXfYj$n}+|Eu~B?Mri@_kKn
zSE%D3cXQCo@XEb!-Y3esqtpKYXFrEes_PNzSAJ2^?a-9EhGmu|w^O;?x{|w``rv>u
zip2Qs@nL*t;|tAkGP)Lqzhu%a4<MZ5f7G6SmGcgh;*EFVjs58GuBm@Fhc0(Q(F3#X
z5tjb|mWU41^&g#jhLx@9-YC{Ii%mmqn?)DeO`L7G7wx<JE6mokl0LFjmJTwg^)@sw
zJ5JN1Nfn;vLS>z#AQK~UXs(8BD&3enqR3os1fFs4??rm*F^aM9^Q$_NbH3->dc~&o
ze@|NBHC-#tO{z1^dUm50^24tk>y*2+C9pWb>tCOye#()=OJK=`+IU<Xj+LI4r#T?!
zrh8YQojn~_a2W^hdVhsxU1@4E0331;OjUD^#}h0ka3gUeH4s-^F(c;Yx*MH8V+Fy-
zq3ci!C0u-?<@Kw$tY+HoGhmaDdI4P~kr|S92Lvbrt4C%8fbHk50IjVu&u&Ovq@JVL
zRLPdMFRCC2K*1*j<F9(@G^7i(k)5Nby<*$Nz0?nWYicv(#2<Gap2M|f*obVjcJXJ6
z?KHxFbtI$AKkSZM^cCZu@KHb7L;e-6mNX3_Q53eZ2x5s3238zof3l*!i1<Yc9}?Z&
z$<`ZZVY{_{KYreRKe+Lq#_du$B`*{cNO7L!&-4Pl3K~VrvCC0&NksYUP}J_VjY8u3
z^5O0-=5r*H$a2ibsjYtx_;K|~1hX+nJ$tvX6`7-HlIoYSK^hk<K*0TL*3#EV@b;R?
zsLt@a!r>x>4cQCO1CISq^EG$2tk&1q<~4i!O)@mHw?VN+HUr4_tb2>aw6(T-OJf^O
z2{JesBscJnPTyL%x<XvrnGhL1T=~Rx&t5&NMI&pATPtZTqG;{ah<J$xOr72M<EI$n
zv?F7i^L0Co?9Cw_fa0c$PSi*av$B)VEOINaw$s{pWq1T5%R<FoMOIu9oC0!x4|>nM
z60Dnd>>vz`cc)+=u<=KVbSDjKsoX`5n2?(6w2#_1$GSp=Yvs8?+IYiaxLKu?=jT#B
zl~u{f-2VWTA=y38Y16-KFB4di_KPcdzyuIe0E5?xh<?ys57qBY&Z*!_B)V=|;>|=%
zPp=psrFe|KTV$Qp1o8P%?+fzhf@n?mA;K*BKHK6?!d+wrzR<6&-}}hl;a09LJ`CQj
z=x%T9*#7`~wS2s8%CBlse9R7USI*Z{Cq*Ag4+MNP*2@cD6PO9;X8Cblu8;o!1oZJ=
zkMue9ZwPBTUZrT6P^~@-9A^NGjPqY0TWTo^{`E^c4hZ0ndiwtW_I~*Ft$0)ReDSY`
z^lSMnCR=fMlR)3S3!F=X{5b?ul-pVgG`U{oKiP-&U%Au%Gu}_*_OjA6pA0FwyZc4T
z09YTF9Q5u<udg-#0NY#P)rIf&27@)usE`wf(&7mT1A+U-``=tw!(RcuB8^AJde)tK
zTH+mNOP&T2Htp^P;hgo(59MA}2B|iWtC?XWJ+zF2JxXU7HOnUCbf~VU+&>upAYS}w
zj=`?C+34C(A7#@-yMmMVrE~mFM}FO_#e8$(qoLlzEJ?Wv02k90&S}3C+DYS9kogxV
zBP01&3E~ZUYnf(AoU_K|c9ZYLV3TiCbktUY__xNlI@GVdw!}{WS1%T?C6foXnd5Kv
zPAI*#W(@vfoQ&=Cs<5`kQn7>F*8Hg%b2qS>j-_{hFPkxUY<=eMRO8eoi#cKxo@=5{
z5KCwyZz(o^5|9l`s7a=4C?}2;ha+<Y`^)Q(QB5R?xyxNxfWlaq#(Lv5e?iwTEY@3j
zgS<0>fxr6w>sH^v*B2APBvVB_laQ}6z6@aZAo^FScst=tJ}(WaJVFf?&;1tFSmW&y
z590R&sr&_Y(xl}VHgeReqb^gB(61X%v;E;yEGdoTa#uJ%k3&^2@};-7amN+Y__N`b
zsqn`7M7jcdVuifa`=cw!KKUc^t_NAvcJZ7NI&`dNmqTi9=4P*}n28abk@{9Fx~3!C
zyfO9Wn|(g!B=xMx4p{nDDB8r0e=|5c&OxuWe`0@#_g*QE`$4~UmF!~@#K-%Y#w+4`
za){2-41lBo?rZE{_$U7Wh2%;6L9aRy6DnGxW2x$EUWYwfTT|@d*i|I+L6)JB_@>6>
zt!K(}CRpFaF*wa2Bi9td#BuLH?hd<wH02vj0DQEP$+T4<WUSx{ic$g1LeiXOnJlEC
zH0&b1+`)QvHJK#kdBsa`#g8>oTCRezR(2;m^{U22PNtUa*d7IB%JOBm9CO;aq~6C=
zQYXE)TeRGu6I4I5L<F;O>?-BG!iX?ilhT$$y=q)iR}OaFjPWwW2=}MMDUh7>s+yE+
zhk?hnEv)Ml1(@<_Mn;t)+R+z-kWMN&6v%KnCWUV?Ln+8LZr%a76J0A)j?bx!;~PN_
zfc3-N+CTEInTBu)=jPy7)V@3C!5ZNI0H;U(Rr1@800b^r4}Yyak7EPyzf#thqw1bB
z*Q_)>F%8D0sLqhWor^KXPoN!h*b4ij;dj8_1pFiT@u2u)MYy=W@rCz=wI$Z)d3P?>
zMG>~&;0#IHJvr}QBl`n*e%D|9qGP+CVUJ6(EQnV>JGS4{0y(IB6Y-+sz+NN$q5diA
zEEmO4v+0^~l<sjW%;7QE1PYvb3gfRQvgvQc^r!E>RoMBqdneOxZ*=Q`$#-EqmXf24
ze9^gF{$HMJXTqAiI&P^HvL@iC83Q%B@z29|GHc7@J82<y)}V&(P=yZAw=Rm(I*yoj
z<c`(B*lH5Oq2#fj2cc2HKF7U#$vMRyeBZP9kFdNM;%i|9HxPt^G5{liUqpOA__w5J
z)~d2y#~fsQ#GH(G&3<j!c+&a>{mq2VRB$9b3f#5vca1e{M`(*`r{6gItIwWa?Oo*T
zb59#(%LIOvYoD}+i=x@jJ)Cg5pbhqoK!{*L-0nXy`Y7hS&%r+&Ej)1tihNDt`{{>*
z_s6AMq;}z>W|7YKc=-hg8*q6FJ!|5(@P4`CJLH2@)g-*Qa5DlqZG*XE{35-dU9*o&
z{gM1~EWdXZmZzr(-z07^xh@C4IH;?ZVBIH5albXL-~0}}n$+f{QoTv@N$7iTkG>83
zJ@|DB_=ZSy>n$4nlIF_xYnf!TZZZgrWU&LCy(`50Mf*i)9vPN@_(iS%0JU|8W;1Hq
z(g=_cc=GMR$rv~Vh{iY-?H>#zwfN`z6<M^f$#*u1aAjEcpEF?2KyRV_E9Z+33H&+n
z-SyXpV-IP3xDj8>_?ih$MhHK61D}<v<|*)=7NuFw8B%qWU6P!eNpxE3*RSGzCL13t
zJvR(IR4>0OO8&Mv?-Xl(Eb(21&b5AS^$k5$WoZj}h`jE~fx-dX1A*4Ocf=aaqUoh=
z7IA#hJKfE2z~zrcLC?$3?mXmIv-ktzCae23c*^g?{v<<xr?YI4TLw9yCjvz3IUT<W
z=)Y(mf_@kAmw~Q+BYZVjn$w5B)1WyKkK=V6L0{rp`kGiuIAozp3Nh(qk}#=?oZs&q
z&y)2#Ne_pur-;dgh%+*;48-8O4`YS(Iq6)i_p^&Mjz@VSB>wAm6o4_GuRI#qp7QB*
z;b9zvLnQuhlBWt^7|*Xy!nymfzGzg+ssI2r*r58H^0mu4EwFjsVnOo86b^*?(RC=D
z+EPY7=pWXt%7(*ARUAm?oN-jH00|Uh0Fjb9)iq^eN`Md>r=UL6+es#bm$!0~V6I1}
zdbJ{z2LKb&n9zjV^PFcS0a=4rH@qtX+zHi=>04;|k5Sh+9`*Eh!<4wO@fU=2=&q-W
zPLkB!$jl0ElsU@o^dtdZ8}P3|yVm?=X{OpDt<A)tBMxv03!ERw*U_FA(rqogE28T6
z=3!x^+)FIB(#ih4g8>vfhCtlsJbwzB@^P4tWZYBLroR0V>CmF8LhWpQH8gU*ljTbq
zhh4$qje+MJVE4vz^sg8A!SLSSO`g|T5&fbUB516O9sXfnqz~X2Em=B;#krHk<y4ky
z=8hQcLuCjkS;+KYeo%XWD=*?N#df^#-uvQjVvgG4Cq!||90;3^K?ikwmLCg<!`4pw
z>U&K_O5M&wOYnnSwf*hY(s@xkOeR5{$`?64gN_FkkD(>3*ODc)A|lJ?<nl;8&#h(n
zZ{plqZkwpvT%?FDfwnTojBVO}==#-P33!g$*JM&OhG`qjX=Lk<>0eHjTThvKA3upy
z(~l#$?w%cypoT9k*%*+7A;%y6dehZmxU<vlH9xc@7B<nevdJu*Myu4G{{T;|VfbQn
zk_ee_SRC+k*PgY~T!`<q%bR!G<;xj}tAT<TVE+I*`0UEPN-^e#b$?6hYle&)y!JQ~
zsHcpy13!m#DK4&>ZzelCsQ%8gkTbO#CoPPSr?q*mtEgY;*Vl2~TFqm5Zskl<$A%*v
zThI~H)}-*|)A)1Y&F#<+1;(A?1=}wq0*X}Q)b#wSj5_4EIu@-4r8UN-dwB6=%0>tb
zSjbx$$iM)5n(o2QtYtNRXsKIW`hFIDkE5wZk0e|dE4BQHY~D$h2;@I#MN&`$i5I!r
z-90{5`F@n>ufEeOt2DB1%BxL`pG@`?lE9YHqHGA$j3FI{dFkuZ6p=vdDywu~nV9z-
zl#bQzrFgk?JZa6TB)(+LJ4yj8Foyu+Ipg_z)!i;nvL(&a^@#{par{ld1K%9|D?U~r
zWz;J58SFa>k&;`OZM5}QGtS|d1KjYV{sO<wrF+^gtbE+Da(kY;`#gA7ORJaFWK$%U
zQfvT>4=qn$#Mjy09`KxRZ<z)gZ(M&5p|2eMp8PC{6c|!%?W0gfPn2>00N1ajZHSE*
zY;AyZCOT)Z{01wJ8&dAb$|}Yly{GN5Fs}5sKqFYk91hw3b>{ljvm1h{r0u|5*Qt1x
z^o5QWAQB12{0vv0-n4Q);4)Mk;kwrpZq}o3sl#iqiNlr4Fi@+3SvPulwzgQXLonRO
z(z>bb)Lh_>zcI<_SenakW?(?}9qHRsKZ)4${{R!nVuNBHPD^9Bubcib>m?#fB`0yr
zj-!rG1M;t;d{4C!X(VEy5V0-T0bd>X+gm!;w{ddMm6AdNlhId!{cGC5!tB<kxersA
znE7_|AnI|z@69<iM~2(Wk0=jQp0zYKGT+<?Eunj<pM{PkEXR!W9@(lg+X%1;ApZb#
zpVq#DCzT~;CiaoHl1V4iHL;*br%+;e!s8>>v4fQv+qCD5bTwm6ovm%<m+q2C&w8h@
zmg4xT;bg)CkDQw3_vegp>sD^>4cv_*=ePiK%~uFKvfIXTK2g;B{VO!>ph?}_l`f_7
zOe-M~7Ti~{KBlf|hsu!4{438)=Ck7)vIkSp_N%e3*ohfjec{;Fh|RLZQ#aa-j5jGc
zZ*fzLgFCw*VvIl@xUACPCIQJkN$u<^<PYRo0>3LU$g12EyR$lN31hUKr<aC_epLrQ
ztyj9eYnk3-l34NVDy_-c3{o!%aDmhgfb=8ts_4a=jFIU{U8AB&McJ28y^Ch8+s1M`
z)mvz}1&AK?sT`>SWA2XLwJs)iMvbMLNT+cKxaxSV-6nYLTVx+Jrx|0{70Jn}<czjE
zXZ!_f=@%jZr0}2Z9)B9ol+iP|oUD+E0}|bMs}{aqnu~ER*(W0?y@2{wMxUxJ#1ewM
zFWsh1Gfu)rJw`KGxSGRPVogGMZf=rRWR_gyhRz3|sXTdo4yB|)XJ)R}kT`kfD}wS7
z{JP?_d_Unkoq3i<4Goy!&VKWIEopo~((SbW01m~c+Lex0m0UzV^8gN6sV|8c;+pnS
zen*nSVCq$(iT#c`)uSe%7DAa5{nf6EPSB+hsWOt7>F8=_(B&sSV952&bhydte22G+
z(HxU3?nBKhu^-+Bw~zuC9J2sw@X%x=<<R<rNj8%$iH*To5cc%;tBPpnoTEJumh5sf
z(yv;wwZfi32dU{)qEXk8p2xLWfEJV-g&D4FR<J!BvPu<h&hx{vDYa;sPa{7nh6CH$
zx?2I6VZ$>YP1(ozR~_N|ksbg$Kp4$-8bJ+mn8N}5L#24ReJ314XTP3TRj`R2{)Bl^
zj+N?O6B|^weEL_JcyR=hFE}2)_3C;Oh;9B}KgPblq4{LposR2Ixc~v2`kLtUyUa)i
zS8SLWbHKsi*9C8Puwytix2NB>K<ikz+=VM@cDMR%#jfJ%ki$EZm8xi2SP_DI)r}9u
zYig@+c&6stupfCop1(@oI(C_>fhCRgzT?l8AP7PC6@yMYm2YXPXFRfHOM#H5IU}W8
z)9jx5bWjP;R_$78;TM67mlrXlKwz+8Npr1R>CrSAb{<prr|dqJ7ruxTIbG&IU%!sl
zIP};K0$|Pc>?@hQVTO}9-RaVwG{hlnmS*IW+}33Cvx21Kn%OQ>8OrR;xR9ALdYbdE
z5n8mqD7%Oz=D4z%RCdT+z<=Yl=<>(Cw-v^E)^)a#ZzqVMyP7i61Z0i#fUJ1u@ZzpA
zvnHR$u6!N)SNJmG_SZ!5<)ktxw%c!V88L|Slwv-)U&g%O!v6pewQm95z5b0V&0%=B
zp2~IK6pP$4{(O!_edqC0L(^Ux%_~)ne#3neM9AX-f$jD0^{;?DY2aJ`0Ee2?_DJUc
z07>4)-sO5Sd-@U3*F+oF%=$>Q;+IwV5%{wAPw_PJ-q_jwhBo{5`}b@<$GZA=uRqmv
zlO$)$!H|CsuhO?8yYd{oUm*jgDjUR(=!Zt~q{l0s3HCqIvzocnN^TEBh&h4q<G(}w
zYNg$>^6B?aBo3qh0Igju^^e*ZtfOm=kB)yD=dL1aIf}UFu<S>zdG#S}Pi~D=)rwHf
z<s3Io$GvEw4YZC&tx=jKCvfZNDl~V1_4#x7R&kdqvpU>w6LumOA6jMP#m>a4kX?sN
z^!BR|e8(q~?_Cy01ijtH+v(<606cdm(xkoH9C%PRAo)l2tIzwJRa}ko9!T`8+xd4U
zHuV@esFsKz?#yLTC7r|f-ajg@ZZeX6D*m3LMZAP^Fc=S|SGSAzj-I4fJ+@~}&V#}l
z?Z<@tL8WVUj8fxIx{l=t>y==CmTT^>6zVn_-^YCy&S|93e0iapcaBL*#Ej{Yu}=y<
z?@smbWRUE)J+sAqd#1{D&x2nGyhwz7uSB|r-bqR2P)+5xJ%%>cp9?u=_GinUE72}|
zHR3%}MYvM4*j&qM+5Z4xii7$CUZwj&d0LD3KgE!u#=8FihwbH;AH_T)$&md?QBM7?
zG-&j{ju)3vF35FDcyER}ZY#C_09E~KH^wjhulqE3BJ6VdKCgSGG3$hUNTd1CpDVe1
zyz#N)$mF{m)n}8-Nn56VT9Vu}(yVAQ?!i=gl5hoeUM2W#d*JCUpx5TNv=<Vzd#$IA
z;G+-kDz-)s`Qn_Onhh&6gm*jh*FCG#{tWzI*1imQ($B=&5?S0|#|*aKRqe2zCyrIx
z;Hj@TzO!o^*)5x7jSG}4jk!)cf;x)pyb<AR4-ikN3*BzpLl(AD+{V{R&F9E~6@?1)
zX6#P_v5ng<Yq|Ajj6ZGNXT=xVMWohNHhO*KjjOG-q!?RSbNj%fj4N&6at9T|Y1bFf
z>-vkpCSfE-^<DmAyCe(>bIuL{9jn6cudcO_l3htCOB-c^BO-uSH6I|yey5)G@4vJE
z0PPv!{{RJiA@QbzuJ}7i*5$hhQs(aE;JTezfbs*VB!$M|jAOlY(|4Vvvpo9rBMLH9
zqp$g&M6RZ~=9yuvTgm37#41_Qc+YZi)13a5yQE#26>{Tp_RlrPcyGnN0r-#bb{j8-
zdVTJZ;@dexUt9g9rC|sdC8TTv8E$}J;8&@5W8oFGzlF5TOG(#nyg{tazv-3|LN-L>
zB<*3)a7IpZjw{upCr$}HQSjKz3g@3L>!FdTyCiH|C`rLN&OgtsW9eG`=ZDi$`z)jz
z`)9MCe0zgL+>!0|uEWW*@f0l;znN!zfbX`G?{UGv{*}pJ>1H{aJP{}d<!{&9@vho)
zjE|bD2~&FL%|YW0M??5~s%c59-dq0wX~`Il-Za2<QHDFX+0X)`uNCH+ewLc9r4lTS
zZ7h-jy|Ppj#&A79O6~R7-(R_f@9%!je7NFio;-{UD&#Ld*~dL<Nqh)2+X*go?OktY
zc??PRD5L;^j;`1pPfS-$;<gCM^4jl1jPZwrFMKtr8!bWrCs0{rU%Knrm((9x;k6rg
z4oM(Ba0nmI{A<;GSE{zD_FH*>dbW^=x3+#}@7B0YLcv^@P!10mtY34ci&xy`<TAwD
zN*p&C#b{ba@y40zNC)w&V#u%ol(EiN6!8=Jqxid%o+@@b*Suf0lho(>)!S>Xa!Q2*
z1e3*OI|!6_Bd|Sb^LTSoyzuS*vm_ILW!NE-=5<Ce!Uy+=sdKTcC_0Ya&Qs%$#0T*A
zgZ{;)ODf%Ja+_%{R1&a9kC+4f<{W>!n&y5Y`13{6JQJn(cf;3yXN2^qk>-{YgMV_%
zk_V~JUqCCb@u$a63u>A^nc&|D*~Y#R3mn?L=A&x9eycgbV0UtgqdQ8FdRGVFUxD{O
zAN*N&bOLxw;pUk<miEn^*g#8pae{LsnB|!C9ZwbHRh*|QuBY2zaPY%XQJS+_`kpVQ
zz5f7?yi)_pv%GBKL^i<fjxc`eDC`I4U7o)s*M>FeFKpS?Q6a-fe(nY~=hq!awS5QT
z&x#)h{t0N0p?E`0hg-fy*(~ArLSG#Dvvp(s%5m>sDERN<rT+koZ^YP+*$|M43IvKV
z*dmUD>&<hi!kV;Q+3KiDz2!Azdm4OOym+Ezg&Z;il6fAW{yCz&OGnhM8p%tz1d@8u
zUV^W6bHJ8aNjGGE<X!SkJvvql@chSv&syEMICI{!E>%D$pG?=qT>BJN*sB>vQO|MM
z)j44zGn`{RMOrPK@$50)s$EOv3><Sx<rvp6<?sUaJRUi!O=#(!{mp9U%$opY^~E5!
zEbG8N%i6b0f~}2<O(qDy;2zbj9kSco@OqL#r`p<;0e}mE*BvU^wIVzYgkqCoWL9Hf
z3;@Z;9`&I!selZQnBuE!X!*bx#UYmn5RCFUBi@y+#C4wx@gA#jZb!{^Z!9?j9OELt
z6aN5f9Rl+~_~qhQF42{~(%39I;E-$fbHjIyr)pN$A+XW9xd8fg{WD$%{{RI8{gEtu
zf8o3T02z3G+)1oRi+LGdckLNu{{Vq6KT%!u8gX$jlCx4jJU$e76HLC*A=PC4{n8NX
zrH&P{Kp5}PR}12s$MIi{<&8YYMVwnj8jqMDIN*Adll3*}zXI4#rfPaE<9QAo$r*k&
ze55zNc<Eks;_nOUe-HdYac)a6ak(yI7z_!>Aod{sqNO`c#horcyHZ3NcBt^(&Xe41
zibh~v<J?F9;1%>F4A9z)g5_Nprv&{yz3VC%TIm?O#}lqgNZ-VI<J{(@kT7M-5;l(A
z{XPD*(3y%w^K#y610hKm<Th{t{*@i}r{>&9LM&Y4Y7#)l2O0GvKaEA@G?5t<N~(|x
z5s)+QS?*4922VcNr*N?|-x0tk@v5?4gVl~R$fRZ&W1Q!nYP9SC$K#ROlVcV20(b!W
z=AxGVM~ETD(bQBdGZ{G`5&<5*wHsq<V<eBgnpa{UB*llODXQeEobyV{fKLDs^8R$h
zQh6Yd2<uBwCYt2P%gH!UIj=?dSFcN^TWXfgAy~BwN$nesyO%ipz^^utrT|qThI9QZ
zS3%Ss)>T0GQgRrchrLjao4{TsZwmOsN!3&wTF$QP>b)3#MAqkwEd_^*bvV_xK=QLj
z{h1r)R{V&r2Km|x()b^Hw?qE`)~=@7<@N145}~D%**vgul29HL`;Ok#Ep<ahc(n(F
z1l{)zN3o}C7lLcyC~S#9E%mB*5tyVQ(1YJRR*r|^tw+Rn0t-8LlhuI86-i%Y5@{`2
zoMmEhy^ekBLOnM7>{48tX+^r@GVCF9Uc2D`0E1B2yjL24H!i1?0}J%Wt$pk9{{X_i
z2=IsOcVd=4AF`I~TDDb7f-xjbjF$v=2aHy(lj?GNc;3-JCGH)_&RdSW@lt4dZSJ?K
z!D(is%PR%s@F|+zv)WA25XA1_diSp1!@eB4(QG5Kwgr4ati`$Azvarz^<(JHKM=oK
zDN*K?-o~{ma>q@?w6?SRZOx22ky}_edy=OUH+4hMcg0chZ^cXbe#@oUTs^&=nuy~N
zxm|!BgC6{o-m^S&uUcv9x-ONEvfF&GH+kE+HN!>8Cp~LwwK-QJGp7!3eV)6n{B+hl
zN#KY)Kd3+2{JF%xYmol*pd2AR`8myb?E1S+!5KcLvjdUEC^g4+&V``oY=UYW)5*Za
zGdQXks->791ePPEew=>Gz7k&s{2b9OnN@DCRj#MUUCe)=#eQ;lFH1U}y8$bhp<yGQ
z)%u0-wqG~GI%p#vUB1#iD^#?($yw`Tt9BcRrYr5oI27qI)}wE_4J#Out7e-ce99@R
zbBb#m%CxL9?(wvWhJKt<XSF2F)`U?3KPl#-ocx&0If>0OUew4b%sSOs1zMc_>ZI+*
ztuQFxqjR)Y6~q?~#BImDbh1AcN_ht-rYk0mRIFv%*%--bPfl@DdC}ytI0CER-Lj)&
zKZRmTc+iY7!99(0VjEGiI<V<KdPcq7<QG3613tBrdk>X>4+FJBYa+C}yVLb+kgLa~
zal-o@(rKc`pLhx~?!ymy-LRczhdIj*4RBCgl$RioYoO7t_C85F8~N9K8(W@rMRMnk
z)c*j&53YaLtN#Eh`E0q2=Dwiu-P(V`4Xq50ZC@Xie6us2O>Xr&@Jroq%=J&%zs2{y
z1itZRwWq)ORB33aIu&J<$+-J=t)JT?!&-i$@ap5@H-qB|q4=u#C)Mp@!wXCGQm(8A
z-3(4gsjo5kbFRan>l%a4^IA!6B$_P!<wnGxj>lvF00FC-PsJOLf*v&2{3+slQQ|KF
zc!D*K?#)hCcpI4KjPwox7$dl$QRkAH-5!+MP)0q6iO-9?B?2><U|13<5qV@_0KxR>
zSax>T@Ji{Y%_L=ifpLO9hl-cSzX&x?0WtA+!~IFTPvE<fUMtHlo?B%+`ByQHQGmeC
zFl(FBJ}_@8*ZOjK>BjgUW_^0`UC?gY98&l9J8KE`hnXYU9FB4Fi!knfgc|mL1AHR!
z2Dbv->X%YznszrcO3M)={l(|=BD`lx{iLMPq1&eDu_TzlasgrVJ*%GaSI3_h_`5dm
zcIpqzVU4f=_alHlwRW7gIo&wk=e_(^_^shjg&GCMjp0^`^H`Bcd!hp<Ki(&~teHH0
zJRh-kqov=09qrBftVFlXKIweT><?DtkEM98f_@y_c$VcgeR}L#X|bKgSvbx*h}88y
z=ZdxQAI3+<x|W^c?G6(blV$Qp5IH393zOWR@u$e5k)ndppKSiYUME{$+2_PsJ>-TX
zs>i7tRp<U$MYxgagNpDUjB+imr{T{YX~)U_-mpn7?SgIFc8UuV>E;X{sIOA}mZrB&
zQ&G~RlR9sOd^>Xt(gT&2?qz0~htuU>z}FczkUlN`%qe%L#;L7%aBaAW4<y71Lyz{5
z{{Z8(<E2SUy{<%En`VBbEq*zCK3kX`DRrww^I(%0Va5tJdF14K5Dj@x!hekV*TKIM
zU1|3MNwkZ=&|hiNoP&{qK<T)CpjSEY^Wq)ffILSP{j7)Vk{z<4956i(>&UK)<5rPr
zscJU*<R2}q_zY!1ov)6E>Dc!bPgk~vtzB81hl(MymtMMS2O`=@cR{zJgUI|waa45;
z3FW_rMZ~eL11cYwW4=3zWU#fB%otykoOA2n6`^v78Dx!uNKfAf1ONvjwRYWwjD0{Y
z#lT*nJ-9x#MgC{tjyS86Yf)WEB3#BxhAJ*%>VBiWR{#<Um5^tKU+}DFYYNvP*^}3H
zJt{*9mE%PNEJE|h=Z<O0f#y{QC5RkWt;MWMZ=p>JsA-wim~p&`0`pvxl7*A%Z3sCg
z`yW7l$-V*b=ZrolUg`SGx{O-ZnG)UEM`J3aS4I`y(MrwG^uZlQLG`0u)~+V_Dz?dK
z=0@84NQLFtLpnQy<)Olr=PKOsn)Hv@H^B?9{CDwwp(@HF)AZRO4+!2?X<N(OGl98N
zJl5y!Ti{O%cpu^XT1JJZUBh^8qP4LM+nRaR2x+7slFmaBxbQl4uaU&#DMuX&61=4*
zt>tZhtDe3Zg+`~#rIGVzfVEk5?~EQAQp#dWm=RURLj~)fty1{Cs6T?dQ{xR9;yA7(
zOY4g$02o*moIc@>yc`PezhoZ<ORIcu(Qe|jhgY}KNVvL=W417W6fVad7m`2P7_Pha
zp!k2Mc-P~Osd?dO?Qd-F;Y*vCB$Fux(QUGYVc38UPp1{@OOmEJ!8@fDr_SV5tvXbq
zxpjAU{49K>Z=+nvV%KQ|hfPWRkF>DLEQDksJ&rl|uDjuWn{TT4<4Co+g|A?F@L(Ty
zL5_s_XBn>N#h(B*uL*c}QoQ?hpZ0yMv8=Y%(-tBV_hlpaxjhaDu1`bornhNzXt7)Q
zS2LouxDsuAF~X8Mu6q3|zOV0S${gIcO)n$Lp!sUXYCPPo_V51y4tu4PwmMzLrD+Td
zin6qJ+awr_2I#<yk&Zp9ThDSl#T!~bn;miQT>hQmZC^yzn?%$t<C|27M#t=OdFdc^
z$`xVPJ?rTohu^aAiu_Gu4c+Cp_9SEFD}TckJ;IJi{{R~JEJX^gT#<v4(_3$*{{S<)
z+tN+R*=qj#?dW{>toSb9#hMMLhxA5GDsy-*BPs`%CBb2zHz57y+s|W)<TbB=-X^!Y
z)uWF~ops$BDIZYLgsUXOXwS+NZDJD-f%la0$TjzloBKRze+xC;Kg7DezXX;RQml^z
zQRK+{gi^R9jl>)deqU-!e-wN;@NdNJbH&<^j+gCvrOMk{>9L_`)??+Qhh{fO`DGa=
z3LgMhJS9^E)i#`4ce7Wz>#5YS(y19~%&pU-=zeWQZ8UbW?hhkl%Lt?pgdWGC2le)-
zrqd;kz+KAfma;fiyMz*+Z~XFO*@vP0GhXxiO?()Cg})PZ>n$qYVWM8|)a+jb#}dfc
z@`o7tTt6O+E6#jBakkKz3M}Xc+H56^HLc2ysUr2r?8D|$(AT>f&81d`z1Q`#?f8`k
z3X%8v?c{LrT79zVBQ3cU8D9q&3);F*0BJ9*YfW<Ff0Keg-eZOQt8(Mv&6d01{cpv-
z8oU<v_RAVw=z5HRTv&$53amJpTc}~rt$MG(KLAYcrO5C<ohu8u5mTykOmWXYHa@l0
zN~~s`O}bmCR8hKwc1N=ODlzDKR4|yxmE0Lrj1!LEPipl0eR3;7C!3jJ9G$%z@TxkG
zhb5FdJ{_ZQ8RbwM<mI;!)OvbXC#hLo!z_j=K2<ou9Yu1_mZYu`r({d6>)+Z}4)Hh(
zy8#!l^sJ4BQO<eflg0&OYY#H3nQd4&I8r-)Ra;Heu3K^jd~_iGwJsWH$(pj*(3lhF
zN;xOxA4<T~V=OV(13B(%s908Ob#1I8&H?o`opWxitsqUroD6rYT&<!qr5DX}=Wp5z
z!~?^A51t@=yPMPGrVc*8g?yVFA)|R)1~7K1Jx{M*_3PiXzl-AW2aJxN6sXqLN6QP4
zL2>uGd)Jm)>DQXEipNc~lI}(!BxseN=5GFgbm!W>n+FePNjYxA6(r}%x*eCpIA?`!
zuc5ejRmok-N@hNU<a2}5gVY+@_`<VIrC52A+7*d;kH|vDBP8w$2p+`N&8~?qjc2BK
zbsKHM;qFfxcW`mXPf|Xe>zMetcC7aH@TpPdw&o|O^sizIL(7)ukcgt5M_uvTk?O<`
z%bKUCNU*}T(Xey;ZPKM|8<iw{tI5qpC4|>8O?QA-t`A&PPDvFAx3Qd55ssMYO<{!}
z%A7ePn!jSdW3!U&fHKMqP^tGHGk#T_wI#G?LzQukO=vrngn^TuYJ@`#fuBs(@HY&5
zR?A}-p;uu{1K*0#xIiO?eFaszjK#F|1oKxejy$uD*{fFQP40CX)}D}fQ^R)v;igHy
z^eV(hRN;67s~XhtCZHci@a3@S6`60NGaBbi^>#@X<k_GuL}xk9In6b19!JX_y}c_p
z%p{(1{VDNE0SCXmNL<^5ndHDY+B3Vbrpu{3k^}^VUry$;j-atbVvFvdyxq+uru&%^
zIMlN-zyqnJ%3aR?0K?bi;!)-to}#@M!hRG>))!Z~TV(l!WZ=KK^cBH)a>my{gk0Yc
z(-FjDu6>11;xCD^h)ue;8-LUx;eXyg(y@}hhR~I_JzrY!&Ayipkz?hJg91W6Mfwi)
z=Fw{{t!spCGP|iJNe)Y6`B$FH;#-^PKX}B0(E-OF(ziSTDU-y7-;g3;J-Snb5@#+l
zZP1^g>JY)?TdE$tYBwMgnzfvuKGj*r%z9Sr_>{=^Rfs^5Kwubp8iQ{&W)Z-k8!+8A
z%F0LaAHtswbDa0-)K!VC&zj#aImSIGj@`o!IqZEanzVV9k>7@>p)0djNS6{WIKlPn
zUG9iWTq`y){Katkb(E2+F;W5PitT(c69i^{ojI<2H5VD&>Br(!>2jWj;rpK@mpET@
z(!D24)l@(Np1!?Jc@>S$2>w;nY5L1VfpP)teJk5X$5f5o&qPg4k%kUE{VQuv)jm)&
zUUdejBmfrqyZhFqmZyw-{?x2tW7472cMRu>xR=0SmdWfZgM(SdIp?)sgH<vwT84;~
z&fYysaUV8)vsUB0B(4rCjknam!Q^CceQQDuPdtvkf~LgY=UDRQJm-_gH2{x3;QLig
zI(C>~ka_D{5kt90U!bg>gWRAk^4yL+O%JDu0Lb_JMQg=tw1JO)dexy~0^|+{H6DPK
zozEQj=c3x`dTcRW?}V`39m<TD-1N`2e3h*DR(&s3n%7Yojy$oDHdTlBv)p^v-G38w
zw7SsLI_^;BH~#>1j910pESBR=)UGt^#5XG&G;Nd44+M{5F~xdRHI32QsykZejCi6S
z3f{})>ChQ&2M=(AY7bC7YlFPdl(d2ZK=f})>92H2thM{Q2*w#CKQZVRJQMg<jM_LB
z>Zqk6U%e^gaqso3apvfHFrOu?&o9(;yPH_|GLnP%haguystZeq6>%gp0rLZfQhmFM
z^(|z>X>90tiB9G{*!C6X+KY`w?Vrm~3E@R?RHc0nQiWY^a&MSVKX;%by*JEP9*j+F
z>Uu-OrIdz`)3~mA!A_hHdh+I_4H@rJnv%R(F7G7ed!F45U9{8{@E_$qgQaE`E!=%7
zR8$~t+^42$qb^>hO(|bPd8P9?&Ojc$Dy8fBppp4iXbVC|I3AU1I7_!Z4Qq8Ia#=;_
zRMR79?_G%bNU81ZK)udBUQJtyA1x0Y^s83Z;9>jXvr(|z`jB4@kO5y)d<xZ8^Y(SV
zf<;+Im#8z8VUSgdw@<>pU(^O780%j|{>kdvuYo*uq>q@mibYfEL0TuV&r#k5U)ws|
zCa2)NL><BH9uUX44Y+^kRl?~`H4oXN#n5to#jjgjueRvQlU|GaVOukJ7r_?{4)|u4
z20n*i<MgjD_-%48J`em((@r<}7S~EY-Es-6WcnTWNc&Dyc<asw86f9^J5<+K0^uWn
zG1^5NSh>zlc*Rnaxy>wB1~hcQA%-~}Iqg`s*}PJYjK8GnOQq_Tw`K@##4;?NMZ)fl
z@N>0?bv?y>HSp*5ozbs;A$Vi^U*X?{63S@XPr3UppXCLPW-85&gn)QIuNCs98PCcI
zJxBwkUWQrSpp>y;77-X8nB)r9omoRfa#o`X(P~Gre$!VT8Pa|x>zYr6G&{Wx!ro7_
zY`MAf#DE-zUOrVI@CI?~URus#wSqI0IP(~E<+&fFZTv{_8hE!<wz~s2ep1IF^$Gw`
zeZZ_(gU@sUT$LaZ*QXVueDfxf_H>f$dgs9pi1*(KrMlL%tGBVgo--Eo!8=Ig0C>r-
zOZb2BQ^!~SFI{Ti#8;LYl(Tt_c9N6kkM59uLUG4TdRNZ&j`z1q7>qh0!t=rP&3a{y
zzBCOvHkj992Fc*BJOX?5>&7dlt$s*y#Lk8WHEKGOTc1~Wdr{T(EopTN-A-HE>xmT4
zI~Cr<vFHE;&~aa0X@3j+JE7??YQ88BJl6`OS`b31KBL$4ug>3q9yhdaiW)4M#q4uw
zaZZmUGY#9i?Z-|x51_?+XYG0MwtLSN`92@lp}ny&B#71wSr<QfOXmjz2LN&SSITA?
zjy8oQ8VgqU>{VQ+2Z_a~HL6F_`lpFL9YoXFMW^XjXc)$P*pw51FbC<}SI^%PE#%R)
z85Y}YmQw*D#T4f#NhB*Sarpj~@!qfTS5}!{I*FPuFb7D=jmI4G`e(g$x<BnFcKSS8
zex+w`d!8dg%%w&!ql5F0)Ykb{FEWef#cLPX^D{h?4=?w2uDTwrKDi@6rs^nHs;k2L
zmAJ=VtbNB7U+rml7ia+g0Ck85@)h%bz43=p)wO8^5~4^()*Jv2K!Y4{?ZEf1OZXw-
z3-2GzCFRVrMQIc5G8Q>@#!lZ~T=uW9rHiRjR!c+YF&RZHH9u)mR%fCeJ?0^Tfw*)%
zO*i`*UpsczC(|82#Mh(P=^AZ}A#Ks63-X8Lv*Xj(>(Z{<Xc~sS85TNY%$dR1xyRJ>
z{A<6M@uJ?*S2=%&e+;DYj<F+N27^hmQ6<x_4B7jr537GH*U_E`@copUea+sZ7`5>A
zq8L`%6SX8<V5T}1Jq~l4%kbxhWzu{trRcFQlc?%){ibvSlqZf)`zk+L*YVD%v8A_$
zw7{(d03+l0pZqH;bGvsv`V>@RoR806+h5^Fg!~%zo+R*vn`rauGtK7d5M`i-CB`OP
z^0_AkdU{vRZSix&$HH-Vd2eKoPK@Bk56C4Rjz``Du<7ky`}=HomhHYGUT8}&SiFcM
zS5vz&<0Ns@=FUfcE6OF(JRjlz0EZg2j-zcV_|743JTv*Nu(Q4-wq)t@p$t!4@m@wI
zFpXYVy$`R#^ZO*_spYd9i<S)|lkJ>i=}(JKk&a5@KIOO{txf%;$+==k<Hy})T=Cnj
zSMx4)8G$U$dy`nTPCpK;=)C2*(%Udt5HBhR1PUp1+cMUO7m&PU(OoF>-LJX(m3Nhp
zn#s9*5}^G4pVqc7Mq07uL*}y_WbyQ`muI+Doz^7BU@Inhe8O3>039nz^Kt2qYRz={
zdLEgf0b6kZ<0{#}tJWlA<{%!%sueGTpUbeTwvdyKJJq5MYj}<`f%L0!+h#C(9GZeJ
zkghSDWYxbj%tZqu0|0wd0;R-%A<t5KR0b3CXRdl0(VinAjdPrV{OM(YWBukd*5tCZ
z3PBB=eJjyCC3O#ow0XQgse>iEj6}!yPp7xFaTf6a#B}to@8RW#n?B}HLEw*3THlva
zIeD}{Gyec=Z`q?z_#NZPHT^!!$D(R-?n&WKkz~$xk3xHj`H#jr-R`Gw_P35Cw^rsS
zT%6;RpTvND>-q)p-h<)YU&0r*o+`7pztMcG+(mkU)POzv*T9}G{gi$Pd_ldwZ5B;K
z!%~J}7Tb?95>HGc4xf#6Ki+%RB~RJ=oe#|0*gv$UmgZBs<(KBh26_R|W1#&j72`1@
zZUmoR)%T98`wQycAA6~7^vE^+Jg3PlAf4X(MhG?Go;UrU=Yr;Bw2oU#h{+#jz69<6
z0D)saoi2N5&AD!_j|7fDw(g#P%-CR<IB#Q)GgO-dl1_8aHQMVx4|Sa#h}87ilp>Xm
zYiPqp_#ZCfJ$s(LE0mVakVeHBLuE+XN#m_-oy^SkQdA?8v=e|2e)RWP;kqwraQ;|Q
z7DipHPTXT0^~n`6(ytihE#v`?er4p=g|~z+?%|ZjXDN|;GT;s~Oh+;HDPBcTgq9nC
z1a$uZB8!+~RT~>1s3#+x-j&*1crQ?ER)wr5w{&g!@`wwZf(ssda6YtK_$F;)1dHsg
z1b_l=9$pSlQoXq2kH(7%=-^QUNwtVk-1QXk8DY*8X9VZ<uVK=^XFXR_weusmN0ZYb
zaqHKU`BF>p15mJw{WA8>#1cVgn|lyF0IF2v%>Ar8o<Q&vl%_}4tXb*v+?DddJ5B)u
zfnM?8?|}Mts+YG>UCXL9j9z@vh1yE=Bm>tK*!Xis(=DzR+S^Et2tO&2128_s^sS*o
zn{x@`V;vFZ-weJFYo8IXV;6Q!_8>s>qy<N&0s2?kp9ek%cw^zgjjwduePRbvv)aiP
zcY;5?*yp`@o|SWXXL6Xiniq_Q+Y@dW^#1_$SFBig#^%>Zj!}CX$f&1g;q#Jl#(I&|
z`&U|}7^BXqUZkDkX!wKSe}c6nmUw(A6ijo0C=3UuY@B+YwdCGA@FuaRN4vw8cGj_>
z`<B5`kG3!~UA4!=+2J^p85<;Ea1Ybh`Bnz6@oQ>H9QH0y9&v&P)SmUz2qn<soa%LF
zj@$V6#6BFn@^zara;kH-;Wo%Vgb(qrK=B{N%dZ{yBEv)%vQ4L1tc!01zj_ybN*}~=
zQE6~qd{Nb8vA5E`*#Ro8CC~2G1Me?I80QDCVMm014O^caYW^9x`%L<}u#U;&xDmm;
z5ZH2gApOz!*J{8{vgV4=^J-?aFpb8m)oWwHd^h2gxzvWKsM;IVjm&m-L}Trejxc}N
zUaWnw-l_aU@mKsJcK3Gi+<mLW8i_A`b~=E6g?$us1KPgi@ek~Iuj#%QlV7;i>@D?M
zkRq61D7RK%5hED^Q=WUCJ!|tX#Wt7Ic((S+bM}kXxOfOdkXcx+2jVM(5~GHuqX&O8
z*2}RgF{tILOZrA5yGtP~a!+$p#eq>e809M9g1lh$6ySOGuM>51>8wsqG==d|hsRnA
zXQd%ne(Frp6M<K>JvvQSP=*v!^EW5hb*oyt3A-z^(|#3bY}QciLT0^~DvWcq8vPpZ
z29^Dn;JrN<67r1y0KjYIKiR|K<eJu}9p99G&u@sJ1aY|aKY-1BpvsU(91&y5fdpzt
zPCC~_w2hBByUOgM{ECf^6pESnH4_SsgBcuwdi9_-Wdw{+P-iqM<?~W9bZnga)OQWf
z6(WtIpp@mQ4UuM=&Bkfj`Hew4<067cCVsU-XTjpEOxP7hOq|v9BxH}pQJ<XEvNEEp
z%(y)&R^=JD3+-5Xg^_tnuQ})GTlXvg<kmd1e8dyM^sMI#yPC#!Rx@qUk%-MD`ZItB
zCayC8K2QZSZsIdpDrp;gX7n&5kXgx@IcmLOd`pJU9+g*Fw;^-JM_SLex$<1EIUOn5
z%&XlS{xZ7X;LThU(%Jt2m3-YSzWpoP{7tEs@Xo0bKYHDVA4>4qMQF7>EEmB2MbmF2
zx7Do-QTfi&Z1TV7pkcRX(1J(hP<XRZv%k?TJT<P$57^5-gW?g)azMoX;!;0^9DsVM
z2AK?o-trkgy1?hAI`jJ1mmNm(&8{q1tk#jaQ*S8Tx%M58xUPzNG<8&Omd8`!KN{=b
z4)qCP({JUp4mQY&`Hn?B{RlmYtNO0DY2rzq;`2+elHx=2Nx$suPpQDcuQq1M`54C-
z<Rct`QYGB+50>SK>ww%>prn^WjO|y$mbU`~?3OBdBwjP0`{Y*EohFCjs8U7^sLZ^9
zGWk${-#uEq>O1SXoB?vlf!yMT@;4JK&pYGmR+`*LW8!a!+J}goK^)A`^a`U0qtoe0
z;hjnJS&7r)Y4v$HjsEb6o;Ly9`d2q&6~)^Gh7lU%08ew<KEF!!9|ks;@bcgLPVH9f
zRy@WS1`!`$bM4SpGJVQTp5Nj9dN0}E;9j{5vV-EE5X4~#yXQ|Ze-Z1Pa%vCQ$HhKJ
zjx`?-T;fYTcTZ@{PBNxg@>h>i2^I4vio9!k<4s-euBCYvB_UObAY}FDx$bMb_)X$T
z{2k(b16qxMdpnhL{?O=G@f9&zwkvzB7&_LE6}O7KPo^M44XyOirUweUjyU$M{{RPS
z(Q2ABz8==gHI_tcHgn{k;c@Py{#A|R3yl0n@jc{fJiD3N<N?zTas_O7Ys2XWhbO<i
ze;`N?m$>!zrzU9?rl%+3T{`nX)Xtr#F58P&19G30)bwRN!4*eQzYiMWN{|XJJuA~M
zwDj@shE~n?jekpwn^{T!0FQ`AB{Sb0tI1~=C1eCL@ImSgPcyJu#YIi5ItBuyRphr$
z_@<Cc9m9^mcdH2zY!Akw{mJwv(}DX;L=WNzJu6zmBXFJ>h>}5_lqNHb%m+^OLees`
z1{fJUX1a?lA3D#5tnD*+-g4n%KQSg{AP$_I9C9m*5CHu^0>Ac0Wf->!No}X*eTDlK
zc#7*q@lS~+B+YR%ksJ>u3xO<woS*g>1tj)$><O=T@pr%v2mEB#?r%=9twDbb=@r$D
zwaTj5D*phkfDho#a8G~6zC8FmHQOH;-&|;u%WrF8u1og03=FHW%Z3ccki-p`Y=gVr
zzQXv;;i*0u>iPz&rRq|p=CI~V+hHgT<pf1UJ=iJZC)U3!;tU=q-o@Z5`#P~(^!eQS
zoVqT(cwu?@T}>18*IzHo(DVNQ6nrNi34dqb9k+z;-dQg1mMgoH3G<-S1MQJ;yo>|$
z5$<c{PY!s8P`A_}WV;Nyt9gwoL{Vcv04*P1lmG|^*WSNEeji;;qIjn7NS$J3zk~OZ
zO>w!m78p{Vn|Q(d!oGO;gQj?Y#a|tKSEpJC1a|jRwv%gYpJ%suQWRH-j#LFLxZzLm
zw^Bt<3upBzQIoCjOL<GWFEqg=TBIA&+|hkHE}ohmJ>m}^Pj}#ZjVI5Q{@peZ20(8k
zjry_1Q0E5~(|9Mtdf&yL8r$AOr={k#e<Iu6i+1vzE>$oJ0l)z{Vfc!z@Vmfv8t=pl
zO*dX!X*9*Sb&ZiqhLjvLeBUrR7{har0k2<MpAY;Q_@m*g?+wo-{k^Q7bi&bENw+pZ
z`9}em939;hWLN9->cf^5m9^DhW8f-9!cLt#-1@)ZXYBQF@b$&th;1w{^+XFDozNTP
zg#gJ-$J}sgAKJgbJ~r`Bg%*DaYS-d8)uXu5>|JAdQGwkcazP|6aDa@Qax2a?9})Pc
z#5!K3<9i);YpL|xf^Co~XiP=1xfOba$<BG=yI&Z1kHmi#eh=yTj)APnr0aImq~&aw
zs?Q{3hKt<@=N`NsE6m5^<CewZTf0kSzb*d$hB$05x2YOeNjGc0kIeXE;(v|2ZLRz&
z@a%V*p0j5^fHgKvLMOH{%!<gy*5$*Uz$qSA%t&U>IOO-=2)-qFcjEs5hhTkcLypT^
z^HHvv(PNH9n3V$QRfz^RbwzMK>#lkl_x}J2>sMbBJR_;tc#6|Yxz}L^*HyW_v?5E(
zO^ip#KJg9|5J6GYV!tx<e-&!~01tj3*y`Fu;%oSP+i%~*Ue4xF<s`&>W;Y6}^<3A_
zRjEUWFcOssd&_Q?f8}<4MjH=bF2&0aQ>SjeYrjoTz5X=(X1TENK8LPqUIo3^G#goN
z7hY)w;|zc5(4%w>xCZIB9A`Y&%RV{qRgRl&YvNA^>PJ}cozoOIPEVU0t<RAvb&f?D
z;ktv?vV0@^Mf^q5JY##Qom%qh>7RYhlYJoj5nb3S)`zTEY+!uLfYtOgUJ}zh3#M6o
zZt!J-UU*ktQ3r_p3P`z0E?j`VUI7Y;CLbq|I&f<|7FqQeLy_86_qwTi`I%R!>CM?%
zE$R23GvSX8KaG5Mq4-wXPcGIc`%a-Yb`Z~xH5o@ajX+V`xUaar8u%{PNs{{F7GV^x
zAK8_@P?5%)20xkiHRXQ=J{D=VTCCnH(pKe8xhaxJbvTdOZJ%=sg^`#H;jk2s;rdt6
zw${sLr6WQNX|*`d9jdC-S}5pkO*eMPXH~bil>+&MWQFx#>s)@LWDyHScf~KvTNT+)
z9MD^DW)gA00~q{&I>VX<Q6Z8&ySj!yc>32Ir^~6=1aP-j!K5GnLvhVi8Z<MrDh1DW
z0|Zw|cWeZh8QuNTI#p?zZ7tck-Hv#y-DhJ-K{K7zZPvoXNx)d}pegNMeXaPfL({CS
z^?w=aZ=!fvl(bfs<(BhQBeY09Vgbj@4m#(pdNtwGGBpixoirOaPqeWFnF|~wY6mBC
zaB<Leubn<DYucy9?P$xdLj&mc5<2O6Vnh}TNKXM&0g`YAN$P9S!P8QF)lNymPU@QV
zXP0W<w8VN<%#rvH;r{@JbO>_bPkUgu_iR|7oV=YI9jh`g+Lz)4T61Z0UTM<Wx)ph1
z(<6da1oX-Q{{ZT(={^kTx`bPywvrZB$Trh$QSY^@7gG2#B8T@nlg2ZF_NZ~k)q7Xd
zLUdJ<I9(TevzD{*#kIB3nR7kd&Ud?=+3&lG=R9ex+iTM_=_+v^)KK3yzZJ2rd@$5)
z6ud1Ain%HyI2rHP82oFA)jTJwS%5AqCXi>0AtV0!72O)KdmOW*)~6{Vl$7-Y80Mfd
zN?|d1jpgU=_N__nUK6p>NLZh`M?aCO^4h8DIrUM3KBAPp<0W;mma!{Js_nH2!yiH3
zrN6xsOiBzO`6FNMsPFjH)7ZM?J^;sGQB@w_VGEp}TBK&vxY$=Gz7G|8-v&^G;GP9B
zVi86O`ihPw#lkneQ4ezFqZVQ+M?G^-+9g<l!iHU-p2nYd^KHjD1XXER{fa+Oaf+zE
zgrrM%aMQAG<zrBT@T$+i&uV;_P%+aL43Z7F?^+Zz$%EFMu3V4~DhDBXClm=e0Yy6u
zwG?JYR7EW8Z~(5Z$?d!|60-qwsYp1Ce+_d`-hXDU<;U6>V0YV0Yk~<1#iU{L$a?46
zlXeriuX(LqTugBxn)Y6-4-pSR?^g8v0^T(e>KA@UJecAeoc{p4dJ5<C<R>z_>{0#R
zrnVmD-VKiDEC<Rnn$L2RHs!n0VToY0)74>a8sbBi`sCHW1IOifmPm=oSlEnr2N<qG
zOZhJ3RSV^>y>_1uEbiyl(&kv4Mv)lV$oWag#Zo!(Pxlk_Jv~qE$I_r~b5^8Xgn`g;
zRbRiYS@EfYE=^svk2%1}tk%0rbTUZm9Bb9w;19;MJRhxTy7NmNo$MD<Du>w{WPwWS
zU}{Y{C9oStGP4oD?nO@xvZ>B7(~jr-D`smfWUnLnR0#*SQ`6Rju4T^dTD4rNf;w~q
zrF1$`@{?x^n#HoSnnc{Oxd8IYxfQi#4#V>#na}rlt5!r*Zc9^U)xK=^_N(hZL!1%V
zRTx+Id%$kE$Wz5wg-Jbe-_wujSImw%BWlj+i0hwv+OwU9zH6VgjriK3iS#13Y;ENy
zk;v!`QjyOs&3k!C$IZoP+D^#a9uI1x3^-A|@zXV=kVlm<l1@EnxtldD=O-NV-n6V3
z<PJ|c;<GK{QaR3km8)z5k%3Z0j;BqZVy-YxOyat08FvrE@UAw(Q@KFn*A>-R$`wKW
zb(E27Lw?#I;2yoXtv0ce6b4w?0qMA7HKC!~Tj;uaYB#ttI_{1({t?^pKc!vOwK<-~
zCw5)Ve8C!!KvUR?=H_=gITtH!>To_WvUxl|atd+sw<FNk#h(_mxKqXi<e7E@GWHl8
zemLu2dw7;uFMK6;ENO`#fXLCu_f8LQ@UPDwiu$DbZJpk!Ff8|yx<Fg_vG>0&mFi&{
zvrNLI9+r2x<hteT4XEn3WC+!z+RO*u0ALT!v(3b}x=@*+@*;jBRy>eDkMrPHUEn_f
zOXIIGCO%e|X&JiIF2MOxA<jSEW56Eg(!BHHXNnNt&uwO+B_AM-KI~{XRUbiJl6oDq
z)0LIAIA0U$Ydq})M3hda)YdJflRGn&$OAQI-qhVrqq`iO`&GMuAck_kDBuBHPgB}X
z$we7fZ*3LWYq$-}ApFtq+q?0qcYYhP)MR#mJStBa=L6oZ$85zp8P7F-($u1Z$*lQQ
zv}XMLn#G)){sOvCTHaKfjO{!TS@ZaFQnPKcT1XFE<3Fu>g``B0p69i6I(DD-t9DX%
z0CHnIhO8bM4`W#4XCH|0gpNp@tcXJn#9-Ep4i$dxee2jYKL=W0mNP~PBd_s(eXEVs
zz8=YMC|fy!$jA~Nze7^!xa@1~)lGIeUE7<bPLl;(5Jhc6@b|=1Od^9zdzWrhLN{iv
z+I%kYMa;qzqs;>zGa%XqeREm3wYjpp*yJze4Hz8pUqt@Hm$A2qJXK)BbIWBM!AZtK
zz7O-P%`^5*@dlx$iLRuE-bNoHBN9*Ic&i@(ejoUs;g7}WpHb7~io)jB=FUmf<w#^Z
zPv!<IIYLcnjU6`P(>l-EzT-^s7s5rm((fR?hQnG&-R<9UqcSimeGe7$-@&O3)&Bs9
zG~h`~-6mya^(*rV_1_JAB-dx~AH;tO+(#IDxgpf734DUfAqPH(X{#Rzzh(ab6!=TU
zDXHJRq`G;vnl^Y`<Px~3R4P<@BiPHRO->xD==myG$dT0l013rD-r6|<EhfSa0NPG{
zYv~w2WOjRrrN7q#qq${UXuU~2PqlNu7W@gLd@AsW)9mM-=JQaT64NsM)z3HswsF>!
zu(c_tDd=Y?J^fD|n@{@*WpTaQC6NlSAObVaJ-by~akk_J&H<_BAL!6Iz&K%yV?6%=
z_39MdB`h+;`gQy(n$|h*a^$fViDY(Cx%=wFKZvc1REFv9u5E;*#y1hPw;1J#9Xb3f
z8r>e|bd3}NAqoKX?V9SmF|O<047Aj(>>yT6Rug9|zDRZ`Pa#iYd;S$_PCE`Ui?W-t
z7edpd)nsTbq!3|Dg`DG%2{}EweJGDvH`Wu{%o*4aM1<rjj>P(NSXMff<%Z^g+Ak_l
z@3+(p44*;GOBDWhnBi9e_h5_+b?wu=EAwu2rjm?d^E$l~#W(s7lX9%{+bzIZA&?&_
zg69jk9ODdl9WzDOj5QYB9mJ~2=X$0=#!h+n>F<Cmr`7d+Hv7S=Gy*$D+OU){mfJ1?
zC5QlLw;AKDd9q%{brGCgTMxR9g+U&e9Dqsf&2d3{d-KNDI<YX09Yw+2v|pa@&6EPe
z2ZOX@nyNKvEsXO<%r0DcQ9pEE>^%t_el<-cZSt`>10(}Y+P^O+8#`B}7+F~!B%i%=
ztoUc5jb~Bv!h^_N%rX}~p7r_-@YllHe}jGqU29hn381xNtr~IWF#Z$u<2CT7!7mqo
z!Wpkx>X62bJ*C1<zH06pl<|?=X1<W{_s46EM$g0=HmM*Qez_sIS106#Qb{-+fzB{H
z*RzL@x_owN&pN$r*!5qB`rJMvy@KOWxY-MjI_1a9mFf|Hx&Q${71j9H$5$R8vW~+`
zoauUARy4A-K3mJ|#d#wghvIAIJtM_3=(bZ?tcPZQcOW@9^gfmC-v@r(r&&d-$tIb9
zcd8>QX}8=<_F=~O*!|aI$G8=>T^5HO2->2)i0=FsuY5p)`e{DR4cnOh$8K6Z(gC!M
zi0jngdSleqo$jRm6tvYYHBU0jR=4>jg5(1v1g>&$NF<I%E75!<@h8I`74R+hh)IGi
zb5FNuBGO3NHMu+m*xZuGlEaJ&_}AhG#G8wwacvx}HluYb&v!T8!?DhN-=H7e3s}&p
zCfjPr^)XzO?HE1(00eL!8vYx$v#ckGrMQX>d=n(r*Q`99KG-NR8Av?N#O?{y3}>48
z!uvskQHg%ps$5MfyEKyd8-_{ciR;f^c{TN)hkOKYEaZJu`TltsSq|bD;1)UR*Z^m>
zbAK8?W{ZCcYCaFqd@8qp@R4h?`H$wvN+NQ19y%xllk97k6^G_r{Z8DsGEtkL(rImT
z=Em@EhgKzXu1-z}z_iIH)2(AAq;RQJZWlOWFi7{V<KicRH6I7~t5)#+`lbG(rLk!x
zvs^^c5(>ztjH%oYJo8-Mlc!0nYB)q92u@_rPhe|`l_dweZ2AmFE7~N~6jJSde-lpL
z)W-@cDGC4_QClM40|OtG72nE_1=PE_`!jpjHJ3W|AbxeHbMpgKrc;65zDW9mBIZ7E
z+NsG7Sb`2SirJPe&7W)^YO2=w%K&-c4?*id8HNx?>7RPD2g;;@fzqV%Gr$<<9Mggg
zpaGuL47rWOeLK|g02pJbs18Vs=NQI$si7gCX&BEZHE5Fl%rYOnf&8hj9~kz<T(?w_
z@c#fRjB`tQZxXVI&gST9$z^MJ_cu-jdcVWFAdxHM99NoH+mh-r#%tO<8)hKW(XrZ%
z*ps%WpL}$OQt?H_kB0Q{{#>KXorm2cH~`V$&w$z&gmm#W;-u3mmU$)Uz0G=g)1#YF
zYqvnGycQ#<uO<DU{Ao6kqQhq!6kD=K%5ZQiz6r}ja?a8Bna}vI;$EMp-L;;JiIoXM
zE<RC?xc>ke;&1NsPZ+}S>l%H%;~eeYgLodyD~XQkN##wa1h?Jw9nE@w!ruiQe_mOn
z0NTLWVU7r`B@~vXF`Tb-c{jwb*{@8GKuu3ec`Q;;ac?<0PeL$#>*P-rcrMFHzM2c0
z1)fBHu;32;$Gv`oe$)OsO*-AQn<i8?;2{H!PeJ)t#NI0r%JP&vg#?`B6X{yhrkW;o
zrOji>rhPsq@|M+;89yvQo<5_T{duis({)RICLLMy39YpG2m<DKaK3<odUW)~T$<!t
z-7JAr%NYa$1`j<jdey%b`1e@xTmlQ3@TxY1Pn8q_jAIzi2VbwXYwVH-T1FL~wDHF%
zxs6&m0NT08>FM9xRLgY$G5wktV`9aL^6)c`IOr=D=n3Rp9pP7Q^ccy`f1O8eS(q}&
zbJL!LdK%s|;*M$6lXf+4HA~H`(M-sMACfbH@Ay|OtXebP?lNUq^#_q#k=|{!mLnQ%
zEC_6J2<`MWk#+m{n3muSjxn58wn)PEJs0+K)U32$h?;X<%o-b(HkUB;{mg*I`RsUL
zbK1I1Bf;_LJ~Z(Loj($3_b&+F&qm;n!ylb^_lEC3vGwaak(B|*EI7g7*U&yLzFVIf
z>RP;Vl{eP#O*+XFAz3`N-_(D29%>S5ML&e&uccJiCpiSHgb<@1*%j4k@vf-|CPW1g
z<>Uu$&<@e=eJhx^f*G!i(qBfBef$tkYiCTeQEFn3QBtM4D>oIh)WtG{`iRSk6<L&H
z*YU1@Q)ps{kXXB8a)X?GYjaxDtYT)8<4{Pk&O$NC{HiYxFNG!6U|Z?0wRl5r`z3^X
z9zr?Wy|_?&oMx&KZKR2*D8dd>JCB22CD(M#DprR{x1Yn1$>&_%TquPWPeSMKHa_a|
zeXH8^UmbXFOws4@ey5^_k_MB=x|SfVB1g=QeR_Zi?_39qeja$w#oBI_ajbu0_#ffj
zv|{4k!rtsiQ3oW2US2{ol6lQ|pU3|Ii#FF52g2SQV`XI`M+OAj<B9?EA`{bbIW3P|
z8v1zGP@&5=xAkN5ELL9(em*n9tzPe<Uy)n(zWt;8Me)PH7COF#d3R%=Sl%_hp{K|4
zg^--VBQWIdP;fgDUk|jb^0TuN<m8fhCccvRAl6sDA-(X1t)J}A5kl5dNPg^79$}OY
znGP|ZLTko6Vc{!}hkiD>(lvWQES6Uvyos`+Wev#4<O75B&3sNNx{sS=e!YODPYn-a
z4xM|PvNl%)^PUAfRh>=XxzsK$Z7qJqXDqDH!MwkgxyW3eP6h^Qu;iNI=z5?=X?jw1
z?Z?)yX<9Ybt#KSS@jR}_0U}b&XkJ3Sim=4@5(JHM#PP*@kHhbVb68tjdC(WtwIZnv
zum^l-Pu?T8de+avzkwQ@m+%GCP5!XQFzLuQEPwRUR_HO;9V_hr0EJ%xZ9X19_O7ou
z)x1Y2%Sd^O-#@~hr{D6e;?x!1=OsxwzEocLTi`VC4~8vmwt;rZ^RWIBe)r>DhRjAs
ztyrAoj{fytLs87+?#x=Yg;yUrrWMJlAMaz<ngGEx(1s8IIi{vM(nvrR51+Mp!1bmp
zmKpe}@*kR`HyP?ieF>=AQ@GMj;mkT&HuF^(w~W)Ka&exWz3K*#bK4bm7UO3>FIuQ+
z^UI9$RUUj(5a%j7oK|(^(%^+TKD67Z&h!0iHhFMPKN?96<8Cek1B1z|$#O_M;-xN2
z8jfSZHBBL*6^1EE8ONnid%rEmpfu*Y&ej8w`qph~Uc)btI}k?)u4|pt7s!(%is&w6
z5|PhZ&(v*%kz{f}sgf=-vC8YSlcC?|l45Mvlt}rnQt>2aJsSHX1IlXhc6^*vtq)@e
z{nkS{HRj$Ww`(nS=O1=a=B|6@y&hA#xepy#JKIBgqdrr4Y<&UgTS4k}y<&2);zjwo
z`cuWa2N+xv+#FS5HbW7?#Yq&PfypNw>$y9ejTaUO2+W(v^vzjIh~xQSgALfVXCs*a
zC+_-If=__E01xVGLQ7HJy}XL@G4l}aUA;5vDJH+T)MRFr%a-AedsSk=iFP5M`5*A7
z2N1+I@>{MeHc};kEJ$Mw-vp7`yBP0RPqng$ee?64dK1#P*xe#{m5y21aD6J>&Y;&9
zvPcgd00*$5F2x;|prZH1MLzOP1bey3<E?c%rR+1TaP+mdPr45T*Ap$tT~5wbY&o_m
zl6q6_wI4JP302PGM{(&>J<5)ytIcuZxPVz^Ip90iZ1ShK;q6(ro*CCXVXR3Wouyss
zce0juj?NO?kGyg1SG+x^E~&0*@J{Otb4rOUkNhfF20!o>$*;L|p8?O|zZW!5bERpq
z%>rBK`b>9frR?JvR|J5p^0o=X6V45IIi^0QUx$@ibll{Yi(k<9Fc=DWxv15jBJO-X
z{{RTQYXTRW!dk)=T<-fL9zPD0**p&}kD*U_su{Ga^qMH`^xLaRR$HZP?{sW!0WdNT
zt$u-B=@+_1oyG0dh154I6Dpua#BGjA1F0UI<39Bco$yD*-ZIu=j%Y6Q%j*!02C(xO
z(jw#LXq4r@x(a|n!Syxqm|iHySDmZjTvwMh{{R%eXTOT!{2dxnoG~j-_P5~fetKEk
zl|bi=jmw^Q4^VwQ>$&)GVu|97E$*gEaRtP3{Lz;z17)xWsK^7pYutZq?}oky@PCT5
ztA7pL!utKam|b=3;#gpkZq6cKl_fmnFBtc)J@7w==AXy5HWFL3>?DJ7#=#pKaz10g
zAP{-4ud9evC^_4f>7OS;n$fg<>-!aaCDC-aJY97bqYsHJ?(ZOyF!6-)loH`2LzBH+
zgPfe=toV;nPl<oD#qEKX6KSG(alA8Rkg<HlkO9bNf0$?fS+8p8zhv}C#;|Ri3tNJR
zd*Ni3Scy{XgQh^uW7_CT@ay9@iuG+u@gGW@&vwlyE##wNnNHq#8Toz3t$c<ilg40W
zO}%BM{{Roj=%G{C;^#i8JAKE?{xSHOd*fdl_?qRlGo!_)B)3-0G4h%yqGK9F&IjHk
z7TunVqP(w4GU+~L^CVH;2&H99yp>1al$9HjuDIN-*C(3#Kf(GJgMKA=*WyK=ffrxX
zZ0GR({D=Elk?o0<#u_-!&Y_B)8;o?td2fq60=iC#;dVFn_d1@TZ#~qyV%$a^+A_gn
zMt}CQG2CLkJT6;H2Srn(8ziiIAKDSbKW6WAl1cQ_{sDuc`MxdJV7`Xy%+?<Zi>^$Q
zJAz&lz}>+tq;)wSm3LH|P`lL{&c<lIQ5<)7fD~(;jC!!f5DbnvIIK-GO1~>_8^G?c
z+;Jiy;qmeS2h2#$a(aF>=-M}dwS%%N%Q?r|(<XPy4Tt!Hb?MXVThwZXlApS^M>n%l
zcZ_}JeKYW)?tNq6Y?@W1rY&JD-C1E`cg~!Z4gT*W9@XM<{ABUZgAz+CEm*;9B@D~t
zCQ|_9XBhzh0G_qpd@u2gUIp-N)X}O)uW{y!Se)%`tW+UAOB`qL6weoUV@LRprfVJ~
z@cb7xa%pmre`B>#ZX{*Nc;s$!IQd6kOxK-{&#)MZk*7`GYe{JAowmE*@;sR3*o-Qb
zN|F8Co{#<k=RXiWFzY(6gEXHN+p}Bvt3%aR(rdeCTXT1BDmRqvA0XmDRfaKz_on<P
z_z9x^(3-UR_2!j#XC|1-cYkGP5L;bBR*Yq$a!G_S&iF<{aKjbP_<F~}_sghwd&E=6
zrCi-8I&Q0d8GC5w55#hSe&Oc_qPJda?_Yqv7t{O!;F~=e)(c%k#`iZ-+AwJ4T=3z5
z!wdic^sY=aRBJpXCCI0JbnDyC{t4Ayv}sUI_o)}Br}-Zm>OZn?fc$Ubs|WDTqaTFU
z-gaxtOO|OD=L8ptU*_I60p-8R#(UL|9DF}r2gTYKfjl2I&90x}*qT`{(&3QYUpNM7
z9YDtL0)Ro<H)9<u>MbVo;it!aV_u)aH@5R@7S|q5phN`>g$~f$SPX-L1cQz$)|22k
zqSt=gb8WRD%uKfk3lpB19k>}i_^!B7#^teeaTDiDX+8dAeP)&}Nvo@L^X@-D@ch=g
zT0PvJb<~Ke^1Bu%jFL(9?@w!gAZ`+p;N++{{OhE<K?309uOwsF{{XL8TC|A{6pR)*
z+`UD4{?)6U)ZnC2xRZHI{Hg~-{P9@Ygo`0yr{xEg6=ve-1tfJHOI9t^*y+Z3CZbbp
z>rhCeb&zw=4%LNte$jIrC<{D}0^o6yc&%IczCt>a#%qxAXNZr5JT<E7=zr28L2YKI
zjnl{oQ=Y|CGY@bpdJuJLC@!Sce$Jbb<G&nyf2!X2l;|2g%D0Ae-Jy~)tS+uFM6!Z8
z!EQ0_T<(o!CZ%O0w~mqIpOLao;y4TKUSn|?lJ_q1vj$}j81CD<f<5!iU$*gntEHBX
zJwXQC6^iCP$lu-@{{V;3ezo>9!ph5&p~FtHrxy<Uo}StT(&8CoQZ|P36}ZQIp4hD2
zURXde+PgA!X`FKBxBLWpde;Ya@uKG5Nn)PL2Tw2Odr8b{PJItkRbu$FqRh+YNiCUi
z!c7+NPu8(`ddpi9s=>X@+lUroQDag&<Z<<_og>5c8f~%EwK($%a|~GFk8F>BO6Roy
z02k?cjEdHmc6VeD#xFMm=treq*E~fow829}B2Yk%HVle;nu||oph>AOcd3Qqy-w3p
zNmgq+d7x~MCTN4l&p<La55S7>PZW4$Y1hU}c>e%#A;KUz+JC$BuUgi11Ob>f=+t3z
z-=16XuQ=B=AlGd8<}rmGf%UIOgx&d8#}#UEgl_jZjZIZ$agGKC2YSk6-#7=asOef)
zvTu>k2h@(0pF4yAfs^fCq)#qdkY3BV!U2FPtSy%mj&qVvH5hC@YS|508?q4+qaK*5
zGr00V0Dxa39je5R9^DUI3Z(Mm3+QM>A@>p;xa8GCCp*2dR;FK<o<Zl{skp+Dcpmkr
z#<YcoC|iLGZak1U`p~xW$f{8ZEHjU4uNogQi{pdHqy_kiRrBRJk@!VDz3NL?+F4a2
zQtaIT>s1z0GchNQfO~pY)|YbjaWg~|JN&IKGQfXyRNSOB1H=GR<qnx{zksfr!~Xye
zH63m>pUN;mPCVKETH5$)@XK8B#-i6&8e27t8S`)Nq*S+y9dhT7z;j)1jJ$E7cr1N_
z&r@F#Y6>>D*%Yn1A3(>GRp>QksUC&!*GbTdTwcm=<{cFLh(E@)q46%6eXFjSX>$uq
zzc-T)Awc&Wcdr=Md_}K#p|;&zHLbbn1U+$GAHq#CJ8u-^M!<+1mvhPKQnyDYE?>Bv
zyB@zRStq?fZPcE5Jw;z=>8lodb5|qd;;xSf_@yqZXRaGNo2h5hn-f7BVL=D6$6OvO
zl<?-eb>V$t4MNHGPYdo)dLLCj#BpChd_(YA@os`v-2-YCq{#v9{Cvsy{{Tw);#*b{
z5J6TP9PmN!TvAO+=g{G&3Y7V5eU0I*O5eqLEtaQig9X%K!v6qqzll${ITf`Ai<Zl6
z0Dl*H@n6}W#Kf+npibE;e%Ame@NhpvUqB`}?_CN`$In)!?5envTSk$hs;<=?1zm}O
zBRrnMsi@CGQ=uSWaZ_@1G}3Hf41N`H92NPne+tQg@Nw7Pt)q}RKZP{NaYmF<B4YtY
zJpdJUElnaACpCj5>nOtTeX21tXXYpA-nEI%I47~x$EhrY4Dfi}o_>|9ZGRXbD~>%n
z55~BGcZ>myRqK1U-T*l#@U2nFDt%7=-%w6+*~M&G+<B@1IIarK+&KBTJ^uj0x-B~O
zV<Ab&&m%wQnyVNzcJ>YP4r`~=Wmc9YbMlrM`e&Nrbo=a#6Z3W+mAhqk@{FEyP}S9m
zy^mq>iTho|O&e~sXw;!!ySB2fZXPDKZJ%gwC@XyjK-Hb$zZ1m`quO6?lGyywwln3Q
zLF{Vew>oCr$1#bs*|%rg=~!~Lp>lVVn(AM%VXNtyos2M)mP>aS&j*e%>0b)|(cTfb
z*L;1T=r(b&gv7DJi~Zt$Zaqoo@UOe{O-{}`vt?=X-&{oN=QkMnPBC9P{7}#&w$~xQ
zi9pq1h})>&$~$mhKBKqSHPwb~Nv3Nj%NWY%f_!7~2A}Yw;e_5E(OnF2GK;&E>w9)b
z{(_v~SH_61e<QqvvoS2rf1P^A?Gxhf_(}X}7l-U0eU6WqnTPzfoc{o<bL-Co>0V_y
zOJXhIIV=9L{ZFNKw0abM*skm?;t|IzgT5me^zTlR!KFk5V4s_jR&AgXG!Y&E1)J+z
zH+KGBNgWVi3d+YvGe<Y}CBfr?$29S6h6e{D@~suOIKdyCNG(ap#~ADBP|<wsrw$N~
z<kse$aM3cl9y?;FNop`K0X?x(X@nLqH{g8@Bv$I?yW9fY@(&#PR(7PsZEw_7zajTB
z<CUnd;4@jV1C{jjs<FIMI)8>**Vr|^K6c>}WmC^%#eG5cG)vTQK1Vxw{K2n|{3AJL
zczQVpA5qA!to#kE{{X^A;d?1h`sIi#6aD4?0P9pVk7Exnb;Ry3pjj-VV7bruR~M;T
zAk=MT$4M0a73dLxK8hR+?Hj)e<+Ym@V6K3Ci~%FkvrgAXqa@^vOT8ESJ6@TUHt)8P
zPi)}iR*s%!WZce7UuTh)5x_q$THEnN3oAo9uMYt^_bLAX)~<bT_aq1pD*;@A>c9PJ
z>86{|`#ieTzjpe@Uy4>aYo!G_QOW2&mGOV=6{<eF@iJ>jj$yJyw?Y2bQ~H|wLr|Sy
zx0-oARE-NaumeBRz8Zr~o8lLa?w^6@+&uC*=o5GNVzI@iE?6Vu@l{nSJkoYY8GUIj
zv`Y-=(a9T-Q1-~Gk$@PnABeAP@z=v^uM$d<=~1iS#H$RJ`2nR0Fs!-YoNz^XmZ{;5
zQ^VIEWz}^1iwUv2WyD1Z2h%-8d37mD6Xj^~@c5c_;H4L@V=~fJiR3YthGoOV@tmF6
z_3MwPdeJdzR@PAJCg6oKvW?G%$vGsCaz8Ot^s5<eHAs;{s8hJ_AHq2&n$*|fX1I+W
zRba{-f(}D^lj>`l(zVXy<Fnk#ixQ&^t+|f&VtNk$0EfL>vx-g410%Dr0YDsc*Zlid
zg_ebNC9IRjZz-}Ws`57go=N^y98Ys^c9X!)moXAVkJMv<oQzfSNk-*HaOPJfWVzCq
z+Ffs0BPG^0Y{poOZUeqYe`==!u?Xx`<ghF~eJa+iuit90HMQfbktY4@fDqgQc>Y}}
zgHMht2%bCG%S*Y8Ns{S<#^KvOf|O&}sm(sRbSj9)mh8OcPCY6%Vo?a-;8fP(&H)%+
zquc)g*G~H%%kq=K^yn+CN2%w|NixQha@Lmt(Y{l`0QDc@E5Ek9`)#P$+scsG6~RMd
z`-VXbIT)`(@Hc~RHOW*Yh6Q1tGt=<zUDRVIqC9L|>92ZB>HHfWp{{DyRvu<yo>hg5
z9s-le{03|2e;n8e0lu<ru|&&n76t+;jjA!*rfZq-SA*iwbsJl$Bo~(nAC}Je3`X2=
zJ7YQgE7480rMHDH?d~iE>&qgDbq&XA0oNdU`gX5EoLnEge6Aw8tt9E_ka${YwGRYp
zmwKwD{FaI$f@z}+$j7D|t#TSKhAr*y9_Pta!Q-6d_ixPNy0-BxrMH6Z^h-N-Ov+}u
znYj|T@T&U`wZGu+3k_pa)3n&!`HsqQE&~=PgY~5@RjrJ<ru4q%@54`q(rP!FOmo5u
zU0iPEauOsvn2)9gYt}qB@Y>hLelqa~iSGyzOj}H}l}|8224*UIZu^-X0X5h7U&9Y!
zZ>U}f!cVYiwrYLH?+6{MSH==q{5AMlt2Cr4!E5C!W3lC%>(}zHE~Q5~S|_uFq3odQ
zPqY63fPQOyRZTY2<JG33+H&ez&ZVis9t2^C?UA;UN$Bjn1JD7UwHL<!01bFg!ANv}
z4cJ@TPjPVXGTSj_NWml)W65S>c<2>Z*0F9SxYRsfcjAepVJk6c3H}4mBit|QYn{}m
zn^L`==Hc1pknSw!ByrR7uEzxjaVoQ@qOYPkpRx$}J$n6UtL8N)EWDoHlvY+cdD;5g
zCp8&B{d-f3W34x-uZuqDjKCX;f=IG@aaY3NW2QROJg|86pk{QCe(bgYKXe+V$U)!|
zD{@#t>Ckk|S7?Hq{{WLxfTb=%DB`5G%L3TO2vPN<XmW9!0(uIb7z(Hzz}?rsYSj^A
zP6A5fJm;-g^8WyMIPKWgdtsGdZnbhT^5@i(Oqmz7xIF8)lO0WaW{Yefv5b3Hl=w>5
z&(z}YiuNcY*#d$wfm+j7cO=c$ZrK_;XusA<dG;0ZzsJ34^=}kBAm%G!<u5(|0R46J
zhN-BniPSHoZ}m}?TKPL(X?4AN;r98i@)8HRHR#gQQ--s?h98FP9&4!A?;PX*0M}nn
zcxuuQ0(chU{&iU0VlSLvla7_a{4($Yz1%N6Lvl9f^e5Bsr~cDEG)*r=hefuNZt<1?
z_NDwuSe5QYwtS`WhsE-1I;F(2u`h8ka1XKKynj%TGi?C5VmJdl^uYD5i^O_})U|hT
zcSji7Gv2uS>w|6?csUsiI)D#vLtcfFR!1fJnL5}8;Z`&ZgzY4*H=k4QoPJe-HdxOi
z%D*VW<3Be&exKJBv2=dZ8gsmJza2e?<<M3e%A#<unA3PHV_q@qj`Y$KV#m8rF<dl}
z+-2Kz3<)C~ayoK9l|!lD8LxiUkjHM>0a7*jRfivs{{UXKbl0_#&ge|rrhk$&V}hjf
zpQlR1xWa6UX7&n3T%LL7Kj*z|v5!!_h{X|~%9k?j`I&xEk&;LBtZ06B#{o|q{{Rha
z>S;Bk7RfEbNd)V>q;hU#!NA-x!Q+AHS<qOk$1I+5;RyLoL2UXDY6`GiVPp&xoQ!~d
zyMC4QH^7_wd!GXMk5;--b9<y**<7yw0G;+TG;E=dAQO%aeE$FniO=s_U;vQadB;!2
zy(i&jlYbwCb<I>xBeXX#+8bT?6S2$Z9rliXrlqTty4>_jy<@=|rmsEy#)&jCyB0Rq
z@`OeDFJOHuHrrLR((jGcg`{xWa(t#c@V}3G^J|NyH}?#mD>IxfbH{FLz0>ThAoxkB
zU1~af@#=EjK(;q!h*pf`$N=CGf%mHIV;#=g=ff5rGtyD6*c*A5ax6=;8TG|<o(=ea
z;XjA^W}GLED_fmf;jSgUxg!u-z%t#l^R@>aYs+;102bfc=n&dKP!*48U@}ynyIJ2I
zJV&Tp+Dk8#(MN9_I%FARhD1<Zy*_5mbkmo!YDbBTp;H#UjqMgt{iOaeH;?t*J}oX&
z{gG)0mk>ugN6o<ibSEd@1Xs@Ycduh}8$*KVtfMLj8OT4O>spd(2^0?=!qc(Jv0=4F
zIU9X*j(w{R?mW3Xox>#hR8_3%JD!#g2@EAQP9(Zb+t^-OUc&xdw(~0jeTXCZSJ&Sb
z{{UyLGvn@qW#St-R?||rwMk9D9A|GQ{h^xp+T98HNY5PmSLhGyudIK<L-3<ZNn{c|
zv3MIfCjjx*yu3u3N38{Cbbfn$ME#vSW$@=v^0gf~+S$L=-b6P?Ao}*}-mvt~g}T0_
zDn}F&GqJ|iLe2RK{fgAIdo5DT+g@77b8f?HM<Rl!{QA}|li*(u_<jEX&@`)ScI&(*
zTl20MR)w`kTa?y^!9NQ>Vf}N(4;{v=hflLOY|q0H?6viGgTG{-2l!IfDC07h1Lu)2
ztly3+*^)IF?~l^0DafTNjqFOiw=%pr;Y}mK`f;|>tfRGDa>_nuKCInMGVpFaI##<3
z@M@LXj-sJ9-lbT?kYlK-(hfMQbGOV!YO5TRTcajakl>1DaGcZTVVqTWd@WkkhF7_p
zXgu(EtgTn#_l5o++s|dK=rL+~klQ?OE#ty`wiwz1gUBHD^{ZEG2N>P$UMKOt#7p5#
zY~C_vOWOezw6h>67&aY-cq9YTx$zQlh2nczE*6d_7^zlz9ZseDPk2StZ17kIugj}T
z$NoFbRnk9cEhApntTf0xF9Zhm;yZCEyODy1U@`}-e7h#08%3YA#^m({!TfsE{u8@c
zd|{za6oyug_7+u(mM3`N4_fk7O0A!JUn<L~(vyl;*QxhLp|&|(D8SDjO3ac_s@d;W
zby+bsGtW5~=dEPRIA$WgiXRHO`83_8(CMbnL-|!XW#g?YzD_-Aq_S;OS`vtECFJvy
zR3>mq<kRjY13tZLI(v<(aC+5PS3vV%oSrH<E*3~wDb709ZOqKs8NjMS@S!RW2c;_%
znIzJk=dC(BjEowo=I>)o-?Y><W@uhlC9#^ZpOkmVs!_?34Ov6D)G)n{SH|OJjemi`
ziNBS2sgpjV-o5L`R^j|9sXjSK82**>^2kX%{Iv}p#u45Ok%mW6Rj=)m^3H2{0cDX&
zk3sMCsW<*q>~9zx5<khI9*kM!S2piyaU?D<(Xn2AsxLDRJu9^FmV#DPzKuwYhS_o0
z`u;V@&1z7Af&d*cUD5bXS*trCR(t{g02BDts3qKRdSi;t1gXIpJ*rSxa0y%<YSYln
z)Gy_`5IA9iDe=sxazF|?W~u$5!l;oxZn-|4sfC(G!8jz)%0;1YWOkIEgXH$5S^iMm
z<&QN}f0#lJ6n`pIkUZ+6l~a$xvc9I4J2@@`i+N625sY9Cy#fBUAtpInl3bn4M@)KD
z77(nONS~de@vwb<TC(vxm|PYtc?vVQ_8<Lf(d^qdrxwfl8;LFvALL{>80E4@9R7m7
zyZx9xD_m)JdUk=WK)RlaKoHv8N9Kqm&Pzt;nE2{J!R=oyX+9Ixk>X>fT)yT6FpQ1e
zkNNFhnc&?&SJP!OrIpNheafOU&fnqscg<p*HzmyUJ-QU-7+ItBlf%0I0D^6!v`1||
z*7&ne63X6eqj<wPJwZN|^Iz?I@!MDM7P&8n{3&+@j2G65(p=g_wsz+QT2Oc#^cb%M
zkH)?xj4aXFUqyEj^J7yVP<xZu`c@x`Vb=V0bggl3aSVt+GTfm!&m<F7T80kJN>_a@
zVH_-}&Cbo=x#kxP+Rf=^M!23v8~80FF+31DvuB(Pn&|u!X|3y;ytlUGU+U9K9jrF-
zljcnBqyho#0)TxoDfV6h)@*!4)}m*!m*$x@8+av=V%VhbU4wCMzziOFuBYM9yBdg}
zPnK(&iJl+yEC}UGmn178>~IfIGm<^)=W(?o?4K>RutG9bFC=~U`z&Zzx^9tm4yC6(
zrPOWnZ$Y_87&#-UJrB~l{{W0$4^3J(f3nFulfcj_EdE~RJoLfmfJr~yz^#vgc2nB?
zKhl=vUK@plM3Kz<rBXNnk6e#R?=>4`n`0cYMh67($G3X$FtNQWzUP-#tBxVb9~9~z
z2kyK%;6IAS&e~IXac6C4hfahE<+R_rOlPYco;@oYOz<{?;GK6(O<zlhY8LkK%$Euo
z<9VhVaj<>Y7{Si~4m#J_+9jZyRn+d%6@_E~I}S^c(>{b(&b}(s<nSlP88um<65HIF
zBWO^3@(k||#GC{fW1Mh61FdbCRBsNWc9)Uggs&%BSL%3{vGJ2g_<MgO8a9s(p>k9A
z(pn<N3^O?;u;-J;G7n1Uyjk)0UDslc(i>Rqt)+;8`_DDhcgu7j^V8bBN8(?=xpW^F
z#o`|gTit3pCD=2YbU`yr!~^9NKXyD02;}Zyam9Ss;$1Z9my#rAJ9!=6XR&}_*`aT`
z7orTD0g^>|v~ak3*0;RZc6#d)-K3SjMtUd1pNR2|drGj>wA(A0S&gdR+{q*nq6Eh2
zLj1C0ByZ|7&P{z`py?K#7t?KZ&lBnvt!bu`_V%#@4WvwRK}i1q<+Pa282Xy{SHkl(
zrmd%oMf-G?^9|Ft%(0R{`GyJLuqT7XeYD>NS5)|Y;q3!a)Fx}4D7^9AM8CRuB;B}d
z&nN?N++w^u)`k}ih=pfl+q2UD06&qO>N=5B?(O|-aegK6U&2q?2UVE(Z&iN<_^ww9
zk=??p1><8pM%fuW^0+j74e)owZ-bs7)%+Q)+fOtXI%VXVexYe8mfR4zO^n=va&m{h
zDi8Q%TlwDV+CLInuHn8JmL}e=KIsRjAmD;I;=YsAJZ%Ss^zRQ_-rvE0F`gSMo1uUj
zQUW>lJBTEE*T09vP{s19GE!-;>;4$=F$vLJkavq;cg*1QzX9p+Xx<;x{4W60{6(Tb
zkXuYtA-IojWP)Hih9!?Z8x`si>Je)8w)g27RJe&-4fmJTPq;Z1m20ERrRrO7u}9~~
z8w4HsVYP=jz|Jb%v#eIJ2>4QkmmCc4ImLOHd`2~4s!+#j<1426rrE41(sfj+bwyj|
z%C>&#>ZZ9ZT5ZllW3GP+-@LlSZVx%&k=nUi*>FibWS^M&RvgFD(x+r)T*`UD!0Vh=
zY||IO0Dyf>PkAxPz~I(=%)vn51L@Yf6H9Yf8<p<ga3cg`J$*fE$9_EA3w<Wa26GfE
z47V*Natr;?K<+@`=DlkF08+Ejtj+eL<}?BPsBNHK^WDqZIG4FS&p;~+OZXRa@ghA(
zTk*`(S>4?PYb`p}u$C|>`B0DIZpS@-weRJ$c_r=YZP4-Y)z|Rj(D^3ATG9MZYNB`h
zB<jI8p#u$$M&5_DH&*bS!rT3x#%o*nqg=FY@=3S8ayy)N`d7L5^Y%j3boUp2A80iQ
z44b32$Cq)8^(Ub|pOCKv@lKnlcqVhDY7lA~6o}?mONiM_sq79%2fl0Uqc~6CXOYa&
z<$g8OuQaJKC8@Vn7$adSGt>jn52>yoN-pAY8^>;8QzMCAZ<;fVFTb^VezmX5rFced
zR_=D5;ivK&Pj?F*-iwjfIOjFyHrj@tu4)0NkL>n?=aqok0Q3j8eXFTzu3<SMg~FMd
z5(>(saLRVzd;KbFON-lkg-B+a1yhhCW7F4;)z;W}V?xy|UhdZB=`Eb*FzO19JClM<
zK5qwE+uQlJdZMnwxEWJ}$4$$Sx%c9ft!+q2-se4Q;%!9Tt)%w`cz0#Xu6|MIYdZGm
z%<T9$B$4!~CM#{`I<QkA*c5j4todA?PuJ;K@}%yNN>Y2Wr1%5duj5$v3|d3ZD^BWC
z0~3taRL_uiV;wVEqbIptj}L*!6&^u<O#0Hh<8k|^?t*<QLr#X$(%9WyIFeS+l>j{F
z)X?3CWq58_0;eE>-lI8rLv$qb?@_z7(y?MPPI4#=z<_bZB=s{iCSC>!BNd%FF2QlM
zXRd1eQ;8x9$L0p0L_>9%FcEnp*VtC+s|w}#gWGA+Ox<07AS(cp06gUL&1h;D&2r@d
z#7Fo=QJI8=Mn3O)o86X{mo8~mX6jHshdh(~hpl=wu9u^DO4bYC5XbhJ56upp8(_Dm
zndm*MkMQ21ZJ^uzvC6&S$hL`wLLcv+=~_A#lP8HasBblSoJTtu7bg!m9d@7LUcWcD
z9My9up`CO1tHj?Hrh>-e^yvBoCt_JzAy|<|K(P*|>-d`J?LHQGcSbv;znU-xOI*f+
zQ|*l4eig=euf;E>%$7PtzuEeDOsd=CdoM*S-A;cx<0ZI~;qxQP><=bJCBIt5$Siw>
z#+l(eWC|x~igjr&6e;=vQ+P^!5?ytpu!<F*%5GLo&5ko(2O#T}Aou?OJ?psm8}oSR
z^Z?q))9SQC;yH1D?(_6L>+JcbC9%gfabnL}S3oPF>U?_8tsdkf9D9z{;(r*tKW(G<
zmK{pjqA_1(r->L79HZ}LM@)`K<6h(Dh>s&9J*&*MKL#0mS8!psx0c%P=Z&t@cvkMe
z{<_bS=S>QVoUe1t{2i-a=zcJ?)HJvxxOik4iWMx*hlX4aQD0hWo+z-;EbZsiu2#s*
zp_Um3??59wr&0LVo?3V(U4RJJT(#37Bbe;q+Cch%Kb>^`1o&N|F0hkovrDa9XRX)V
z&#v6__?p#H=)%0?S*h(~tb<aHSkgqAB8&qod4OadL8}8)g$M@X4CkNTjw{3Ud1ldd
zy+co)W(L}Khy_nW$LWfuHMa+7X&XK81t&=-t2oqU(E1+VQ?_CM03R^_0N=qy*N9=a
zV1bdz8QiW2uLHNcnpOZiaq4TMut`|PH<dk3deo&Cu;os1)<>+{>pp8_#<>3gdauQD
zB~Rf~)MVEaZ7EolLl6f+S8eYInEc-`^grkGs^#o*v#nJj^V_}#D%?^9=s`Z7l}6ii
zks}o<7<DJF_*K|#n4gq%=RTFHOy_RqlozhK3`spU_Z6?E-AH5r1DeFMv6goMK_k>x
zXW{)1$=YPYbC9|HYek{WCXJ0boj^E0?;7Z=<|TTa-jx-em=!l11|2ildepI6vbov9
z_x}JIO_&<?w<HV#2X1QNx(vV`ap_qx#?gR3hOWZmFc^&c<NW53jol*r*k8vmD1|yR
z^cnT^73LovJV4qPg?vHc45*P~F<FzrX;=_j`SV?Np?2XFRe&zRLVZt4^3U3KKkc^A
z{3mbndC~@ubt<5J<MigPjCragDwE}=+u0ujc#lqR4`{2X+`60XN=7h`;t8LbkJOsy
zE$x|ANgTfw)O<_4Ulx2&n#R<Ruvj<QVgsN=<0JJn!=yjj@;QxEWOx4nC#7{%cauo=
z=w39_mhWGeL_bBaSKV<T$|wf|sRN+=Px7g6WocyG<P+Qqy*072w251ZU+2K%1N{F0
zDys1kGDZm;3e2}Fl*D1S;~CCPPaGg}PC@CB#YST*^&HjoQTBCihZQz1<wv(>JOhL4
zO14ML1voq&wQUqiqmp}n^{ZCG!+1=bx4kPGrbQG!YR4urz+mIJrTY)>nGgE81#Mbb
zY-9NvzPdpmk<QbQeW*HM?V;!P7O$mh>OdoF@_no5PlH$fNV>CTI9Kx&dmsL_d5)u@
zhmmB)4?}@nUx4phN!IO~b~g=;nrQSW%a&=MOZa1L#ih-zazQFT8mAO-BH~5LWU2a!
zhr<_tXVj%q3hoNnHKB6=d#y?Z>l=BC^saWS&!eQRB$uD$W$W@Gu(u_)tYwouutx&9
zJvQ9Qv=hn#pm+PH{A;c8Zj<HsGE1W0yjh}<#hhU|2DyzF4?K(mFjvNN$195Nqwu5Y
zu+CA&(4|r~{o8+oXOw(%){lq&9o~q}E-wJLjQ8iM{Hw-34ERsVwx0Eu5=_AcNAQkM
z;a!)<jY?k;c*fI4gcgp?M&LWh-{32z@NR}JZDNWPK2%G9K;WNRAq7&eKV$Md{{R@{
zdJ#0Nye$*L8PRt!KQZc0_}2ERp<4KZN`lW)vx`!<jdq~80N{O3KGmnEY0lsf3poSv
zK9%TNC&5n`c!xwyPS;74>F$O^o#aT>lz$cp&-voFsNkQ!W8k6fr7P3D{EwP-uYg_}
zX-(C>o%T7faL_S8GwsMV%W8iCZEmH)%+uS(fGBU3)2~k0^%eBc(`+s8V~X}^=8YR{
zg4QqzIOA^qI`z+PmEL#{;h%@KFAHi~uCJ)YYjQK>$0Vwx3}<qw!D4gABoaMq8PcMb
zsPrgfv93zfzpFkr)BgZxjYs=2T1|9YL2k|D;yua}w<Fa4TvwU+%fa{l8`iZA5iYG!
zt&?=JPbgNIa2yof!5+h@_piNm4+mOLJ56nFVT;TLD^=YznN<Wc1B@`iC3zh4(zst2
z{1MW;U@@p^sc|Ec7FgulEGh?XfDH44#c)@`%2D?j^w<p9Ioh(l&x6}ZCMeiOS(p=?
z;F0U?^sMEvXzl`)1(@Mf90Te-tLs}|+0rPdmeLY;#<@u($X<Ote-GtZ6MoIND=NFF
zEJ69o{Fn#x9@y(x#|t|n+NqDGo~O-L`h1Rn`ECHuBm6yUs?&TIcXt4u-aj!JLuElF
zPoW;Q^)7+%TTQkQv{4I?zbvw>Cm)BuuNCwM!S97X56zwRp1W+8my(qhx|EDl%QzWA
zIs@M)Am<g`96F@0GsCZ$IlfwJ?>+<4ejd*Z=x2w6*8rT4#=V2XHnui4Ff2B{OsX*Q
z7bLmQ1P+4(j1IWXeP!|g0N}TTbloc6O(JVp^&2?CF%snf`#=Q0?g8vT&QCe7o96KK
z#k&Onx&g|M=UvoPZFGF*D!odwahZ1B?l<TOJAVE;A6(SuU6RhmOLR!VbAvLbB*rm-
z{{a1J&$H8RWOYr9N%=yL(zLBSJ8^L;#*D8kVLo0?Phs0NtgY1Hg5=m=3#O~$O-Nhm
z){(W!!b@Dr3p8lGTPKCU?_SUFkHYNQ<-NtshB&Q?xtWQ^@Oqy8Kdoozx_!5T{3zO7
zY^PS$NS5zwWudoUxfdVA?a3X073_L<gIh@WT-WdBtdjiEs2P#VbRUnYuQshjmAjs$
z3Ua9}I%;|ar;l{0E@hD~?SvYAxp-Da*wQ&T+(GyLmE(W3--&0}yg8?MXHC7gy1Kf%
zQFsz%W{rseD=%HV_32!1i}gE?32Pb_nR^t1JC;dj8$gsmtV*$N$A70#E6+Sr;9XjH
z^^5!Iq?Yj{HwwAPZ1T*Q=V{2n=yU2TlASdf-siPeq@5c+rx)?YT^`R`w(#bKwAkqI
z3nrZ~O!5P|XC!sYo)4fE<T{KYB0i+!@aj!<x7NO0+eX`jk&i2DDF;1IUw`Rao)-9R
z9=)UMy2h?#gGACSTH<C#+S0K*VvIMdN%|?Ssc5coMcPRoYiVz8AXd2mJN(PX@b;p*
zJp;fR--#@K)2G_YJ-RoQzBDYk027bSiWGDct*QD}{Lvp#nh41>xW#-LXX$8pZoTSQ
zh5>9~^s4wIfH?Q51D>6`Q*S|pia;10s=SI@=ql43j8re*9ePp%yh*-IxC7kwtI<Rx
z`H1_*t46G>eMrHrJ7{u%X9E>?C4_NtQzGr!0M0v9kjL{6YSOdQe$>f>yOG}={b{B{
zIzNYyslN<IKyzN1EP*0p=rdk@WdyonMnbelcEu{ZfO+ZM)_=tBinhNE{5Un8PUbJ{
z$X7GS%n;+ZUgtOixvrQiBR4CvFU6Xy*YV3^2qmPHf^)_z7T3WM+S|s{B19MGX*T@i
z`_gzL;zxu&CqezA;q6j8nG|oD*@zOw^<4g=BAer16Eqtw<9-PR*bH%7MpD@v^GkMh
zT4tqb;r$+av$Z2dICEbe{Aux|z9ZJIrv<i>RY^S-y+_A>GLq{`TO@zos*=F;bK5oZ
z*NB;}q?yZ-Se$O}U6hw8g-3Xve|DpDXFj!uc@oVOo0x;b3=HS#_~V-0)QOGQWaFqH
zaaHZL`+GeS{>p1*xSD)<QWMN)7$kT8mDsg8UWW}L+y#)QD5H{tfGVxQ2Hr^ahi%*v
zdK_c&Khm}mSGI!U2+2t1z<zU_a=)!ou+n9{lVZfuZ(@Lso2zt;TQgya(Y{0tjyLVd
z{HiHnMnXhPjq?H<o_(v+^bZ4S5Ug^*jMJe3nm@b%_V4wr9~b;NpH0!E)J#m0f(xlp
zou`k&otc-kSFz*XOEk*LJjOxM`Njtnk=VS7fQ)5{B#x(_#=6Z{DIy?5@}rCa^6feI
zBdugKM_~#VdvXu??^0}xPq8zzmH81DE=c3s@vUzTO)b77i+b)SAQAv1W9gi6nzt2<
zdWNSBgt9C!23MM3K2-<x;CA(`*3j*Aok+uLbsh8sf+ma|q#l_Ep&iGq2vT+;kHPoa
zgsnBLjM7HgWdvs!9sdBvwjjOMwX0hX4eAqYvapU<iYt}~kql*kJ^AQ8>(cx?tm(FX
z7GX3l^9ag<4mr(v#=mR*qb6ce9%SAe{M{>iCXN+SS9d*s;Wf?Ag}f`PTi$4venzDT
zDRX%iW>(vPs1IJbuMqgt2Bjsn?u&U7!*L9P>fAR4xa4E$UB8HaA?p4u)^vF8E)oSs
zjRULsVQ?|HlhdVm@5CK){{T_(HQcvajF)q7Wm%oW6OKCf9+jMpX(!Bi6Shh2M?BM;
z4GfHN#yI?H-nnP?OQPyA6r*^@Z&6u~CL(-*Nyk5pVROFM9cv+XEO|ed;a^~W#9F|c
zN7gj*PGPtbs6BxgKdpRgYWo<EL!6HF^!MyLcG38+R)n|9CB9gDh6noBJzY~eAgq1C
z<av<>>yetQ_Wb(RwY9$MVUfu1QOynywRr1BbhzBgSUKraq7XS1J46(d(v5>?tuLUG
zvn$9kS-0)zc>30)+$krD=kMffoOGuVvV`({-1n-oHZn&QGn|p1tyGpzo0{4#?8-zk
zw>6)6bF_S?Bl%RfcL50*=CY@o5uA0c(lBg>^6LuR_WD=4d<^(?pm?X?--s@JO{CfC
zcCU3S#xG!56+E=TE4XpbabG-L&KMr#3i``XWbg;<y=QE=ON+}iTy!PnJDC1Pl|^=<
zT_?)ild<wd{sr*NZn8t7X~?AUve^JodK1T`O{Zwq9v{{%+TPC1Ln$Gmi2y7|1aZZ5
z8kL%9+I_3a%Gq7P6yuC%C)f@v8&`-)40q5HMnfu*(DxOir4cMeN26k2%348|<%W9q
z?^MEs1eWPZCCbQhI&}p12B_R>3XDSXYjqo=ndsRh^NbQW_o_2o?ilY?X1Fd%CyK83
zN&y7(>}y!Tbu)5f<d-1lt!3Ta1YCP{tu%OA7V{22+8sgl_Y~_t5?EMRJUW+$^zCXq
zk|hrC+(tc_PdWD$tEZuwsB$+K8Dq&j)d?jb#dZE8(R6F=0rl?<%QSZINU`4Ox2_&r
za^H28#y7@&P6)0on<qFm&f{aIEV8aRsR7?q(o8qDD%>X{HH?w6Ww-zf$g7axo|R5B
zoK?sVIqg`pL}qy0f5jcFXa0Sk`B%;1jeR5ImLKqjYX|<FpZQnK1m~>>tAx@loetx{
z_vWv|8ZkLSOlLfG=eN?c;bl>r;PI2!CccvWk^UTbBgc)bN8=49^Hq>-lFVDm(EyR3
z-jK`uM3r88jMg}eL@<<{YMUe2rAgFwiQ&(sZ;Q*4+kA2kH!fAj*0}Ew_yw(tN`b|>
zJ=Y}<p&pg`d_Q3igxZ+%BOW2ryyq@0A5z+}XPg1rtbfD0`~fnVEoRj2(>!NXWBio<
zwS_F>3sN_W`~^BPT1i;^in{Rq<TANiR+cf*qZw0F&9taWut@vm+yX20`M=<q9tV*!
zZM7!cInJiT@aa^4;ntrIan6;fDfL}Vv;P2~)yp%X(*FQ+IlYs6Nqy(!g2Y-TA(ZX=
zz#RTH8b-v9Id4H<uGYWsOAmzxbr;%ihys)Mt#vGM?n>6Of59<+1!<CR{=xAb%8oH6
zw!vH2;YDWij5=QPx2kLOCDN(Q%9Z|S<D{En9#nZJyY{BREOD?P<^H0-X>9)hU~hyr
zEFrP@bEQTf<@-hD{l&ziwnk0?^y0GqC4S6*4!kYl9YV_6!`d`=w@By`8$>S>vTfWL
zjDwI4dgnF4S3Sd1N@=(Kd7X4Hm8;ZL)stuCRfV*tOt*xo0XfG*yz)=wUd`|y!Y!!y
z_sano+u6x7?~lyfaJazyYt6L%H|&~(w(cA}4)KxKg*g230k2{BPP15FY7*N|C6)Qp
zZWcMCIAm;sFnV^cdlxI=<l^^7ehMGFS6d%Z=pGjPYM@;<%3#V0l#gyOc^_QnyUjD<
z#7`oJ(gzzzRd500`FmH3d>Huks6_<!zA&_BSf*H`y)o^P*BqA}F~K<%^+toJ*<Rbn
zG$AC8Lb4;KbGNoZugbF=MSMKm9*5Y{n{nqz&$RecVHV<T=8ca5S-}7vzgpR{_*-ft
zK<bhO<wL|Wobl6|>ul~~{nM5N7s2n5n&@EhC5TyqMIhJ?Nm0QS;JjT+zG~d7%;2w!
zuC$C>Is0#iwlG9OS>P7jaB;Zflb?R|<Ubky75I0<8eH17oOaq>?qZ2yiF3G_0~s6r
zp@W9%GI3txcjAc*DYNB%?~mzTJ@Gn8HIFU^W&Q2kVn>tifrkA(Pin&nnM$Qq7|qJ(
zH7Yf#UR7O^J<s6iSMX(o-ebw<MkE45lY%%T_dfml*K$*6=Plo9=O0Sqd=mp`o)puq
za1~>>NcOKdz`z-<y4O>Q)G2L`j(1ntU~#Zf_7iq%x$;<wZPas)*DaV6$||Y`?!vim
z75qN1@l0#>n6WZrY{|oc=nZvCAUuKy_NClAn)c-B)s36b<&32VbkCjqW$^pO7CI!_
zKZXR?CP@i<%ZB-pPIoF2c*s4mTrQvR1H<<I9h%3)NemZ`0!sGGLyQtv1+oDIo`atG
zudS|a<CA_e4l|Hz%sfrti)+nhYq{rji+1r4IKd+g&!O}c@)-MN^!e#&c6*p=?z`lk
zhX>&w3iwyTYZ7R7aY;Ndh?ef<L(bXYmV9@vhfLMAAd;ocgFHzoSmv}Pm6O~aK_885
zLb`sGs~xN&IKJpukn#RC1nu*q#UjCQ>`7+jWB_mv_=@=HVm`Si?UmP=>GH-a-A5Dg
zhr~Mmm*D%0jV4iGrdZ&4RluC9{{VCQzJ|P$N_)5PZI!EelF#<Xl|hi;Wy5agoa9%z
zPvD!sA1oi*(j7RVV2Cz4&Tw#`o<=(#b5{H_@I+5?lHCWkNXS=&=jQd=eJk`Vnv^iK
zlS@;`#ne(xCoGySyQ%p8%f|5;<g<=AgsfeJSlh9aZ)}_s$5CFnuUSc^X?k<-GLXAV
zeNKNaYqHUFICO0-Y>)<pGr{!qIqOtBMWm(1k8y5tMhhW5R~+KKtYj5hGQRfxZgAnM
z`#RCK*F3)VZNQh{big>Q3&#gJ!3WZ=YEt~nhk!m}JJvkY5>7}9*oyeb-7Q)7(T<6t
zrPkc!jt9MB>e{3lUAx>|$+jgS2to_U=kCU%yC2fD@2;Y@f@_)POM6%bWtLNbzmKnP
zsTIj+T6Eqr@f3I0aebcB*Y}WHOrtO^crDlWjo;ItHSXYWp59*4e+oRzW+_pRvySZa
zFAi8*Pi1YY&1D;1Z2~x?A2dMr2e-K$NUMG!@KboY`5o?Lw6}&S6&@!j0bzzC9dbCq
z<GALwv@J3#gjw!ghWw`18+L$r81(n4b$vu1^mqtVD~-eWK<F#kbL4*R$A=j69Cn_w
zcxlb4y1Bo&kSk(o35)^4gkRy{bU#|)e0K)1;(M}n+Z%tgF)M8ZTTVuJEANcvy*^o%
z#aqgarDnh^!z|Uy>vq%XkjVZ~Gr&qI$8YvWt#nbXCo7|;*g+%aKOTGx@ZPUGwdR^2
zgfLW=081ZT-MjHzRqyO)p<L)!I?kJ|u4A}JjhwONn<uNE;lTXs>YMvXBD7&7QZ!7t
za!y+VsVmfq<ma@5K#F_IXi|GubwCv3B#iO>>=RzJYt89$9v%$vS^ofKU2ZqEn@iVp
zyQ^@ePno(Uj1I#be~o!(#gB)cCh$ItWxuz8T1hSi)tSsB_xI<}*VYhTU0Mi5t<Xsl
zW9Gt0!)Klc*N!WQ_`~8I8(8q>t>Nbj9sKY~<t)E*%?r5X_B?S~#&vns^+eU77{RoB
zLwjpFG`A8dShu7@aKQwrY!AY-r@7>C3Z6p~*jA3Gc-J-|neil-Lx$%YBLgJXP4?ZW
z;$(JaV930c0{|2C>MON7?s)FeDbFc66;4S2Bp!#ZD#6ZJ^~a@ElO%=b>s50mLhdlx
z=N#bEah4c39RZ|#9P$b7X<{U|9DNN-yC6~XFxn2>dQ-q0@!FGaPDmpjhxDm70m&FO
zNCtr22X9J8EW;nfDovj@(aH4{Zo^BPQo2|n0H3_fcnH1Ok6<fBu}K+c0P|9aB$H@W
z`I~{0jC<Ev;m-$Z9xT?P(R?YT+-usC92p^xaL0D^RUh7B1$_zO@7SOF9?~1%9C(it
z$^-uZWB6Sb^6vz6XCxx`>7JD?QCb6OA1*DGrkqvff40q@iT<n~bNpY1E3nhyxzhAy
z)NX+D?ogChTop|40PGEYSD}18_(S2+kz0H)@g|`u$ZbPakWRxLN~j<M$Dyvmcla@J
zr$=F{LM-Jkx#88U8xe!R<#Cqx9X%;2HKRgWAD?$uFLQS^vJ7uK%VUrS_<pra%1J@E
zt6=p#y=&=DiT?lzz7*Uf_g)Y2Ez~Zk<+R(uz~`)TcgOb)eEqLBm}Q1NHr<jamt#pW
zJLPWfeLaU^SIGAs=0K6R0>rl&0H4;q2jFb+Cyw&%v^BJfH+toGt|I+y78`K@W|*mI
zn2Z2>D*F#w-uPzwd{2-z{^kDwn5`6fPJBwe8LuPiUpLmHdxg$74m~+F%IP}kE3)aE
z%U%Nx-&)eUzk)YbNsyKq;PkB{8=n(4;#j@73J3>i^%a!bn#YllbBy7Kt$Sy~jZzz*
zgPt<EnhOH@bQy;9+h%Ck_uY&i#=LI-0L3xhp^jZh%7>B!Xg`wHoTh44)YpGL=Hy2T
zw2t`Npk$o%z^zRdYmGZqM4C2^Wo@k+a8&!(0^TjVvYjNdxnk+}U6+yR%~ZVccCUKu
z^4yT^_nj9EPpLGyl4o>RG<Gk1T(^!-8{FJNM%$RuL${n9V2{$di%V6EjPm}Sg-r51
z{{TCZ@9M`Ur?s_a18z?>6MW?*$V(A8#t1&#R*W|f08p|JG0qQKY&ulZf2&f%1d=@m
zy*e9WtO*1Tht{)Ew!(5!+|jqyE~7kLKOHm8bzT{{X(w53TX%3VTzoMd<0Idi>a-Rk
zX6ylvyheXYv~g5Q+8(>3i)+L~oz20=_}5o!rd#1a`GM>Q6~g#x<~9LiL$rP3PxCeC
zT1-MkTy89XFT%9B<x5Qs3p*&HA~itI=sH(NWo6}saq~Y1k-+}|>(y%=J(L2<GyJQe
zwzVM!dJdST<cw``u}33vM}CL;)W?I)2hy}5w~|0Ixm<P~YB}uK=jG$m6*nSKYf`dE
zLC5LoR@WqhoN<y!s;g?q+2*8?7{u}C0l2EU9z-@fFZKAz7=qvTjh=+|{{R~BFNvNk
zeM4IDF1Zm8DnTW^p*@jQ?rQd*5nMKd;fth+vvqJhrN>i|$sbDbBciS6_Kk`z_lCkn
zEB)tXIKeg1L*ltg9M)1kIPi=n887uag!5*Rjx*JXImhFg^cegx_8ly07xF6u^U1X1
zbM{5`#&Az+)$lLDol{iTA&TsMn@)ku(@IX!-2NMVeX762zY$yNdYQG;R~l`v?&CPY
zlO0>}uD(?i?s^lb=Om(WS2qoHa?&y6o(K8uPIdV+gZR`i#L`GR<mR<yjXbb}0RI45
z!6Qas$&yY^YTdcSARM06elHG7a09L}*MZiyZ2UJ9!~59T2^it&G;e(@j&c>*D9=p&
z+Um4RXygPYPJOtlE8*mh>U3aPV+u#jlapKi60r#>ACP05RK1O&^)%qqj|Z=>9cxM}
zAO<nh`BtU1nzH@Vnzi;I0O01aZ1iIer#W+G2N>Yh287zWN|_9YKb3U%Hp85Xs&wNz
zI0Gc{%{0x@yU_N_8_4dU(e)G~{VLINN2tbYtkYoMt~9#=x44UApGL<O<vs?Uc(fa9
zdAH&TT@XGn5RL%+Yu5Y`a5W1gke$rGBX=1UN*u|geuI?ZDdVa}6<ws`bo|%#sn2+W
zI-2S>K<+I8g;TC@K_AMwjVDc>#@bUoyTKumNZc`=Jv-NFs983f;u*whau_hrBL=E?
zno}x9tCm$gI+SkXK9%0mrsEZ}K9>iD`qP#owEqBmN18V1*585hXNYv2{{Rwoi-rN7
zNytz-gVMc+!rmG&Sw@P0Ke*_9tB4vy-~33qj3M(L<K<QuB>VbT+n)`-0|d~a)iqnF
z>}=XS`{W%*sLw&~S;}#t9Z%;R!kqC~%8^|g`>xLhk3+bC!p#$r(*<*r{(GA4wC@!7
zuFpd{UB0C}8hB4H?9AY%JFh(Bn)JK<YvG@UW-(&>T)0jEi2Q{8c&dr<QqdD`kKlVN
zXtGC}cBV}GoK_yvQ(p((d2W?DH%_g+FZv!+W8gc>>xkVg&Be22n1PiT?~~IWhOO%U
z0Jyx8&KOyyk_8!9A}(>$XyA_htG3quFnE&ZG1{0kyM4PvHdg96_9s8Bd4G&Ned2v1
zMAUVUwOrm@*g+-2z|uMic8rBPJ9p#Kt<Iv0l&p?#XHHE*joWV5^*xKiS~iJc;A3lf
z;Y*d*GQ;ink}4^3FeC+tz#MUk=I-t^t0+;fHC+vEk$z)wsYF<g{HQ(qSI!<B_@Cf^
ziZ>RMYWlFe5`;3gs_bWpijC{AxKg2s+E<b~*R^;%;n%|rX3RQ6;n_+C4kovdJf}Y^
zl3N+TuO2wMT_&#2q?TV^lZvZNq_ok0ko{xC`UQ-vTUzk!uZ*yY@HR8s?)B|a!>xE?
z;ly9?llW%ZE}(s_nB)HWt~=uY0E0XY;oCT8io-*=xVX8BD6KTI@yj7*!8|ePpT@dh
z41UgD8Pv5{txk><%ElEZ%nHFqUP3!6<Q?4!I2~)E7`if-B9^Gat`em*olDvJn|4}<
zh2ou6d}pA_#O+9u;00e(w;3k3?=197n>lPxiu7yGv@)rS&tKhhk&kkE=b)-rzXpC5
zcs>&oUr(sq%A*lzu`FRY`^4ar`R5hGYo7!BH{#7QX*8`zRPgLJ@UzRQ$`o70&)#sr
zWO3>U_pX=iTIY|#y0eR#lxjLI&QHz1$n8JjPR(^HEEmvRBlIHS4o|Lgnk4<6wJF`l
znQBlRWX_|Xerx7^4&v$bEjsF3qG7m+i%km<7zZk@InLGtJ-YE$EhV{i@*&o4V=8`3
zN^l2by1HQ+lZuj&g+)#ilqU^Wne<cs&05-8$tnv*5%N|!mvHavTqGVJ9v*q4l1S~|
zPu|AFk{k{IQ}yFK*POl9tq0kjYx|!rTwv~DlixV%Yt{Y~_zwR7#~SvVr<vGBShS-#
z5g|EZ2W*elt@G21bCtxQ8jVJrT06gizX<2?*0Oa=xSroif=`&ZAH9C16p#FP{uS@v
z5^gOmptfH%qq5VpD@79kuOf~Vj^i2Q>07(3yd9z2>Dqit1+*w5jT0Fd{=LbsAn~@X
zF1zO2T7~;C9Lh%<y+^1A_2#@uP1T<>*!p@j6$)O?JH6T9zZ)f+#QN+Ss8<?=$P46W
zi%3*{<|2Ja%Qs=gcRvpAEqp_xjXzOr-}!2&`*K8mhtngc;wtC-A&aQmUE8a!<&Seo
zA><DGea%nsM$1mpd_gXmtXs5J>zO0e3t*#+;R!uK894Q=Ins%wskyX|pMEcVGM~e`
z>_%Y!0AhbJgsKTBxyy6iy?qM-RPP#35O`BT@GgfN?!T038mx?PtjaMQME?MCDDcGp
z0C<yMaQsR5YX^*Uo4q>cbV$mr70%91KI!@o;vL5|^ZY*y&xCZQwVK$+t7>89x_#T4
z-o;yLMW63X!za~8wQmZssFUh>6{yvq<d)VrTUXM2HDzsM2D5dg+N5B@AYNxuPw$`6
zoaE#CyJ)4GPi;o(;%JISo?%_m#9MhC*ZrQu*=VYDQg%6`PU&iWQ@5H;y((~>X!81d
zSHh-#sze6U)2%2tJ!!=|;-YsLJ-zCHQUQF9gR!Qfmw%h0@lQm^>6(;9w*wgNezb-v
zL77emO1B;WBh+A1Vg<VJO-=T74a0twQPhS@0LiYS!?xR*2(D(`Kz*R(pGxlhGiFMm
z$6hMzPUm9<y2%2Kpsq23E5JW&FM@iv#gBzI`e%lkKkamy8<A-%?OTP6ljVchl5z>|
zE9h5oYX*G>O6K*O-M?v+HwZ$E*JN}^$+UhZTk8HN_)Fr}kNrQx+Rc;fc$+av86Kn1
z{zu-vnD}@5T|?qji%;=@NG`GzYLKoKzw4rp<~?ibkBz?r{2}qLPYbST%D3v6i&MA4
zPapgXxb`*hAH{#!H^m<hZf-8LIfscfnDR%GHqRLU0BSz}09DO(MoFCVS3ZyNEV5jf
z65IVTE?5W*2X~p1#t7@0=6)dfufcx{5Bwuq_0$(WF1!Q%5fz=Siw}@H4iKD<rnCP5
zWbFr6@Q;M`Pl$dnl4x{YD%MM@OG~yPQBZyTr=vSzK=jGSHSt%B_3Qrt8hC?L*Yycn
zZASX#u4Ryukj&Y~Jn_IEO4pmQpR?0b-nGk?GP0H-m9R#8_xjch%yLOQAFXjd6!@d|
zgj*ecP4cM3SI4OXfF%C_;=elVyu^_~RwZy7Bb<X>QJQGwlv)dE+SX){1(2f#+T(B}
z9N>=Pv~(-cr|OYU<|;!glO!lP<D8M{$;D+c$5IY6$fv<N!g&WHis*|u<gSmjz6{&x
zSK3<NI-{{(xX&GhW_)75)2+j{Q2_oCj7k2<72qBd_?xNtZpp1|?o26<^<Z=zNfpd^
zyW&2n;u+)<xoo$|9(Z5=zk1clSmKQ-DHXg!aJQGT7|;YQ&O7tR<y0cHSS`+@JZ#XW
zL~7k6V0Z%_{{YIm{{Vu&4=?^N`0md_h_JU@+~3cRBY{sR>Us};X?`z_FnH4APSOdJ
zN7I;yihb9|$`8LF{=T)$<*koRM$x-Cn>gAK1er5Q`^j1Q<njRat$TIaF_LjqZ6Rq1
zZMn;Ctx%y8BO$wwPfS%KIciKMlLZ8f5cba%rKR4%ZyT)Y7z~gJ?^yt|43+!9ahjzp
zjO=hhJdBF8`A&ILQdSQUT`ly|_EJ7gy#;yJw`SJ&BP@22kb2j6@CP3@S#ii8&arjf
z6`lr{YZYGp)H};kpp-N@f9$cY+qIY*XOI_xkNorlr{XxS3Dwle8>=gfsRVKeuSPL7
zsY#+{62?QOMleS;N8(?H)(uAAShi5Mx9GAujsm$H{<*ALXH+DvaCYcPBn~U+uh^pG
zL-9ImbJ6XX{d#|ecpF;%l~kM%<8a5NdiU(Ls0}mX)|UAyCJhhgSV?Hkj)&=8k1*pL
zd(ux9<vppodu9-G$;VorVT3%_jV;;QwMJaa!R#s^z%O2vVtiwP%}|m<7$=%&k_)-E
zE;H7!FJ%RZAD(LU<c!6-^sIT?oS(+7Bt1x7w>C-09+gJ&Pyr{eY*eyIfKN4*eK{yP
zeQTm9%3BpK<<3oJPc9Vu)tiUN{3||kHu7>lxUQ3!G%XJA#a<@Uw0kJbw^y?)YCQ&i
zVts(gucZF~X=}8x(X>Ae#Nnc}w^HCAMF4*)tNS{9KK}rO7ec!6_P+@iw=wy8eXB3b
z6!!zMZ_rm4@doz$#6BRny}KmKI~LkG+6O>Ag$v%5%4?QBr=RNZtlDzyh;IkyJwMKW
zrDDr)MWBdqn~CaAYUpP|6@fe+*sgx^KbVC$&*5DcIrT6Ej1l7mmO^%n^<PTOlIIM^
zuf1trLh(o>PrSWr8&G8p<{-j9QRr&o5?x5{Z$L)HSJK<<8k7j*bbGfrK7bNO=~XTz
zG8}aF6+V=<=Hlz^GMHEF(iqu$$+zWS#MY?E>}cKIN#adVx0!EXg-<e#i4Y&wxlL++
zE&P?ZS%>`e9C!Y;YVJWSE0@~ai2=Whu^&-do*}xt@jiuN66u!Kk!hC60B94-^EuAu
zJ=pQ>*1GRxaNAu>4-VhS5Y?l1%v0G)j;Ff-7-#55>s<O?u)rMtHPL9cvD#=F{-rvC
zKF)05AKsouTw~~=iTaAi({!KT2`(_GIRQ>{?d@EXS9%@Po`t)3%(ytf_p8crh5YIX
z?#BxnfCqYPavbt&o=E9T(L>J_I=<dJ)flDcHDx5-&2vjbswRJpSf2*#1OA<#`B%;(
z(!PxGlWy>itUu}9zm<IO$OClfGeSLVFT%!yf&f2_eVzLW=<=?KCGxpX?F;rQ+5QNI
z-@}r}@UM$5B@kOpF$9E+w?4J{Ir}2`8uriOu8j@7v3rR0q?llZ`K4xJea_M=;`p-|
zu-SdDZf9fYvh<x-C$dLl@h`^Ho1(GJ<*rV4aq@%3c=Wy})EarEgHXGKvFioH$De);
zdY_Jd5O~K>xsel3R0j(iwCEQHoPBGWkM?WvpiD@XP_bO9{h%D5Op5R_%x(&X9yhHN
z{ZBtIom#bR2)<t<gt+*z;~gg8MQ`G5Q51g)-L-4z{{R?#ZF?iD_@h_e2H+at@G6u1
z74c4?;@f>4Ez;ih;!y&-HdRUPPo^u<ApMQLBwEULUg+}Ufa2kokESc`rI=tT(~6x)
z-LLCo;Ab3k>q=4PaZl6rD|lbxr;4PXY~CZ(r9w0F+#2d1$6D0P#!X`KKzo@<rO^KX
zWe*)%NxJIeO7!_|PyYZDT6cf5r;7CDo_Viy$gR^1%-1Y1KkuF^^8O&H8O_d|{o8lY
z`|O^)p-D!%v-8|d7g)Z%``VSzaB^`m$oH-7ZpCh_(&|ji(?Vi$#IY>DFgf}IT+fAk
zJ97rRZ)-j6pf3Ji^^_0~V_vbY_+CE_czQ;;y@44@nB!&xa|6&1eAmK4b*e|(#oiHm
z`5$Lb6B<<^ecAa_@h8K-@QM6X))dCbzil#X1FFA2s3yC=h5A7n>0QXkk;%POf-*6{
zJ;$M~kJ__AYcCxv+?2S0uPfB4$Upbb@~Zy;3Fn2irj?RGn2rlzoP`{mdiAg0{<SPd
zQ5wsvO2W{LGZzbKd!Dal;5a3}H+x8OsvCep0CQf=;NL#iN?U2}fQ~7CQp7Ge&N$|{
zogP6LN5CzTLLLdo>0J$_(m^Y4Knl4(D8^0)AFX~zjKox_r1ePqX<=g1P*y!Q(^0oR
zQCot?DsmL@-ns~ItkQYWFq<X5f6lz}-gRury%h^#a@^+~NUeKoq>W_#=|0&6SI$at
zyXKGLcy;i%+5QJraeHBI3N^|M_au{_Ojn%v%UFu*P_Tjs@w)L9-}s#m7{{USTGw|9
zkbqe3JuoYc)qs~#wv4M5HxZNrsOLY28LpfSbgRWiHYH%vr71S{KAwhg_DB^#+vTrs
zU^t~&PH~+5YMr#6Wu3kO2Vm>lj`Z;V0H3Eys`ctlvu&RYpq0_31>3{-9IvmX4|^!d
zEL+^xQ(p*93k;rsp0zQ!3@~yJrA{Jv%5?6e()BRqgL@Ons6IGsA5d{w7ncR*X%L;q
zuKsgYWRP=N*B5FJ-QDe8O)OR_Qsk=cj>>R?)+KqZqPcyO+?hdb!3o$C@~lg}e?+#p
zk}Gw2Z{uI}Fh*I<IsgwDtrtUia1WS~A;^5LOIG%mWn&I}rH)Am8Iz62A6oj1H?y1O
zY;%)>vplMA75L*>@e~WD=(id@qBBP<_b339jvKEZLEgPnNAZ32h2^?xmb;@PZ`zlr
zB;X8p9`%0aW94l?10U~w+RY+3%9%fY>-R_Siux$#al`1k9Fl^ntkKnv6B~=u1--Mp
zNJx8!ARs=&si<b0h$Ifh&PgGUuVL+6`fCzIxFH{&!xcwT*CjUYHwcG3gUu@0+pP;2
z!wG2|ABpZ!HH)>6Xb9zgU(T>*l!LTzan~FQj$5NW)beKlzXR5x)HMXstgbHQUH(dz
zBh>aEo+!g!v}10p`&j<!l5erg_`2ll`ph05w;$+sL+uy2JWDTCjfZ`?&O7nxUX$Tn
zK1h-;C-+fD8C$1A_sao*c<eZ?2f=zn!DzEwF6+y=O~mhv=fBJK;B#K5Zt+{`k*hF@
zQ<-wz8T#}dhQ9s3V@@=Y@$>!U+}RGXum1pN-2Ij}R9KVB<Y#LhK+mo^)_sdi%Hmm-
z+ColOmKi*PJv}QT!Wpi_uX1FTMwF02&QEin>s>yXrNw0$y3M?*9Ah56&r|r<H!pKX
zB?X}3i*07|5(_T?FzN4H{-vmFf?1e=2>XgM!ym*eXIqfRX$qIx%y%_&VinXbZ7rl$
ziDi5i?lH;1^}yn{rmV)bS7s)mtR$Wyyz{M!ptn*X2V=R%Id4pLuPD`~)F8W(+2bp1
z8S^8CDFG4g<#2oTuUxZ+8}AV7vb~Z=ZwvtdLAVqL=3slb9M_Xw%8v7mt{zv(Ip|6G
ze;_N?pwp4nhk~1CD@-GbY0G(Y#@=U|(T&Fk2WajOL0%W*8R4JCS5qvaMM<4PvXT5g
zhq&ZdvfpcwM=$RXK+69BD`zB-a7R5e#eDmJbZ&3;2AWwg(%cfofNb{qAA0p@#jBZP
zZ4{!964c_e{?5{tQ!2pI+{rnOl*x_YF3@-!4)x5;10dKipd9nb=a17B(d&UD(X?pD
z;a2E^!<NjvWcREIpjg0F{{Ws~l6_BFsH-cVJuS+t03Rn_qXw$9M>}zm)C1q%uDHO!
zIPcP+kp|;~fsVMUGG%+aqcj$<c~OBBTSKV9Ic(&gTCSML&V4Gy`ju#ufB=a(ABn2)
zV{zw-OL9q^Z6lua;Brapk6K_LZY%?yi-T4zVr!}6w}L4yW_}hJP`feo0CQB@TZQ|L
zh-98J+j>^sn-ur4+h}$XgL9})9CA4O&RK%0f4T|D^{-a&PwZpyU&fYcsK17j!#YzS
zZ|x5f+@*|rPjIdC5syJ$kKmt%zXd)FYt~xVioPX(!d0)>%(pOUJ{`6{W|h#85X&IT
z0K=Y{?NQF2h)!B)dIyN2@b~P$;d>tkSn0D|YSP-QnxxAhgv}!l`D}S+0Fku;100(2
z$i6yVTCw|mt*(h<X$O!Yw+2~1pB;Y_=qg`~e-^wi@ta?}y|GJ+MVP0NWpvwF--Rr@
zI0Saj=UxxunC>)#vgF%_V#yfuM7#o5a}IhA=bGr!D(M-dt*P`Th5KiHp7a5vO$D4u
zu-!Qh?%4i4pl5fdYP;gU+Hb?YD6~N?lfC3SMawvnLU0RYk&)K`bnV){K$Gng49f2^
z0##W-DsVa5?}~hPb9qs^G=Rp$Z)Myzj+}PK^{I_T)dwov^`95~K+?-ZbbS#-vQFx<
zPJeP!vg7;1jGXa<UTLiA)_T4B6Ka}#qYBFl7ERsTCnRU<iprBwxs^P*6iQg~Sw`W3
zj04)LTw%y;9Jg#>0osXFZ(`Naa^~HjCntL6haD@T_-Lf~oc{pV2L58XIBy+MQORh)
z#yu;t_*|j9QgA~r+1%uI<W<I3c4j!2`<eP4!4%4=6ihHWk}2eQvpb&L3aIC6o<~aC
z(V+8fX2{$QFvN8CuF7ju<0`*(GuVC?#ANsd@jok!!ni;18t_JlFG5E(^hd+1RkipX
z<GB3T&)S(I!!|H>*N(~I;B%Q2p1_e*qTSAARx#4YCdhhprU<?3q`1?ewj<7EkNx9S
zBz-z0KfaBacFt=zNu6=VB&}mU2m=rfE7Cq2d@$5}bEu1bD$;#Ost1*A;wQ^o^SibO
zu6eE^E88uj<^=l#Urc_^{{R?&;aGSNS!*jLmg`gxZj+O-qdWyC)C`^~T{$InsT0IT
zdOZ#=QSct6;k)~Y#hJLbQ?tz5v=RQ&_Nunu3-wK1%iHL85}{T*;faEee39C}L-d~v
zcz09K^i*YmF0Jhrc;IN1ZuvnZk5YP{dh%KS0BRcz3jA7KY4Q0X*$goN6eu|B{VRpm
zn|6eDIwM6zDN0+P5!`4`r+Jah3q<HhJNV>w#Y?8aBRb#`Pi%Gi*ROnR*EG$2x9M*b
z(8(#d3krE^n}<0bwZr%ZC5vCQIND5saCsR&(z@ezuFh(Ymu72dk1R6At_W4eMmf!U
zzk+-{J@%Cy+WzV$J7ZV-(>J1y`0v<P0r4T>)3qCGXzk;Lh}=%yq>_33E8P4G@eE#F
z=~fyvu(qA1+Ap0e!m0@aEu8Wce>%pqlvaqsjVM%==4xo30KB-{BSvseI2;4?uEsBg
z(uEt9fg_wRO3~B)AX{oSvRqi)LwOu0d$qPW2dN^NZSk*41n<34Phbxf<GELs$o7>C
zG#0F)d<8YL1zeDO6{zk!CSs2}nL+gxx;``MOsc4U(B$Br*sAb+Pt$eV^QS{%#^P|>
zB{&KQ8R<<*zhO#-8LiGURMJJGj*dF=FhyfLSaXBNYU#AE73rGwwCg-A9MLgH*f?%a
zPkQI?N=Lnv9fK?h>x%T~&P`fK`TSJ~)0Z?JhY_z^uC1xGlguKrv{HHC06A>el2~iA
zE!5hTnuq&T+mtsyCUJx5#dbdy?_^Jgz9ybIzRptEmXMygKXiMZmEb-E)vhM7ca~`_
zw5v$`%b9b?PwxF#@IM;%D61UNi{<7!PmLZZkHY!^)9t0(b$)@QU~t}Ecpr(+;49|W
zCS1Gx#GXm5Pak;Ne;Rm-;_=Stw_+t8gMs(~Sv$Z*1%W4;*}i6P2slRl$skZ0uQk7-
zXz!`s00uzo#dGjUxrPo#e=77Z2Rhr^qYyJBn4u#^`Au)^7_A~a0fEBz6_+>=z6ii0
zis<!Q#ofjS8O3uqa;wS5>rl-j+@e=~-PR6H<sYSJ_+?66NDuc$KRT@@oYP9uM=zGm
zf(ASL*0zaY_OiRkz?MA}a4R--!+V`|oM3@hOdt#zsi(*Y9+jmS3_a@>derqsRm?v(
zN{TcH`TZ*O(skmg$faASYD-3GvpR1Nl_A%m42}omUZe2k#4B}f2P*zx`V-B0M~7wJ
z%t8MEmSFxyy$iwDt7)#<h9oS=@v!whtFo1oKC>#sJT@;6QSG<ibaq}P(qxlVwz`dX
z#UK#42cf~_`w?87)xVb{kgyC&;d$eZhP@+CkuDqT^Nq;KjGs}#u1m!lVAI{U;IwV{
zk?U1fO<8t7UcqIwY1XSaH_bh5(^Y40Q^bF=ypJABKio4!z!mJnjw|cG3H(g({{Y0_
z2U_@NUA!7bqY^A?OYKQYj4K1s{SAD@;yZP;@g=R+&?7Dw0o)!B>sB>2@~(&g@H*CY
zAggHf>W}5z>YT7R>WbU8n)#mdt9%F5^qIWaBO0}yOC<K(oO`cd%BQf@wOs-cXDzN`
zQMb#ugBTx7b6yps{9y5Cfn!mn>vvGbye+%F_|N<Wvt8be`(${LB3QM3D@@iC?z0%o
zY)|1>&Nz!}X&*eJg`4KPgZs{x#1l=f6#oE;d=|zrkMwwU@A~4ipzxX0<2I5-<_VRQ
z@r(j7jC$3WziKZJ>LGsH;g+-`ui0b|{OelR<EMabF0II0+gIMCnIQ)p_aeO9U$^!X
zN;=v8M_v_Bjc7}r`mNWQ@n)6q`{6c~;=9idc<)QS@YbbiKa$th(-Chl$=dh;uq0!#
z9`*Hq!w=ba;hwkQ*SPS{h;B6-J^ui_Xtfb^6Wc7HkER7->mLz*7wWztyVEsW`wdZA
zMs!VL)JUqOxe?2bIXyGQIB7o!>@xaHhe>G$F{0~6*&p8-!LP#b+(k|kQO8rM2E3i0
zn$Z2b6@s23pDh_FF4sLLO#PdD3E|B{?NCo0#q16XJ2JDU9m@UF_-3#8lE+H;d#6p}
zt1VhP17#eplR5($jkqYvNDsS#jE`FL{SU@I6Yy=kv1t=b=4Rg{*BY>u-M`vl_rDs4
zT=>tU-rD(6S!wq$;44LP$tY&Y=L|^n>s@b&@YJfQQ08mV-|{%&!}6;e)wkdO00hI)
z{{U%^h(0ybWtT_r$AvB0A&9Q7(c>o_@Nzz!*N=FY;b+GG0FN4rMdF_k#c!nfnI;Jd
zw_Ki^0;pm$z&WpTxA?>1oh}I$d#h~>G0GwcA|HpA2_KDNc%$}}@O}|(?j_X2y04oY
zvNVcEJAfb#zx{gFw-;k^Fm!N~<fQg|uE%8@n=isu_SMp9F1GvhvBUW1;haAU>|pTx
zleDreyna-%AHD}h_8y|RlVX$190dVD4ZQ5mJqK^8^si^}2a5b9@fzrOisw|c(cm$p
zNet{v%eSf6fRmqJdh;I~_`|~1;z)FeZf~X8v8SCx2szF|^2hNazi`Ux)OadRzsF^N
zQ}DdPo0c-Gr||PKt%U6~wrIsRFU$scTy-S>054kl{{Z%4(~=Jk-|GbuHkJIiq#bsF
z^8@(T!a8=jD_yUc`}e6#$a(f3g?;nk{{Ru$cwgXlm46yU2(gWrM%V^M7|Hj~KJ~UL
zrPYp%9R{UyOT``}msZtew1M*w`B}&CuTnoArm~*gPc&AL1l~>+o1)|RkFXWuR^Jk2
zxMndqJKTm+aL4?4tJa?oC$~-WTr@pHbIBgn=E_lzQ|Td1y^-qvF1fVR?AFTCZ?v3(
z?s~f(=g@i%!m~U(q}^)w@mj23YR((aBjt<(_nXuY=ZfI8zZ=+HT(P;FJ2@QZjC=Z5
zt7u*~(d;xwiqYa(Wd%VTIrkpJwPzZ!mXT(%qbTbndLM?oRc@DSpx!)_+9?r3aVAR_
z1C}S#{{Yodd`IxEmp+w$bA16>Eur%d+1qPM6cHFcxZFn`wc?tWjSPwQYerPOnTo}{
z-~c+D{=E0Dx5wWU;?w>NYf);^J=BruuL4d6=G<_x?)*kcspXQA<uI$sD77QOEv*8p
zmYHpXF=UC3N-tqWbG{Pzi>GQD4Y9kpSS}$RON9djuW_GRE7Cr0(mudZ#XD{nr&^IC
z%X?Fdc|O9vA}8skS5wbgsw3Rqxaq}cNT337Rn!)2<BqjJJ~iVw&faNNGn0>VPi%z%
zan4T_EJ?L9{72T4V<wJ1a&U3QJjb`5^(rsj&nLYk+yhk-Uc|aIA8EVVdS<;hLx3Pf
zQNXSf!uH)<_peRS?RG+~j8{Ze%=!^{hU{9|o^Ve>D~a(9!A}lNDPy~qVryUG1^UF1
z!@CSlKo!V%i%|1CA#s+@7d70YqB$omPFGOzGPRU(?;G)o<S%26{IEYe6P^z>&s=Kd
zm<`xv<Ye^~X2((wCoTtl#+>2Gu4NhXJ2U5B+b`nH>t0Rp?@F1i{5horThA!q-Q0x*
zna_1nKnv5oeB<!v!S;STy41A$nHn30wi7`-e1&k>Ja-s1?mx7*!Toms0OBKRT2++V
zrKgB46Jd!-+dG}6HY27^2q%nJmwX_GZ5QG(u((o_`7=W5$L_GkMt-2tP2WR#FLvjg
z_{ZV)so?wHBT}}+1d$!oKQv4|bKB@@JttMU(PD`#r87dS^T8qD`;SkheMREIgDg_S
zlUu*aW0VF2eGPc_zwq}>xw%KvB^qgF8>5($67}Q}k_Wdn)S}WkJ))AcC~2N7g8nc_
zF+X*G<yuL6k_?8+6P#z7;<X(D^GQ?ogd216&l$@8ohqu|T*O3b#Eky{5w3{48Mvph
z=$~mIpPY^}&q}K$$MKZ&NEyNP{{R~2bd6f!ClR;;C;4|bJ5RNHH|*K)?0ir7xujU#
zMyE@%Ci_v4jusgHY#*xtKRVIROPQ#)?0deCU_WOMfRS7Lka*w53HFPZ{ySStbkAdh
zf$qk>Vbt1hwX}qgSnebqIUPSL`YZO7yVP~R6kOlmLiVk3v2CMe&zEeu@|PWf<&W09
zO5aMJ?)EbYeTOaIj&gEGp&8H8qMFj>Rd&#&W@AW#l@8dzIl)s_)&@Ah^dx~@j9v_u
ze=27Jd)H#p%FU46clYm_*0%7JR&Hdv+Pr<pmn46Mc0!Y$$DdZbX{)o(qtj=(`D2Zo
zk-<?`7s7Ihqw`%N2k&I$uN-80cCTIV_k%9A%b4J_gY2lIC7+*{x#GKvKZOWwZkc0w
z(Uc;xD8W739dTZSaP;HxGsd1eq}`yjJRT1Q+FG6SeC1aqx)1)nMXLCIE4?|<Knhgz
zUhyx%PbW{3@<|-W8kWNjqv>98bEcDY@kJYBJ3`=^Dkxh)(y2Eq$)7#y+B|l#B4w46
zc1xZ&dJn?3Zu}~BzXbUx8s%++J8WFG!^>meqYt62zZ>{@tnTIW<9L%N<=OXshOPWH
z*VfAG>{oBH7CT2jFFwCQe_Ha{?m_8u=N7V>31+sO4e_}LBh&nNHQIbA{{V?+jkJrw
z^7(fNPt$PCYW!F5)K+@)-rFjy5|X7;I0v`*eifVHm|D_UE)?zgz+4_l70oqcxk~9D
zr*;jp-T+2fly$3D0I$l*RFRMvV>R<1i~j&^7<?t9rj_8mIdy*$c`CDOT7-YNw`>fE
zqCa?!c{mk!_IsD&uZ#8RyiMZ`cK+i^xy+tj=^2#c<q@VnRT=GpUM#7#c{X|}t0U+e
zgbb&)YcA*JUU8b&xB+=Ttzhc+a9BcSxs*d5c7n(6A3!>j>^ZEH>^Gt_?$-hP@xTVK
zF75)7rzMXd0m=UW3eleK=I9(l_DD}dAk2(AaoBx%_N=><<b@c>$tRM2wX|KgGqYxN
z@)qNP>^fF$?1iLmYfeSj`NvK@>l*0g&(^wN?#$+_ij%n&>3_4&!m;re#f^1+9{m$e
zA?6t3SwDq8sjo1)Za4<M&Hb0IbQnGt*;?t^g{{V*0h<0eqy^%Ux`!P`1_f<5WOLI=
z&dlpRA>CQ&`dE%<iWS+lOoS?^#xd<)3F4b8V+4{$W086zDGHeWwYTE!O4r995N2D=
zNo}lPeUZe?^0qkoWYv!ld^Cf_T7{G_-^FjIM3`?O!xi@)^r`5s$(*^{w?o5qi!NAv
zrAH?`eznNl#<Jy<ci?+h(7q!5o-OU65?Sj6xZ`w~Bz|3g8sV?L3f$X(W!0^4Il`uM
z`BzG5qZXRn@*9gi#Hh#>$M}GKo>TJe4fln6F0<gr(^1=al`+eVkJ7w@#VT6nQ=PyX
zX^9-vt%7>yq|-H*(QgvsIH8z=GaT`W5xt6@wN_RuxsO7BI?k4GTN{9?QfQ|g4V`CH
zmsiy8Zr0x3-*DaKT1BxF=%b}vwechBlWo=SV}jNggPBSK?X+M8Rr}6*4E{A+Q1KiK
z;X8X2xtQDuw>%*fLeb-?VaaUvHO@zRjQA)C?rWt_Ly}bWFKarp>k`Hi5o=>^2)4B!
z!U^S=_jBC-6_(Dyj$7sTJk`%W798YfwMMZ<%m^7ixUAfbqYG>{A(iH&LBQuUG$)*A
z6$p%u?OC?FnxxB$<8aBvT7~n{v*J&iB-QvpBOuo-&X{OW@q1&!dZ7ORr*{5T^Yn~K
z;=X|Kc+Z8kFgw8C%DfSw8RCmv_3*mInfjVF+)-La3rTe{KnJ@p10S7!oOtHlyc6IZ
zGS1#dwDLT-)uM8Q7@<|_J?rzg;f9U>00~FNUj*q}cdWXkEbH59HuL%|ef#@WYb#-(
zY4GmBcphwLySE>pugti%^>Mg7GVa^_kEzRT3Uz8f>qF<AOXBZ}r)hxk2BaelNFQoq
z1}hGKjXp6=vA2nJEvhgO+#V~Nl0)`(<ov`8*Drf?WV~!1Gm8B$CBk82MP8JD@D=_?
z;QT>Wyz$K{jVo-Mx8{1@k@366dQIFnw%#JVTZm5hq%sF6-n`b#{{Xdb#aCRvjrA!3
z&PEV%_$aR!x0`B#nt7ZyMtWDdTQ0$3KZ!;!)i3fsVwNu#hq|X$NqV#0bU)hX;yO&Q
z-1wI2NEm>{cB|v-T{gS%^H9`nK)b)xFT-_7?;NT6g}|?ul6|fYaf4fO>W=oS&umxb
zdB+cB@~Wjyl8>f`>^wD@;$?<vtht|et39Vr_?q+2x;0B@Cx%ha=hnTC!rn8w)ihgp
zXSp-P&OvASo9VdoUm3%xJW@yV5xbvC_iuul;nZ|1yJmgU+ejuD{_bnyaoG(B)4Gm|
z$LQI9K2yXcGobyev{QAY={kgG=3FU!axwn^>#r5?ZnrL%acw=g+;Qbe8E(Yz2lKC|
zJ~!zDTKGpTzy+Gzq=1f~SIL@Z?jB93xo@8y-Rtc#II5X%uzfC%R})px#;U0ID{qN2
zTFUm<J7Amu2$CiZyRI=>wx1E4zWHvgl$YOzUUGdo>0dcrc-lpdNp2@Qpl6Or8S>a3
zr>Fo{U5CVLTd7?kohNA60kOCl_dU2Z<W$PzE3Q-7dCG&=srF@`#He02^KWg5Vpn4!
z9eBvDjvtIWiYwe&%HS+umIUm^0L}$`eWUzCjw^pMSwY*8nHcgtMOYsdp=srL3^HJ6
z2N}n|w`%aJWSrMZ=^H%Ggwk~)nf5FELbbScSBGvg*@?FU{{ZT$8h4AVuY6ymLb$kT
zu9ZBp+W^l-<CC|Oj`i}5uf$kTt=w|8uJ$FBB`cBBx3~4K<KYgwE&l+;?FQOunZUY+
z<y4&SQ@AN(=sjzmo<%G)om=gxb#n?8D7tWso6~djV@{q&vW_K_b7YXfbBc#kyiGPR
zS+>sSmCrvxQ21g+mfdDKRS}E<*bM&wg?OjNABeYB`jOM^j>+W&$iU-j{{VMAsLArB
zkE?54pO@yi*x(^4$m={w<J)^XpDiac%(={K{c6sI@j6KW7D4I05kA>y%Mctmf1Lgm
z;u<~5pc79gnB;Q3&*NKv+eD5y{#E*X@;og{YL!^r_!#FHhdy{aCVO?)#63T5A}k5&
zXBfxjSk2=hb3O#fLEtFH57YkuuU<^r?mD({&nJqjajPSrF=S}lI4loS^shbe-VU4`
zX~*w93VFSG%JMzl%Tj{if#<IA+0H9kEB^o}2HqR#Up-lP)=OpD>`SjkE!X_|*F$6R
z5W0^rjIZ~;U*%sHisAJMH52M+=3MD>9@5&>NWVO&qrO_R?z~G7ZIQ}{jN~3G%H;8r
z%-e><U^AR&HC?<%ASAP6BLf4U{{U5RETKXB$4}POwml;6#cLX_>6MQo0Z-+D-l*RA
zi7suVA@^~f1~4&RZrZy<WMaf;_`_rH=bEW6h{QIo=1z8z&q~sHdQ06$MJlPuT;8(0
z5{Gfr$Vle2wHpMzwJ3Sm>fsxs9>DuoCE+Whd2SdTz{!rCE3<|_wLbOS5DX{>?;l$F
zOs@wjaZ{P>;_)%%RN^wq9orR?_jc?FAYr$Toc9&GG>d&6*$Rc-oDML3y(*Yy^75yz
z1+m2p#7XmzGI6+fuc4^!?s#);>|N6B8rtEGR!!ucy$B!F``2Y*q`<Ml+oK+6=V<ig
z<X0UGtYiQ-?z?f+{VQ7ERghWS$eWTV6Ku0%k9^eGnc2wB)})sH=2eL|Zf%2ax4m<c
zNS0SOmlA}V?6%B=jGVbV40pv=)_hFdM3O3QJ6=?70PFOxJk>R=v0W>}=G?lmk95qj
zr?~^~ab0xjIUPza)<;wJkL;fT>UR)G;jbgiQK&h=CwJkWO7S^$X$7YD;CW!=jsgB;
zA8OyzZY7G*pxTB>AZ5vRJGsd*j*4-?$E|VN!hfS#z^r5uuJ16M7En(?*mH{YXwu|v
z=UyH1-1E;C>JnT`@~nthuFR>%{2bx4-|5o4hgsBq!{O`qB~nI>wmgR?mgR@8copgX
zD2^yCoy=H*AkDe^dSG`0wRt~@tmHl!xg@DZwvuP*)Dez<9)i6!q)$H=cG>3Iyz4!V
znxHw`6Xdr{mhJ6Xvaw~$bASiG(z=fj+PqepKEw^>>9L*=9ByJs82)Ct$t{RL%D*eV
z)m_<Q-j#}#(d14pF3h0Kh8R48GCvxIQVgV!K;#^D=B$`^7bB|U8R|}X{HmP6r6*|W
zeK^j0)fkZKdt<tIgsE_`uwOYc6pVHiSTNqj6h@@vChXvPX1dKYz&9F-8>^-<z((vs
zHx~MTjcjO|FZK<>-CF50PjVYE$#ZUI3I6ED(tXZpCEb$KI-l9=_FwTo#@%w>6Ky7m
z;kE*6&1&^FM;uCW#ys-uKp+9%uS)s{#9tGwegemJqkJRq-j}1<K<{omK`cu>{mUxz
zBS_gf>TsY8SI=$nug2OepsuZdXDMB%vnB``<SMWj9=mFnh<-V1ekPB|)#9?fB?^6|
zZB{8%9;7<+*RRsFj3V9Jxg4(gmA)<fWAV3-uH`!Kh;FqgZt%?)nsz%zNXn@^4CLdz
zc{WU~mj`hl!^VB<qt%f$R9Rz?nIclmvJgn^!*|K-d8{d;D?gVq81aqUha`PF5n0V^
zpjeXERhHFCs8RB45~mIZdt<5X%_XL(acmQ7Aq9hsfJkp&u03iH<`rfMk-r>_c07Aj
z4&!b}P`nNo2mb)oQFco-{PelG{nq)MmOmivjtE~|^Tkw-?no8VIUPL3B0x9?o|RZQ
zMjHV_JCn%w%`W_&x$X6#Jw}Tx>dd)6!ZWw7eZA^7kjVS8bKf-*oN{_oiyot;XpInt
zKo1<(ckpGm_`IP52HF1rn63!@F<sZef(MOMbj`c}05MxFEh7v<y%{ef*YzkolaMja
zE9h_8oiqq^eO_&9+}%wZ#$t#lZQzbUC#85kvj*lw1Me`eR`^=2W2Nd@92r7>OI@(E
z^ttiwHfZ-<Tg843)pTpUCLal1HKdX&ylF5B#B|OppRm!q2_bct!0R49=b3fT`;PUU
zpb)S;r>hb8)F0Z)H8<ns+(mR!qMLEl?Wu&k6-e&R5B7@C{5hofZ%)6_bfnWKOOqYI
ziHSx+K2y)3uK*X3vj8!k_4Mz?eJOR11L`p%ZvsC!5$ZY3e3fG1&pm0?ih@?y@?lR`
z8Saj*&f*mVCl&7B4ei$J!uoP43X<J1F~9K*!<_aN;65L>Tbt-2e1Ps0k6vr_BjA_p
z;o;lQhq`>1w(Bg-7-0U#wPNb3YV6LXDO75l7N<*lu2|UW5SF{Ngxl%+W8WYOa=0u{
z<VAdC@ebW}jb`TPy4@Lm^QwS2C!7wMueQ8R`#Wh`_M7%Ew6th3w3Hiw9R}{Cdsh`V
z?5P<mAKIX}jehIMgg4|Xo|S4)*5>n}T1jYp$Kfpk%U$r*@1BNByKkKUJZ<Tc4_~0K
zRPe{aAMJk+S?U^1xJ#rv<P7*PfxsE*_}8XI`zl*Tt&7w`f4Zat>s>#G{{Uy(O$>aR
zoy=tNe5N30^P*Mh?wHmZt9EGlpW_TMMXTCG%43>1LP((S8`%2SU+|YqyKez_f@q-<
zO)Ogg3P^5xepQA0M|gHG3HZhfTW5KrFoz;Tlb-dTru;?KXVUHMZ0B1wglyevJgoX|
z1HZYdnsc*>5nQ)Vq3VAQAzcGaxdv7HRm4pggq&jl=M}`g%3IFJ#Hm)=P;h@*lS#JK
zb)b!LsYy0RCp@>**Fhh{C^aTB++9yAvB76zPfT;`S=m9h-sZ{@npzmvdiB2bjyY33
zKJ|3o9q|pmo#HJv@uorL$^pZ6JQMoV^86QVHZ{4^+Y%}qx*Qy1r{h%gzXL;MB26u^
z45SdC@;!;Ee64O*EaL5Bd&9AX)^)v@X#CbO?K}~hmtL7|w8=UaVT@<BUeSC!*P5aj
zqc-Z;2LzlB^<U%FmYv~EGfG<-=4+XmywM@blmoA~K9r*+&ln?)y&1T=oe}d-$FCE?
z7ltkLAGD-WYPTw^3F#V~l26wa^LL2+Q1`Ygqu4&%1%2`|@#jBGpJCWmm&Q$MUle%G
zM)M4d3W(SHsG#wWa&ugoz+ysJfN|^7>t2d&D;d+KqZG<<qQZIQvy7h9o_Q|Z@_L`?
zOFUso#Gs$<uUahalXe(BJxyx68tJ>-yJpPJI49}eyGu*2?CX6nIUKWY9oLL=T&<3g
zaWTcU!k2IPXCd;*{7)6u*!X8#)1xw8==Vz#jETf0<3I2v$f~+pn7?{fE<U1eWO&ht
z3VYKS=cwb;)}xHZ<d~3|2M1`${Bc(xQ6XP2g&F90{&Z;d3)q_0CUpc4n><$RDYPj)
zNc^a>)8kmdKpWf~{{SkwuoBq8KT3%m6t3AdtL?z==~|FJQ#F|!WpRxAj8({AV~mWF
zalot^J&H`U=m^QI`O_{)A4=2%dgON<YP692y~Rz9-lms@qsuTO8HN7<opkzqhB@rz
zRr!)$KRV}pIQ~?DwtT4@g58a5X!jR4daaebE4WD@BOjSRm2FN=F6Ytmrm*m*t8ev3
zv)f#Gh8a|C4aq0puj(t9@wM|M+zcBVi$@?MsKBn5ONY#jbIL2eap|5bp7G6uMo7V9
zy=G8X)G+@50bLVm!Cd{*_lg{}ou~f*1zhvrA0kI(YcN0RoThWsb*$|}Q-({%R+Mho
ztU)~i8Kn5Lc)DJV5o{HA$0x6~d6$eehlWcl==hdsR1iCWzsjXuB+uraEswU$s`{Ny
zvmv;0>ex^)zSW2>{I@*x<E}I9R4x2K(%Yz!vgBmusNHI1r9FY~Ly_>_$?SBieZf-S
z=imG)#CJqIWlrJ>@$H)AA-`DylEtyypVqGpMUA*|g#&?~=Tj%6L*-FvzK0(T{{V;d
zUx*jauH4UWr^u5GTBtJb=V0f63F>_-%ysV)S-kQTB1H^#g*^pS_`|4q{x7*;NCDYE
zJ-P2(6prc`=PT>$U!7J-`;BdL^$Ilg_OW}UZ_DB$$ACv{lg(Sxyh<Q&Bqh}KTH{2F
z-mRLI-Ju{z3KB9i*A)<@N<U>C6LV_HM{TP3sWzmM{EC2tjT;!xbKKT#kBV<&onwmO
zhIuMU;QIdnPkQFb`7+0Y&QDrpu`O~JCprAjYSD7v-H8&5ard3ir!);+&xUN~X$+fz
zQu*#Gz0LfSNN{ku$sYBcX)UBWHG~#IXN4`}-{&9sYFGD|9-F!!N@GaC9F4xhzRH8O
zG@0-f685sUb4uRq<d(=7z{hWT_bor-PN6S@<I;q?p-937&k8ZX9^8Jl^WyKCNa8?Q
z*np}ANk2;T*>qdI4?}`mgee@Va7KRc>OQq~D#9tLJ$ecnRf1hvb}tdVvpa4K6#eXW
z?Mrv!X|7^bXDk8EMRAL8GRn$!`yQV4I?J>u{KtCv8r55D{R)LR#wSm0tFUm{KTP%d
z*F&XvrtvP}W{?padlx4s2e387+NdE%anD{WO5L+EkfV+=YG~AINOj_zS?E3+@f_NR
ziR=x_hew%C&@iEMjDMcB>^~fr%lKj9G+4}Q<ZK8|8#{r|<zGDbTHv>c?9>gz%^3r(
zGwEMJ{7aZ>Uk!XjrHd=u;4>UxjDEHBl<w@$ki@jEkB1S8=jO&vahfV`C31IQ8Y`~n
z0@(dg34m#KVy5!?{V8^PcCU;+_@^vel>I6g?e_e{XFOuG8Z6*?)0Q4T8cZxyTXN?Z
zH9AGULB|87JtH9V+NM;%HJ!j|TvRi1aaKsG()`i@D_Cwsj=RE^`{G_J(CySiZZp_d
zE8!bZw($}>3hJ((%CnIADX#iR#!o}QzAbpE<MFkPnh}kJxc4K9&%V_R{so>!JnR{+
zL*ibhZhj=(IX^5;iJG(Go71NFIpbbKDIELPrOTn^w(faF{-R@_DrxrG*DHcXIjlWK
zaMKbFDztNz=s*>tifYWsJ07v%ePOS4n@9r!Op4eit`2MANG?X5`#>;k0Q!Wh&-+Z(
z_9w$#L(B1oo0j0q+v#5pc#`|b__c9-q+&fqn4fPj^ZM5sk1a#cgt@7-&v5aJhG0h^
zfGdO46Sy_i>o>yfTL+BrYn`4w>Fr*k*z*h-;~N+-45xw!$gXEq@VrxQlGq4j{_7u;
z{A;b3<YR&D%~W;IZi5tZjWd(C(Py$}l1B{V1yxVVJxQ;>{{UiNhnD^l@RZuVz2(_l
z>$@&)8f+<!*>mMblg?b_i1e?COll<zdS}+X+rj?;6?FBQIdyA^8)(T_JG%fqf#aqs
zCmv;D<33k*d(X!U{R>&uZ#7B0yQ{f3EO0`_(m2!&xg3q$5$S?OeC9kwX>WAG8!6V@
zF4Y1beuF&YwtH8g>VF!%JN719Rg4b5H&dT#^8It-R-DrxCf+mEgJboprxg@sRk^#^
z=&ig}17Si;fQ;kGY=NGY(CK=l1d=3mQ_f5uG5qVsm*Sq#;b2!e7zE?BTZ7{J1dK_i
zqu};9uF7>(I-E6d?x#cOy$8j=*$551tPVDu!a7%|_$S6NT)Ql<ER7)O(O?hCzA%g9
zG%-l5EPD3d27gN5(f<HxbELjv(!&w!COQ18Z)+F3IGz@9*&kDUM)6pJGY^{Zp~2f6
zWwTx}f8w}p?iqs4Wd|ULz!m3yCjFtVHP-W_)BLd@4j0WQ1EA^Cr%L57J}iltk#3{h
zvB)$^q-<R*JmV|K?fhZlDZEO>TGj=?!8ZhwbJX!(QK;HAjleEB4Z&J2ymHqr#n6&_
zkJ7WI@e16NxDdqSAXgi+*z};Ci`t~)R?;Ma`hnD(<NTWB>|`)ol?M$I*1O8VolZdL
zN2O#=#u?Zil6Ca|06DBAbwg9@PuZ*CuZ4B(2m5~eN{a7Pj^)t2=XwG;Q@5u|`X=V;
z-oh4=SB~BUURj7$QC~BD$vz|yS!x<o#FuGAhzD#8b6-c>=+^Siw|5!3oe3(^fwn*Q
zSD^a!_pcKXbt^N`rK&Sz)i3o$AK7+dC<({g^#v-Y{<+1&H`TU}t!BQ99i7u%O!8je
z6~Vcjg3IiKq7P*Q>T9mMkSu$$TcJJcp41<Ewhl4tSf|+C#K*XeoRi-;2l>@0ZTAs?
zSajpkwAL>=K%jDeO6auD4Sk{_SxXQya4U+wAyal|p@+jtYUiA47UmFgJxz1BBN_Fs
z`(Coybs~YsL7&pNtws{8e?P5v;cF?YoOqf}48go&V<2Oi`ZM-H&^$bRF|BIS+bsIQ
zyLe@DkirlK)%qNhUpZX@Lmq=1``6TevvcVBSHjy_V<9f?<&@l&$!B$Jen4Wqh(p?|
zA3avDu%9*3Jp)tleusAcbXpypLpcYKbNE&b*pEe!q*`nT{l_(-sragU31EWR7HFjz
z<n{IaD>@;2cYPB^13=j<$N5*O3>;Ftk>R|>$=VvT6HxIX!SgiGI}A#H!m#f3t6>>u
z(-`D`&%+AzyUPmznUVn^6~Oq)=GrHY-rn9Qn2r2(Riv=cdL(pD80p)Yo^9ftcFh9;
zpg}Rlb2RS5@~;~4=ZYDu+7z?1u~`{oQ3N0oN$HyOKZrX0zM8MS{%)H)n)$EBej=Ax
z@y~~B#u5W|93b``O?A8qjW;%q2<E=m5U9oK%i2}Fh1ZGftTkKFrr6Ii$1Twr*tx+q
z_XVfGAAtV=3p_BsA@Lpcwf%>dvwf#hy4@R$a8(zt(+0kJ{hj_8cst^+#A}@|#lrGX
z>P@Y+yA+N_87w4W$m1BtTKYfZ2aNs+e#-tHvC{Q@63rx<7W*chYiKTJbQ~5N$-vKZ
z-oAdP#G;R`ihR7x`D5b0?bYzE8|gG}0DL9VwRlq~x7v&;ZmsU9qjo*cE6C^lt1n>A
z&xigKgLfx-?#KQ;UpRQbPO#QIW31}-v0q1~U0zExDDZ43jKqL($K_Po`$*G3EoU_3
z^-asoY_9>hsoA66H80x-RvAIzj|#%SclS0skMrqV)}#Aa>oZ2kEHv9?9Bs9+0Dn5<
zbT1a^cQ&FZd_OGj8RRkt$KnMSzACWENi;(XpW*WGN8yT=@6@D{dloOhX-#TH-S&80
zTZ{{39R78lpR~TO1CcJBY~!vSG5n2Y-Op#KDe`pbLGC=hm5em&eeN!7+<$e5u5Md1
zb;|l4x8RT3fiCr%Xf+=R>DM-lz(`a`8y?s-_ve6oH}IFn9vy=E+S&v&AZLnFRgmMF
z{Ik&7O(OYaga|^g{tngj=fcmA+D4Ua9lonHFtUxMz}b;iYoa8z(ChyIX#ERB@Z;)G
zXfiWdwYz+*asl?QmJ(R^J2>YY4E6ONQ@|gIuW$Ib<Lhr2+TD#7H`Hy`HZx5GWTZ*l
zliI$2ymA2`<8@q5M+)s0dY9~RryXPVi_>RiRu&f57a7Jz>;xaHSJI!hZmuHnG<s-m
zE#ZzZJutuhb6#8j00ivt@7w+*Ynq#=A{e5BBznH~{T8hL)BYuGYvR_Zn~7ND#&ggP
z)%mt-A8(&WwK)E4`g&{G<??^b{{XMZ^2_U=v)X4l$2H5qeA|{ifvp{8=(K4bdYqc2
zVw+=TJ@RYyOr}4%N8x$pmM(A5xDFKa-kdT)=e0BdBNV|V4_?)_D-+2+ShIr-m&&y_
zQK3P=2i~m-34nIRP!eQgzG`7F?vFL|xW~?}eb!4KWVua$5crl&UL7(c3@{5jVQdUb
zg5VtPKBm1d_Hotrc*SiUzFU}?8G9Z$uN9g1V4Che3Tnc`$GTxBmT-eUroM8ldFeR!
zKGuxAooLA)aq1dy)&3dkwnP?YwQrZcS3mt~_=?bZUSMttF-H9Oue&@8GU^&erEb_Q
z9L|mObjYuTHRU>AiS=7hF}6lvzwWUWjzM(mQGFTpxr_H{)m={#@t=w%O-la&ONm{i
z`$Mxc9I0Fm-nEIOtVyO^Jkd<iDhWK*U>6zMfS#+82+y@=c<SUYi*<t8EM?$>&T;&!
z0t>c|8P34bIS77kM}9vl`lm@I(nrMA#L3R`=yc2D8>J4SP1{(>y|P;XbkAW{U&dZs
zO7XK{lqpdCd*qK?*AYFzd;x#}#~n}Bp=k&iG`);_9N$~0v`qB-KN&p40IVc+2ood)
z!RHz4_4KcC{g}LKE%(Pi3PB@@S9seY1UBBd>(lFBA;mkZjjhQ&Nw0JLk#2n7k3Jb@
z&e1m{b{MW~RT%pTq`E)I^zb;T&a@)mRDRQVVttV;Om>;!V1R#kF*y8GiuoVo=BaU}
zc;q`X1ezi95zvFpeG#L0p4b~(1)N8D<t*@$K2eUjK8Jusd57&e;7R;lpvAB0mab;D
za`HsKGl7qpaqZMs$?(Nhg~jq!(&rZ!U@BI^%2f5Uzub8Ko8qaOAUyI(=LZ6{6U7sN
zKY7J@U6tuo10)c6-NDUi#i+JE_AB*SL0tUD4;LsS(4WN;u)#ZrdWE&Z{RtWMJu8b{
zN&SCXk#!sqn&q5VQ`Chy^g83LGn0ui^5dUg+}5qGw9a?#<lyZGr>$_2U1S5ENa}i4
zjjhMZxg#UJc(|%jneE}+lGN`Z)(POT3OO9rv937zatIwOhmQXMj25YzP%NEzJ!{3D
z8YuMRCUr7vO(cPXl5$Txk?cJMX6kyle2il}9@UU9e7MWxs*hg&!m(%6{OOKK!*s#z
z+OztyJLt)_k7xLHTxhYjK`kNY+coGmzy^;TWRZ_b;k*rI)|w@xB;m>e;Pxk+S92L-
z51IhokXJls728glQ9Wr%wrWFaV}uMaY;X^8>rb$-YwbcCaVkPs<z$nfBCJ^kxU)9|
z23{CRf)x7T`**3NuzPz*^*C@a_{^P;r7lOzQ<7@Pa!YQ$Ah?Un7WS%2Z^t;t%tksA
zYd28w4Yrt?lopng&u?<LiZXa<3hm1i&~uuq#752>kXt26pS%g=eRJNH^6Fd763R)H
z+_v)>Ok*f;0Ub%L6l_Y3w>TTfWVf`ohDXB!P;fFa)Z@~*=d!ibCJCrTb3MF-gB6<u
zk+*!R10H~r&2&B@S!}gPZSP}vCX6UMhu#O~0D9)Jn&MkRb##hHu(Fw)xm*cfKO{h&
z2`2>lSA8{W9c}x%99F-l-FSZHeLChDtu-4t40j8>9IsHS-7$=EI@c58JA>srJD)VJ
zA7g#e56Tw1ZytD>{{Y3FA)isy;)>q%szY%cN<tE&1COWjt`Aa$WLEjYtLJo?!;m=y
z4_fr&rsdObm2+-!z9Z9`^Gmplw1_*A^Kp=P_a{7dt{>tRm)U$Vsg=sHw^-SB{{Rsn
zmE*P#=U$I-r`%uNJ-x8|Mbx+>Dt`GqfsWYi>s|hh`#ad`R{sEL@nnF+K>q-VY)5G0
zrt|6SDNdwftjMVqG|!7QOXT}Z3o`K)tXAPA+*E<rARM1yF^b8a&ieW-*4EO<sVhjo
zxx)kX9mk;buh0m79e6WEpMI6$jXpA`3oXokV{z%4eENI}q$RvD4@8MneuA*}(OVo7
z!%9-SXTkax!Y>nel1!$bBybVA?j}B3s3$qW<o!C=(f<Glz7Y5u!qzEwci~%q7;5p2
z?)v@Aky?U3xe5+{;7oJrUZryq=X0ny$3ewcl0a00jib<1O0$sG3O!lgT6}ufJSC^X
z(flj$x@qb@U+mjK!uw<zzoUN9z6J5j;ycS9_(Sz$!cHNKMi0IQKb3ivrl)hK$XeER
zkBn{3N9A0XjkGI&5i3vSEw%Xhw??>Gexuf*C{5XlSYB0lp40IzZ9d}2S$NyRnoo;-
zJ2=dGgb*sufb?MN%Kb%rSNlbN%}Ma?&3s+p>zTE`1l+j=rRL_sL;Mopay=WTt$RPh
zPl-30UCf>m*3X-#PmygTyoem}j(>#oHP~9sZ=uV5q20!^cz0Tf?dFA!L{HF<s-B*r
zx*(h#jtN1t^MTn&`^*b)Mmki{Mm03pS2Lj{TRgh=ABQ9w>%K1d0lWp``|A|Jw-U0%
zSJOHCv5sqqf_8>@<z7lK;GVecU2fJn?P9gm%_Gdta?IJu>rllK$s#0<u`hRFkO$#b
z?vX$YFi9nFeLt_(s+1v#QGtwBR=Jx*StRyfOnTDf4xPGYmkO+UU>>8cJ*hK;k=HzC
zrEVyS0c`ZmCpiPq(qxQ|nWI0QXlPx8CB~fN{#wM_hq$fJhT4eIyjy8IACr}H=ngBG
zMgU;<_pXP*7V+I{q8DG4qHLc-k<BZzsl(dMQ}3bljM7Iwqi@}E4o~M^m*FNhnjbtF
zoXCF~@_z}(_JoAY;o2l9#CIcre;V}5EmGXwTiYba5;X;|*{@dz2Ja)}F){X4{fpf*
zxxCcIjiEy+IP)b{2i~w(QfZsb&q+>k=~+HM@f25<nnjhl#w9G=^{DIvc~h$$pbY%S
z9M`!@F^X*WZs$eD3jD2pM}Iu)eWkP#oTP|UgWM5{`5(j@EBJ53w{|QzM>`4br=@*6
zrCcTUnnFyj(fqBn<kyURXRr&Y+THLaBg=1II#(Td`?6h+pT#E%>P=}Qo$%8S?K?7m
z=c@i3*X>8b6x;CkNXAF*6#kX?pW)Xc`q_u~!ToFYx8Wld_;;c<LCvxs&{u<ql2A>N
zNiS?iuj@1EH;tK{K{*3=YReurnIztlm(Lhsn!n@QQ*&<epvK-!c(j`FwY+qZkF`W0
z8Ka@rP|)@<<JT$!0nfEaJZi`IhCRi2>{{uN?lN)SuBNmd0M182M5#p-u@^j__N}@)
zhmAD((YZF@^u~GhuQAc5Ej%bsmc)Qa^seXQ^qzjJrb@$s5mHBPd)Eh}NCngYg$Hmu
zo0{}CiaeXAH*?)QEoN=iNqyjeeqn?3uSC*+cmkZ`BZFKnhlyj00FtT!_BGt;5jg>r
z?jx@?!%}H(dX#K;kSJJ@dl8N)8%v3Xe3<(6HJxvKRTLcO+qGL=N9GC*aLQKMoqp>z
zblBszaj+=Q2WYP){jTjbJO2O+X@A(-WZIq5yC;`#jldDXIU|hMr@?Xi!S(B2eev(&
zu7&X4-$|QYX_6Vs1bAZKjCId6s7d=|%rPlSi;3{PkL=I!AH*{=%cHHKk+&Hx=2Zj!
z`N^)!P5p~*l@d=8c&6IPN#-jkjE;S|{{RZ?yj}ZD{6DyKy|wXHu(Bxs0FO%u{!)Fx
zzz6ZKDA#{$Ng`$%9j>75$@8ygR44FJpVqyb)sf`#=z8vl`yTir+A}rJ_O<*<M)|KL
zTUfs%wz^F}_I~(ZXv53W^y`p*Vp-mPJ=h$6wekMB`$%}(SeJR!;<Y30vY>Zg_!E+A
zi`G6Nc#B+C`*(<Mr$*|ZXkfqh%HUGvsn$huc^`4>o*Vc*;G_17lcMQz$I>^PGX5wx
z8qL(dYj1~|9im#<-2JK~7!g@oF^rFF;C^Dh3~8<R?rqXB>Q8^f^Go*9c#bq3@<!wO
z)~YG)?k{5Q?CgDSss7j+5O+F1gXOuKB$srC0Dij^n&P$Jj=nJQfDzefI!2KkoCk`?
zO)>gAj=w=(d+?V+{@m79YuR0GA^Ybz>*zXWzJs!#L(>`?Ee^o6ufyF=#oHCr1eUfr
zswlyz^*pmi@K=hpIU45X;^N!Q`|?isC+MU90Igo74v=Dj6q4+U0b|skTIe4DXB_u6
zUFBThk&056G2Kp=F1IlzvjFrN?e(d2{{XaET}M5;7-x)!jnv?&>_vJ9gnk$5ek+MC
zQrY9OZUpj;h}{1G2|s)AuV3)*?CKl<@(1hwt}?^(HN{gCP8v3Pw5Uz$eAL#L6G-mJ
z?;{<=?dnfc`O_uSF0xbX@4q-JjzO=pw7=QH&eLnY;=?#Df1s_1KWF_kVD0=s#d5rC
zHb+{aC%Zl`()4?{ZLVT5kYrq->yj&_@CKdsYm1rW3@$j1HDQtg>P3A|aqtSt-Zo`)
zjTCScWRInDC*cj1wy2XuF?i5oRZ)V({{RhDJY*#KPpQfuvBlA?HFXRAKjd=$5Yowl
zYl&2dyv>oTak)oR_?qMVaVpELTu7uW5jZ5DP&xkq3ihju3u`jSyto)u9aL92seC-r
zwV5r2%vS}BVB#hKWRsKiuF2x#7~P+wOEp+!D*UcZ?2nRuJlfprI#!k}V=|*k(lH$e
z72{Vr7P8C?@XIFNGRL6k0QBa+Q1q{WIwqq628GICU`fqrU4G2|9G)gqYQJ{?f$dyN
zF^1MZ5zO%QF_?+gZq2ju$soCgFkKl~@yOz<Pb7g`Ds%5&qTVI_k+gYZl))=PM?le%
z2e%*2yaUHSvpl+W)QP9xSYJvGJd}_tlj*gw&20>AIrTgl<<(=VS3YNnr7|E!!UM_a
z&$Vnf`j!VOFi$)Kiqi2{gs;3Ic3{<X`Rt|xAWDTZ{(Y+kOYb%*5!(Q!RIui{)kM!O
zwHfxY;9nCYl5Y}TWDNPGln+7%YnGNXAQ;ZxoY!6P8saTS#P_B&WO$#;U`KJ+9@Uox
zl6g$chJ19#L0=16?rz=L`VARcoNw6~aawPHlmyA1K9sjml8DZxMkA6>QC$|9Zm5bj
zjnaI%C$)82x5MdeY*J4vDIf%7V4U|KO7d#r<3?&T(Db6B;(288Zh0U0PfT<Mou)#~
zak+>ASOfU_*I#+yi|B16n84u^jsgDwCb?ynTM1c5%1_?o)B4v|9}x<#+;=dm4|Sv4
z^i_ZCXb><bZgNLIh^d+Y;43y5dhie9UeEBu_FKO34}&yKS5WZ<uAXl#<DG4i-K3B!
zbjip7;~ZD1X+N?jha;5Fcj8Oiv`|mkqeOtY$3on;K9%=yzbbY=Dyxd43bKoqsk6ZT
z4)|SlzZ~i2R9PW_nV&fW3OoDPt!olNb)xEa6Dl||d8o&+=cl2r*5l#*m%|+b{(TEf
zG2A>Q8Jbe=SL)o?j$P_)D=*FF8F<J&YvpslX&R7mJZWO*R<e~Xk;>j_amK`B7~|_y
z{DB?-Y<ddS*KShYW=}FXNd73!K=#dZvR~&TABVkVB-~SH>Um{al^~qWcnFLGk;Q1)
zxW;&6p7qSz-DR*%Y};I@3Uh;wzV&n|$;}>yDy(A$Z}@8DMdC|2Fi8_2Byo|N_kCt(
z@iv!zrA=_w_LF>s#6E4mL%=ojzMFDw@8X*U!aAIE&JKFlt$)H}_Oy&f-b$bDrbzs2
z?J)Elm5<MHn2R>dar-^#B%Es(pl6_s4Hc+$Y4;K5T(%ETxT3uus(Bta*5-Yox0+qK
zIK@3jJkSRf`HODn>*ERx&=lsMigJ#$(8$(28gX2o!kk7Z9+a(cxeI)`H3hPH*F21j
z)rrm*1Fa^Co^ATCT)3{7D;Y`Mo~Pm4e=g2E4#KYMl6e|j?0CYHPST^<YJ_@Le~qq1
zzlX0MIK)hI_*bD#TE;aVjQGpMcZU1K_tPJnea8prYo+*#G2vTcPv5{J@UA^iovD{O
zIgq!fdhLEEFkc7h;0~+^KD<|;`W{)}_wo(I=bE5qIO|D&FYegqkVq!8o_&e{IW?;>
zNb0;PsT(g4SjrD8D91JN*Nc&t#X5_X`Ih0onKktupK?<A*~$B|DLM4beCgv!ntfkU
zg@zIvxk4cAmgo5j=cugHJ1EQDXzZfY8sgF^81U$?`HExDxp=G%4*vkm(IFZ6qZ={!
z@m3b^Cl%WrQaI(KWY=fCQ`}E#nn^<pcRgyek`Em!gbYWwAB|^f6BUqP_T!%Qb)UHF
z=~>#UF^yS9N}QURXpJI3x6GUpS$nA`nvJfSKn8bX86%9;G7+9nUX-JbNFe)mtT~Q^
ztXYCXxROV=v8!ux6#d*|IOnI+^r|d=&Pf<zameTCPe_{wf_E<&ZoijP#YZstTiCN5
z$xZ04O1J>?*XvR<EJ_uhmhaN4No#kfGA+a=0_l()Hj=;Br86?8-Wzk&)u}DfGKIG_
zh4K-zXi_^ds9td{aG;Jlb*h%ulH8oejO6uPX9L!+!=+rV35<|?;+v~teVcnCs}4+o
zatCkzy;?{jl41e_+zz<(6><$0;o}=4X5GlZ{*|X6g(oWv?Urm~<lupxDkX?Zc3Pn)
zvM}eiWdYf8F~IikS@&+L_Qzns&U5NGJpTZnTJ%Z0GSFMFE0S``8=QM%>0ED$>^{SB
zZi&cmop44y>S0+_m63C!Dr)TZU)g7KYx5}OyuHACWZ?e*pKAK(U2R>x8;+gp<UiRm
zXyfrz$lSZbfC1=1udB5w;kA9z%^W1>3RnWiz5(m_R}LmE&qGQs#J3lJE98(dyN~N!
z-R<h&Gd9KPmN@)B`qg6M^_`fa`AA1B+>`$R0sjCW%Ccv=j^0l&eA5HWj&euS)9}S%
z7#q2r2ZYuQLsLb#Ep2a~_0Fd_$56d%)h~QWp?GIcHu`Y;Ma!<%xSt&lqWag*$5zD0
z4B5_pzl|;Sw<O?hEtKic>0A|Xla!lXPg0gH9H~p8x#B%S3!AwLDfb!gTu!E@LVmpc
zDp_xi?noehCpBiq#Og3y6TU~=l?rjsy>!r{>=vx#uZZ^T@}rNxPu=$iJ!{!M1Zyp4
z;H!&ImRpu~Ft2q&dJl4PYsob@nki(Dx2Wi7(*7ygd>qqt-xlhLJ>`y_rv__lY;R_1
z5e75I@pbyw)8Ho<w&k($dA%x?6je2?&!udI-J2+AWM#%lZg(F+U1p6tDzvv@ZKDI^
zj=n;_{kpV02gCYqv#I!NU$fIS1Puh_!zOoh2M>|?R63vSODBvqBWtCl#ifcwBoj;s
zb}RuKx{meqwBWJ2L8IfVPOeoIQlun(cd2TY@=9Y_kTG4#yT1Pb(!9UMH_dCPPSPEr
zfZd(|U#)SP4feIFX*M(87tE*RJFv<R@)XaDn$xd_BZNrPT_kPhDC$7sx}k<pl8Sdc
z%Gmj0VC3Ut#QrAuw%&bqcZPFvK_N(7C?3`0c2_4*_?6-E;|I=Tf!l-4dPjwHdv6zN
zT9lfsuLYfqCRZR}pFnHO;<`ybC0WUB8B2IP!x233gN|zPjVcm^OUdZTV%14*evbae
z-w0bm(e6B9eiYp5YEsb{08$;tEuO`{m3-~{czjqsC-K~RCW+vv?yq#9U~OLN(|m^~
z?#53WLHgIZe$Kuv*8Dl}5>0O6{(H--*&01IJGU(q5hob~(>=S_f%xA0P}Frx%}Y|2
zd+Uj0%iLUW3G`9v=mmTPZsyxRU6hi7Nclfg@RyFfO=x1d)5}}`0C1>J<yd3jcZtaX
zbYuh6kzZ2Bs5uzrN2V%hu3e-9JXb?^+|5bbL+4)({3`LLkH75pQN_Q4OA6bz_%GtN
zlg)crH$V5Ac5CR}J~VY*+kX>UQOOsaBSssCJKTRNhS2$>c-DjP=J!O=nq5~+iRI@D
z3XVl{x~Ib!^~X06X%7ZYcWxVh99PoUdXz9(yG(*zs&J%mYk^;g)}cIsX$r1RQZbL`
zOK?q|0^bW{lTc`+RksV#$R{;hQ}|&Hn!-t9N!R<szt+8PR@YkgS7f&U5_#)bH*G!0
z<uDHzCYnm+^D<aIE<vDtCfEEeYvH>nmiEp_8tx&rlo8gxY`!=Ox!^5(hlwm2KMv|l
zcAyZk+v~-6ZlS{-cmki=Ry`~$e60J2{t3^diw^|&vt4z-`$OT62jxINsjr{>b*!Hd
zc+*+b8155XNZ|Wt1Nm3eU$f*x;ot0UtZHwFrk*yFaqGF4Ab;b+zC@9kV3E3#I{yGV
z{O!luW>}itnvdLlUM2EjDO3Km{{V+GF5k?ViR3Q=rbDvbK<`y{mIR*Wm8IyTwSJ#~
ze$Nb<_*QWqd}~OvB;7qKCX+esQApWy=qZ7?ub;$JZmVb5V5l#uDD@>n^JhF%;Qr|1
zswu(FYGr(>6_rVIMC`&-bmT`QoD9~Ng!Kcb_>)bXHr7joIsX8G3dfQ7BAX1RDHXvx
zm4OxI@}EGSsh^|23AGU}t*XPg5p!)F!uypZ8u=gNb(zt=DC%(q+8S3m^#}g|uU_Z-
zJovNwRdj2XVq}aPIftlWl{Mv`wKb2~ej!aKC63;{gmwJZxpK(5d6eVQNc$X3s%5d7
z=+W^n#1-4(yYhZqzeDd_?DMqGx$%$y`BV-7!0BBV#H3{LHOnyj#<^nLX1>Nt-giGe
zs{Q6Xy&{H@!mlLs6(39*x26&H_=)M7%|`6!e(G$K>^CVM_C^Xi@m{(6EY58{GHCOF
zL4-S#=mMT=fxl$?3k-sKfu6PAe$7AfQ|;Vy`-ynZ`xL9gM;{}jE~@8IN$IKj8LfDc
z<gm25hE_@K$Y{!(VBj2m4R`+l5B@M&c!ma!DPxaQL9**kxcQ8urbzVkBavPdE``>G
z516IKGQm%N)yHcQxS3~$SGArqxg|y`f<4ElTJWWo!n|~RvX0FA3fQ!z%O!Zd&vW>L
z@S8#LRI}<{8eykjlbLk+SPZvvJ7T<RQSf%Pq{_)Xl%NjkGAfApudeCR#FsM1cmO3?
zmy?e96{r6I3&{&AT_waM_KC!+`E{>KjcjHfDixPEne$chj2&j@MqIu|JBe-c<|CjT
zaw!*3hXePR_Qh=f0K(4G^#H{$Ezh*j?IhGs;&|m7RkZm}B(f^tR`SHkeGXXVQi9Rk
zhT8B*{{SQJ$ME2a(uzr>I7N3o!mmo9{lEK?%8&t$HmC&twU+)RniH_f{{Sx~l<gJH
z+mWimIBs@OSzSmsKG-AbT2{qad+Qq&8)a+RhiJwSel^MI@#<R3kF!`@-3I4yQb*=%
zrLp*TuH6+ELMtRAfbW*a&{l2<?rx~q&%V}L=a(aBIT5Ey>HG)b8;wfgH4Dg!S)m0H
zy2#zX3fa*<9oT6lWw|EqQG=G~KML+GR0Il$2OCdiUr)xOQ0Wt{l{FW3bb3|3$v|$U
zPB`t{kEiKf6bPavF5>|(Li6+|HI1h_M;`F&xnt@_Ph5MN(uMrU08$DMnMOE4=zg^=
zSn0jl*XZ-w>E0N)h_ZRpD=SJeV_fsla0&fQVqNO7ojEL2xQPHSxv_zag<<{fwT~tu
zjIzqgScY*Pb^<*|PAat8ovN7z3@B3tgYtdXJx9G%RmP%|Ivob(H8>~Ht?ZCqT@xqU
z<Za$WKypI$$nEvbVcd9{$HP7%{>hr!OZgxkOcI75KJ1WBqi3n~tQ)TqY`e3P*`%{*
z)wYz7lK%khdxKo{%iQ1Gti{wiA={IN<o)dWir<sDtYLQVv439IFFZwid2@XnNpot;
z7nd6yp?Uy4$E|00v&8orj*z-i+Q$qd%PS;KLg5Z_Q@K6OS6KYHCR8%I%6Ew*Z<<f<
z1MgD9;qx+~+Uc|t&n@(-x{1@v8zgX=?X=p})#ce=v#hvnk(`hKz`->`O7MQCtrRfZ
zt1m{on}t#Q$@*8VLE-C2lXS?XK;Z}7Pw}Tu=BkjJRTLe?M{i26ZuTuB2S)I=o#6<#
zK%(~ifF!z*0N?J9PIJ<obvc?M%<U%Z=RGt26)oBFL$=oU1mJ)As|4IjsoyQ7g88yA
z=1z+k?rSFQX&Cin%b6we$IT!ex>Z})iW9~TJJ#&>ap;zwN)~H^GV+nhKhSzrdy7r5
zFPd`1WbW<lijGqmO7BCFxl@zZ1J<*qY-b-WZ0eAekg!rY!K_<>f=_YZHKS!*qcd~g
zj=8GSO_Vu5FM5*VKmk0SKZmF1Sr>O=BaHO)=8|IdI4>7l{f|}?C(E_B1|Em6>&<jF
zULbu|J(EaTm@95>PT~AtTEg*!qHH+?<BG$u)EeBt$02!4{x#{rIYAypO(xPZe`u>E
z)wIh^MKEMV@?<-C$UiA3^RJv79mc+u_<?e|hKn?8Nmd{LJNnnoQdqqDqgq;_mD){_
zs`WmVs?ln2)Jo-2eNH&XK9t3ZN!en8H=0)GYAGl1^`$w;$RnI^dF#+usC*dsg?I3u
zRJYdbEUoV4lO}ksqgN6|{n4IR>N&1@({O{7Ro`M^4{0l2#~-GAEAfW36xeuUQk-@_
zXCprOHPcxB%s&*hK4I`}`3I)$qxqWqXW=L9m+<|3MvdYdS?#PYlglw&#fdGOyOtj-
zuctyQ>g^N8z8UdFkh<`et)*+0jmSt!8Dk$pb6=U~9953W<!6L#KC4yu9@Q+mjYrC=
zeuw1Euk20nn(0-O!Lu%U87`q;(6wKS_6+#nD&;&6G=u$%$dB?h`qgi#MI7>}y0Np6
z54kK08~uJ^oK|cQ>Ke7fL9FXGH)wy=CYf_25`HSLiufA7C1w=doT$brcT&H}b^FoZ
z{f|cWQh$H^1M-gN_7wQ#YzaRPbpi)JB+)nZtoXlYZ-~DO?!UEsQ{h|f8&1K8?H2_~
zmu`!}?}J~U9~XWg_;<rTAh^)9>&fEOEk0YTo4rc%K@*-rDl77F&Uy;@%i|A>zqY(H
z<sTB>>VLmhTStk6Tm$oNVsc0Iuer$hk2cLJ(RC=O!fkz-Cui4om&}@(Wehbph;=oN
zwmjRzj`sF?g~IS+ha++0HHV>i+1mU|aR604e75GHpT!W^=(?S#Kfjm-09P%fXYES6
zfLs%b{drfZO3|jzfy80yVlgwTRd2EATDOkGw@6msg*jc$LB&MB6Qu~DY~wBtOGzs>
zeXESrB*eV%cGh^1p0%}1WmwABI_Reee(Bo!o`r4kOIEfs3rC8`NHD@N{xw}SDK&_1
z&CSEyOv~m6k+&WzgDgf(ZFp{7mlpw;jnJ_DD>-B3%kL|kpV@Q8)?U)5clRd$07I$J
zllPYG?_acE0WJok@W;dCwnJJc@~_4%4S#b1$9nyA{gJg-)V>pFF5ME?FY8YqNqZ`9
zyFMnCD%iO}v#<D@ax`rYdBDhC2==d({7K?%_cE|teBD9k9<}uc#;nQVYwQvL0GuCs
z`5VND1*$T>cjS<xzpZ&vbEO>_^7y>YtHsWedLwf~){V`!;uzxuu|2a^Zv0yvo~;G4
zu|VyDHv0bnTI2jPe4;6naUrAQYW3&!70Y;wR3pYVu96%EQ^`Dg4{==7>(rF2&tnaV
zl<b=s`%ieqv>y>^*69$EA#cAsxz2if*9YKlAKGgAloss_f>$K(AZ-}+J%`q){?vM@
zm&W#r?nz4|$0Q6K{{WDGI_G`{NdExa^SImc$Bb;w?s7-B73|Km(@{+NN;Km-4y=}^
z+!_q9!{Nlcjg*BLU;!97?OUE7@f806WqEdjhXD5k{{Tw!&jigU{40%2k`yJz*wmbq
z7#wl@yH=luuB2@_a7>)!jORG@9+mUB+VX`<9wWs|+j6FQb%%*#n7Aiu=j9b=`(#Nz
z7vxp%j<v&Sms3qEpO<mTVU9ucIp}>cSCwUlILX1=w6`_Nv2c!)m5)M}DsI<1v9BZr
zKs=wyzHt4z?w0pV@UUUHW?sCGgSC2f-lEYlN06%~I6VH9;eWN>t}e9S3?U~e=A8A&
z`q!_6sR`6h4(Cg*>Royt8+eM*qGo6$D;{}4)K@%MlY&bhbDHioIbxKONKiYkf30&f
z+C%0YgaCVHzJljV(^Sd|#Yds`r-qTS?dek#jRDIJI|G_&yM`c0U_PB_x<a&@(^1jd
zyPWaQsI470OM8jcIu{@eR#4Sq&kou4tCqUIl<LTi*ykkTm$ic3ntK}ZUhGe6@bX*e
zHc?u{o@}SG@>#v{UYX%bHqziXcc1b5DB>^q=DvsgE9YGg;*7o@lt_XmAaIJm&%JIx
z#@%M*X5Ppg{{VS-&*MsX`Y~3yG;*v>N%J`BeF-)DO5IL92Lh7Q4}WZq-@B4jRy}@F
z!#}2L=KH^i9xl2V`$Afw?1P$vP1ZHRt?74Hnw7-S{jx9=$Ur>$R}}L)R^?q!W*Igj
z-1E@=h4@b18r5_=JJ-E~Qo5HoApYtg4g_dA0HXucn)R5Cx~LHa2s&-XEAoQd$6B|6
z{7HRrqU!e=M6!WAkORIM82<nfjPaA|F<zzMzuSMrT1zB4zL#aFs2}SsWiY?j4Uu0z
zDz=k7SYc+MowPqhyx7na@_;&;f=l>O3wf!>y?n>v-`gX@8r%pkwCz&GpRQtAm-ACs
zwf_K(`Zlp>7e&@#iQB5)!6E=X1_fh0Okl3fsNv%|s7_qIr?zUoGSg#}TN{|#a&fd*
zC#`BRK^%_(3nZ?1&re$Nn@<_}rZ;^{Pie<2P!YT7n!R`8#Bv&EhdAJ~fn3nc>8+yv
z<LfA3u(T}IRQu|6)5C2QxR^?=+cz^4$K}?sQ^rm&t?hKnAMHuX5F$V~C>-b5S1~TR
z_Ul%7pcjA=0ftXA9^Tc+$#*;3yIP{OT(CmLBtxD`IOSTZrf*I*=S?Hrp@fDq5{){F
zQd=g^eY5zPf1t}0TE2mIWF%1HWOb4*dv58UYU=EMD`@ut1Kg}`dpIh8D){dA;xCEr
z7t6fWUg8M%EMM9MU&~YK6oBM*W)+Wr@kMQs5__e&a5}~^PxKYRiOs3OuVL?8-iHTD
zwKptEP?GHQ`<eFb*Tl^-Hp2mMt-$h}F+YLDc=yKJUj_K*LAaLY>+Bk&akP-M@ib>3
z6Sr^#SIFKg_{^U>TVGyTTrO~^Lk1oGT=wntuQJs>GI*O;Hx}A{t2LS!nIw{VK61VG
zuR&T-#@4L93mz6XBf(&HtlM1`nc3?f6K22gcZmEuE$8;dt7$hEx?kEsl1odNKyk6T
zY%=5(&MU(_HKqJY@jI;gSAn%HNprbbEG1~qez*d?Kj3z$rg&%LpTmo&T4QZ3^mg+#
z&|HRaRE{{>axwbX>8FT%TWjFY4_)gQts%9v+DuS=<|N=K<N=e%JuB$2&ZMdQcCq=+
zZCaKdv>T(%%zg@M5?sx!UOm^za`U@NxWNy&?ilTZo+>z&b#t7Mdsl1mr$CiHC3ueZ
z*2-usC%&3^B9V5yNbXoGk5QgaO5+8@>QrFic);fs<xSnbS97v0NhXzzoj!GWQ{@N-
zGEYHXpP_i3IVXhQv(F4ge|F@sAa?ezDzTDzK`U>#=cyc4fz)uoW&TmgJ<WJ{+8<ca
zo~|8RbK5+7<6S=EO)%-PEH*L#(n5=pr`OuPcGTmIwKDMy+nXaJuRfKpd8=LAST(qm
z>~2_RhV-i1Z-#YC*;UpiSo4;aLA8H6jIRW+k+fK1xm0g057AG6e-fneKf_Hb^HZ}E
z%!*($!}1@K<pUo|^!r_NN|-VqHr0px_g+6SUI+U<!+)cEGt-jAns|(oF>s)Q+qvSs
zYALlVnHS8`;&6G(4UB8{EIc`=#lFY$emDL0qK3*(p?Aj`Wd0)X-QB*QD$OLRgb4f-
z?ngEAz3+nc-3S)B)9xZ8ks^%Xdhy!6n$h%l?j}KVZv~QfKWV|wwRKD34-Y&n#pRT)
za7_FE01Dy7;px@Zr$gJzu$5}bMhQ0F-F`>ISGrlbVzS0J_bf1ey{j%y3YTH9l{vuy
zvtM&h;Xeq;AY02_v4hrF6|<kz^H`Uk4ZH)WZ}PmmM^nQm%ah-Nzt*@Rhow_RE1zbb
zD+e1+Hva%Wneqj{fo`Sw)msPXKO<Ky{4H^BBx+Q##zE_v`cnJh7L^E+O+QzU3_ewn
z75Vkx)*Rmr?-EWg?PZN}Fc)DX9<}LG!cva6Jd8#=5jSR!5<VTdQHd8|9SfX(b=%qa
zO6um;H+!K33ZwU=oulcrWBS)*z5<@kba>;pnb(j&1xfb?uUcr@gbc;Y;2pvxIE{Xa
zIImtDwAPI9rB-u(^jXIL0K%w(;FlWJp#K1R>ACl!yW?Z1NJ|J;AMGdximz(_0GPv>
z{{SiU!9J7<y81Qt8*P-%LED4?oc?@%H7&n}by@HQph5otO=VR-kj;K+$oikOX!@Dd
z;5g%@be8@DypU#D;fVS!TCYd&{{TF_;Xm%CD!rQ((m36;gkzwt<NgpBl=D2T^uVUK
zfnXW=r@`(D({!fQgPj}&!(ln#Rlf?uzVtRR^KfgRyzqUbg^)ec?dz2lK7rxOeLg@$
zlQGX5lnSV{w7D)XdmY7~*=-pA0A{$~i&x^)!PoL<gBc&Cb)xBFgc9LKI<duh$Hj}y
zH(b$L$#r`xc@H>wy8czzszNrqGMpzRWRI4iTug`DS84H7=fWCE>bwv9cC+rhISsAC
z2z5&tl{wFw*Yp7IKoGyHzCU<%4+~mbM{x}E#&SYR4{j^1SEkyuk;5t#VzrAtdA}Lo
z@7l6w8()sq(P}y-qOz6yBex&ILH@PR-b-_2Wel>z$OVo-K9$;>WwSXPTz5mFB>c<N
zdRK{jNU-}Cj5Q8FcXf+|J&y1Gy?aEu#D{X+Y?|Itw6cskA5ePw@J)H2#amZ+EpHAP
z;Bg-195S9%(yQ+yN>5gCItAH5EGMIKKBBaYnB*G7RW`RV%A{;;FeBf!Ys#498Y#Og
z1lG3~&i?>vpB&^I6IFT|a6g@FWLbRXqkKr1Bd=PFxo-5i$K}?PhMGoH@RVYtbUu^}
z5T_&Dj@6?qLRIItLsZN@S;71(2_tS{)=)`1tTDv%%_9wgJN50{*HNYEk_&5lsUTI7
z-MQSoP#lgwI>nCYz=^Y`K=|bd=zXhp4-idxZuX_tJBay>ii5k`KBu3@>t0?WQ(7a~
z!bT5S9p{X^P@X)n(=G0-BDH%oF)Yxi+`!`_o|xyoaw*|>Bf58wbg`*BNm8xX(<iUi
zvGl(Z%P!y;Z${edkL6l0_|V$U7bgsJydI<J`B#}Lm7u-T+Eg(0HIdX>_-W0+x{Q-E
z4aBg|IO(4KI@YvW1m|nX8(@w@4EL-XPl$v|wm&a+%x!)F=B}@g1aXawzFdcB8P3m9
z*ZlEZ6)NIc(+W6f=x^F+2sWcCf(AkF`G586p-&IRBr1|AV5)klBl*`7{wy<yw>a2G
z-Sj#3{3~Ng@nB%^2}2x)^B90X`sT8}A?n#}Vb+Ng_>S5;XfPN?BLr{8bI&-h67f8u
z=THR1GNa|=2OYTWUYFwU5~tXu36>Zq<{)7DSCUNaHNhnAW9n<^^6G6znn%lK6&E)o
zPf-1y6HtE>O>d`6w)c%0D<j}-=aZ5TIL~iN`eXYZD;wm$x}W<)P({F+UUn0~BRrp}
z{VV35+2-%;zCVm^DjpYJKi#jT<hWS@LPEQDBdvAU_>VrDve2I5?73&%+l+Lqd)u#=
z41tgVJ#$MW+p!+(f^n1Dsa+v61P5_r#w!S<if&IrDQ4tj?{v*Z*Bh4r;fT*2s=e@Q
zSJ-~=Jb-hGs@Dr1c{s;Rb6axXQ#;#2cDozf9C~2$S1r6m+I))eI1HmC=Y#lHD|@JH
zmx6icilq1ClgKQ6X!Z)G<tnb8#_p@CeXX8G!V{7`tIB_9d%yUy{xXay8?^Wb{!GPP
zl-GNqlP4S;W7fQD<NpAQ<dehxJJ5}~h()YwlA~pg+lY8!+dZqLMN^AN!t$$8KQqVv
zBixz38|cce#bS;D)4tl)@K=j<8!c~67G`_ln2*^|7ikDy2HyQM_*QSk-6Bm3!EL9@
zB2O#}Y{p2(%E6E}KLFUL_!~&qt#nK87iiZpS*C`T@=Yq-s!oOeRb)fUdLONN8C=~b
zB$>~f(v4`vNnHBVO8Dcb>oeG2OS&`&Vv;!c*^UNRw@T9S*Nd9^$_YcUy|rvfvp+U@
zQ~1}(T2G7heIm$eQme~tdo!S(Q^UA9Z^xST%{Rv$XM|_f;loV~Xm-CjDsjzy#46Hh
zcRwh^<<+UpyXe__TlkB1`bF)H&A*Z^SLQ0dTc1q&iu0KKX7g&6G3s;1TFNgX5|OH%
zq;Z^e`kJ4{P;0&|)ikNDZ>`|_U&^w`b|M^S2OUN*GhSz=nIrLSqFxQfF48%EY=8uS
z5%~jNCOZ`u7Lq=@D$AWaN=*HV_&?xXLQmP_!wG#5ia4yi(vo2tVI_e+oE+Dj>;50`
zKZ?U8pNMRd@-7NVaD^3pfjw!zwWq}GbHjfF?jzFm8Ejyi$hTF1?E)gJ^7`P4{K3?G
zU#i7<AX}Vdp4Iu|#>bhYewR-xok*!4q!D}=@Z-!SZxrat{*88+2kE-DtiBHD^2@b6
zQ=&LO-JJ;^t$tR|;>}c(o0E*^jw-S7rmYtu;r{>!Vf>9!#N5(ZWS3L*`qSWkny3l!
z2Ad`Zepyft=T+`L5oofrBKVg~MjgEM>-pE@CC82Rq@C1W&J(cnpZsb+_)Ya(^K&kJ
zeAfQ}8qrrUd3?6H`oXOH0|g0R7HLsNMnP$marxJsYkvatT{>cJyl17cBRN@+{{UM2
z=bOixw6X0scLa37nTY;%nLml|o%7~NpZe%(mLZVvGDqk&*MK}HXrxKwojpI`YyB&Z
z*YuwZ*s1b-W2a0&&l$~pH$RB4*qluu_AY9x>({I~o%(-x)q8dyQ|`~Fd{3<Dk85#f
z7PoJ4JjlxwXn)ooz*m!9Y0}SnkYI<9sw(5DCm+n$8Q`0#@4RiH$0PZJ?MaQ_TK9kK
zc~``?a-ykoGBT-LgU^3q*191sZBD8b`C`)UeP8irc>FQ@CFxphK%l(7OKf*4ynpPK
z@*VKo0mn|ig?p#RABh%MUkAJyr!Cq_w)&Lt&fj|siMSK#oaVeLc~<h&Ng3RXG6?&p
z=xgMu;im~sQePCC{{SP|#!X9=C$ch?a=dbLRSOc?_pM@qEp}l;usC5^62{Q&4bD5)
z)zies*s5s!#<l`;#mcOfjC;3QXgL(dY~XM*M<$YIFv#hZZH!lPoP+KSc(t2yKEn-3
zLJBOExEQBI*$0}Y*Ch6*!8S=IqAl9Y-;yca4v?M46d*p7xGdarpGuBv%X=VA#OVnC
z0BH01Qsr{lT-NC=S@cKj)vdj|Piit-@2Jm0yym(u+REHduh?Cx?)z1?Gv7QB{cGo6
zho2j4JO!%Fa}~A4{DM42iG@~U{p|Lx#%~|^i}r>1gQV#?U8K(h0bW4eV2pCk$DmQq
z)`~bvvBpBPR!KA7qm81n>XfA|dL!oFix$QY5X-n>m@5JJR}6E-dk5_&qW=J4c-3ur
z;vf!t<BaFMc?nTpM^##E`3XDfG@7w!2s@PWywo9bGup0cP)#M;`4F_yfLT=+ao}e-
z9-V3xlewGKSh;a=mhyoty)n}y`d6g>&MoD9LsuX!$R91AVk^%tadWAZw6!+-l~tp8
zH*<lw<n+(!UZwju=+a*J?cs_b&m>HXgPaq<AN_iAT-vj2GF0Y~^;p$}I)9WLkR#6?
zr>%3^uB+y6A%GYe0P-p=SHu1!(g&4sq+Wp;z%j;OKT<kZ8F_s@tb@;nXl2eD%~B8Q
z1}Q=-qdjT2B27a0ZekxQ+ymS3^sKoj+M!pL&mlY2S?<KA3AMmKc$}ZoqlwGhxIZ>O
zgq^sqi&&0c<@gGHrZ^fk+arA3b{+ne(%<QDSm?1{2vp9<r3OwLEsmATTDzmgEQbL}
zBp&|&=lR#E$*0@RZ>O!w$ZaD?<aLi5xa58r>s+;(R=KnhlP-q^v|9oV*}%u~S6|?N
z3r5$`NG*iYD*`tS%0D`R?Bf32(!vy;)@z8!W&5nG*sp)F&~0?B(ZveJtKCO_)x}au
z);CElOqg>Ls)ckN7z2Ttkw_(hW<<sY?nmcVF8M9@LCDF)QE7R?o;$8O(~K^RoSpVG
zE#!>0$84zpf(Ah~r30{I7*N}O>GbzLwTl1<56!)?fC=a;R>+v$ou7XnCq4ZLt8pJO
zzh=OYLCcZ>$MdbI?jyz;P!>h`vM@NVcG7+732>xx4@`Sh)82s>A1Mb44{_Z4RxURk
z%eQ)rWA}<wafg#Q1Kclq=e1d-mS%rDe;+T&#y_C;s`qou@-gz&l>Yz|eALJUmi&LM
zEOWiM%~WAyPH~nt%@!t}Fib$GF`dK=_V)IxUuHvaa}SgY0yF%-A4+08hEs;X86PR+
zk6&tnTj*_H=DCk{bC!IO?@ip=Gt+Zs+hietUF)2mFmg>w({C)w2_yKA09FHQa0X%w
zmj|&?#Ua%%?&R5TZO>1f5>Me-vvk(xySQAl@-fr3X5Q=3+N!DnBaEDkWAXY`%Wnf~
zx~ArqQw&6B<{*AUHMOVs4#P;<HPgp(;1QYFDOUdgccJ>$O(r)|5$bSJ=z7nI+1WqQ
zt#i1_LD~W9k6P-q{{Rlf;oDy(+B~r%%$)T;_^y)kShj@uV(tjlR1C*51;?O2&am!$
zN2J;_M<%Oeicc&jQbG4QtyEcztMh5lr#_V&GBlC}3{K$EHhJ|U)4f1^9j)sQ!%fp}
zuQ<q)Zq+B(C$(0&_=BO^+$oCc(*8?#4lP<x=0FG~ayt=B)IVvx2G;Id-weIZlIs5e
z=^h`|*EZz!CMg$~iTB9%tZMVcd#J_LlwG2(Tps~EXQ&O$ZD(YlWSOpkZ2bVLw_mfy
zw5R6Obmwexi!0;w71e0JwkLpYZS$#IYJMBO1LuzB#N9>@s>v3253Vawe`;Tc_b}!!
zh&C$6kY=%-JP*ZdkyXTA&KCSkC5frg9#M1qIBU036<t!s6*<93rECw#RU41lhUOI}
z>%~yS*!z&!JAQd0y-QH|{{Z3bNEU<Q4X}@}g&jxcQn0T+E$Gtc%kdq)p(*|od$&KD
zR*rQ)q)*mrGs65w`#W!MEO@#F$2&!fkLz59q3{a&Br9dFSltyoBjpWyXNo>P=(<8Q
zcY3CuVE+KzSsGag_E5anlz5-^jni)9bkVG@HH(AyWSSuaVE+KSjmEpFOAzd2=gq4;
zgU6o@JV~T!w`*}??y@0IIu{Oup0)GLx)z}&?1WlL8!0Q#%0D{#J6!Q!h;_T1-%lz7
z^B6%2iapoU>s(#ktU82-Lx}-m1d4dgda=E&GZ*h3Ul)WXyST)U4Wi(Dk^J1&oSJpM
zn$X!u(mF<m6F>)=g!+TS4{Gf6zYpBp>JZu4#pm8!%8)@2{wHR@{YWR<HR%5U5j;EL
zX=l>>8)*tiG>DC~nys)Nb;aA9&gB095{BRd&@FM|>(ikd$r{q6=Pp?8cz5;<;kPqd
zd0Mrfd2IP*_H;imHCERCeKO=dw6_*Bj`2H2!2TGoM%F$A_=`Z3?RsmQ*azL3(Cw5D
zb<S&*yOUb6VwZbnVlY;C#^H~suQHA&3yN2xH2(mu$3*GXqU|`?+tI!}d_&NcW3us{
ztsG<Sqg-!M_%>^yv;C@mC0V*Z_VjZ~youzqi+Dfp01bJQ>M1ZN_Hd-Fz#svip{$87
z*<6N2#(GrM!C~aR=tcaO`2;IYFGkN~)PHGT5^HxR<5aQJudYumIEP~~^(Vb%YCjb&
zJYQ&Z>9mVWQMgOzK+T>zVDZjt%zoaZlDPVFT7Dn9S?|Jtaul9MD{M6=!lV>qWQ=iA
zZj76K3c9SAynOIUJXPHtRi;pTt}4HYt;*gbfC9MPf!d_fa_JV{xdOe1q?zVPb2R&;
zZuqB0I&JcyZR+uFQ(WdNsqr1#TrIF;mW>K^u4t(|ii)2RttEPK`HaSEXF|98Rm_SG
z7(~tWz^r0>X1X5<jk<y}9jei|=f4KL{6ySgrsvXR*ky>s+H2j`@3xH{UWXU_OOj7V
z0~PwQ`xQ<b;O>*2p$~sTYx1*1Un@$HSoM){U#Oq4dwD+xbihyWG5t+_Rw>^NPn&=E
z2jI11$(U42qi3OfT)0`hBdKQx<vmU_?O!DLs$u<=8yo_Iso;WtT>DqjKN9s;9u?Ha
z<GFWoI}U5*FBx91*bzLkM;K+`4D*j;Up&`6JHG7rNl)A2-29j7OW`T@%Q4uXDj%0Q
z<Gy_d^R7Quc7GGb%3I_hoM(ZaneF)3LE+XA88<l$W6s6phvDD)RxGf^apE|^->~3-
zzs5eT$9|@>r3*;+@N#_2@-Nyp=`X$^vK`B|)d$Uk&*A#kpTNB>?e4DSCF34eH%H44
zLyzJE(z)-CQbxZME&^};+)<$tAXWqqZV0ZU;eEoTtjt1^0+6H}<NcC45${^psN-5G
zAERO7zq91uW9hvKR@dQu(-45X<mNRhgnEK`H6Dwpd0IlQ2>|B=WN}bk-wSVq@a5Tv
zIXkil$?gwN!nnIH6QWo$vk-b<w?F6JynJ+TRQ&#RAGL4D^$UGt%hRWA#apQSdE&V(
zd&DxC!HP~dWll~B$6sH@xw(8=zh^2Pn>=KUkMobpu;=jso%4_ma5(MtuO7WFb+L@<
z&dSX7jSE<S=}mbg@=K__z*LQ#k2*3rCyacj-<t9N0NT$}NNs#JWCa{~prBp7hAV49
zzVk04jz62`91QbW{{Xe7fj*^a;Z}+r&at|b1P}qnIjYMPu+_Pn(E7ZgJZ)IhdLJj<
z-<M;uBhPXxI#!vw;YD|_{5!w7;n5=Y$zm$nd@LmimTZCS0R3z0Cyjj%(BYNQi}$3D
zFOUQG<jMN~06|e0iTklJKD$kN)P4=owDP`v^h!rnQVRk20&2m7!f;NJyR3|QdGeGW
zk*;+*)t!<%pVaW`$)^03&my+d?=EHtkv30c8;|m>jR(Sa+V-^;kz;P!hN&(}+eX|V
z9a-C=f!v-2b>1}buA6D1JQ_8zg4JY5t}m>j+^~b&C#!YiHQV?u_rm`G415U=wJqnH
zVXE$yINJvMQN~mT1EPb~KH%y}JXF<uirdNPOe#6vO>)I|-LE8LUwkCB@V$-E7wM>K
z&{~VRphygAu@i#mSmdhr-L#&n4P|N8%`Bc;pkc-W9#6G#{w?^Eta#Gu6`oTCg}K{p
zYYSv18z8eDUA@opuT}78k08|aO8|7)?ysZ<H=5qt5xWD7h8%r;>&wT`uZO1^mc9JW
z?5{GTUZm(nZL>1t@SV)SnO|aqk+f&^rPQ=Gu^L+iz!#DK06Oo&-rSP1K^o((!;(K9
zm6hXrSuM0Eu2$McSSLvyL|woDdXwp1EeuShRxy+6d)4FnsXKHq{8!>w<?)(ESZw<?
z)6NOUYRbRVRyhXQ0QCU$THrn<Yl#nv1o28vQ|}I=HHD`5k)B)`3xY>1Yv`&|YFx<i
zXihGdFJtH*4)~sH8w3d;&U$A&S7~A5XS#_1xso@=avax#cy|5O0WZTJ%DWqFS~y1k
z05oKF&3UzH$-NEXR-f8P^=&^~y^B(VH&Z-p<14(J*5}5}dU-rnu%ynZ6|5jHLCOMa
z&O9gLSTFA(K<Hgua!)ni{A;$2ykF)lsM)oE@N@qFEX8rkEkAheq~j<1yByY);#=4>
zo0wp`p5G|Nj53wnaqV2c!e5Enlj%CflX-V6*1~Pey}n_b<Y%pG=}Wd7p_M-)w@+&F
zUx(L;XQ*o)Zg5>d$6ltXQ;k}4i&wRZh9X#eT{*?ubJnjlxa5(MOQ~*|a@9&}fP-|+
zw2^_9&mT&lvfIrWbU5`G=}{Y-P*0gE5<8PwLY!8}`bwFNN{il={$)Kz);Um=EC?j-
z+IId~u1`_BipBRCcM*_Af0bSFUy9<kiPk9w?wCD)D)QY&ReOodjN~Zklg)b+X||7<
z#8&2!*<SdW?635j>uHQvR(B~nNKf5mV9IOkyKmZiRMh-Ad!=}QjY7iiA|!VMc?1G5
zae#BrJ!|vTHA8VcaAdxDC$0x`UqN_(#Zh1Ah~@;EMT{$?ZMDDy0O0lMUVdRxwi!lL
zpQ-O<Ierr|g-GIHrM0}h3~hhJkZV`-rQ-RYF$p1VK<GdHRnJYL>2fYvBO#9rIv?<=
zo+O^iOSgMdEG$k;a*l_-D+}$?N72h?^{k;q&2y(_7@WTeFLA3_>7T<kf)FDzBA%i*
z9Mwng?wbr)cy5ahcWxN0c(n=K{_WX)gHbNCe;#uziaK-6H%e<p&o}!X3*eD#&1t8*
zeUc_tKg7$p=kckn?(S~nU$hBkiZTl!A1*x&Vg0IU7?UdiIV5%bE80F6d?cGzf;~%G
z^9&3(MQH3uK83%+y*w6QQoOy#Z!_kzPpjor`RhJuJ(qLUz770-@Lz-cEf$+B7yf6O
zWH7-CIFCI*K9%iS2gh%RUL=+~d9_Q+Cw%UT<qB@kqaO9;Ha;A+u!UoW))<s<0;td9
zTCw<g(q<D~#}~>*6emjib<IKw3R<J`=w<YC3`?)<?3>rI?AEe;FuITg)S-4G1*X5p
z^R2rbH^T16`#R<cl$?B%lb^&_g3IC04yYzuXs6@lSm6wF?oV%VRmX=kYj(^KG%iTm
z@XFET{s)X#E5%pUN$7o4GEBY>GNC?MdNuf?=qdEQvNw}-!cM@cbRP9hblX_kQ!UlB
z%*Q{vZvLk=;deS*s!VX}dX!RjMZIie^X9nE7~3y~?G4rDyL$)sQfsCKkFe^03eHs}
zJt)S^%MpQCcg<R_GV^Wz4Ekc?*4>F%<w83TF={o_pJ`<)x7Vnzh~MHgmQ$7%*D|Jl
zS#Iu`0sjDigOOS~XYD1ZczRvJE11!Gc^6Dq@?nEoCya{tFY`WXvZChH547R3lWsy5
zBdNtmu>{^7<H^9lY>{3M;2+v=NAa^8Dd$UFK%6Se`?DeZ_UGtpveZ0x0sxGMAoIJ7
z){7HPI#jxSna?~nCQ9_8k4S0Zi0x~L(2mMzuOE}fX}pzUp!0>n6jt9`quk>D&deWU
z&%8OTd{xjjSGmwUV)pkkuyzkSGRf}Bax2_C1@Y(OT$jmds6nIaxnMFRXdL>Sj1SVh
zXF>g_{6DTw8|!`t({%|u+-UbE)tl<0BD!1O+G|a+iKXz~gQ(fwh71komm-fq4_feQ
zQ+0G`{X(4MOI5k`eV4?48^IE`rQr=mHrtZcH)%5d9OkV<@uFyCDRFJA+55|EPuIPC
zVt;J^02^APFO?qAoCUW~ML@r{Z;3A^U+rBo=H)T>-~n4c+<#i-`i0%^_?oPDc>V|7
zHoqFQ`+^X7g&{cmw+PX<<V9Pb;^a{#f46PQkK$HS{VU>Ve`~u9iVlITJb-k6EOYr9
z&Aa`kZ*?rRJ{x6T-)WR*^b{<;xADKc9yUE7zT@os&x{sRk{`qts@*}}7P(zd;$Si#
zBgGK1w;v_I;=Wn8_`|R2Br<qwPQmD8p0{j#Y2rD?;o*G^vU;@DPb`c3OiMA6`zZEn
z&l#oBZ#P_N&!b%v`ew3a@fCvPo2Z+%Gh71ryTm6M@V|$U`k6`owT-BJQNOSZldp!X
zWd8tS111mn)c*iwL#t)~0BTKak?GR-lE5z#hCl3#<MgWSs0-HFYiV<I=Yi!yNhZ8n
z<Ktef33h{4vr>3pBTrBP{CZU;{iD1`r&|ENA-w1B2m;5STxXHb@}cF>hNwTZYEMfa
zMCsoTJV~x8c(ix6R^-ci%MyY0Jl1E#PllR5hI~DHE$4`=rq_0XZ>HHdnsbr3x$HLQ
z(!NmDzi2IQQJk&Tw{v;s=N@FufI;9Mdym4mJUj77Y8sV_q(x(ue776+kuP4|LH@PT
zPMV`6&U&(yXtd6qzk6&5+TE0&QP9^J<K#CU9J>Gmv3G^;FhQtw82n>zsmfN`<>HZ<
z;tB=|laK{v>ah62;?hWNbc>S}!j;1(QQrhtvCSDR&lB0oFH~~6ji>xm`hCy`p;jZY
z?OF+CaV4Zuh?uJ&IZ>LiKZm?eZ>B|IaiC8Hsc+q_#B8OHu{9!iXT_}uhC339qY5`S
z0-AAJ+-@w$nm3G)%2(<!Q11iYwO_z~F}Y$sz%%%oYWyzol!qT+IQ7RLooUY==57)*
zM3nC}NacX~n)EFj_F4FyuIZ_$TWGgmVRg#Rv1M3P?fWu#qAY!+NC-GTDFZ)^B{;sJ
z$k#*7R(QbXRUeH%_b!<d$DjAB(ogK?s6<GTR|E|6w;#@;li*d<OY=@R>H`7|E@*^e
z`yO+%%5@%U6mLQgHCU@c%J8{xTZZE+&3Y`K2K=&E+>?=x20*A+;d5I`CXRN<82ML&
z_#P{gtYag#jGP(g8rYb+YS03^{z=phJkL}9@aOZa`Lz>tn^j7m;RNRg)9GHio&!lE
zni&*2N;WnJ;tmg?$Lc%PFZfi9z_Jhza2wasyy+_*v??}vBs$!%$b8ju%hw!!RkLTT
zmHA058FF1t2=pV|SENJmVnevNUYvZS;<YV)4#NmmbFhxMQoMd0=|<0SPEzQ3#kRE8
z;aSPXF}HVL{;{nXJW(Vw<}e94EP(yt+<Moh>Hh!(AiF^E->S@8E4?`5{P(V>O8uRm
zQ<f=K;uz&u%DEdvc=zx4*6C7Pvz}6`(KF}!tx8*z%B<LS-M0jf!<wlM=V7)f3wHqz
z74^QG`y#!=dDmCAOmUyx;R*NaitKcM*!xSuv)-=kWr~r5U2(5YD#@HPqf$=oA1(YH
zS**NcV>`y^oQ>|d1B&`GU0MJ^Ylwz>?-{R2(*FQszYa8N#`v!6bZik_4X^Bn;aHe6
zUT*cv)KsfZPWKdP&OHwWpIx)JB>kZqJmYZUstsPvLrp6@nI9=oNUx!MVemWQ_ros*
z>b@e?8_cw@gv!c}22Ww%sqbGfC+#!vVXX(+?XRqE&|tu@tjVAE@;{Y0Q}&CQ6Dhc8
z$*(ooq;3BIGb!WcLs?fIAdLf^#0+&H{{V$gZTmoc5x$s-HLKeg+zw0t0DJR_?X;g8
ze$M)oTkbw0O2Yt{p5Y-s`UQ1a!C1p~d2IT;GD*9B2<S0Ue`v%&JNd6_pT++GvWyns
zNvP@46<irkr!jZ;IPY9XiGC^m$`&B`QrynKbh4h|v-tF(N%kKvna4$S9GrdNdww<K
zf3!}Hbk7d>rf)F0xpkBi*fh}{zd_o%KOg?l9|!CsW$<>H;wF@@4?N4Tdttdhg=PNH
zvq!0X9+y;58K#2Ul|758AEiv6Hs!)DEi=ZxBKX%&@h63@8Wf!+mM9ejfF4m5PEV%<
z*A_1fmaryUh=P&OeFaz54m>px10<+Ftu_|SNuEAYUD!zSH|AxHyHPZ)u4Qu!4J#Ox
ziBx7K6rQ9G)#-i-@fdpzS4NqY7V_QJ-CK;X8OD7{uO?Q>&sxwLgtG14#FQ9Q(2vHw
z7gfVlc`~A^=`(vzOU-s>h~1*PM9e<bF2JFBfKT$Omec1@v;(w=4o6X%&htw#bITq{
zJ<W7lc8_l*wWaf4#Mdu%<qJj6ng9UrJ%XNsy!x|l61zPbP+ZE#=s)d)Kbi0<$mcmW
z=RUt8*XI_p%C}FBdCh)_{??Lj_!Xy&e+(b`0@vqOus_lm``5rlUZjuGu-Z_IGNx>T
zdQ~Y7{DH+;xIeq90nYR3UC>$?)Rv@@jG=z`0;}&h=}%_LFs2@rsz)Pw+>P3+$mg-C
zr6#MtLt0qIPUJEZwAFJ{NPb#}JPOeq?VUHm^atWUhesUV#yR@e(^~4A9}!!+bLF-$
zrxo)r!z+lbej@mGX^X5$dl_RI<*}Z_*1nq7ucEls-r@+^BSV<hToOK}yJ=q96HTp5
zFC6HxYFc|(T}&-xGMkHeNXo0@2Y0a?SCKWwso}|48QD-A2aUI{)E=LWde@4ixYjL-
z=bZ4><NDr{b8RCRiS`4MdBt`@5tMZ%q^|WXFO2S!fqW!hq)wkUexjP*FqN<iLUMWS
zT-#Z@leRYe1w`6({Ql|x04krW)9iCergV~coRiWo_z9>_h{?bzxcb*L_L3>VM*ONK
zw~&7eepG#0mteWnoqdS=tN#FfQZI;JGm)@;Yn1Yw54v;oqsf$yy2O1d)*g@Y5iNS1
z74E-rFktubz=!x^Tva_AMbvy(dbYYIlYgz;vA|fOX6Nw#02<4QS{GIb0}g(5$o#9(
z{4x7Md_nM)u(r~CK_!F(7>YPzRs9ya6)upnF?No-o{8XZ*w4n=>cu~Zt}VPTVAw4J
zVoRbuHVy~kYwE9rKMQovgmFP_qwRtbD4yct<nET`hA>=?c*hDj9`*75nfp}yKr9yX
zTKh_MWj3kvdh#pL{5$(d++GB0xgBN4PyYa2VJsy|jX6r-@jRDNiR9n5WwPD;UB8w+
zg+kymDN;^yL0*amcmliyELoV~9FFP@eGTK;Y`jyhOL~6QF>~`tETQmzy>U#_JQbl|
z6%t>hB#uO}IT`#otEo;ZTAb>nxpq7bJvL9_h?ee6SdPkQh-{dj7zKFGuOhyt(!L02
z+L8}1UeN}7TuxN}L{~v+`!8AF6*21m8#(m5UO!re7<FWGPWQ3$u8X6FmvCP(z`@_b
zA8Pu?_Brrear{fut|e<i`*J*)QQTnk!2L<C$HM;rvG$8BirR;XGz)0IiG<=$<zGX5
zGWcWg$Kh9(G+rmPxVju~x3)ZfBBH$E?A&9!E|Q<LQfH^!_&Uc^#Qs9g_~1rLk@c=~
zQuupl`;c2k$NQ2ZI3HY!+r9Xy;T<$$E6eGD82!vD2_B;+ymQB&wZDUWDRm{hkWTV~
zdAhq;_Ye0_Zm*ezJXP%!&!JVR$kx_A1n9b?XYD>8(q`Skc7#UXfl@JCexv&zcw<qH
zZ?nrTgK*sqz1N<<_w1=ZlU6Jrx0Zymhlj&2c!RaQySr!EpM&{UjNkB3IoJiT@Ft|d
zam>-ge-IYBy_Pv_iuTme<DNPFoV;7%3+WQePQTW)*w=U4;;@krf_`kPbH}YlZ^VDQ
zBkrO`54aWTx4*Z4iLWj0A-C`;#L+r_{{T=u)PvjJzE9NrS+3jPJhJKQ7UEl(js{MB
zE6}G$5m8#I1zHK*_OA#^XYnSNU^e+rc@xB^vlbh%?O#%BaY<;6Eyb<c!Wf*xZe%@u
ztMli>pAxlQA6K%}#gq>A(g`C5Qr|E+1EqS5U$#e!B8+)@bP>ohLxy615k&B{9Ik`w
z5V7w1{{X}-H%n%XZ1pFYoDVkNVyOCc&2rQHL^7B1w5yaYj2BV|ABTGK+dtax#qnG{
z#m<dub{nXXK4G5Xwxa#Bt`;?G&j#2wIm*0>NbiHwHLIs6^!!M5uFkdoAnHvfA+W&5
z3L$nm^~uF7Ul6=ZxDeSqdrp2>27bJni^D&)Kf|vP$RpIW+x;3r)uVAdS^i?Z3O^M1
z8t&dZn{7%{6OG8;M^I}v?Yu|NbaBym$Hy0~v?Re#0GO0wrx(6CkcN_1Rmb}vk?qBG
zzwnj#P5=VxGcz7X?t9jK{{X}f4?_V7Jn|8oe&-k+`kJ0@$WBb<t~Fm0z(UJ=B63M>
zz!Aq<=QTeS>aZQBRGL`_VIjc$X}&7>kE7m5v4_5mxsZ8L0$V=)>x+|I({Eg?c6ycJ
zJd?aA#dNJ!s8?p>o;tr>z+67>rsJQjWL)^aQ_B>*l?FNn#%eL7z2biBO_mM)8cU3y
zZYizdpj77B<YUmZr}3xmkDA42@wTrK8?_sfhp+EQ#WDPMsfEY(oya8r0MBOs02-Em
z46f$?07Fu7_gRlLL*aO3APqV_-H0r6MbZv-XI%KhRFEGxR=p=b!!v)4W!&r2%KrfJ
z5WgqW?GAsfb&KIxo<|F2yNB@PkPTa3463imX&5WXI}Q)FYd0u`<j*>Lx48LF+vX#n
zX<Hx8tU{L(E?U*e2Oz)AkI(Y2LK-x&6)K3M*nc{5Y0iq<({m2sVzQM)kx`7BJeN`N
zR<GfI5n3A|4ZNN~mQOK8A3cfmuD4A1r5kE?cN(nI*j!%5Stho(Sz;LbqCk3re=5_~
zJ`xWZ$oCQ6No9KsMNs)$(T7|S{cDBQJR#yO62MsBXj5LLyOy?|-ZLVuM>r$5wQ2jh
zkxQtXoEhvM8~vd?Gj(Rs_>;i4_i_XB+-Z&>C!A%Mr#S6WYF`+>8qTN4@NT6OXOHx?
z#eCR_uG##N=@(!QaA1G%q<h~GM0Xt$@DAic-{DyORu{Vb$#og(UO4!d6X};X-WAgu
zPitVCpt>JuX*Uy+G0-Rj^P@`m(_q?lt7`fr>v<w>k~KRQ>H*{fUQ9eqs#;CCVZF4I
z7~&#7rBa(yzMglLqLyh~{Ld;9#<X-I)tM@*i#~?7_{ZQap{DSq?Cp+ANQa-srJLh_
zhGUOrnc$%l5rET5Fh01io}jzbiTQ1upZe#DyA1lUz!u~m_Kqnk5nu4(?JjM6K);j4
zT7>g_ns?fce7M|*fE}t24%sc0p|BH0F;G~vM<Ge=Sr@jtz3Oe1Q9wEQnyvU#Olv(s
zrIJ+g$`c~SdI8Q&Z?duoy-uS{@h+X<=DyUP=r4$t5f+<nM=H)XyxHhBsUQyI*Cl0l
zHKd6=bG%a{?pDtszq}1_I_9kU{=Ip0T&fv1rcVJ!0~MiTugPrtW>Y1&;c|Hc)PIF}
z*s6T@I%!XqS3afipU2%cZ8ke!5JGG&F}1Gk*D%ZO%hgZQit(R}-YJX4-YJsm&QSV%
z2q3l##z5r1)SBfa)?t=5G0nRN1&HJ6S=U;EnK>Nv!gG(tzF!HG(86LKB9dvoojymg
zTNx_U)Ft<s)4$Apd*-8H2|4T9opEv*kIKC{C%@9G`G!Her=@)&S{_GI)s_UT2+7Av
z>3k}x`mhB-0SU+DT&xlW$#1>uSFB|+&Hy~)wM`{z(SBrNKi%ZKk3iRSYfIZyeY;vC
zxPZfv+*W3tqFc=x8#}aR3_eLt*Btb$xU~7*K`P9r>w{OVH0F&J*loe?I#;J!X-eqw
zFy7KhDLbTJ@eZ2S?xCbk{SRu#TT6TcBPseCxvUM+g2#3RWcdv*2)~K0I_Xc|bFub0
zCNeZ*>|umeB;QhNfecZy$BduGv~<b7*9Xin&<uuAkHWJS)Wid`sVm4|Q&uc*&7(WX
zgTL=6>0BJX*GhV{eUz(92&GdU7}124xmB9AoxXQHZ^3>UpTwG6No|3b1hE}>?O&!J
zvi6rQf$+A<+Ev3wq#XCJfPM^oYw&l1bbDCt{4G88$_`JPv=i&ieG8%f-`XaZY%QSh
zrPaJx+l85NUqGdZl1h4?&iL#e8dW7yj{57|^}mUl%C>>3DH$l+g7nT#I#<qqD7q5q
zv6NtPfPy;uAO8SVQt|igBWL3nu2Rm%TR}O=-ouZ;SDhanODtYfHuTR!R~$R`e6rm9
zkLnnzb=<jhJr~1zkw^r8JglJgP*q1_d;T<~q&_E#Npl|d18l6tLC>cJRCe{Q4%f$5
zcCeMUPn))S3Vxm9tyX<XDei6~hBaWs`Cy;^y=4k64^!I1;p;*3IG%0s+syv}g_(aN
zj0VO5V#68g3Fj5j{4;ZLrQR%CpD4LN;BM#By+!d`PO$Mlx-aeTV@Y2D@;cXBr(b+I
z@NSc2Ep<!%TKVGw;yD*&Xv6!b>sZyI=T<FrKI#>%m$+M79-ZQ?Nv!@I+Q;U$Nl~%b
z=b-J<yrx}T#TP6#gWr>yr{gb-P-|Av*x83$MHwk4Jo_5rZoEh)O`)(l<1|&mI+D9S
zTAmVbS*^~_UlF?u76g0qP(!X2kbK$aJ#$<k)pOsEYJ7UKq0Vtw{Y7lemIU{Ihqp&i
z)Zky3<%bA)CpGGSw9bohY2jZFYGNE1$CTUG1%Vw%_Vurr{0H$JJU6OEa>3nHf>k=6
zee2x(Y4O(I;~t4)r!AcJ0@DY5%!=FnFaYAX>S5Ly^SzIu%J8#{oMN>+MoBeWnFtQ(
zhqP)CR&DKs(Z?gqV`J$&#fkhY(r$b`Z#pP>nTK+@1lJ|5d_S?9cG--0>A8uhhGj}p
zveV>!jIyjkk2CzwB=JX#;?eIcY^AyTC6w|3b#Z49-jM_IZ2tggjw^TKCyD+D_=mv;
ze+_8Y`bUZFZN$ja$kM@chQWet@?4Lbc);yk7mvI+VbmVVDF@2g4tfgZhl9S(ttHLe
zL}04N8x|~orF%B#v)vwjIYy4}RA}3LSc*$~xOGh)+fs`FSzC!77%iB87nZ+&Bi5=}
z>Q--~r1GjH>p2BeB(gU+$4``wp7o0dhS}vKI2q#un(H)g22Xpn65NCz#78xB>Bfq?
z(9acKShv<qm{Li3tT&MaY>>AbSMN8}_8!&hKMn3A@Ct^z)c0HLII%2VH7BVH*!?S8
z!+r#}w}=Tr1ohf0ztQ{|rP=vXNoxe1$s=(;T-T9}$Ixlhapyfe7F&r;N;2e2w&pkd
zE#Y}GN2cjIhO0046-q{b0bYRL6EzJl!L}n-@VA!omlt;NrZz><e!O<8&8}K$dUutk
zs@O<BjcwHPUU%X<zY*&(N~^BLDQ-klN9FGJ;=W36+s-N!)FPSn^x&w)%B4p(`<@H&
zJXzoC@?2?VNNxK=Fy#;TYmSWEz~F@6{^74f@g9eMFCI(Syt$81Ml<w1>jV2fJ;Qnw
z?dxAxO|C^}bHJR^leU92?-slYsV4ZaKZ!`KRq>uwe)u=ip@UCcU_L<g83Lq_O1HQ_
zyowg*uHChw<y)hT)0Fj$jgJrb@pS9h<`NvK>0Xtn{CSU4SlPo%Fi^YP^}xk>DDd@&
zIr3C}9+&$f&3^1H$NP1h>CmX}D1=rbnrbt#>63hGMAB}firjgUAyCBh&3W&Kb%}KQ
zTZ@Sf4A1iCsrRTOvxx{1TW({(+<2?0;c0G2dA2z|yyNh!r3z4sj8(2{Mm6U(DN6P(
zOXKO|1rlAOrU|TjzZTs}a3y9QflY7Q_#y&x!2bYsaaNCnZ6J?xoMVx>vs|3iQ_&v$
zBUz)*uCH#cSm%CEbH!163#rG=yJ_r$`qy@BAc4qO)7KzWmilX6F<9Vv`RN)GKT6fs
zP+BVvG<l!A&Nocb8_z|GJ;jgR?oYjXZ-v_HPPUR8h4SQ4jihkH+t=$|FN8iCOkHj-
zpxCNI2s6q50C@KLS8SdEwq%r)^8|-KIQzd(O8FeaH^atUl4sK3D$fk<PC6bRsmZ$P
zP06`WBY#@XYc!dZMhJ3o^BVNO6Z|p1TcvAQB$Z!`C?^$xb@0btx3_7Y##M<1OOaa9
z%&^jmwM2TdsR`>ea#uPv^`PEu-Qpb{H6&GKxON)=EPuFby3sWM02Aq3Ueq(e)W*x^
zCOyHfzfQIA_liTOh;(D8Os5}ivm{{pkzS-K#-`G_F7a1Ty7K=3BLm>yfZ_2C{{Z%l
z<a@0xTnNzr02})L74%|D<!nfk7~lozUT33x9r3@1EPu2-9jU?MJr&7xfri%leQM>W
z#cA{_ud#TaMY_>02g+oT4#W7@=vWs<odlqNJ!g`{;&JnxZjQ-zN2l7&FqIV}9%(&5
z$*aC=#k@LXcUZ?X&kF1MCB~j_Z7!p{VltvO9cnW?l4^c?$jYyrsPy_*H0vi(s~+wi
zil;&?JNgy`@>>|diZK}k>>QI$d#|)eWRq`{IxF#kP+VNCv~4*e-XJ-4;F|G2ieDY<
z^d*wdN0BaXxl-ep;y$O;*F0A<y9(9ZBI#1qokzs)i*Wcl(A(*Y3u+P!N4-H{J#pB3
zn)&|w#JBp5%$M4Y#m1p{{bXWYfb<{T70ADfmg4$r-9+48&mrE#k@KH(+OsvE8tRwv
zn5{zT)`x=}A9=p5)Yoil)1#zS+2pJRYWH;>=H8{^twc+<)&U9r`5T!2Ca`X1)UO=L
zFrHF<Yo3A$WxI?QIdTSEk=Lg`*1hk>Uktn>s7c|?AHyCm(yg_tZ8kXdCAgWA8Lo*e
zN=x@n0SbC@D;k)|H>v1S%HX4Pc<`2MQV9(k1DxQ4Ur+wdJ|VpE=8dUoxA(?-tvD;g
z=O3M;ZV1nC4S5a6z+Z?q4t~pT7Nr>KZSG^0NcHQ|y$it>+P}fu>x<}gdEvRZgUgW5
z6Bxkic8=WFEo@yGD^y17EUvETG<$n^YFN(9?c2r%Xs;)e_J!Bsk$@3^PI0rP6~pZ?
z{h6oKvE9-0CFYfRq{ufUwvE>#BlQ%6#QNo;vPG+1DxSh)%{ty)Hsx@2ON({%k#avn
zS0K@Ji-|x!BDRM-W<+6K#oeN$UN-RrL~PV&Y<1k(&0B-whl=c!q#Bfws~nNRIs8Wz
z(Aj(^T}?Mxc%Ma^Bct2}Kdop%@C#m;gKw*7uEVA;5Y;H%k<DY6Uy9x^ewU3f#y)7l
z{*}9+d`t2Dy2<6)TVEp%w^o7n{W%rT+I$Z9gK9_FZnUQ(rX)jJ_TLJAF556@O`HD!
zx>VN68W_s0&VJ9vzY^W#dwY1I-nil=QTS)2Y*_fC;?u?D#>y0o5>+c&-@`u^^yvVW
z-B0_Ve}!m8FU0#{wCVf3!(;kYC>gqjeT-ihu3R#*wnjd1q++z7_-*k^Rc+H=!D*hX
z0uSj{7Wd+#0gW|(ADXShH^pXLof#Z^L=7B_k7?Y^PubhWGwyru5_x$%h}-`FuTu}%
z#?sx8>YgE*Mb1esPJb%dwbXtlTaaJFhEGt@oL0nN5Il8d6D7xlA|Q>sys~)wO=T)`
zdm6^7Pg8s0pX}f8_Wu4mdH&aXDCLV~Qy=dd`q#m~3%(ur7f+My8YPOzc>-8Oj!$k|
zIQ8PbGSL1g>$(v{8ZE*F;CYf~71G7=`^1v3mNd2sIs-5N09v`Ff|U1W>lmY=N7xtE
zJ{HzwRk_jQP^c9e5)%Ws9+g*FO)tZ`UAz_=66z4K%Pe3&mG$Q*-x#lvkH&u$HcHuT
z1BJ_zkJ6G)ik>aMQijyBu-e-tarz1t0!jQtOBkfA(e(=Wy4Gt}g4@7$7n8dZg7#gh
zpP|XFV_NX%#2*stmWy@a-7`edvGb;gk*vfK)U)TGYVd0x6?nqggTbfD<a7)eitD2I
zyW^Eb$Ax!DqiB^Sl~3iIS4XgvdYF4;_dQ=i@E3<|wGuV27iy4*#v0Vx#?lUY7F8m;
ze-~MNJkq=;Inb8(R@I*i9@Z+#f4tu=Fg+{hH}S`fo-$Lx5P5}*d9NgdlT{nW9z8L!
ze+@#oJF}PL@~qtn-s(mBXfB7ZTIzl-@TQSrHP4J?&~*ET{{TmZd*Kvr4`tjl_;Fq8
z>AF1nEK^;0>U)GByFI!i#&gLalh4;R@@B2#-EUDt9;c(nw>jA)K5D7;0ke<7sQ&=M
zL3pF&Sl|PqM-J6f>?<0$YVuy>tZ7dXNqlj+_J4=;KLmKH+}o|2-b6Aa)n>U^2kLqH
z3gG@H{4DVIhSSf{{6uepftz-WGB3GsK>RUZHMfcNM?dlR@}8ijM=q~@bl8&YoOA<>
zR@A80inQI#>Q|>X?yDY!Z!Vu4g)i@Il?l#X62~6Oq*mUK;q4#9F*KGpz(l8+Cxk}7
zLGu!8#$6g>+nk?pI@6&`l)#Q@S3G>8x0MFbjB6y%XuI%6munzdHCq$UUR3^72lhXQ
zp#VOwXijmA5<iuBNVmJ30@t^!2OUjTxzpJJeb;Z>=6cOx6Zn~{#k9TDdmyyZ6X%Zp
z%&$TF#OL#=SL|wxWR})N>J&(v``3hs7B=OI^ZY|U&ZdHQh>tGdsU0@e{YIZwN7t&Q
zq3%|GDbqBa-c_x!Va7l#->r38FNkBhD$xCsGEQ*sx1XUE;krJveQ*SFTzQG;V+7Zw
zcwMa~+a1hI?0W9`f30ct8Wz;c>-d!PJ(Ee*B9;>OmwJ3BqO5an{Q6f#Vd50Gx5Pr)
zZZnmP{{TGKmM))cVc8X}w!{20%0Kw3B~fk@4`C1=!Y+XRwV%~#M`E5Pw6&4zBjO#*
z0I7~N1HQriO*UVNx6GN0!b<-D2vs%XcOD^*)ym6%0w?=3WPX6vE8mD3PM--P?ir*5
zJ4Yn`m1=}L89LK_PS@iuqp18{_+hK~U(H)9o7LEptYiSd_v4|j&Hn%!e#_oH@U_FO
z%quD>AUemnPvR@?iyzttOSbY=uM!7F$Pz?OKPtbe{C3lIvcvmNN|2GynQxQ&*0J|A
zOF?TbTR$~F;Y)a~QZZ!=nd}z_^`w7*zBI;0gCm2^SG|4xs(8Odx{gU1^pPp&%uu60
zhI3q#{7g+g;}(x&F;kpJD<<rIgjSbKcWC&A<KUl+ZMl<0k|F-t09AGPH{yACdv%9C
zubTUBSNMu`yJ;i2XyEfR#48ym@gw?Imuj{V3#XPV+{8{oF$DfNrlOfV?H(hh`~kRN
z8&bVw3<%ns0meF-^lyz?yt+SyBGP5IO+&+N1SsjWkc@z1=3Yl28rEnKkPAdcdhe@t
z8f*qcN$#!E)tl!@WCbzq4MgcFt<9-daF;VzN6I(eAkc3hTXAPJPz(UE$<rR>^HsmL
z^nIaLJCGNF$TjwMqwpW#&be%5n^)28?;DcP+?e_LbgmOx{h7WSM<cbLh~bq#hh^4l
zl~Ji5hDWFMx>DYnK5H7@lsYZX<BCVu?3aaEPo{BR>o@G>;K<nQ*BIn<^9v8qRAb-|
zh0uVP^3J`n8y;zLRoTc+y)KYPg{O$3-NZzJ)RSEogmtrjtJ}ma3o9@s@^H94>(=c4
z3+Psk#w%Bu)B~AR<F9(^Mu>F%HSH5mw}*Na-2&}BI6W$5PIla%Zx7{1xBk$+KG6JK
z@M7B6PQ8w6n35pzlZFH$=lNH`+Pi}z5PQ{c6l%UD(5|JjxRT)^40eWIF`uP!(rY(z
zpPE2V(!5oMlw{jJf}S>-rlO6Df(qk2)e;C|YLV3E1DrR~oA&7&_-mm-q9<Mw)iU!O
z15G@Nq;H&cN>hGMRXF<AiY&raQrMnRfmNRvrb(z{_&-{yE~4yk6BSA(KGNi}eia#9
ztyX)-&)ou`lHu6*s<Se3Js03-hf~G>02%xlp}or5CH|)j26s3Klhm5~+g<&XJVmBm
zNMbQu#^m6m0zW+0=NH47q=(1445;QeOSEA5=s~ZsqWIIVG!d+RAiIH(U@FPLt+{5^
z;>m=p^*i~0XMJwes4$^8!;{VjwPE;!_I{nBIJTI-t8x$FSTOu>@kCtz0K!E%KkZ<O
zzYpyL;>fug&Aajc04<ek1q$iknCf*YwPubZSokM(9E8hc&fcMR<MOU!Q212#ip<co
zL%vwouTA?!_?4I>@oCqixkVL=sr*}gWf^XC=;9}-j|5k2Wml>>q^Z4+2Gx8xE#1`v
zk}CpjDrIx`kI+}KU&(EIBR0^a^Nd!vg(2{}vW;KNg@MU*T!II+b}F`7m~ykit^7NE
zaEkFEV0vU$lwJzfq)dx)N3PR~`y;~=d@R$MWsc<@QggcI41hkB+@`hg`g90&jWtyN
z002O>=hesZJ4OwjjT{Q9uPTWj0k45INBi(2dVq05cnalmWVCI*y=(4`M@8^9qmu@&
zrO3p85mEY@rDgDk!&;nshOeVS3E&Im&OVu~>Eo2u+Ofx92;kkNbK=N+7$W}wmJGS*
zM-_S>3$o)YAi?XNwf1Gtz?}_aZ!1>Pq&Uy~ey8}?2d(@%(sbzwk5RLdT<{^CWFMt<
zEKFX<5~R`LR@yX47#>`)9^{OJSE17O;KEt))CC~dv)X(+Y2jH_pl;dolpo7Ab}t4`
z6M1G>2?x1c56Y+7ru8e?%c13(E}HW&5(PkeNWdTEU9E?Nkbp|S#~pGH^{(qk@ORoI
zjZoXLEr&amvVWy}uY`UEXcxL{65lPKn#1NlY>qO2TF#~>l@)nA5~x(0v{CbQkA${u
z@}y)wk*##L9u>PlWcys>rb905en!5dvi+WX5}{eHY}5YYi~cpQ7r}1<TLI=?*yGdt
zR96(UJS$%jUsSEw`5#u%HA_fjo++igl_M;+=>ZKP?77bsz^}l2pB~)D;)?*Uc%u_p
z+#Rm0KZl+{9`*PAm&0F%)-x=Ld!>g%oXe_iACWaasiS-}F}b_eygO=7-kKZLNq;=^
zT+_}l@c2I?s-7CEUh(mVhQ1&8Q$s+WA<#8grXVxz8C>U`xUW>wPr<ebv+2`Y&FLkn
z0R3^^yH6A!hE~wXu>3@_W&rI&H=3*XfnHCn_=CfOSB~3RvD8T6WLZvI@FR-e71e)w
zE&eFe+1YA(3TeLxJhm{y5y8rawqwuZUTyKaL-=i~crNlCD$or{!ib$Mis0_&n&$Og
zV8nlMb!rrh^CvaI_?yI1M-!yCkbq8pR~%PFE}iZ!RC4|m@K%AMc)I4>#WBnMkg*aP
zHQL*cLtdL};V*-CK|ar`UBp;_%TfJtUL&aZi%y>NH&^CT+%r6xHD=Gn+8u;Sv0cW<
zy%h6`)_l@;yDrg7W85bAN#JX^q!wNw!m!WIfQrYL;60f}HoD!cfb~fR0Ut{9sD3Ev
z7T}LIHsii9YRY_P(xb*5Php0v?E8hUt&L0XfI4}&x1Q$u&I4{)fVS$y{{VSM-lqXT
zBfs1Y0d#8xwT|XdZxa<#-zhxxAaoU$za8}G)Q_{IzWKqd>+g<MGAYHH<D4Ayp*byC
zNi=ju!GJ^NZzG}JYP()T1D8fsBZgtld-Iy}?L*?Nq|KXn-X|lFr-S%cJ$>WZCqpI*
z<nZ{wHLAM{9+RtUtig(y<J`FO5PaGC_N==<7t5!z?qD{KanRS6+<4MiGxKp-@8Ve`
z7~E+|FujkV^skB*$jBvu7%xINuFp>K=7DyAGa|M>j917>uSic?(YNsg*cfWb#H7!#
zhln(JfoP_Lucil2tu$)ZJAjSjQO*kbSIV~DCY?YVi2Q2$c)IQ{yD1-9#imxh52<aZ
zvn5+{2F7aY+-ZdYB^moJabGF7#9c}yz?|dhRBk>dYC!RD!lT*T)l}t2*Fx({MPnky
zFgj#ZGhAu1W6O0>{n3M(`0`(gI;MEKlj)kPC&X=7^t_RO8h*uh7qvdx(8iK4k><R6
zHP14|f=A)ry2oD&E*N;8UcbX`0sef~=Y*a=)!#YPCLcyMN^coza%Xm<bNEVWeO~=Y
zby@euvYri<m`UOr$vOTA(NF7K?cST<lB6!ZspsBA#e6q6i0-7xYsuR_vbXrvC)Xw?
z0(nUHkpBRUY^htRlc{s;ZAVV<lQ~hWUpLjGtV^vM!XS`!E4KQSualoyo3xBS;6|c4
zrNe*AXZ#q}>IB~B)EAm}gP<9E`@DO~pdZS)%}-MBg^^(lQ{0d5z0L)E{BNgjvc7%I
zM=i^ccS=X&RHI>~-1fP=QQ@^x%VQdm$L}RAkEL_kr;Bv??QWuh;j}VFP#DG!HRUp0
z0Dk#3UGB*Hqdt`RQGC%aMq`PZBI-zP2t8`T-85r6f$TZyT<o)#$l5=`q1=6IDMm6f
zQnS$N7UiSB1+Y256sGmSI9;RLjMp-=O1&MwI(psQ59256T+W<L6;C5*$VWm*^s8aV
zmO5s+<-54*G5J)Be{`S1r|_ec(NwX&6YVEFn%MBXM(b4a)DO7FpTfAc)UL<yFXdf_
z!ppg?d{R<SHvOH>0QVFqa=vcHkhOWLUGzO8!@9Pg;n`+gOGg^4E;1L)D9i3D?x*n!
z!joD&HvSX2hH@|@7rZI#af-5ZsBVy#2l4M!-@-NyIYX1)w{_yxq>f4D6{ys5rrP<O
zY1S>S9hvPNf%STdONa@|MnLp6*va7QBnN7o`e&NWxX~?b`VL29Svqv4?_;VNhCc@v
zPBXX39JI1K1F!-0sL}}wXOmqJ(%w^up>M=h>t72=E)1M{fn897g7;OA3lWH`ik0h4
z$@-Z(OG^`vn4fCzbe%F=^xhPLy>cpru8a034IE=0wZEob%WW`LW3=!^Dc0qVjNx=#
zo`}><1IUC)B*_|%w2LmE2T2T!Ke{NmGR1>}(LZBIbYVPxZMui`9+l=)$3L{}j@nst
zUbedlHKbL<F&y^wsu!At<im*e4_9mou(mgwN~C0c1`p>{8%~;83P=?n{+i_0QN=CP
zjb(th*s+gudi#hDr1T*MYdYY&XL0A;;<_81BK|ZiuM|n^7!}N3+bnVt^7)56^Ii0@
z3V#bZD&br{BvoMHj$4X)5*U2uqIpytC_jxNi5sBD)K{dcD{6W3p{|CK+$#45hr2>J
zT-6UEL;OeZ6oT4ZWNtrNiB^gfaL=)?X>`jfkQG<CHQx9`Sx>cui8&*y^fl%XM>4l0
zkHWhD01HbbS1~aF>^aHeyqsH8gS$O^99=o4k?a;1S6Ye$Eb2+>4l48bLTlKuB;;qV
zIjA&^LK(I!Y-Qb#yT%Q6Ho86J$_Zo&Mt=2tHCmIV^w9hIRH0Hn{-=fbi{RXIY7<EE
z$U!+7teE~0*h7Nl-?=?2==r=GJ;*=3hw5bMT;8?dW`l7z5v~Ss7@XH-EY}k+a_1FX
zrwb2xSn?e|;r5qpFUt}Fz1yvHT6c#nM2nU3*VuKd^4mdeCIfF_#_EwVwCpF?S7WJO
zwvJ5XQN!Vt(nnRI_-9P=)U;>u#woh|Hi*I-X<Y#S0G?_9tnDIOjsk;<>g+Ta@6&X{
z1dgaHm0;_0UiU7xWrtB(Sn~Z=OC2SaOS`vKX4*58lT{#-OTN_&h-YiPyY@1WLi(EY
zyWa`4Pq<pG;5>jAc0Kv&Rx~dK*~xy_H#&KS<%Swia@h1I(z<@DDJHu92P4d(7N6Pt
zPd&c)SbRZh3@EaN=&SP;A5Nm6hxS;x)|HB}v~m66!LMl_4$?{5TRZ734mb+fKb>^C
z4Thtp0%X=2c*i?j1JmBRFx0D3_hyej9gLw-bJCu@nm$1BAMEw2cy>pQMU?Hy1#m&G
zIsP4AE0Mbx{`VF4FNw9QYq^%*ctn`#Apn7&TIDrQ4{4W6hl()R<+3mbrYYiU%7(k6
zOauDLN?J4I$bK5?a;ZB;2fb@s{4Lc=0>A}52q4$eHvR*O;xKLY?axwp$I`cS?}ibb
zyO}`8L>;T=>Ss<nBkQSUH0AL#<#qURaS`T5JoVX(R;8cA$q31LIPZ=t?E5bU{66sZ
znzQSWYI>xnhD%7v+w}srw2upT7sOe81#jR{fYDvSjD93n6tk++S|hjnFAGI;%}U+v
zr|0?~9ou|2bwzebWK-E&kx^dw0@@WmbP{Cr$<Ax;Nd6yb7O<GgDH;3TU-hc`uYt5X
z9WHG~T|3T--3u^|EV0L6E06n5vUid;jXVsMqS^5VpWy2`ZWbtP14>T>sX4D_(EJmk
z_-{&#U8G4ggMLDQbLsT2QHJlsvDkT<b(fcWBJN0}i>IbdMg5HqwP(LjiU?&~;K$JZ
zb>-soNn(F@dmhCMx-nkTP-;Jq$nsb&Y@Ys0X0&T&&QZgMMju~V+0&u1e>GiRGN&6{
zb^2DOzZQ$3n4z|r4bz+=M}{EQMWv<Gx2NqPRu;)dCmB8SUk#WkPEP*n_56?0GVHep
zDqQ}~-lnux5d)MxdICACQdzaq7-fkii2ncy83wBeZeuA0vLoC(ZB@?)*jJ$Vf5S37
zNN*u3<Ya_yGhaNGAvnqxciVH$%`iABQFC^`;(6S^3kBQ;Z8{B;<Na91-=-?BgnkCi
zb2N{uK&>Ol8PSud9Wh@)SX$X>fmljGUfqpp%c)sGn}=mqBh**wnF6mEb7EURDaT_f
zQj>-in4TE$MZSw;9j&}66e=>Rql&HK?*Zt#^{ZRzb}-#39V9@;lTrAc4UsCWq;vpR
zFRFY`w<NO99Bq%|JuB~}om8bBN>Dy(84JbF+G1yW>dzqYhwScrJu~WF51v~Y7v^hC
z-@1PI$7;&4_@AZedQA5k;QJn_WAaOE!{;sS?OvOI@qJ+gi>p^8V+_K)6XIT(W8=Mf
zwJ8iy%FDEfT~s%@uADAwI@)4Yyu%Tb{aXJ3R(tM`HTaM5msate)UjKWXJg8&ah{{!
z-n_~5JDaA<BV_k#_NYD_#1x3aMnAof)jE7ThyY}pKkslW(VlBfD@4MMQ8lAHR^P$}
z*kl)PKr5rtz7pJOfNfQbAMdcQNnI)}A@GY6W74c9poEe3ZKtN>;MbF1GpyC=9_<Xi
zgfywC=K7bxd9>&Rw>NM_5#x6!Kb2*#!&@m3XHbokJtIaV`B$|^sYx7PG!svro?-$k
z16CKer1^0{YagnNes$+l%&~HIn>{(;FzH?}(DS&wFq&g#4My=;cS%`mPSeDGAk!BR
zc!yE|a05-|56IQ|t?XdW+qGDuBm1OQJ@<+{F{g~IlitWZ1{d?L_+oJr(I);!B(S)6
z>n>QYQ(iBKJ|R`XzqgS80D4<E*^kuGTy(z{bogAXT66&i_s&ff*L-jCen&f5eg6Q9
z`JO!>z|ZSYG6x*<Rvhpu0&!nVd}|;hY2%NjNH6A$fEoDitVJ=#4MxWXnJm(ST)d63
z73JDD`!hHA*4CfnFA-TVXmy=N9lbGf!}`}F-_obH0p$UKasKsElOd#fzM1j2#x`iG
zo+G^=afX+7^R3%YkA5*MF{=rP?Id8=j=+(FfPfFo)!10a_dZ{strVYgHwD=CVe$9I
zDmj)H`YtJB{h~ZthXO@U_X!o`&|OW(&VMSq6wG*M2k@;_nUkiG>(;-t_loVv-7#-t
z&-1M${i^lXTul;4JMLaTO7TFd-Nh%6r=h^7%*^>C*!Me6+Hsap*BTy~ckMbjAI`g7
zEB3X}{H05ehW;P8Wx(8$GLkFhMz$xT)Guh`IC`&W*5vyr^ghzlzib}|TS>a-;kC3;
z_FGBEADFI^-}dMDeQE(;0o<YXWIxir2Uoa@qp&01rCmnh1`gKu#V=`n!BV!z=mcNz
zP!9#5!gxOANBhhE6=^@<puP^=nDCXz1Fy`#;a{I0+K>E74No4fqZ^YSN?gf@BS+|z
zU+_^M1sL|fgznXSnSa8dy8XSp19=+*px#1A?T_%UfFsq>hB*AGPNtrVK&*Y4sQ%af
z1HX17LcC&qhxk`BsQ%Kv3%V!D(C-wVVP7gXt+f203W@c3vKdGpS{%lpeJdx&?E*<Z
zb)nr5eMl6G@xsDF1bP&M{<=s7eAe3K<~Wq&+lqnpd&lcD{&agr9?c(DTYPA=L7(hc
zzwh_|09vib@zM}>rjZJt`s%;oUn8cYFITydN3&FeQj4j%mA~Lr)Y!e5_1gSqAZ|Kb
zKzjMhRj&Rz*$9fyV`>zCeo_4^<mc6`{{Xz?`;$;ft-%_f+=kC%Po0y<eKB+K*G6V5
z_FH5h?76B^d~DJSWI+xOy?olfA%$4C*;DeTj}=4+-DZV;l}fF;1EbXX3U7`vMt^f@
z!RX7zDWl_jj&Rn}GWv$Wsjhx4{3X>s3`+&ho{;#OJ8-eb6l6)}0m09ue9$~qI}iaY
zp82YEA3{|d?tLo18Xzb4miIXJb6Qqk7_@f)PHy5}{BEz74~cFtPn_94hc!~;S-ZI4
zUdW%NFKWy73A5}K_{HHA+$O!$K>q-ya4G))*{V5yFZcoQUsctY9Aw%l{{Y#m<3~%m
zc{dVCp5e1m?Vd6IwZ?rxsP>Y%ud#{s)}!%P;qfcwYWl2zpPzN2gO6{TrnmTw@Q&?3
zy73Q*ty>s7bjX~3l2^y+bs_`0VA%fp(+;gTV}=;?0-@A>N>pNfmV9CGS<@%Rn*Koh
zucNB|yUkHPGx$iZ<{ueqsylAde(&SHzC!-il>Op<yjK4JhcC3Yjzg#_$o-U@RmOMR
znNv$s>06H%d^uu1&EspVdP$|fKluG;%i`Y(haw*lM{=LvXpCp_B-hNEj9z8bjRI_h
zpOr&&6#<4tQJ?2Sn&=9OdY@07Yr^wj{{U@7eU}v~>b?<G$-j3#qChiWEi7^#Lbv8>
z^646xK<fm-Pe$!i?HGF%-1<sS68JXU0wj`F9Z719ej<1>+ziJj9n84JeBN!YV*dbJ
z=KU%O8AD)({9818Mu+TIbLltuz2M05-b{>f!1GQ&N`w9vw}q{S*fjg2?w1TJ=Mf<v
z!ZDvpu?_Huf2#xNDSJlBRMI|%KeRrL83lA%<txy=%L<AA0BNfT$P3}+!0u;bUJkm<
zSi?u3!lQe;SN12=RJpcNrxV&WKiWG=)KIM67q^gf`PjGfu1`?-o8fEnOZZORGyTJg
z^QTx_B^h(+ROMz|;Q{&@#jEaYk~@2!6?j%&SHr;%w3U}}u4r<cMW%Uq9Dee-R-R1%
z0ITAdW+ZgRYdb5NFC$`W%Sj*iaFiaKvr(Ifal~U4mF5yTrkPK*9HyG1QLZ9Czy(C_
z3H}lJ)_IWoQw)F~nw5aM?H8c@>N3&erBpKv@l7(5&{d!{bgR)}soPG3BS#=$-j&nb
z_`*k!y3etR$=lG^0<O&}=ca1jYo?2|Sy7KRr(V7-xIgmn8|rCCimr3Sqy7Tb$L^&M
zj)t}7FWH@%_`2<SOxXI=<Hq+<@Mn;FX1H-PdYWG~k5Xv3`406yAWkv5c0W;9BKU~S
z_oiNbQoPtoedzO8deplySoIb0{{R>lKym3$PajB!DSn{)it@id6r$p%l@yO#Pl>mw
z%HPHKsQ&;8nLa9NBN)Gg{{X;6c>e&}IO=H(+3!n{0ql?QUsHU8dk!(^D!e}w^$TWV
zlUue<a6rJXBJ;f|MCPlN3q#nC#r;vRb82z)r+>oa)jR!|pZBp|HfE#Fs^vg?GW=uJ
z6#T7hgRjcue_BWQ#i~9@n%z&RkpBRUc-NVp^cgETQ%`gr^4Xre-xqaU=k1XGE=@JQ
zF6zW_c>u?uHROA<N&_B>PvKMSA=wqpKIf{x#eHG9`+6UxM1Cmh=l)-YeZa3W5lDZ9
zs#xtOZi>~SgjUFoX&#K<6?MRJFNff&DmM73uMkJ}gO1~96~iu_KU0Ms=AHIJoHUBr
z><v<|T<b3WB5PMnUQI?vBx9{)-T0?axF5W{c0XL!8?;~_nDyjUS!ElJ)5l(E=mwP9
z%%q*t7y9?9ZEkL_2@$I}0Fp;eJ?jL?xRdzS!``G?6_vbXq>RkcFylUy7KxWr9A&eb
zc+|k}SW-#*&y(&BDMaK{wk=KZs&IMzX%3*o51T)YVGDy!6B2571E#arR>CGN7-xR|
zf~s8WQ{8}!<V2_CPQtMsV#hrxM8Az%ObsJ3J<SU-&r?+rO!cW$<kUSk2cfA9nKhvd
z{{Ssx!uxU0TC|dm)q`N2&A4tj9M!QfW$TLP*5hddHEnK)PPnNwxoYC;9B!=nu5x<S
zYn#LMsxn><b4!^ZdY6l|t`$y6IH<0b?I&oZ`CU&rswVSqAoT7jms*6EWHX(_3<7G?
zOom!%s=Lah^dFs3mT5gsDgwmyCbx=YJxlE}bjbYbC%8e<suM;<U5CqJ!MOD6LAgBu
zs@e6W%^~beXLH{b8xm=`q$G~QnKrnDRBV~-YC>~KLsu)65cy#Cr4Wq$+Ln}JqnEf@
z5eSrj2%}&fg+02`QDjw;*l56~15Bk9`B2AxCc6)XSw1Vk$tDNs^%ck0Q}EkfkHfwu
zTgh$y(I1q^P#h7GIUf8}xTdT`syZv$M|G+17SLTk%8<y->(aLL?+a?$fm}u94^}ka
z3VcY=@18Y|NR`;|lMFi7sA<~#0YUjacoi!Qo~MalJjCB)7T4gb-P^2^>roeOu&itt
zel=x20LrS;#Tl6wtINf8*8VYq%{LMiy?tuqd{MSYN<7<_A9S9TqNmENk1nn|5n11q
z@~uPR#MUlXV#9;qy;WZYxncXj20DuG^*<HEtPEf>e+!e%FNeHCr)byWYwI<?gCFYi
zf(<y+l--*<{ih`*Q<qcEB=DJtjq;K@a4NHCYiRj(%8uf_8&lM+wVg&ut*yS#3o+Xe
z=M`RUIxU4{Qm3&a6~kVOP2XeCp<PWSZA?8n!rJ=J%h)##K{XeVCY1RRjlSJ0V#>{7
zau*yLwRdilMnmIxuO3(}9=AukPP)3%$1!_5T(0<6Z>R#WZ)~n+>RX=pu7Y`1+E!fg
zR3_BbI0v6v;81j3qK>(^)K`-^xqLT0!1<h&`r@*t@WO|~9R4^p)ZJ<?D8SBZDtM=p
zkgVTY?Wa;}cVo?|P7i3@#+O90fZ!Hp^dg{=)(}2evGp~z6`bvp<!<&g!KFK&m<9Z6
zp;fl>Fm!LUOi#0%{Hw|BNp_Q2$|9X4h4G!PyGS2HTbA&BhUz&>g=p9^ox6rX$jI(1
zN5|eQ);=O?Fy89M$IOBu5`CH>7s@(k*00*d?##WMld?xKV`~(jym=HSa>t*_wyo|2
zAbBi7^&+$Mt6Mm=D`wOg{=w#`x$-!ZXMMi)v*MBA>kW2j{5PoGUuio-M<w)IC0zAV
zJLA@|m%MaG-8A&=YFJnR`%h4D*1HQI6iX$rSugHC-BsXMms;vk#K*7v2Dj{W>vTbu
zWj{*ut59!bdK4?hI!T_LrTBx*hhp4bJcFhx9YXr)_XoI;6WzegaRbEBs6JfmGycs(
zCbuQf+OjkLHHA7gFQKICP_@pE^FqD0wvPFe%(sy345v6g^`WKNzSVu9KDEwiI>XwM
zmrpT>^RZ4x<y|CxBZ75R7>Evl4!>IUXj7lVBj+)dB9}A~*J*aJMo<<Tx{+NysZ=3Z
zSPr>Uj8~pp_>$7!m%+z9YUBJxw`B5{o;a^!5sf9O^A$a|B3dKT6YQ)8Zb}k)QYmM9
zf8ih#(bVv3$ENXYqDFLfBhxiz)5li`zkFw~r_Dj=^d-{7M`U|MHqprd^A1TE45tRD
zYW^M)TFAE7PXwf8a!Ic^hvG(|ZbrkqKf*b!OK*xYODe-~s*+>n1Yp*6C5CagIaJ|{
zjJdfRHuTF23soZW>`g8+=5jh#4~Vsm7S3@e+m)Q3^=Na7vvJ}m?hxh})=4;KW#HE(
zePb(6`>3tfJn{g@BD`w&dd@tlT^~_^#Nk}|>q*7-Jv+cR(9L%+ywf#H_|h*iAktx1
zcJIwk;SY_z81aU)eWcph-f5RMZLy%f#_)5`PM=ERJU8PlSHm}=eLm&wBn#z}3BcP)
z7N@6acb7Vh8f~gx-7#q<3G+zDsbgO=TMHOnP8y$899yfbUWd2M<2@%!VZZl{+hg)-
z{h!584z0USb#QH>PFdt1DEw>ZuLYO!$HZG53hL|ZaZcGGzCcEL17BVEhv2q{rg&*>
zq_ww})UFslK5xdko*I=kdmOlm)fSAcrg*ze)%4MGr@KoOgDEPm;6+r_V)0I=Y2@hI
zjA^^?EV1K`_^p2l=+@djly>&;MGPpoRbRRmw{G>QELu3$NY{DVy|R4;d9`4tbe4v=
zJY=xbRV<rb8o9>kzYe9-ELu%Wt;O`8duYz%{V2NlTYs$E7F{FEwg(Q=F;mZRiu6q-
z-sWs70!ifh*LkN~Gsoq^<VHt1?^)2x=qBeSuS3?U#2C7*xVwK{51cfw*{?&ib(-$j
z&2r7V2*d<FgV*a_Y+nU5&Og!6C>Z&<uX9fhah@`I8l|ZCT5%h+1zEB8O-%B6PxqB+
z+|}_GDt3ab*Ub2z_MQEfHLY7qu<<W|uH>}w9+JV5);WqStDOA2<92;(7s1*e$BVrI
zF7+F|GRsl>tz&keu>+wjykytb+NPU!*$*av*<<=wHGN||4~J99>b28L9O=Pf<4QJL
zT`QhODwS&Tb)`|Z{Lh;&_3w=u0~WsUg|4d-pY@iL1E0dPXZZW7StidAcq3M_^<{;f
zN8~H$y+!AT198FWSu<TnkKgSh(Z*}trI<q2of!T{%hjb(OG=Wt^WDG3&jfgzQHxKG
z?n^B~W(u~~5tMBDb6Q?3@a$TBft=dhT?~KAy6z+Jt*?kb4SX5m{Z`jdywg%GM%l!P
zcW;Q_Bd1Eyxz_Eiq6j4<Du4+-mCKB)O0$c@t9M7zWmT(CU+ou{spGbO1Bb-=e76>*
z;FKH~r$VPcjdD7N!yQ`L+ct%OJU3%r$EtYaO|v9D)Nq{i-IH8ix$%=i(^a!_WntSq
zS2a4An7ct-_G+lsmpoT0^*r)#2x@U0Y%#Fzj#m|xsX?ykG46v=m2vzs{KNC6c(3+`
zw;RQ}#|rxOCcKNpKN>D|24JQ~&v!T#@8M`;Dp}M?BZjsk4^3N|@5t___=&1%(~)?I
zIL{;jR!_z|SYh6=qA(nQE5P*+8s1*9`L^fMvSrm~l;p7|-o1x_rBQq9?mX3+RHc7r
z{`1>4pN+Ru0xjRmA55C!wQq=*_u2EVrIS41HV5Tg6t_h5W4o_4586u{cl534VCYi#
zVsTW=F*UT5&C8z>O)C?-zgo<>)MlGIq+|HiLneFv6&pTD=bG(Nw#O5cp6phYnB<a8
z6%O^!RR_LB6?P)##F*PvituymMHNV<GETJ-KQ$CnxhyM6&IL<w*yXt<iYcN?00C1(
z#2P57Kw6BP@m3Z?_|ZjZWg<nGj8v&J4wO+;jLAIK^`zcUerTes2DWq0G}5$DS+cH1
z%15O_z&F>{iYW&`R%JXInPuELqKYm{9ti`Q3GYP|*oN-njGpvSQWqLfIT`ItYgo>3
zFleHgNUbaZP60I$k7%Nb28fwyj3}a=^aesy6Vj0FJO#xRQE(fN){_L1GlNAGhA|w*
zFnK=JIJj=ieR<}JD7i#4PIHo=gX$`@imnOT6#G#{NFoei;M9>SoNxvYQPj~zA>7d1
z#YfGb!jKb16-1AX+#IVB^sN!7-YiN!-7p<C^%PM+ikB1I5O+*u8SXJvHmPDj!S6*B
z6gzJeg*c*$1Tm7DR*ESVaXg?w&su4S3Mi;3icP?9dQ=W*qLFdjjP#@j0*WX#Vb+>g
zJt(52kwkes(ip+%MHJAU!g+gGA9{IeD6QeKwLO^9)NZ1Rvc=p@J#u)WiV;Y(wlWk9
z8G2P$`FQlAin$M9$)w*!6$z7lK=VtUXri=5A=JI7qJi}T#Wb2IszLz7QM)~8qL8|i
zKt}FHYeg<X0fYjIC}hY}e8V^bqkT$DQAK9ioyza1NNOg!cH=ZrMf4rWnoOQ~G}73j
zie(`;VxPRQ0Q4OwqNx;^nZ*=QK*R@{cw<ErFk!zm@YGRF2$nO=T|?4}D=C@Co;K}M
z2jYq=5~4UZO-U@}W5DPrqJ~+UJ<9=uRArd%$)bwqmg8(-xM4@yTO4EZqKdfaX4u!A
zh~(5;^F<Zk9f|_fZ=#B;g~anSOTLOJ1;q_Zx{4?m()4elicA=9ph#$<ku8CI(@RlB
z0GPtmp@^c2mcrtPHKm~)*T3)+MHNWSukhq_O1Nb`2;>pkyB`gFMAfuq5VWT1+<yd}
zN6_`6inQ&q;Yun?Q`7X15Vp5%9OHVc4vcvfaz-rONyn`eRkJE_OJZnmq=W#3gNmJX
z85xG+PtuAi;%8B}H*;@E@ig`fg#xxZZd}#PPsO)7oPnfOQ;*>oqKZ{paIjE{RTC@2
zsXlYgMHkwONH=l~6jzBYHa?mT=;!b6510wbs_-;ZHz-bhO%zv?SsslYcQ);9Ao7b8
z<P6qj?xwcnqQ}yTE6%4CDLbB(XUPT1WKXTJCku+pxYeX2<BZWoda#62*zsyqpK~_n
zQF3x~nyB~Ee|XVFbSR^l$(5e**z!pg8J&p-CW<QTiMu3QrX9^o9sH}_iYfA-Y_gh^
z-`=Yc>eDIVj%cEeNcJJOV!-hg%8YKOH;FEG^)yjQBxLp6fa~`MoyVVgmK|<zKKp)j
zQAvV|TNk0$W-X1)`PRj*yDVU5=K9e^VD&nphiMnm9andl=W}i&JPPil)9keyxb7ws
zdE0p^*aQkFtffvbMvSXgsRwww66*Tb*)4WLtPVKc&0$C4Tb!v=w>@!173|^C<gxQp
z_nVX4vt{BLCvdH}6WmqHjZKJEW!wqC%@kLYi>D54kE6g&N()i3YO~tPk|PlDoU3<V
zp{}mW$J*Ydc@o`fcax$Kgn5YoQAK?1dU03K`if7RbLd}(I?UJBaTy6^01PqBZ~Q>;
zbbc|??Jd69ac0q6EXy=4wUmR}iYw=)+D^qnE8DFO+e2xzu?LlLCg7oR&79-byWbCZ
zqUY^#Lvj{C=NVChMHRzFwMA7~^+&5;4@uK+l=PU$*a+$AO2<D+D6i6>-zCq6d)V?{
z6wEHK?p+H+OLP^=%Xs#7^0bc9d3h&{(M5i3RbIP(Z}UF$43naf=lcHu#XUCAIFn75
zLcC-UKdpK0v+;(|hw`+?Tyyg<qKfwE(u;y^9Q7#E=Dm+G@i)e1wh~DVg6`*+=O3+j
zmyCaC9a`m@FbrUJMG8MdMHTH~C{vTw9C(?^oPE}LR<rRJQPd?lasztT8LW7Q=3MS_
zTR6$0iu!7F;RWua=IT|cKZ%^p%<@d+$mNcE)k1ch`q4#fj$#&OR#Bc0T6uD4qMH{%
ig|kNU`B6m{0X||n`_py;eRD+=${J$l(uydiLI2sraQM>z

literal 0
HcmV?d00001

diff --git a/configs/karlo/decoder_900M_vit_l.yaml b/configs/karlo/decoder_900M_vit_l.yaml
new file mode 100644
index 0000000..02a3530
--- /dev/null
+++ b/configs/karlo/decoder_900M_vit_l.yaml
@@ -0,0 +1,37 @@
+model:
+  type: t2i-decoder
+  diffusion_sampler: uniform
+  hparams:
+    image_size: 64
+    num_channels: 320
+    num_res_blocks: 3
+    channel_mult: ''
+    attention_resolutions: 32,16,8
+    num_heads: -1
+    num_head_channels: 64
+    num_heads_upsample: -1
+    use_scale_shift_norm: true
+    dropout: 0.1
+    clip_dim: 768
+    clip_emb_mult: 4
+    text_ctx: 77
+    xf_width: 1536
+    xf_layers: 0
+    xf_heads: 0
+    xf_final_ln: false
+    resblock_updown: true
+    learn_sigma: true
+    text_drop: 0.3
+    clip_emb_type: image
+    clip_emb_drop: 0.1
+    use_plm: true
+
+diffusion:
+  steps: 1000
+  learn_sigma: true
+  sigma_small: false
+  noise_schedule: squaredcos_cap_v2
+  use_kl: false
+  predict_xstart: false
+  rescale_learned_sigmas: true
+  timestep_respacing: ''
diff --git a/configs/karlo/improved_sr_64_256_1.4B.yaml b/configs/karlo/improved_sr_64_256_1.4B.yaml
new file mode 100644
index 0000000..282d3cb
--- /dev/null
+++ b/configs/karlo/improved_sr_64_256_1.4B.yaml
@@ -0,0 +1,27 @@
+model:
+  type: improved_sr_64_256
+  diffusion_sampler: uniform
+  hparams:
+    channels: 320
+    depth: 3
+    channels_multiple:
+    - 1
+    - 2
+    - 3
+    - 4
+    dropout: 0.0
+
+diffusion:
+  steps: 1000
+  learn_sigma: false
+  sigma_small: true
+  noise_schedule: squaredcos_cap_v2
+  use_kl: false
+  predict_xstart: false
+  rescale_learned_sigmas: true
+  timestep_respacing: '7'
+
+
+sampling:
+  timestep_respacing: '7' # fix
+  clip_denoise: true
diff --git a/configs/karlo/prior_1B_vit_l.yaml b/configs/karlo/prior_1B_vit_l.yaml
new file mode 100644
index 0000000..159330d
--- /dev/null
+++ b/configs/karlo/prior_1B_vit_l.yaml
@@ -0,0 +1,21 @@
+model:
+  type: prior
+  diffusion_sampler: uniform
+  hparams:
+    text_ctx: 77
+    xf_width: 2048
+    xf_layers: 20
+    xf_heads: 32
+    xf_final_ln: true
+    text_drop: 0.2
+    clip_dim: 768
+
+diffusion:
+  steps: 1000
+  learn_sigma: false
+  sigma_small: true
+  noise_schedule: squaredcos_cap_v2
+  use_kl: false
+  predict_xstart: true
+  rescale_learned_sigmas: false
+  timestep_respacing: ''
diff --git a/configs/stable-diffusion/v2-1-stable-karlo-inference.yaml b/configs/stable-diffusion/v2-1-stable-karlo-inference.yaml
new file mode 100644
index 0000000..da867b4
--- /dev/null
+++ b/configs/stable-diffusion/v2-1-stable-karlo-inference.yaml
@@ -0,0 +1,74 @@
+model:
+  base_learning_rate: 1.0e-04
+  target: ldm.models.diffusion.ddpm.ImageEmbeddingConditionedLatentDiffusion
+  params:
+    embedding_dropout: 0.25
+    parameterization: "v"
+    linear_start: 0.00085
+    linear_end: 0.0120
+    log_every_t: 200
+    timesteps: 1000
+    first_stage_key: "jpg"
+    cond_stage_key: "txt"
+    image_size: 96
+    channels: 4
+    cond_stage_trainable: false
+    conditioning_key: crossattn-adm
+    scale_factor: 0.18215
+    monitor: val/loss_simple_ema
+
+    embedder_config:
+      target: ldm.modules.encoders.modules.ClipImageEmbedder
+      params:
+        model: "ViT-L/14"
+
+    unet_config:
+      target: ldm.modules.diffusionmodules.openaimodel.UNetModel
+      params:
+        num_classes: "sequential"
+        adm_in_channels: 768
+        use_checkpoint: True
+        image_size: 32 # unused
+        in_channels: 4
+        out_channels: 4
+        model_channels: 320
+        attention_resolutions: [ 4, 2, 1 ]
+        num_res_blocks: 2
+        channel_mult: [ 1, 2, 4, 4 ]
+        num_head_channels: 64 # need to fix for flash-attn
+        use_spatial_transformer: True
+        use_linear_in_transformer: True
+        transformer_depth: 1
+        context_dim: 1024
+        spatial_transformer_attn_type: "softmax-xformers"
+        legacy: False
+
+    first_stage_config:
+      target: ldm.models.autoencoder.AutoencoderKL
+      params:
+        embed_dim: 4
+        monitor: val/rec_loss
+        ddconfig:
+          attn_type: "vanilla-xformers"
+          double_z: true
+          z_channels: 4
+          resolution: 256
+          in_channels: 3
+          out_ch: 3
+          ch: 128
+          ch_mult:
+            - 1
+            - 2
+            - 4
+            - 4
+          num_res_blocks: 2
+          attn_resolutions: [ ]
+          dropout: 0.0
+        lossconfig:
+          target: torch.nn.Identity
+
+    cond_stage_config:
+      target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
+      params:
+        freeze: True
+        layer: "penultimate"
\ No newline at end of file
diff --git a/ldm/models/diffusion/ddpm.py b/ldm/models/diffusion/ddpm.py
index 6090212..bde253f 100644
--- a/ldm/models/diffusion/ddpm.py
+++ b/ldm/models/diffusion/ddpm.py
@@ -1793,3 +1793,58 @@ class LatentUpscaleFinetuneDiffusion(LatentFinetuneDiffusion):
         log = super().log_images(*args, **kwargs)
         log["lr"] = rearrange(args[0]["lr"], 'b h w c -> b c h w')
         return log
+
+
+class ImageEmbeddingConditionedLatentDiffusion(LatentDiffusion):
+    def __init__(self, embedder_config, embedding_key="jpg", embedding_dropout=0.5, freeze_embedder=True, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.embed_key = embedding_key
+        self.embedding_dropout = embedding_dropout
+        self._init_embedder(embedder_config, freeze_embedder)
+
+    def _init_embedder(self, config, freeze=True):
+        embedder = instantiate_from_config(config)
+        if freeze:
+            self.embedder = embedder.eval()
+            self.embedder.train = disabled_train
+            for param in self.embedder.parameters():
+                param.requires_grad = False
+
+    def get_input(self, batch, k, cond_key=None, bs=None, **kwargs):
+        outputs = LatentDiffusion.get_input(self, batch, k, bs=bs, **kwargs)
+        z, c = outputs[0], outputs[1]
+        img = batch[self.embed_key][:bs]
+        img = rearrange(img, 'b h w c -> b c h w')
+        c_adm = self.embedder(img)
+        if self.training:
+            c_adm = torch.bernoulli((1. - self.embedding_dropout) * torch.ones(c_adm.shape[0],
+                                                                               device=c_adm.device)[:, None]) * c_adm
+        all_conds = {"c_crossattn": [c], "c_adm": c_adm}
+        noutputs = [z, all_conds]
+        noutputs.extend(outputs[2:])
+        return noutputs
+
+    @torch.no_grad()
+    def log_images(self, batch, N=8, n_row=4, **kwargs):
+        log = dict()
+        z, c, x, xrec, xc = self.get_input(batch, self.first_stage_key, bs=N, return_first_stage_outputs=True,
+                                           return_original_cond=True)
+        log["inputs"] = x
+        log["reconstruction"] = xrec
+        assert self.model.conditioning_key is not None
+        assert self.cond_stage_key in ["caption", "txt"]
+        xc = log_txt_as_img((x.shape[2], x.shape[3]), batch[self.cond_stage_key], size=x.shape[2] // 25)
+        log["conditioning"] = xc
+        uc = self.get_unconditional_conditioning(N, kwargs.get('unconditional_guidance_label', ''))
+        unconditional_guidance_scale = kwargs.get('unconditional_guidance_scale', 5.)
+
+        uc_ = {"c_crossattn": [uc], "c_adm": c["c_adm"]}
+        ema_scope = self.ema_scope if kwargs.get('use_ema_scope', True) else nullcontext
+        with ema_scope(f"Sampling"):
+            samples_cfg, _ = self.sample_log(cond=c, batch_size=N, ddim=True,
+                                             ddim_steps=kwargs.get('ddim_steps', 50), eta=kwargs.get('ddim_eta', 0.),
+                                             unconditional_guidance_scale=unconditional_guidance_scale,
+                                             unconditional_conditioning=uc_, )
+            x_samples_cfg = self.decode_first_stage(samples_cfg)
+            log[f"samplescfg_scale_{unconditional_guidance_scale:.2f}"] = x_samples_cfg
+        return log
diff --git a/ldm/modules/karlo/__init__.py b/ldm/modules/karlo/__init__.py
new file mode 100644
index 0000000..e69de29
diff --git a/ldm/modules/karlo/diffusers_pipeline.py b/ldm/modules/karlo/diffusers_pipeline.py
new file mode 100644
index 0000000..07f72b3
--- /dev/null
+++ b/ldm/modules/karlo/diffusers_pipeline.py
@@ -0,0 +1,512 @@
+# Copyright 2022 Kakao Brain and The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import inspect
+from typing import List, Optional, Tuple, Union
+
+import torch
+from torch.nn import functional as F
+
+from transformers import CLIPTextModelWithProjection, CLIPTokenizer
+from transformers.models.clip.modeling_clip import CLIPTextModelOutput
+
+from ...models import PriorTransformer, UNet2DConditionModel, UNet2DModel
+from ...pipelines import DiffusionPipeline, ImagePipelineOutput
+from ...schedulers import UnCLIPScheduler
+from ...utils import is_accelerate_available, logging, randn_tensor
+from .text_proj import UnCLIPTextProjModel
+
+
+logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
+
+
+class UnCLIPPipeline(DiffusionPipeline):
+    """
+    Pipeline for text-to-image generation using unCLIP
+    This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods the
+    library implements for all the pipelines (such as downloading or saving, running on a particular device, etc.)
+    Args:
+        text_encoder ([`CLIPTextModelWithProjection`]):
+            Frozen text-encoder.
+        tokenizer (`CLIPTokenizer`):
+            Tokenizer of class
+            [CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
+        prior ([`PriorTransformer`]):
+            The canonincal unCLIP prior to approximate the image embedding from the text embedding.
+        text_proj ([`UnCLIPTextProjModel`]):
+            Utility class to prepare and combine the embeddings before they are passed to the decoder.
+        decoder ([`UNet2DConditionModel`]):
+            The decoder to invert the image embedding into an image.
+        super_res_first ([`UNet2DModel`]):
+            Super resolution unet. Used in all but the last step of the super resolution diffusion process.
+        super_res_last ([`UNet2DModel`]):
+            Super resolution unet. Used in the last step of the super resolution diffusion process.
+        prior_scheduler ([`UnCLIPScheduler`]):
+            Scheduler used in the prior denoising process. Just a modified DDPMScheduler.
+        decoder_scheduler ([`UnCLIPScheduler`]):
+            Scheduler used in the decoder denoising process. Just a modified DDPMScheduler.
+        super_res_scheduler ([`UnCLIPScheduler`]):
+            Scheduler used in the super resolution denoising process. Just a modified DDPMScheduler.
+    """
+
+    prior: PriorTransformer
+    decoder: UNet2DConditionModel
+    text_proj: UnCLIPTextProjModel
+    text_encoder: CLIPTextModelWithProjection
+    tokenizer: CLIPTokenizer
+    super_res_first: UNet2DModel
+    super_res_last: UNet2DModel
+
+    prior_scheduler: UnCLIPScheduler
+    decoder_scheduler: UnCLIPScheduler
+    super_res_scheduler: UnCLIPScheduler
+
+    def __init__(
+        self,
+        prior: PriorTransformer,
+        decoder: UNet2DConditionModel,
+        text_encoder: CLIPTextModelWithProjection,
+        tokenizer: CLIPTokenizer,
+        text_proj: UnCLIPTextProjModel,
+        super_res_first: UNet2DModel,
+        super_res_last: UNet2DModel,
+        prior_scheduler: UnCLIPScheduler,
+        decoder_scheduler: UnCLIPScheduler,
+        super_res_scheduler: UnCLIPScheduler,
+    ):
+        super().__init__()
+
+        self.register_modules(
+            prior=prior,
+            decoder=decoder,
+            text_encoder=text_encoder,
+            tokenizer=tokenizer,
+            text_proj=text_proj,
+            super_res_first=super_res_first,
+            super_res_last=super_res_last,
+            prior_scheduler=prior_scheduler,
+            decoder_scheduler=decoder_scheduler,
+            super_res_scheduler=super_res_scheduler,
+        )
+
+    def prepare_latents(self, shape, dtype, device, generator, latents, scheduler):
+        if latents is None:
+            latents = randn_tensor(shape, generator=generator, device=device, dtype=dtype)
+        else:
+            if latents.shape != shape:
+                raise ValueError(f"Unexpected latents shape, got {latents.shape}, expected {shape}")
+            latents = latents.to(device)
+
+        latents = latents * scheduler.init_noise_sigma
+        return latents
+
+    def _encode_prompt(
+        self,
+        prompt,
+        device,
+        num_images_per_prompt,
+        do_classifier_free_guidance,
+        text_model_output: Optional[Union[CLIPTextModelOutput, Tuple]] = None,
+        text_attention_mask: Optional[torch.Tensor] = None,
+    ):
+        if text_model_output is None:
+            batch_size = len(prompt) if isinstance(prompt, list) else 1
+            # get prompt text embeddings
+            text_inputs = self.tokenizer(
+                prompt,
+                padding="max_length",
+                max_length=self.tokenizer.model_max_length,
+                return_tensors="pt",
+            )
+            text_input_ids = text_inputs.input_ids
+            text_mask = text_inputs.attention_mask.bool().to(device)
+
+            if text_input_ids.shape[-1] > self.tokenizer.model_max_length:
+                removed_text = self.tokenizer.batch_decode(text_input_ids[:, self.tokenizer.model_max_length :])
+                logger.warning(
+                    "The following part of your input was truncated because CLIP can only handle sequences up to"
+                    f" {self.tokenizer.model_max_length} tokens: {removed_text}"
+                )
+                text_input_ids = text_input_ids[:, : self.tokenizer.model_max_length]
+
+            text_encoder_output = self.text_encoder(text_input_ids.to(device))
+
+            text_embeddings = text_encoder_output.text_embeds
+            text_encoder_hidden_states = text_encoder_output.last_hidden_state
+
+        else:
+            batch_size = text_model_output[0].shape[0]
+            text_embeddings, text_encoder_hidden_states = text_model_output[0], text_model_output[1]
+            text_mask = text_attention_mask
+
+        text_embeddings = text_embeddings.repeat_interleave(num_images_per_prompt, dim=0)
+        text_encoder_hidden_states = text_encoder_hidden_states.repeat_interleave(num_images_per_prompt, dim=0)
+        text_mask = text_mask.repeat_interleave(num_images_per_prompt, dim=0)
+
+        if do_classifier_free_guidance:
+            uncond_tokens = [""] * batch_size
+
+            uncond_input = self.tokenizer(
+                uncond_tokens,
+                padding="max_length",
+                max_length=self.tokenizer.model_max_length,
+                truncation=True,
+                return_tensors="pt",
+            )
+            uncond_text_mask = uncond_input.attention_mask.bool().to(device)
+            uncond_embeddings_text_encoder_output = self.text_encoder(uncond_input.input_ids.to(device))
+
+            uncond_embeddings = uncond_embeddings_text_encoder_output.text_embeds
+            uncond_text_encoder_hidden_states = uncond_embeddings_text_encoder_output.last_hidden_state
+
+            # duplicate unconditional embeddings for each generation per prompt, using mps friendly method
+
+            seq_len = uncond_embeddings.shape[1]
+            uncond_embeddings = uncond_embeddings.repeat(1, num_images_per_prompt)
+            uncond_embeddings = uncond_embeddings.view(batch_size * num_images_per_prompt, seq_len)
+
+            seq_len = uncond_text_encoder_hidden_states.shape[1]
+            uncond_text_encoder_hidden_states = uncond_text_encoder_hidden_states.repeat(1, num_images_per_prompt, 1)
+            uncond_text_encoder_hidden_states = uncond_text_encoder_hidden_states.view(
+                batch_size * num_images_per_prompt, seq_len, -1
+            )
+            uncond_text_mask = uncond_text_mask.repeat_interleave(num_images_per_prompt, dim=0)
+
+            # done duplicates
+
+            # For classifier free guidance, we need to do two forward passes.
+            # Here we concatenate the unconditional and text embeddings into a single batch
+            # to avoid doing two forward passes
+            text_embeddings = torch.cat([uncond_embeddings, text_embeddings])
+            text_encoder_hidden_states = torch.cat([uncond_text_encoder_hidden_states, text_encoder_hidden_states])
+
+            text_mask = torch.cat([uncond_text_mask, text_mask])
+
+        return text_embeddings, text_encoder_hidden_states, text_mask
+
+    def enable_sequential_cpu_offload(self, gpu_id=0):
+        r"""
+        Offloads all models to CPU using accelerate, significantly reducing memory usage. When called, the pipeline's
+        models have their state dicts saved to CPU and then are moved to a `torch.device('meta') and loaded to GPU only
+        when their specific submodule has its `forward` method called.
+        """
+        if is_accelerate_available():
+            from accelerate import cpu_offload
+        else:
+            raise ImportError("Please install accelerate via `pip install accelerate`")
+
+        device = torch.device(f"cuda:{gpu_id}")
+
+        # TODO: self.prior.post_process_latents is not covered by the offload hooks, so it fails if added to the list
+        models = [
+            self.decoder,
+            self.text_proj,
+            self.text_encoder,
+            self.super_res_first,
+            self.super_res_last,
+        ]
+        for cpu_offloaded_model in models:
+            if cpu_offloaded_model is not None:
+                cpu_offload(cpu_offloaded_model, device)
+
+    @property
+    def _execution_device(self):
+        r"""
+        Returns the device on which the pipeline's models will be executed. After calling
+        `pipeline.enable_sequential_cpu_offload()` the execution device can only be inferred from Accelerate's module
+        hooks.
+        """
+        if self.device != torch.device("meta") or not hasattr(self.decoder, "_hf_hook"):
+            return self.device
+        for module in self.decoder.modules():
+            if (
+                hasattr(module, "_hf_hook")
+                and hasattr(module._hf_hook, "execution_device")
+                and module._hf_hook.execution_device is not None
+            ):
+                return torch.device(module._hf_hook.execution_device)
+        return self.device
+
+    @torch.no_grad()
+    def __call__(
+        self,
+        prompt: Optional[Union[str, List[str]]] = None,
+        num_images_per_prompt: int = 1,
+        prior_num_inference_steps: int = 25,
+        decoder_num_inference_steps: int = 25,
+        super_res_num_inference_steps: int = 7,
+        generator: Optional[torch.Generator] = None,
+        prior_latents: Optional[torch.FloatTensor] = None,
+        decoder_latents: Optional[torch.FloatTensor] = None,
+        super_res_latents: Optional[torch.FloatTensor] = None,
+        text_model_output: Optional[Union[CLIPTextModelOutput, Tuple]] = None,
+        text_attention_mask: Optional[torch.Tensor] = None,
+        prior_guidance_scale: float = 4.0,
+        decoder_guidance_scale: float = 8.0,
+        output_type: Optional[str] = "pil",
+        return_dict: bool = True,
+    ):
+        """
+        Function invoked when calling the pipeline for generation.
+        Args:
+            prompt (`str` or `List[str]`):
+                The prompt or prompts to guide the image generation. This can only be left undefined if
+                `text_model_output` and `text_attention_mask` is passed.
+            num_images_per_prompt (`int`, *optional*, defaults to 1):
+                The number of images to generate per prompt.
+            prior_num_inference_steps (`int`, *optional*, defaults to 25):
+                The number of denoising steps for the prior. More denoising steps usually lead to a higher quality
+                image at the expense of slower inference.
+            decoder_num_inference_steps (`int`, *optional*, defaults to 25):
+                The number of denoising steps for the decoder. More denoising steps usually lead to a higher quality
+                image at the expense of slower inference.
+            super_res_num_inference_steps (`int`, *optional*, defaults to 7):
+                The number of denoising steps for super resolution. More denoising steps usually lead to a higher
+                quality image at the expense of slower inference.
+            generator (`torch.Generator`, *optional*):
+                One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
+                to make generation deterministic.
+            prior_latents (`torch.FloatTensor` of shape (batch size, embeddings dimension), *optional*):
+                Pre-generated noisy latents to be used as inputs for the prior.
+            decoder_latents (`torch.FloatTensor` of shape (batch size, channels, height, width), *optional*):
+                Pre-generated noisy latents to be used as inputs for the decoder.
+            super_res_latents (`torch.FloatTensor` of shape (batch size, channels, super res height, super res width), *optional*):
+                Pre-generated noisy latents to be used as inputs for the decoder.
+            prior_guidance_scale (`float`, *optional*, defaults to 4.0):
+                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
+                `guidance_scale` is defined as `w` of equation 2. of [Imagen
+                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
+                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
+                usually at the expense of lower image quality.
+            decoder_guidance_scale (`float`, *optional*, defaults to 4.0):
+                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
+                `guidance_scale` is defined as `w` of equation 2. of [Imagen
+                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
+                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
+                usually at the expense of lower image quality.
+            text_model_output (`CLIPTextModelOutput`, *optional*):
+                Pre-defined CLIPTextModel outputs that can be derived from the text encoder. Pre-defined text outputs
+                can be passed for tasks like text embedding interpolations. Make sure to also pass
+                `text_attention_mask` in this case. `prompt` can the be left to `None`.
+            text_attention_mask (`torch.Tensor`, *optional*):
+                Pre-defined CLIP text attention mask that can be derived from the tokenizer. Pre-defined text attention
+                masks are necessary when passing `text_model_output`.
+            output_type (`str`, *optional*, defaults to `"pil"`):
+                The output format of the generated image. Choose between
+                [PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
+            return_dict (`bool`, *optional*, defaults to `True`):
+                Whether or not to return a [`~pipelines.ImagePipelineOutput`] instead of a plain tuple.
+        """
+        if prompt is not None:
+            if isinstance(prompt, str):
+                batch_size = 1
+            elif isinstance(prompt, list):
+                batch_size = len(prompt)
+            else:
+                raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")
+        else:
+            batch_size = text_model_output[0].shape[0]
+
+        device = self._execution_device
+
+        batch_size = batch_size * num_images_per_prompt
+
+        do_classifier_free_guidance = prior_guidance_scale > 1.0 or decoder_guidance_scale > 1.0
+
+        text_embeddings, text_encoder_hidden_states, text_mask = self._encode_prompt(
+            prompt, device, num_images_per_prompt, do_classifier_free_guidance, text_model_output, text_attention_mask
+        )
+
+        # prior
+
+        self.prior_scheduler.set_timesteps(prior_num_inference_steps, device=device)
+        prior_timesteps_tensor = self.prior_scheduler.timesteps
+
+        embedding_dim = self.prior.config.embedding_dim
+
+        prior_latents = self.prepare_latents(
+            (batch_size, embedding_dim),
+            text_embeddings.dtype,
+            device,
+            generator,
+            prior_latents,
+            self.prior_scheduler,
+        )
+
+        for i, t in enumerate(self.progress_bar(prior_timesteps_tensor)):
+            # expand the latents if we are doing classifier free guidance
+            latent_model_input = torch.cat([prior_latents] * 2) if do_classifier_free_guidance else prior_latents
+
+            predicted_image_embedding = self.prior(
+                latent_model_input,
+                timestep=t,
+                proj_embedding=text_embeddings,
+                encoder_hidden_states=text_encoder_hidden_states,
+                attention_mask=text_mask,
+            ).predicted_image_embedding
+
+            if do_classifier_free_guidance:
+                predicted_image_embedding_uncond, predicted_image_embedding_text = predicted_image_embedding.chunk(2)
+                predicted_image_embedding = predicted_image_embedding_uncond + prior_guidance_scale * (
+                    predicted_image_embedding_text - predicted_image_embedding_uncond
+                )
+
+            if i + 1 == prior_timesteps_tensor.shape[0]:
+                prev_timestep = None
+            else:
+                prev_timestep = prior_timesteps_tensor[i + 1]
+
+            prior_latents = self.prior_scheduler.step(
+                predicted_image_embedding,
+                timestep=t,
+                sample=prior_latents,
+                generator=generator,
+                prev_timestep=prev_timestep,
+            ).prev_sample
+
+        prior_latents = self.prior.post_process_latents(prior_latents)
+
+        image_embeddings = prior_latents
+
+        # done prior
+
+        # decoder
+
+        text_encoder_hidden_states, additive_clip_time_embeddings = self.text_proj(
+            image_embeddings=image_embeddings,
+            text_embeddings=text_embeddings,
+            text_encoder_hidden_states=text_encoder_hidden_states,
+            do_classifier_free_guidance=do_classifier_free_guidance,
+        )
+
+        decoder_text_mask = F.pad(text_mask, (self.text_proj.clip_extra_context_tokens, 0), value=1)
+
+        self.decoder_scheduler.set_timesteps(decoder_num_inference_steps, device=device)
+        decoder_timesteps_tensor = self.decoder_scheduler.timesteps
+
+        num_channels_latents = self.decoder.in_channels
+        height = self.decoder.sample_size
+        width = self.decoder.sample_size
+
+        decoder_latents = self.prepare_latents(
+            (batch_size, num_channels_latents, height, width),
+            text_encoder_hidden_states.dtype,
+            device,
+            generator,
+            decoder_latents,
+            self.decoder_scheduler,
+        )
+
+        for i, t in enumerate(self.progress_bar(decoder_timesteps_tensor)):
+            # expand the latents if we are doing classifier free guidance
+            latent_model_input = torch.cat([decoder_latents] * 2) if do_classifier_free_guidance else decoder_latents
+
+            noise_pred = self.decoder(
+                sample=latent_model_input,
+                timestep=t,
+                encoder_hidden_states=text_encoder_hidden_states,
+                class_labels=additive_clip_time_embeddings,
+                attention_mask=decoder_text_mask,
+            ).sample
+
+            if do_classifier_free_guidance:
+                noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
+                noise_pred_uncond, _ = noise_pred_uncond.split(latent_model_input.shape[1], dim=1)
+                noise_pred_text, predicted_variance = noise_pred_text.split(latent_model_input.shape[1], dim=1)
+                noise_pred = noise_pred_uncond + decoder_guidance_scale * (noise_pred_text - noise_pred_uncond)
+                noise_pred = torch.cat([noise_pred, predicted_variance], dim=1)
+
+            if i + 1 == decoder_timesteps_tensor.shape[0]:
+                prev_timestep = None
+            else:
+                prev_timestep = decoder_timesteps_tensor[i + 1]
+
+            # compute the previous noisy sample x_t -> x_t-1
+            decoder_latents = self.decoder_scheduler.step(
+                noise_pred, t, decoder_latents, prev_timestep=prev_timestep, generator=generator
+            ).prev_sample
+
+        decoder_latents = decoder_latents.clamp(-1, 1)
+
+        image_small = decoder_latents
+
+        # done decoder
+
+        # super res
+
+        self.super_res_scheduler.set_timesteps(super_res_num_inference_steps, device=device)
+        super_res_timesteps_tensor = self.super_res_scheduler.timesteps
+
+        channels = self.super_res_first.in_channels // 2
+        height = self.super_res_first.sample_size
+        width = self.super_res_first.sample_size
+
+        super_res_latents = self.prepare_latents(
+            (batch_size, channels, height, width),
+            image_small.dtype,
+            device,
+            generator,
+            super_res_latents,
+            self.super_res_scheduler,
+        )
+
+        interpolate_antialias = {}
+        if "antialias" in inspect.signature(F.interpolate).parameters:
+            interpolate_antialias["antialias"] = True
+
+        image_upscaled = F.interpolate(
+            image_small, size=[height, width], mode="bicubic", align_corners=False, **interpolate_antialias
+        )
+
+        for i, t in enumerate(self.progress_bar(super_res_timesteps_tensor)):
+            # no classifier free guidance
+
+            if i == super_res_timesteps_tensor.shape[0] - 1:
+                unet = self.super_res_last
+            else:
+                unet = self.super_res_first
+
+            latent_model_input = torch.cat([super_res_latents, image_upscaled], dim=1)
+
+            noise_pred = unet(
+                sample=latent_model_input,
+                timestep=t,
+            ).sample
+
+            if i + 1 == super_res_timesteps_tensor.shape[0]:
+                prev_timestep = None
+            else:
+                prev_timestep = super_res_timesteps_tensor[i + 1]
+
+            # compute the previous noisy sample x_t -> x_t-1
+            super_res_latents = self.super_res_scheduler.step(
+                noise_pred, t, super_res_latents, prev_timestep=prev_timestep, generator=generator
+            ).prev_sample
+
+        image = super_res_latents
+        # done super res
+
+        # post processing
+
+        image = image * 0.5 + 0.5
+        image = image.clamp(0, 1)
+        image = image.cpu().permute(0, 2, 3, 1).float().numpy()
+
+        if output_type == "pil":
+            image = self.numpy_to_pil(image)
+
+        if not return_dict:
+            return (image,)
+
+        return ImagePipelineOutput(images=image)
\ No newline at end of file
diff --git a/ldm/modules/karlo/kakao/__init__.py b/ldm/modules/karlo/kakao/__init__.py
new file mode 100644
index 0000000..e69de29
diff --git a/ldm/modules/karlo/kakao/models/__init__.py b/ldm/modules/karlo/kakao/models/__init__.py
new file mode 100644
index 0000000..e69de29
diff --git a/ldm/modules/karlo/kakao/models/clip.py b/ldm/modules/karlo/kakao/models/clip.py
new file mode 100644
index 0000000..961d815
--- /dev/null
+++ b/ldm/modules/karlo/kakao/models/clip.py
@@ -0,0 +1,182 @@
+# ------------------------------------------------------------------------------------
+# Karlo-v1.0.alpha
+# Copyright (c) 2022 KakaoBrain. All Rights Reserved.
+# ------------------------------------------------------------------------------------
+# ------------------------------------------------------------------------------------
+# Adapted from OpenAI's CLIP (https://github.com/openai/CLIP/)
+# ------------------------------------------------------------------------------------
+
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import clip
+
+from clip.model import CLIP, convert_weights
+from clip.simple_tokenizer import SimpleTokenizer, default_bpe
+
+
+"""===== Monkey-Patching original CLIP for JIT compile ====="""
+
+
+class LayerNorm(nn.LayerNorm):
+    """Subclass torch's LayerNorm to handle fp16."""
+
+    def forward(self, x: torch.Tensor):
+        orig_type = x.dtype
+        ret = F.layer_norm(
+            x.type(torch.float32),
+            self.normalized_shape,
+            self.weight,
+            self.bias,
+            self.eps,
+        )
+        return ret.type(orig_type)
+
+
+clip.model.LayerNorm = LayerNorm
+delattr(clip.model.CLIP, "forward")
+
+"""===== End of Monkey-Patching ====="""
+
+
+class CustomizedCLIP(CLIP):
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+
+    @torch.jit.export
+    def encode_image(self, image):
+        return self.visual(image)
+
+    @torch.jit.export
+    def encode_text(self, text):
+        # re-define this function to return unpooled text features
+
+        x = self.token_embedding(text).type(self.dtype)  # [batch_size, n_ctx, d_model]
+
+        x = x + self.positional_embedding.type(self.dtype)
+        x = x.permute(1, 0, 2)  # NLD -> LND
+        x = self.transformer(x)
+        x = x.permute(1, 0, 2)  # LND -> NLD
+        x = self.ln_final(x).type(self.dtype)
+
+        x_seq = x
+        # x.shape = [batch_size, n_ctx, transformer.width]
+        # take features from the eot embedding (eot_token is the highest number in each sequence)
+        x_out = x[torch.arange(x.shape[0]), text.argmax(dim=-1)] @ self.text_projection
+
+        return x_out, x_seq
+
+    @torch.jit.ignore
+    def forward(self, image, text):
+        super().forward(image, text)
+
+    @classmethod
+    def load_from_checkpoint(cls, ckpt_path: str):
+        state_dict = torch.load(ckpt_path, map_location="cpu").state_dict()
+
+        vit = "visual.proj" in state_dict
+        if vit:
+            vision_width = state_dict["visual.conv1.weight"].shape[0]
+            vision_layers = len(
+                [
+                    k
+                    for k in state_dict.keys()
+                    if k.startswith("visual.") and k.endswith(".attn.in_proj_weight")
+                ]
+            )
+            vision_patch_size = state_dict["visual.conv1.weight"].shape[-1]
+            grid_size = round(
+                (state_dict["visual.positional_embedding"].shape[0] - 1) ** 0.5
+            )
+            image_resolution = vision_patch_size * grid_size
+        else:
+            counts: list = [
+                len(
+                    set(
+                        k.split(".")[2]
+                        for k in state_dict
+                        if k.startswith(f"visual.layer{b}")
+                    )
+                )
+                for b in [1, 2, 3, 4]
+            ]
+            vision_layers = tuple(counts)
+            vision_width = state_dict["visual.layer1.0.conv1.weight"].shape[0]
+            output_width = round(
+                (state_dict["visual.attnpool.positional_embedding"].shape[0] - 1) ** 0.5
+            )
+            vision_patch_size = None
+            assert (
+                output_width**2 + 1
+                == state_dict["visual.attnpool.positional_embedding"].shape[0]
+            )
+            image_resolution = output_width * 32
+
+        embed_dim = state_dict["text_projection"].shape[1]
+        context_length = state_dict["positional_embedding"].shape[0]
+        vocab_size = state_dict["token_embedding.weight"].shape[0]
+        transformer_width = state_dict["ln_final.weight"].shape[0]
+        transformer_heads = transformer_width // 64
+        transformer_layers = len(
+            set(
+                k.split(".")[2]
+                for k in state_dict
+                if k.startswith("transformer.resblocks")
+            )
+        )
+
+        model = cls(
+            embed_dim,
+            image_resolution,
+            vision_layers,
+            vision_width,
+            vision_patch_size,
+            context_length,
+            vocab_size,
+            transformer_width,
+            transformer_heads,
+            transformer_layers,
+        )
+
+        for key in ["input_resolution", "context_length", "vocab_size"]:
+            if key in state_dict:
+                del state_dict[key]
+
+        convert_weights(model)
+        model.load_state_dict(state_dict)
+        model.eval()
+        model.float()
+        return model
+
+
+class CustomizedTokenizer(SimpleTokenizer):
+    def __init__(self):
+        super().__init__(bpe_path=default_bpe())
+
+        self.sot_token = self.encoder["<|startoftext|>"]
+        self.eot_token = self.encoder["<|endoftext|>"]
+
+    def padded_tokens_and_mask(self, texts, text_ctx):
+        assert isinstance(texts, list) and all(
+            isinstance(elem, str) for elem in texts
+        ), "texts should be a list of strings"
+
+        all_tokens = [
+            [self.sot_token] + self.encode(text) + [self.eot_token] for text in texts
+        ]
+
+        mask = [
+            [True] * min(text_ctx, len(tokens))
+            + [False] * max(text_ctx - len(tokens), 0)
+            for tokens in all_tokens
+        ]
+        mask = torch.tensor(mask, dtype=torch.bool)
+        result = torch.zeros(len(all_tokens), text_ctx, dtype=torch.int)
+        for i, tokens in enumerate(all_tokens):
+            if len(tokens) > text_ctx:
+                tokens = tokens[:text_ctx]
+                tokens[-1] = self.eot_token
+            result[i, : len(tokens)] = torch.tensor(tokens)
+
+        return result, mask
diff --git a/ldm/modules/karlo/kakao/models/decoder_model.py b/ldm/modules/karlo/kakao/models/decoder_model.py
new file mode 100644
index 0000000..84e96c9
--- /dev/null
+++ b/ldm/modules/karlo/kakao/models/decoder_model.py
@@ -0,0 +1,193 @@
+# ------------------------------------------------------------------------------------
+# Karlo-v1.0.alpha
+# Copyright (c) 2022 KakaoBrain. All Rights Reserved.
+# ------------------------------------------------------------------------------------
+
+import copy
+import torch
+
+from ldm.modules.karlo.kakao.modules import create_gaussian_diffusion
+from ldm.modules.karlo.kakao.modules.unet import PLMImUNet
+
+
+class Text2ImProgressiveModel(torch.nn.Module):
+    """
+    A decoder that generates 64x64px images based on the text prompt.
+
+    :param config: yaml config to define the decoder.
+    :param tokenizer: tokenizer used in clip.
+    """
+
+    def __init__(
+        self,
+        config,
+        tokenizer,
+    ):
+        super().__init__()
+
+        self._conf = config
+        self._model_conf = config.model.hparams
+        self._diffusion_kwargs = dict(
+            steps=config.diffusion.steps,
+            learn_sigma=config.diffusion.learn_sigma,
+            sigma_small=config.diffusion.sigma_small,
+            noise_schedule=config.diffusion.noise_schedule,
+            use_kl=config.diffusion.use_kl,
+            predict_xstart=config.diffusion.predict_xstart,
+            rescale_learned_sigmas=config.diffusion.rescale_learned_sigmas,
+            timestep_respacing=config.diffusion.timestep_respacing,
+        )
+        self._tokenizer = tokenizer
+
+        self.model = self.create_plm_dec_model()
+
+        cf_token, cf_mask = self.set_cf_text_tensor()
+        self.register_buffer("cf_token", cf_token, persistent=False)
+        self.register_buffer("cf_mask", cf_mask, persistent=False)
+
+    @classmethod
+    def load_from_checkpoint(cls, config, tokenizer, ckpt_path, strict: bool = True):
+        ckpt = torch.load(ckpt_path, map_location="cpu")["state_dict"]
+
+        model = cls(config, tokenizer)
+        model.load_state_dict(ckpt, strict=strict)
+        return model
+
+    def create_plm_dec_model(self):
+        image_size = self._model_conf.image_size
+        if self._model_conf.channel_mult == "":
+            if image_size == 256:
+                channel_mult = (1, 1, 2, 2, 4, 4)
+            elif image_size == 128:
+                channel_mult = (1, 1, 2, 3, 4)
+            elif image_size == 64:
+                channel_mult = (1, 2, 3, 4)
+            else:
+                raise ValueError(f"unsupported image size: {image_size}")
+        else:
+            channel_mult = tuple(
+                int(ch_mult) for ch_mult in self._model_conf.channel_mult.split(",")
+            )
+            assert 2 ** (len(channel_mult) + 2) == image_size
+
+        attention_ds = []
+        for res in self._model_conf.attention_resolutions.split(","):
+            attention_ds.append(image_size // int(res))
+
+        return PLMImUNet(
+            text_ctx=self._model_conf.text_ctx,
+            xf_width=self._model_conf.xf_width,
+            in_channels=3,
+            model_channels=self._model_conf.num_channels,
+            out_channels=6 if self._model_conf.learn_sigma else 3,
+            num_res_blocks=self._model_conf.num_res_blocks,
+            attention_resolutions=tuple(attention_ds),
+            dropout=self._model_conf.dropout,
+            channel_mult=channel_mult,
+            num_heads=self._model_conf.num_heads,
+            num_head_channels=self._model_conf.num_head_channels,
+            num_heads_upsample=self._model_conf.num_heads_upsample,
+            use_scale_shift_norm=self._model_conf.use_scale_shift_norm,
+            resblock_updown=self._model_conf.resblock_updown,
+            clip_dim=self._model_conf.clip_dim,
+            clip_emb_mult=self._model_conf.clip_emb_mult,
+            clip_emb_type=self._model_conf.clip_emb_type,
+            clip_emb_drop=self._model_conf.clip_emb_drop,
+        )
+
+    def set_cf_text_tensor(self):
+        return self._tokenizer.padded_tokens_and_mask([""], self.model.text_ctx)
+
+    def get_sample_fn(self, timestep_respacing):
+        use_ddim = timestep_respacing.startswith(("ddim", "fast"))
+
+        diffusion_kwargs = copy.deepcopy(self._diffusion_kwargs)
+        diffusion_kwargs.update(timestep_respacing=timestep_respacing)
+        diffusion = create_gaussian_diffusion(**diffusion_kwargs)
+        sample_fn = (
+            diffusion.ddim_sample_loop_progressive
+            if use_ddim
+            else diffusion.p_sample_loop_progressive
+        )
+
+        return sample_fn
+
+    def forward(
+        self,
+        txt_feat,
+        txt_feat_seq,
+        tok,
+        mask,
+        img_feat=None,
+        cf_guidance_scales=None,
+        timestep_respacing=None,
+    ):
+        # cfg should be enabled in inference
+        assert cf_guidance_scales is not None and all(cf_guidance_scales > 0.0)
+        assert img_feat is not None
+
+        bsz = txt_feat.shape[0]
+        img_sz = self._model_conf.image_size
+
+        def guided_model_fn(x_t, ts, **kwargs):
+            half = x_t[: len(x_t) // 2]
+            combined = torch.cat([half, half], dim=0)
+            model_out = self.model(combined, ts, **kwargs)
+            eps, rest = model_out[:, :3], model_out[:, 3:]
+            cond_eps, uncond_eps = torch.split(eps, len(eps) // 2, dim=0)
+            half_eps = uncond_eps + cf_guidance_scales.view(-1, 1, 1, 1) * (
+                cond_eps - uncond_eps
+            )
+            eps = torch.cat([half_eps, half_eps], dim=0)
+            return torch.cat([eps, rest], dim=1)
+
+        cf_feat = self.model.cf_param.unsqueeze(0)
+        cf_feat = cf_feat.expand(bsz // 2, -1)
+        feat = torch.cat([img_feat, cf_feat.to(txt_feat.device)], dim=0)
+
+        cond = {
+            "y": feat,
+            "txt_feat": txt_feat,
+            "txt_feat_seq": txt_feat_seq,
+            "mask": mask,
+        }
+        sample_fn = self.get_sample_fn(timestep_respacing)
+        sample_outputs = sample_fn(
+            guided_model_fn,
+            (bsz, 3, img_sz, img_sz),
+            noise=None,
+            device=txt_feat.device,
+            clip_denoised=True,
+            model_kwargs=cond,
+        )
+
+        for out in sample_outputs:
+            sample = out["sample"]
+            yield sample if cf_guidance_scales is None else sample[
+                : sample.shape[0] // 2
+            ]
+
+
+class Text2ImModel(Text2ImProgressiveModel):
+    def forward(
+        self,
+        txt_feat,
+        txt_feat_seq,
+        tok,
+        mask,
+        img_feat=None,
+        cf_guidance_scales=None,
+        timestep_respacing=None,
+    ):
+        last_out = None
+        for out in super().forward(
+            txt_feat,
+            txt_feat_seq,
+            tok,
+            mask,
+            img_feat,
+            cf_guidance_scales,
+            timestep_respacing,
+        ):
+            last_out = out
+        return last_out
diff --git a/ldm/modules/karlo/kakao/models/prior_model.py b/ldm/modules/karlo/kakao/models/prior_model.py
new file mode 100644
index 0000000..03ef230
--- /dev/null
+++ b/ldm/modules/karlo/kakao/models/prior_model.py
@@ -0,0 +1,138 @@
+# ------------------------------------------------------------------------------------
+# Karlo-v1.0.alpha
+# Copyright (c) 2022 KakaoBrain. All Rights Reserved.
+# ------------------------------------------------------------------------------------
+
+import copy
+import torch
+
+from ldm.modules.karlo.kakao.modules import create_gaussian_diffusion
+from ldm.modules.karlo.kakao.modules.xf import PriorTransformer
+
+
+class PriorDiffusionModel(torch.nn.Module):
+    """
+    A prior that generates clip image feature based on the text prompt.
+
+    :param config: yaml config to define the decoder.
+    :param tokenizer: tokenizer used in clip.
+    :param clip_mean: mean to normalize the clip image feature (zero-mean, unit variance).
+    :param clip_std: std to noramlize the clip image feature (zero-mean, unit variance).
+    """
+
+    def __init__(self, config, tokenizer, clip_mean, clip_std):
+        super().__init__()
+
+        self._conf = config
+        self._model_conf = config.model.hparams
+        self._diffusion_kwargs = dict(
+            steps=config.diffusion.steps,
+            learn_sigma=config.diffusion.learn_sigma,
+            sigma_small=config.diffusion.sigma_small,
+            noise_schedule=config.diffusion.noise_schedule,
+            use_kl=config.diffusion.use_kl,
+            predict_xstart=config.diffusion.predict_xstart,
+            rescale_learned_sigmas=config.diffusion.rescale_learned_sigmas,
+            timestep_respacing=config.diffusion.timestep_respacing,
+        )
+        self._tokenizer = tokenizer
+
+        self.register_buffer("clip_mean", clip_mean[None, :], persistent=False)
+        self.register_buffer("clip_std", clip_std[None, :], persistent=False)
+
+        causal_mask = self.get_causal_mask()
+        self.register_buffer("causal_mask", causal_mask, persistent=False)
+
+        self.model = PriorTransformer(
+            text_ctx=self._model_conf.text_ctx,
+            xf_width=self._model_conf.xf_width,
+            xf_layers=self._model_conf.xf_layers,
+            xf_heads=self._model_conf.xf_heads,
+            xf_final_ln=self._model_conf.xf_final_ln,
+            clip_dim=self._model_conf.clip_dim,
+        )
+
+        cf_token, cf_mask = self.set_cf_text_tensor()
+        self.register_buffer("cf_token", cf_token, persistent=False)
+        self.register_buffer("cf_mask", cf_mask, persistent=False)
+
+    @classmethod
+    def load_from_checkpoint(
+        cls, config, tokenizer, clip_mean, clip_std, ckpt_path, strict: bool = True
+    ):
+        ckpt = torch.load(ckpt_path, map_location="cpu")["state_dict"]
+
+        model = cls(config, tokenizer, clip_mean, clip_std)
+        model.load_state_dict(ckpt, strict=strict)
+        return model
+
+    def set_cf_text_tensor(self):
+        return self._tokenizer.padded_tokens_and_mask([""], self.model.text_ctx)
+
+    def get_sample_fn(self, timestep_respacing):
+        use_ddim = timestep_respacing.startswith(("ddim", "fast"))
+
+        diffusion_kwargs = copy.deepcopy(self._diffusion_kwargs)
+        diffusion_kwargs.update(timestep_respacing=timestep_respacing)
+        diffusion = create_gaussian_diffusion(**diffusion_kwargs)
+        sample_fn = diffusion.ddim_sample_loop if use_ddim else diffusion.p_sample_loop
+
+        return sample_fn
+
+    def get_causal_mask(self):
+        seq_len = self._model_conf.text_ctx + 4
+        mask = torch.empty(seq_len, seq_len)
+        mask.fill_(float("-inf"))
+        mask.triu_(1)
+        mask = mask[None, ...]
+        return mask
+
+    def forward(
+        self,
+        txt_feat,
+        txt_feat_seq,
+        mask,
+        cf_guidance_scales=None,
+        timestep_respacing=None,
+        denoised_fn=True,
+    ):
+        # cfg should be enabled in inference
+        assert cf_guidance_scales is not None and all(cf_guidance_scales > 0.0)
+
+        bsz_ = txt_feat.shape[0]
+        bsz = bsz_ // 2
+
+        def guided_model_fn(x_t, ts, **kwargs):
+            half = x_t[: len(x_t) // 2]
+            combined = torch.cat([half, half], dim=0)
+            model_out = self.model(combined, ts, **kwargs)
+            eps, rest = (
+                model_out[:, : int(x_t.shape[1])],
+                model_out[:, int(x_t.shape[1]) :],
+            )
+            cond_eps, uncond_eps = torch.split(eps, len(eps) // 2, dim=0)
+            half_eps = uncond_eps + cf_guidance_scales.view(-1, 1) * (
+                cond_eps - uncond_eps
+            )
+            eps = torch.cat([half_eps, half_eps], dim=0)
+            return torch.cat([eps, rest], dim=1)
+
+        cond = {
+            "text_emb": txt_feat,
+            "text_enc": txt_feat_seq,
+            "mask": mask,
+            "causal_mask": self.causal_mask,
+        }
+        sample_fn = self.get_sample_fn(timestep_respacing)
+        sample = sample_fn(
+            guided_model_fn,
+            (bsz_, self.model.clip_dim),
+            noise=None,
+            device=txt_feat.device,
+            clip_denoised=False,
+            denoised_fn=lambda x: torch.clamp(x, -10, 10),
+            model_kwargs=cond,
+        )
+        sample = (sample * self.clip_std) + self.clip_mean
+
+        return sample[:bsz]
diff --git a/ldm/modules/karlo/kakao/models/sr_256_1k.py b/ldm/modules/karlo/kakao/models/sr_256_1k.py
new file mode 100644
index 0000000..1e874f6
--- /dev/null
+++ b/ldm/modules/karlo/kakao/models/sr_256_1k.py
@@ -0,0 +1,10 @@
+# ------------------------------------------------------------------------------------
+# Karlo-v1.0.alpha
+# Copyright (c) 2022 KakaoBrain. All Rights Reserved.
+# ------------------------------------------------------------------------------------
+
+from ldm.modules.karlo.kakao.models.sr_64_256 import SupRes64to256Progressive
+
+
+class SupRes256to1kProgressive(SupRes64to256Progressive):
+    pass  # no difference currently
diff --git a/ldm/modules/karlo/kakao/models/sr_64_256.py b/ldm/modules/karlo/kakao/models/sr_64_256.py
new file mode 100644
index 0000000..32687af
--- /dev/null
+++ b/ldm/modules/karlo/kakao/models/sr_64_256.py
@@ -0,0 +1,88 @@
+# ------------------------------------------------------------------------------------
+# Karlo-v1.0.alpha
+# Copyright (c) 2022 KakaoBrain. All Rights Reserved.
+# ------------------------------------------------------------------------------------
+
+import copy
+import torch
+
+from ldm.modules.karlo.kakao.modules.unet import SuperResUNetModel
+from ldm.modules.karlo.kakao.modules import create_gaussian_diffusion
+
+
+class ImprovedSupRes64to256ProgressiveModel(torch.nn.Module):
+    """
+    ImprovedSR model fine-tunes the pretrained DDPM-based SR model by using adversarial and perceptual losses.
+    In specific, the low-resolution sample is iteratively recovered by 6 steps with the frozen pretrained SR model.
+    In the following additional one step, a seperate fine-tuned model recovers high-frequency details.
+    This approach greatly improves the fidelity of images of 256x256px, even with small number of reverse steps.
+    """
+
+    def __init__(self, config):
+        super().__init__()
+
+        self._config = config
+        self._diffusion_kwargs = dict(
+            steps=config.diffusion.steps,
+            learn_sigma=config.diffusion.learn_sigma,
+            sigma_small=config.diffusion.sigma_small,
+            noise_schedule=config.diffusion.noise_schedule,
+            use_kl=config.diffusion.use_kl,
+            predict_xstart=config.diffusion.predict_xstart,
+            rescale_learned_sigmas=config.diffusion.rescale_learned_sigmas,
+        )
+
+        self.model_first_steps = SuperResUNetModel(
+            in_channels=3,  # auto-changed to 6 inside the model
+            model_channels=config.model.hparams.channels,
+            out_channels=3,
+            num_res_blocks=config.model.hparams.depth,
+            attention_resolutions=(),  # no attention
+            dropout=config.model.hparams.dropout,
+            channel_mult=config.model.hparams.channels_multiple,
+            resblock_updown=True,
+            use_middle_attention=False,
+        )
+        self.model_last_step = SuperResUNetModel(
+            in_channels=3,  # auto-changed to 6 inside the model
+            model_channels=config.model.hparams.channels,
+            out_channels=3,
+            num_res_blocks=config.model.hparams.depth,
+            attention_resolutions=(),  # no attention
+            dropout=config.model.hparams.dropout,
+            channel_mult=config.model.hparams.channels_multiple,
+            resblock_updown=True,
+            use_middle_attention=False,
+        )
+
+    @classmethod
+    def load_from_checkpoint(cls, config, ckpt_path, strict: bool = True):
+        ckpt = torch.load(ckpt_path, map_location="cpu")["state_dict"]
+
+        model = cls(config)
+        model.load_state_dict(ckpt, strict=strict)
+        return model
+
+    def get_sample_fn(self, timestep_respacing):
+        diffusion_kwargs = copy.deepcopy(self._diffusion_kwargs)
+        diffusion_kwargs.update(timestep_respacing=timestep_respacing)
+        diffusion = create_gaussian_diffusion(**diffusion_kwargs)
+        return diffusion.p_sample_loop_progressive_for_improved_sr
+
+    def forward(self, low_res, timestep_respacing="7", **kwargs):
+        assert (
+            timestep_respacing == "7"
+        ), "different respacing method may work, but no guaranteed"
+
+        sample_fn = self.get_sample_fn(timestep_respacing)
+        sample_outputs = sample_fn(
+            self.model_first_steps,
+            self.model_last_step,
+            shape=low_res.shape,
+            clip_denoised=True,
+            model_kwargs=dict(low_res=low_res),
+            **kwargs,
+        )
+        for x in sample_outputs:
+            sample = x["sample"]
+            yield sample
diff --git a/ldm/modules/karlo/kakao/modules/__init__.py b/ldm/modules/karlo/kakao/modules/__init__.py
new file mode 100644
index 0000000..11d4358
--- /dev/null
+++ b/ldm/modules/karlo/kakao/modules/__init__.py
@@ -0,0 +1,49 @@
+# ------------------------------------------------------------------------------------
+# Adapted from Guided-Diffusion repo (https://github.com/openai/guided-diffusion)
+# ------------------------------------------------------------------------------------
+
+
+from .diffusion import gaussian_diffusion as gd
+from .diffusion.respace import (
+    SpacedDiffusion,
+    space_timesteps,
+)
+
+
+def create_gaussian_diffusion(
+    steps,
+    learn_sigma,
+    sigma_small,
+    noise_schedule,
+    use_kl,
+    predict_xstart,
+    rescale_learned_sigmas,
+    timestep_respacing,
+):
+    betas = gd.get_named_beta_schedule(noise_schedule, steps)
+    if use_kl:
+        loss_type = gd.LossType.RESCALED_KL
+    elif rescale_learned_sigmas:
+        loss_type = gd.LossType.RESCALED_MSE
+    else:
+        loss_type = gd.LossType.MSE
+    if not timestep_respacing:
+        timestep_respacing = [steps]
+
+    return SpacedDiffusion(
+        use_timesteps=space_timesteps(steps, timestep_respacing),
+        betas=betas,
+        model_mean_type=(
+            gd.ModelMeanType.EPSILON if not predict_xstart else gd.ModelMeanType.START_X
+        ),
+        model_var_type=(
+            (
+                gd.ModelVarType.FIXED_LARGE
+                if not sigma_small
+                else gd.ModelVarType.FIXED_SMALL
+            )
+            if not learn_sigma
+            else gd.ModelVarType.LEARNED_RANGE
+        ),
+        loss_type=loss_type,
+    )
diff --git a/ldm/modules/karlo/kakao/modules/diffusion/gaussian_diffusion.py b/ldm/modules/karlo/kakao/modules/diffusion/gaussian_diffusion.py
new file mode 100644
index 0000000..6a111aa
--- /dev/null
+++ b/ldm/modules/karlo/kakao/modules/diffusion/gaussian_diffusion.py
@@ -0,0 +1,828 @@
+# ------------------------------------------------------------------------------------
+# Adapted from Guided-Diffusion repo (https://github.com/openai/guided-diffusion)
+# ------------------------------------------------------------------------------------
+
+import enum
+import math
+
+import numpy as np
+import torch as th
+
+
+def _warmup_beta(beta_start, beta_end, num_diffusion_timesteps, warmup_frac):
+    betas = beta_end * np.ones(num_diffusion_timesteps, dtype=np.float64)
+    warmup_time = int(num_diffusion_timesteps * warmup_frac)
+    betas[:warmup_time] = np.linspace(
+        beta_start, beta_end, warmup_time, dtype=np.float64
+    )
+    return betas
+
+
+def get_beta_schedule(beta_schedule, *, beta_start, beta_end, num_diffusion_timesteps):
+    """
+    This is the deprecated API for creating beta schedules.
+    See get_named_beta_schedule() for the new library of schedules.
+    """
+    if beta_schedule == "quad":
+        betas = (
+            np.linspace(
+                beta_start**0.5,
+                beta_end**0.5,
+                num_diffusion_timesteps,
+                dtype=np.float64,
+            )
+            ** 2
+        )
+    elif beta_schedule == "linear":
+        betas = np.linspace(
+            beta_start, beta_end, num_diffusion_timesteps, dtype=np.float64
+        )
+    elif beta_schedule == "warmup10":
+        betas = _warmup_beta(beta_start, beta_end, num_diffusion_timesteps, 0.1)
+    elif beta_schedule == "warmup50":
+        betas = _warmup_beta(beta_start, beta_end, num_diffusion_timesteps, 0.5)
+    elif beta_schedule == "const":
+        betas = beta_end * np.ones(num_diffusion_timesteps, dtype=np.float64)
+    elif beta_schedule == "jsd":  # 1/T, 1/(T-1), 1/(T-2), ..., 1
+        betas = 1.0 / np.linspace(
+            num_diffusion_timesteps, 1, num_diffusion_timesteps, dtype=np.float64
+        )
+    else:
+        raise NotImplementedError(beta_schedule)
+    assert betas.shape == (num_diffusion_timesteps,)
+    return betas
+
+
+def get_named_beta_schedule(schedule_name, num_diffusion_timesteps):
+    """
+    Get a pre-defined beta schedule for the given name.
+    The beta schedule library consists of beta schedules which remain similar
+    in the limit of num_diffusion_timesteps.
+    Beta schedules may be added, but should not be removed or changed once
+    they are committed to maintain backwards compatibility.
+    """
+    if schedule_name == "linear":
+        # Linear schedule from Ho et al, extended to work for any number of
+        # diffusion steps.
+        scale = 1000 / num_diffusion_timesteps
+        return get_beta_schedule(
+            "linear",
+            beta_start=scale * 0.0001,
+            beta_end=scale * 0.02,
+            num_diffusion_timesteps=num_diffusion_timesteps,
+        )
+    elif schedule_name == "squaredcos_cap_v2":
+        return betas_for_alpha_bar(
+            num_diffusion_timesteps,
+            lambda t: math.cos((t + 0.008) / 1.008 * math.pi / 2) ** 2,
+        )
+    else:
+        raise NotImplementedError(f"unknown beta schedule: {schedule_name}")
+
+
+def betas_for_alpha_bar(num_diffusion_timesteps, alpha_bar, max_beta=0.999):
+    """
+    Create a beta schedule that discretizes the given alpha_t_bar function,
+    which defines the cumulative product of (1-beta) over time from t = [0,1].
+    :param num_diffusion_timesteps: the number of betas to produce.
+    :param alpha_bar: a lambda that takes an argument t from 0 to 1 and
+                      produces the cumulative product of (1-beta) up to that
+                      part of the diffusion process.
+    :param max_beta: the maximum beta to use; use values lower than 1 to
+                     prevent singularities.
+    """
+    betas = []
+    for i in range(num_diffusion_timesteps):
+        t1 = i / num_diffusion_timesteps
+        t2 = (i + 1) / num_diffusion_timesteps
+        betas.append(min(1 - alpha_bar(t2) / alpha_bar(t1), max_beta))
+    return np.array(betas)
+
+
+class ModelMeanType(enum.Enum):
+    """
+    Which type of output the model predicts.
+    """
+
+    PREVIOUS_X = enum.auto()  # the model predicts x_{t-1}
+    START_X = enum.auto()  # the model predicts x_0
+    EPSILON = enum.auto()  # the model predicts epsilon
+
+
+class ModelVarType(enum.Enum):
+    """
+    What is used as the model's output variance.
+    The LEARNED_RANGE option has been added to allow the model to predict
+    values between FIXED_SMALL and FIXED_LARGE, making its job easier.
+    """
+
+    LEARNED = enum.auto()
+    FIXED_SMALL = enum.auto()
+    FIXED_LARGE = enum.auto()
+    LEARNED_RANGE = enum.auto()
+
+
+class LossType(enum.Enum):
+    MSE = enum.auto()  # use raw MSE loss (and KL when learning variances)
+    RESCALED_MSE = (
+        enum.auto()
+    )  # use raw MSE loss (with RESCALED_KL when learning variances)
+    KL = enum.auto()  # use the variational lower-bound
+    RESCALED_KL = enum.auto()  # like KL, but rescale to estimate the full VLB
+
+    def is_vb(self):
+        return self == LossType.KL or self == LossType.RESCALED_KL
+
+
+class GaussianDiffusion(th.nn.Module):
+    """
+    Utilities for training and sampling diffusion models.
+    Original ported from this codebase:
+    https://github.com/hojonathanho/diffusion/blob/1e0dceb3b3495bbe19116a5e1b3596cd0706c543/diffusion_tf/diffusion_utils_2.py#L42
+    :param betas: a 1-D numpy array of betas for each diffusion timestep,
+                  starting at T and going to 1.
+    """
+
+    def __init__(
+        self,
+        *,
+        betas,
+        model_mean_type,
+        model_var_type,
+        loss_type,
+    ):
+        super(GaussianDiffusion, self).__init__()
+        self.model_mean_type = model_mean_type
+        self.model_var_type = model_var_type
+        self.loss_type = loss_type
+
+        # Use float64 for accuracy.
+        betas = np.array(betas, dtype=np.float64)
+        assert len(betas.shape) == 1, "betas must be 1-D"
+        assert (betas > 0).all() and (betas <= 1).all()
+
+        self.num_timesteps = int(betas.shape[0])
+
+        alphas = 1.0 - betas
+        alphas_cumprod = np.cumprod(alphas, axis=0)
+        alphas_cumprod_prev = np.append(1.0, alphas_cumprod[:-1])
+        alphas_cumprod_next = np.append(alphas_cumprod[1:], 0.0)
+        assert alphas_cumprod_prev.shape == (self.num_timesteps,)
+
+        # calculations for diffusion q(x_t | x_{t-1}) and others
+        sqrt_alphas_cumprod = np.sqrt(alphas_cumprod)
+        sqrt_one_minus_alphas_cumprod = np.sqrt(1.0 - alphas_cumprod)
+        log_one_minus_alphas_cumprod = np.log(1.0 - alphas_cumprod)
+        sqrt_recip_alphas_cumprod = np.sqrt(1.0 / alphas_cumprod)
+        sqrt_recipm1_alphas_cumprod = np.sqrt(1.0 / alphas_cumprod - 1)
+
+        # calculations for posterior q(x_{t-1} | x_t, x_0)
+        posterior_variance = (
+            betas * (1.0 - alphas_cumprod_prev) / (1.0 - alphas_cumprod)
+        )
+        # below: log calculation clipped because the posterior variance is 0 at the beginning of the diffusion chain
+        posterior_log_variance_clipped = np.log(
+            np.append(posterior_variance[1], posterior_variance[1:])
+        )
+        posterior_mean_coef1 = (
+            betas * np.sqrt(alphas_cumprod_prev) / (1.0 - alphas_cumprod)
+        )
+        posterior_mean_coef2 = (
+            (1.0 - alphas_cumprod_prev) * np.sqrt(alphas) / (1.0 - alphas_cumprod)
+        )
+
+        self.register_buffer("betas", th.from_numpy(betas), persistent=False)
+        self.register_buffer(
+            "alphas_cumprod", th.from_numpy(alphas_cumprod), persistent=False
+        )
+        self.register_buffer(
+            "alphas_cumprod_prev", th.from_numpy(alphas_cumprod_prev), persistent=False
+        )
+        self.register_buffer(
+            "alphas_cumprod_next", th.from_numpy(alphas_cumprod_next), persistent=False
+        )
+
+        self.register_buffer(
+            "sqrt_alphas_cumprod", th.from_numpy(sqrt_alphas_cumprod), persistent=False
+        )
+        self.register_buffer(
+            "sqrt_one_minus_alphas_cumprod",
+            th.from_numpy(sqrt_one_minus_alphas_cumprod),
+            persistent=False,
+        )
+        self.register_buffer(
+            "log_one_minus_alphas_cumprod",
+            th.from_numpy(log_one_minus_alphas_cumprod),
+            persistent=False,
+        )
+        self.register_buffer(
+            "sqrt_recip_alphas_cumprod",
+            th.from_numpy(sqrt_recip_alphas_cumprod),
+            persistent=False,
+        )
+        self.register_buffer(
+            "sqrt_recipm1_alphas_cumprod",
+            th.from_numpy(sqrt_recipm1_alphas_cumprod),
+            persistent=False,
+        )
+
+        self.register_buffer(
+            "posterior_variance", th.from_numpy(posterior_variance), persistent=False
+        )
+        self.register_buffer(
+            "posterior_log_variance_clipped",
+            th.from_numpy(posterior_log_variance_clipped),
+            persistent=False,
+        )
+        self.register_buffer(
+            "posterior_mean_coef1",
+            th.from_numpy(posterior_mean_coef1),
+            persistent=False,
+        )
+        self.register_buffer(
+            "posterior_mean_coef2",
+            th.from_numpy(posterior_mean_coef2),
+            persistent=False,
+        )
+
+    def q_mean_variance(self, x_start, t):
+        """
+        Get the distribution q(x_t | x_0).
+        :param x_start: the [N x C x ...] tensor of noiseless inputs.
+        :param t: the number of diffusion steps (minus 1). Here, 0 means one step.
+        :return: A tuple (mean, variance, log_variance), all of x_start's shape.
+        """
+        mean = (
+            _extract_into_tensor(self.sqrt_alphas_cumprod, t, x_start.shape) * x_start
+        )
+        variance = _extract_into_tensor(1.0 - self.alphas_cumprod, t, x_start.shape)
+        log_variance = _extract_into_tensor(
+            self.log_one_minus_alphas_cumprod, t, x_start.shape
+        )
+        return mean, variance, log_variance
+
+    def q_sample(self, x_start, t, noise=None):
+        """
+        Diffuse the data for a given number of diffusion steps.
+        In other words, sample from q(x_t | x_0).
+        :param x_start: the initial data batch.
+        :param t: the number of diffusion steps (minus 1). Here, 0 means one step.
+        :param noise: if specified, the split-out normal noise.
+        :return: A noisy version of x_start.
+        """
+        if noise is None:
+            noise = th.randn_like(x_start)
+        assert noise.shape == x_start.shape
+        return (
+            _extract_into_tensor(self.sqrt_alphas_cumprod, t, x_start.shape) * x_start
+            + _extract_into_tensor(self.sqrt_one_minus_alphas_cumprod, t, x_start.shape)
+            * noise
+        )
+
+    def q_posterior_mean_variance(self, x_start, x_t, t):
+        """
+        Compute the mean and variance of the diffusion posterior:
+            q(x_{t-1} | x_t, x_0)
+        """
+        assert x_start.shape == x_t.shape
+        posterior_mean = (
+            _extract_into_tensor(self.posterior_mean_coef1, t, x_t.shape) * x_start
+            + _extract_into_tensor(self.posterior_mean_coef2, t, x_t.shape) * x_t
+        )
+        posterior_variance = _extract_into_tensor(self.posterior_variance, t, x_t.shape)
+        posterior_log_variance_clipped = _extract_into_tensor(
+            self.posterior_log_variance_clipped, t, x_t.shape
+        )
+        assert (
+            posterior_mean.shape[0]
+            == posterior_variance.shape[0]
+            == posterior_log_variance_clipped.shape[0]
+            == x_start.shape[0]
+        )
+        return posterior_mean, posterior_variance, posterior_log_variance_clipped
+
+    def p_mean_variance(
+        self,
+        model,
+        x,
+        t,
+        clip_denoised=True,
+        denoised_fn=None,
+        model_kwargs=None,
+        **ignore_kwargs,
+    ):
+        """
+        Apply the model to get p(x_{t-1} | x_t), as well as a prediction of
+        the initial x, x_0.
+        :param model: the model, which takes a signal and a batch of timesteps
+                      as input.
+        :param x: the [N x C x ...] tensor at time t.
+        :param t: a 1-D Tensor of timesteps.
+        :param clip_denoised: if True, clip the denoised signal into [-1, 1].
+        :param denoised_fn: if not None, a function which applies to the
+            x_start prediction before it is used to sample. Applies before
+            clip_denoised.
+        :param model_kwargs: if not None, a dict of extra keyword arguments to
+            pass to the model. This can be used for conditioning.
+        :return: a dict with the following keys:
+                 - 'mean': the model mean output.
+                 - 'variance': the model variance output.
+                 - 'log_variance': the log of 'variance'.
+                 - 'pred_xstart': the prediction for x_0.
+        """
+        if model_kwargs is None:
+            model_kwargs = {}
+
+        B, C = x.shape[:2]
+        assert t.shape == (B,)
+        model_output = model(x, t, **model_kwargs)
+        if isinstance(model_output, tuple):
+            model_output, extra = model_output
+        else:
+            extra = None
+
+        if self.model_var_type in [ModelVarType.LEARNED, ModelVarType.LEARNED_RANGE]:
+            assert model_output.shape == (B, C * 2, *x.shape[2:])
+            model_output, model_var_values = th.split(model_output, C, dim=1)
+            if self.model_var_type == ModelVarType.LEARNED:
+                model_log_variance = model_var_values
+                model_variance = th.exp(model_log_variance)
+            else:
+                min_log = _extract_into_tensor(
+                    self.posterior_log_variance_clipped, t, x.shape
+                )
+                max_log = _extract_into_tensor(th.log(self.betas), t, x.shape)
+                # The model_var_values is [-1, 1] for [min_var, max_var].
+                frac = (model_var_values + 1) / 2
+                model_log_variance = frac * max_log + (1 - frac) * min_log
+                model_variance = th.exp(model_log_variance)
+        else:
+            model_variance, model_log_variance = {
+                # for fixedlarge, we set the initial (log-)variance like so
+                # to get a better decoder log likelihood.
+                ModelVarType.FIXED_LARGE: (
+                    th.cat([self.posterior_variance[1][None], self.betas[1:]]),
+                    th.log(th.cat([self.posterior_variance[1][None], self.betas[1:]])),
+                ),
+                ModelVarType.FIXED_SMALL: (
+                    self.posterior_variance,
+                    self.posterior_log_variance_clipped,
+                ),
+            }[self.model_var_type]
+            model_variance = _extract_into_tensor(model_variance, t, x.shape)
+            model_log_variance = _extract_into_tensor(model_log_variance, t, x.shape)
+
+        def process_xstart(x):
+            if denoised_fn is not None:
+                x = denoised_fn(x)
+            if clip_denoised:
+                return x.clamp(-1, 1)
+            return x
+
+        if self.model_mean_type == ModelMeanType.PREVIOUS_X:
+            pred_xstart = process_xstart(
+                self._predict_xstart_from_xprev(x_t=x, t=t, xprev=model_output)
+            )
+            model_mean = model_output
+        elif self.model_mean_type in [ModelMeanType.START_X, ModelMeanType.EPSILON]:
+            if self.model_mean_type == ModelMeanType.START_X:
+                pred_xstart = process_xstart(model_output)
+            else:
+                pred_xstart = process_xstart(
+                    self._predict_xstart_from_eps(x_t=x, t=t, eps=model_output)
+                )
+            model_mean, _, _ = self.q_posterior_mean_variance(
+                x_start=pred_xstart, x_t=x, t=t
+            )
+        else:
+            raise NotImplementedError(self.model_mean_type)
+
+        assert (
+            model_mean.shape == model_log_variance.shape == pred_xstart.shape == x.shape
+        )
+        return {
+            "mean": model_mean,
+            "variance": model_variance,
+            "log_variance": model_log_variance,
+            "pred_xstart": pred_xstart,
+        }
+
+    def _predict_xstart_from_eps(self, x_t, t, eps):
+        assert x_t.shape == eps.shape
+        return (
+            _extract_into_tensor(self.sqrt_recip_alphas_cumprod, t, x_t.shape) * x_t
+            - _extract_into_tensor(self.sqrt_recipm1_alphas_cumprod, t, x_t.shape) * eps
+        )
+
+    def _predict_eps_from_xstart(self, x_t, t, pred_xstart):
+        return (
+            _extract_into_tensor(self.sqrt_recip_alphas_cumprod, t, x_t.shape) * x_t
+            - pred_xstart
+        ) / _extract_into_tensor(self.sqrt_recipm1_alphas_cumprod, t, x_t.shape)
+
+    def condition_mean(self, cond_fn, p_mean_var, x, t, model_kwargs=None):
+        """
+        Compute the mean for the previous step, given a function cond_fn that
+        computes the gradient of a conditional log probability with respect to
+        x. In particular, cond_fn computes grad(log(p(y|x))), and we want to
+        condition on y.
+        This uses the conditioning strategy from Sohl-Dickstein et al. (2015).
+        """
+        gradient = cond_fn(x, t, **model_kwargs)
+        new_mean = (
+            p_mean_var["mean"].float() + p_mean_var["variance"] * gradient.float()
+        )
+        return new_mean
+
+    def condition_score(self, cond_fn, p_mean_var, x, t, model_kwargs=None):
+        """
+        Compute what the p_mean_variance output would have been, should the
+        model's score function be conditioned by cond_fn.
+        See condition_mean() for details on cond_fn.
+        Unlike condition_mean(), this instead uses the conditioning strategy
+        from Song et al (2020).
+        """
+        alpha_bar = _extract_into_tensor(self.alphas_cumprod, t, x.shape)
+
+        eps = self._predict_eps_from_xstart(x, t, p_mean_var["pred_xstart"])
+        eps = eps - (1 - alpha_bar).sqrt() * cond_fn(x, t, **model_kwargs)
+
+        out = p_mean_var.copy()
+        out["pred_xstart"] = self._predict_xstart_from_eps(x, t, eps)
+        out["mean"], _, _ = self.q_posterior_mean_variance(
+            x_start=out["pred_xstart"], x_t=x, t=t
+        )
+        return out
+
+    def p_sample(
+        self,
+        model,
+        x,
+        t,
+        clip_denoised=True,
+        denoised_fn=None,
+        cond_fn=None,
+        model_kwargs=None,
+    ):
+        """
+        Sample x_{t-1} from the model at the given timestep.
+        :param model: the model to sample from.
+        :param x: the current tensor at x_{t-1}.
+        :param t: the value of t, starting at 0 for the first diffusion step.
+        :param clip_denoised: if True, clip the x_start prediction to [-1, 1].
+        :param denoised_fn: if not None, a function which applies to the
+            x_start prediction before it is used to sample.
+        :param cond_fn: if not None, this is a gradient function that acts
+                        similarly to the model.
+        :param model_kwargs: if not None, a dict of extra keyword arguments to
+            pass to the model. This can be used for conditioning.
+        :return: a dict containing the following keys:
+                 - 'sample': a random sample from the model.
+                 - 'pred_xstart': a prediction of x_0.
+        """
+        out = self.p_mean_variance(
+            model,
+            x,
+            t,
+            clip_denoised=clip_denoised,
+            denoised_fn=denoised_fn,
+            model_kwargs=model_kwargs,
+        )
+        noise = th.randn_like(x)
+        nonzero_mask = (
+            (t != 0).float().view(-1, *([1] * (len(x.shape) - 1)))
+        )  # no noise when t == 0
+        if cond_fn is not None:
+            out["mean"] = self.condition_mean(
+                cond_fn, out, x, t, model_kwargs=model_kwargs
+            )
+        sample = out["mean"] + nonzero_mask * th.exp(0.5 * out["log_variance"]) * noise
+        return {"sample": sample, "pred_xstart": out["pred_xstart"]}
+
+    def p_sample_loop(
+        self,
+        model,
+        shape,
+        noise=None,
+        clip_denoised=True,
+        denoised_fn=None,
+        cond_fn=None,
+        model_kwargs=None,
+        device=None,
+        progress=False,
+    ):
+        """
+        Generate samples from the model.
+        :param model: the model module.
+        :param shape: the shape of the samples, (N, C, H, W).
+        :param noise: if specified, the noise from the encoder to sample.
+                      Should be of the same shape as `shape`.
+        :param clip_denoised: if True, clip x_start predictions to [-1, 1].
+        :param denoised_fn: if not None, a function which applies to the
+            x_start prediction before it is used to sample.
+        :param cond_fn: if not None, this is a gradient function that acts
+                        similarly to the model.
+        :param model_kwargs: if not None, a dict of extra keyword arguments to
+            pass to the model. This can be used for conditioning.
+        :param device: if specified, the device to create the samples on.
+                       If not specified, use a model parameter's device.
+        :param progress: if True, show a tqdm progress bar.
+        :return: a non-differentiable batch of samples.
+        """
+        final = None
+        for sample in self.p_sample_loop_progressive(
+            model,
+            shape,
+            noise=noise,
+            clip_denoised=clip_denoised,
+            denoised_fn=denoised_fn,
+            cond_fn=cond_fn,
+            model_kwargs=model_kwargs,
+            device=device,
+            progress=progress,
+        ):
+            final = sample
+        return final["sample"]
+
+    def p_sample_loop_progressive(
+        self,
+        model,
+        shape,
+        noise=None,
+        clip_denoised=True,
+        denoised_fn=None,
+        cond_fn=None,
+        model_kwargs=None,
+        device=None,
+        progress=False,
+    ):
+        """
+        Generate samples from the model and yield intermediate samples from
+        each timestep of diffusion.
+        Arguments are the same as p_sample_loop().
+        Returns a generator over dicts, where each dict is the return value of
+        p_sample().
+        """
+        if device is None:
+            device = next(model.parameters()).device
+        assert isinstance(shape, (tuple, list))
+        if noise is not None:
+            img = noise
+        else:
+            img = th.randn(*shape, device=device)
+        indices = list(range(self.num_timesteps))[::-1]
+
+        if progress:
+            # Lazy import so that we don't depend on tqdm.
+            from tqdm.auto import tqdm
+
+            indices = tqdm(indices)
+
+        for idx, i in enumerate(indices):
+            t = th.tensor([i] * shape[0], device=device)
+            with th.no_grad():
+                out = self.p_sample(
+                    model,
+                    img,
+                    t,
+                    clip_denoised=clip_denoised,
+                    denoised_fn=denoised_fn,
+                    cond_fn=cond_fn,
+                    model_kwargs=model_kwargs,
+                )
+                yield out
+                img = out["sample"]
+
+    def p_sample_loop_progressive_for_improved_sr(
+        self,
+        model,
+        model_aux,
+        shape,
+        noise=None,
+        clip_denoised=True,
+        denoised_fn=None,
+        cond_fn=None,
+        model_kwargs=None,
+        device=None,
+        progress=False,
+    ):
+        """
+        Modified version of p_sample_loop_progressive for sampling from the improved sr model
+        """
+
+        if device is None:
+            device = next(model.parameters()).device
+        assert isinstance(shape, (tuple, list))
+        if noise is not None:
+            img = noise
+        else:
+            img = th.randn(*shape, device=device)
+        indices = list(range(self.num_timesteps))[::-1]
+
+        if progress:
+            # Lazy import so that we don't depend on tqdm.
+            from tqdm.auto import tqdm
+
+            indices = tqdm(indices)
+
+        for idx, i in enumerate(indices):
+            t = th.tensor([i] * shape[0], device=device)
+            with th.no_grad():
+                out = self.p_sample(
+                    model_aux if len(indices) - 1 == idx else model,
+                    img,
+                    t,
+                    clip_denoised=clip_denoised,
+                    denoised_fn=denoised_fn,
+                    cond_fn=cond_fn,
+                    model_kwargs=model_kwargs,
+                )
+                yield out
+                img = out["sample"]
+
+    def ddim_sample(
+        self,
+        model,
+        x,
+        t,
+        clip_denoised=True,
+        denoised_fn=None,
+        cond_fn=None,
+        model_kwargs=None,
+        eta=0.0,
+    ):
+        """
+        Sample x_{t-1} from the model using DDIM.
+        Same usage as p_sample().
+        """
+        out = self.p_mean_variance(
+            model,
+            x,
+            t,
+            clip_denoised=clip_denoised,
+            denoised_fn=denoised_fn,
+            model_kwargs=model_kwargs,
+        )
+        if cond_fn is not None:
+            out = self.condition_score(cond_fn, out, x, t, model_kwargs=model_kwargs)
+
+        # Usually our model outputs epsilon, but we re-derive it
+        # in case we used x_start or x_prev prediction.
+        eps = self._predict_eps_from_xstart(x, t, out["pred_xstart"])
+
+        alpha_bar = _extract_into_tensor(self.alphas_cumprod, t, x.shape)
+        alpha_bar_prev = _extract_into_tensor(self.alphas_cumprod_prev, t, x.shape)
+        sigma = (
+            eta
+            * th.sqrt((1 - alpha_bar_prev) / (1 - alpha_bar))
+            * th.sqrt(1 - alpha_bar / alpha_bar_prev)
+        )
+        # Equation 12.
+        noise = th.randn_like(x)
+        mean_pred = (
+            out["pred_xstart"] * th.sqrt(alpha_bar_prev)
+            + th.sqrt(1 - alpha_bar_prev - sigma**2) * eps
+        )
+        nonzero_mask = (
+            (t != 0).float().view(-1, *([1] * (len(x.shape) - 1)))
+        )  # no noise when t == 0
+        sample = mean_pred + nonzero_mask * sigma * noise
+        return {"sample": sample, "pred_xstart": out["pred_xstart"]}
+
+    def ddim_reverse_sample(
+        self,
+        model,
+        x,
+        t,
+        clip_denoised=True,
+        denoised_fn=None,
+        cond_fn=None,
+        model_kwargs=None,
+        eta=0.0,
+    ):
+        """
+        Sample x_{t+1} from the model using DDIM reverse ODE.
+        """
+        assert eta == 0.0, "Reverse ODE only for deterministic path"
+        out = self.p_mean_variance(
+            model,
+            x,
+            t,
+            clip_denoised=clip_denoised,
+            denoised_fn=denoised_fn,
+            model_kwargs=model_kwargs,
+        )
+        if cond_fn is not None:
+            out = self.condition_score(cond_fn, out, x, t, model_kwargs=model_kwargs)
+        # Usually our model outputs epsilon, but we re-derive it
+        # in case we used x_start or x_prev prediction.
+        eps = (
+            _extract_into_tensor(self.sqrt_recip_alphas_cumprod, t, x.shape) * x
+            - out["pred_xstart"]
+        ) / _extract_into_tensor(self.sqrt_recipm1_alphas_cumprod, t, x.shape)
+        alpha_bar_next = _extract_into_tensor(self.alphas_cumprod_next, t, x.shape)
+
+        # Equation 12. reversed
+        mean_pred = (
+            out["pred_xstart"] * th.sqrt(alpha_bar_next)
+            + th.sqrt(1 - alpha_bar_next) * eps
+        )
+
+        return {"sample": mean_pred, "pred_xstart": out["pred_xstart"]}
+
+    def ddim_sample_loop(
+        self,
+        model,
+        shape,
+        noise=None,
+        clip_denoised=True,
+        denoised_fn=None,
+        cond_fn=None,
+        model_kwargs=None,
+        device=None,
+        progress=False,
+        eta=0.0,
+    ):
+        """
+        Generate samples from the model using DDIM.
+        Same usage as p_sample_loop().
+        """
+        final = None
+        for sample in self.ddim_sample_loop_progressive(
+            model,
+            shape,
+            noise=noise,
+            clip_denoised=clip_denoised,
+            denoised_fn=denoised_fn,
+            cond_fn=cond_fn,
+            model_kwargs=model_kwargs,
+            device=device,
+            progress=progress,
+            eta=eta,
+        ):
+            final = sample
+        return final["sample"]
+
+    def ddim_sample_loop_progressive(
+        self,
+        model,
+        shape,
+        noise=None,
+        clip_denoised=True,
+        denoised_fn=None,
+        cond_fn=None,
+        model_kwargs=None,
+        device=None,
+        progress=False,
+        eta=0.0,
+    ):
+        """
+        Use DDIM to sample from the model and yield intermediate samples from
+        each timestep of DDIM.
+        Same usage as p_sample_loop_progressive().
+        """
+        if device is None:
+            device = next(model.parameters()).device
+        assert isinstance(shape, (tuple, list))
+        if noise is not None:
+            img = noise
+        else:
+            img = th.randn(*shape, device=device)
+        indices = list(range(self.num_timesteps))[::-1]
+
+        if progress:
+            # Lazy import so that we don't depend on tqdm.
+            from tqdm.auto import tqdm
+
+            indices = tqdm(indices)
+
+        for i in indices:
+            t = th.tensor([i] * shape[0], device=device)
+            with th.no_grad():
+                out = self.ddim_sample(
+                    model,
+                    img,
+                    t,
+                    clip_denoised=clip_denoised,
+                    denoised_fn=denoised_fn,
+                    cond_fn=cond_fn,
+                    model_kwargs=model_kwargs,
+                    eta=eta,
+                )
+                yield out
+                img = out["sample"]
+
+
+def _extract_into_tensor(arr, timesteps, broadcast_shape):
+    """
+    Extract values from a 1-D numpy array for a batch of indices.
+    :param arr: the 1-D numpy array.
+    :param timesteps: a tensor of indices into the array to extract.
+    :param broadcast_shape: a larger shape of K dimensions with the batch
+                            dimension equal to the length of timesteps.
+    :return: a tensor of shape [batch_size, 1, ...] where the shape has K dims.
+    """
+    res = arr.to(device=timesteps.device)[timesteps].float()
+    while len(res.shape) < len(broadcast_shape):
+        res = res[..., None]
+    return res + th.zeros(broadcast_shape, device=timesteps.device)
diff --git a/ldm/modules/karlo/kakao/modules/diffusion/respace.py b/ldm/modules/karlo/kakao/modules/diffusion/respace.py
new file mode 100644
index 0000000..70c808f
--- /dev/null
+++ b/ldm/modules/karlo/kakao/modules/diffusion/respace.py
@@ -0,0 +1,112 @@
+# ------------------------------------------------------------------------------------
+# Adapted from Guided-Diffusion repo (https://github.com/openai/guided-diffusion)
+# ------------------------------------------------------------------------------------
+
+
+import torch as th
+
+from .gaussian_diffusion import GaussianDiffusion
+
+
+def space_timesteps(num_timesteps, section_counts):
+    """
+    Create a list of timesteps to use from an original diffusion process,
+    given the number of timesteps we want to take from equally-sized portions
+    of the original process.
+
+    For example, if there's 300 timesteps and the section counts are [10,15,20]
+    then the first 100 timesteps are strided to be 10 timesteps, the second 100
+    are strided to be 15 timesteps, and the final 100 are strided to be 20.
+
+    :param num_timesteps: the number of diffusion steps in the original
+                          process to divide up.
+    :param section_counts: either a list of numbers, or a string containing
+                           comma-separated numbers, indicating the step count
+                           per section. As a special case, use "ddimN" where N
+                           is a number of steps to use the striding from the
+                           DDIM paper.
+    :return: a set of diffusion steps from the original process to use.
+    """
+    if isinstance(section_counts, str):
+        if section_counts.startswith("ddim"):
+            desired_count = int(section_counts[len("ddim") :])
+            for i in range(1, num_timesteps):
+                if len(range(0, num_timesteps, i)) == desired_count:
+                    return set(range(0, num_timesteps, i))
+            raise ValueError(
+                f"cannot create exactly {num_timesteps} steps with an integer stride"
+            )
+        elif section_counts == "fast27":
+            steps = space_timesteps(num_timesteps, "10,10,3,2,2")
+            # Help reduce DDIM artifacts from noisiest timesteps.
+            steps.remove(num_timesteps - 1)
+            steps.add(num_timesteps - 3)
+            return steps
+        section_counts = [int(x) for x in section_counts.split(",")]
+    size_per = num_timesteps // len(section_counts)
+    extra = num_timesteps % len(section_counts)
+    start_idx = 0
+    all_steps = []
+    for i, section_count in enumerate(section_counts):
+        size = size_per + (1 if i < extra else 0)
+        if size < section_count:
+            raise ValueError(
+                f"cannot divide section of {size} steps into {section_count}"
+            )
+        if section_count <= 1:
+            frac_stride = 1
+        else:
+            frac_stride = (size - 1) / (section_count - 1)
+        cur_idx = 0.0
+        taken_steps = []
+        for _ in range(section_count):
+            taken_steps.append(start_idx + round(cur_idx))
+            cur_idx += frac_stride
+        all_steps += taken_steps
+        start_idx += size
+    return set(all_steps)
+
+
+class SpacedDiffusion(GaussianDiffusion):
+    """
+    A diffusion process which can skip steps in a base diffusion process.
+
+    :param use_timesteps: a collection (sequence or set) of timesteps from the
+                          original diffusion process to retain.
+    :param kwargs: the kwargs to create the base diffusion process.
+    """
+
+    def __init__(self, use_timesteps, **kwargs):
+        self.use_timesteps = set(use_timesteps)
+        self.original_num_steps = len(kwargs["betas"])
+
+        base_diffusion = GaussianDiffusion(**kwargs)  # pylint: disable=missing-kwoa
+        last_alpha_cumprod = 1.0
+        new_betas = []
+        timestep_map = []
+        for i, alpha_cumprod in enumerate(base_diffusion.alphas_cumprod):
+            if i in self.use_timesteps:
+                new_betas.append(1 - alpha_cumprod / last_alpha_cumprod)
+                last_alpha_cumprod = alpha_cumprod
+                timestep_map.append(i)
+        kwargs["betas"] = th.tensor(new_betas).numpy()
+        super().__init__(**kwargs)
+        self.register_buffer("timestep_map", th.tensor(timestep_map), persistent=False)
+
+    def p_mean_variance(self, model, *args, **kwargs):
+        return super().p_mean_variance(self._wrap_model(model), *args, **kwargs)
+
+    def condition_mean(self, cond_fn, *args, **kwargs):
+        return super().condition_mean(self._wrap_model(cond_fn), *args, **kwargs)
+
+    def condition_score(self, cond_fn, *args, **kwargs):
+        return super().condition_score(self._wrap_model(cond_fn), *args, **kwargs)
+
+    def _wrap_model(self, model):
+        def wrapped(x, ts, **kwargs):
+            ts_cpu = ts.detach().to("cpu")
+            return model(
+                x, self.timestep_map[ts_cpu].to(device=ts.device, dtype=ts.dtype), **kwargs
+            )
+
+        return wrapped
diff --git a/ldm/modules/karlo/kakao/modules/nn.py b/ldm/modules/karlo/kakao/modules/nn.py
new file mode 100644
index 0000000..2eef3f5
--- /dev/null
+++ b/ldm/modules/karlo/kakao/modules/nn.py
@@ -0,0 +1,114 @@
+# ------------------------------------------------------------------------------------
+# Adapted from Guided-Diffusion repo (https://github.com/openai/guided-diffusion)
+# ------------------------------------------------------------------------------------
+
+import math
+
+import torch as th
+import torch.nn as nn
+import torch.nn.functional as F
+
+
+class GroupNorm32(nn.GroupNorm):
+    def __init__(self, num_groups, num_channels, swish, eps=1e-5):
+        super().__init__(num_groups=num_groups, num_channels=num_channels, eps=eps)
+        self.swish = swish
+
+    def forward(self, x):
+        y = super().forward(x.float()).to(x.dtype)
+        if self.swish == 1.0:
+            y = F.silu(y)
+        elif self.swish:
+            y = y * F.sigmoid(y * float(self.swish))
+        return y
+
+
+def conv_nd(dims, *args, **kwargs):
+    """
+    Create a 1D, 2D, or 3D convolution module.
+    """
+    if dims == 1:
+        return nn.Conv1d(*args, **kwargs)
+    elif dims == 2:
+        return nn.Conv2d(*args, **kwargs)
+    elif dims == 3:
+        return nn.Conv3d(*args, **kwargs)
+    raise ValueError(f"unsupported dimensions: {dims}")
+
+
+def linear(*args, **kwargs):
+    """
+    Create a linear module.
+    """
+    return nn.Linear(*args, **kwargs)
+
+
+def avg_pool_nd(dims, *args, **kwargs):
+    """
+    Create a 1D, 2D, or 3D average pooling module.
+    """
+    if dims == 1:
+        return nn.AvgPool1d(*args, **kwargs)
+    elif dims == 2:
+        return nn.AvgPool2d(*args, **kwargs)
+    elif dims == 3:
+        return nn.AvgPool3d(*args, **kwargs)
+    raise ValueError(f"unsupported dimensions: {dims}")
+
+
+def zero_module(module):
+    """
+    Zero out the parameters of a module and return it.
+    """
+    for p in module.parameters():
+        p.detach().zero_()
+    return module
+
+
+def scale_module(module, scale):
+    """
+    Scale the parameters of a module and return it.
+    """
+    for p in module.parameters():
+        p.detach().mul_(scale)
+    return module
+
+
+def normalization(channels, swish=0.0):
+    """
+    Make a standard normalization layer, with an optional swish activation.
+
+    :param channels: number of input channels.
+    :return: an nn.Module for normalization.
+    """
+    return GroupNorm32(num_channels=channels, num_groups=32, swish=swish)
+
+
+def timestep_embedding(timesteps, dim, max_period=10000):
+    """
+    Create sinusoidal timestep embeddings.
+
+    :param timesteps: a 1-D Tensor of N indices, one per batch element.
+                      These may be fractional.
+    :param dim: the dimension of the output.
+    :param max_period: controls the minimum frequency of the embeddings.
+    :return: an [N x dim] Tensor of positional embeddings.
+    """
+    half = dim // 2
+    freqs = th.exp(
+        -math.log(max_period)
+        * th.arange(start=0, end=half, dtype=th.float32, device=timesteps.device)
+        / half
+    )
+    args = timesteps[:, None].float() * freqs[None]
+    embedding = th.cat([th.cos(args), th.sin(args)], dim=-1)
+    if dim % 2:
+        embedding = th.cat([embedding, th.zeros_like(embedding[:, :1])], dim=-1)
+    return embedding
+
+
+def mean_flat(tensor):
+    """
+    Take the mean over all non-batch dimensions.
+    """
+    return tensor.mean(dim=list(range(1, len(tensor.shape))))
diff --git a/ldm/modules/karlo/kakao/modules/resample.py b/ldm/modules/karlo/kakao/modules/resample.py
new file mode 100644
index 0000000..485421a
--- /dev/null
+++ b/ldm/modules/karlo/kakao/modules/resample.py
@@ -0,0 +1,68 @@
+# ------------------------------------------------------------------------------------
+# Modified from Guided-Diffusion (https://github.com/openai/guided-diffusion)
+# ------------------------------------------------------------------------------------
+
+from abc import abstractmethod
+
+import torch as th
+
+
+def create_named_schedule_sampler(name, diffusion):
+    """
+    Create a ScheduleSampler from a library of pre-defined samplers.
+
+    :param name: the name of the sampler.
+    :param diffusion: the diffusion object to sample for.
+    """
+    if name == "uniform":
+        return UniformSampler(diffusion)
+    else:
+        raise NotImplementedError(f"unknown schedule sampler: {name}")
+
+
+class ScheduleSampler(th.nn.Module):
+    """
+    A distribution over timesteps in the diffusion process, intended to reduce
+    variance of the objective.
+
+    By default, samplers perform unbiased importance sampling, in which the
+    objective's mean is unchanged.
+    However, subclasses may override sample() to change how the resampled
+    terms are reweighted, allowing for actual changes in the objective.
+    """
+
+    @abstractmethod
+    def weights(self):
+        """
+        Get a numpy array of weights, one per diffusion step.
+
+        The weights needn't be normalized, but must be positive.
+        """
+
+    def sample(self, batch_size, device):
+        """
+        Importance-sample timesteps for a batch.
+
+        :param batch_size: the number of timesteps.
+        :param device: the torch device to save to.
+        :return: a tuple (timesteps, weights):
+                 - timesteps: a tensor of timestep indices.
+                 - weights: a tensor of weights to scale the resulting losses.
+        """
+        w = self.weights()
+        p = w / th.sum(w)
+        indices = p.multinomial(batch_size, replacement=True)
+        weights = 1 / (len(p) * p[indices])
+        return indices, weights
+
+
+class UniformSampler(ScheduleSampler):
+    def __init__(self, diffusion):
+        super(UniformSampler, self).__init__()
+        self.diffusion = diffusion
+        self.register_buffer(
+            "_weights", th.ones([diffusion.num_timesteps]), persistent=False
+        )
+
+    def weights(self):
+        return self._weights
diff --git a/ldm/modules/karlo/kakao/modules/unet.py b/ldm/modules/karlo/kakao/modules/unet.py
new file mode 100644
index 0000000..c99d0b7
--- /dev/null
+++ b/ldm/modules/karlo/kakao/modules/unet.py
@@ -0,0 +1,792 @@
+# ------------------------------------------------------------------------------------
+# Modified from Guided-Diffusion (https://github.com/openai/guided-diffusion)
+# ------------------------------------------------------------------------------------
+
+import math
+from abc import abstractmethod
+
+import torch as th
+import torch.nn as nn
+import torch.nn.functional as F
+
+from .nn import (
+    avg_pool_nd,
+    conv_nd,
+    linear,
+    normalization,
+    timestep_embedding,
+    zero_module,
+)
+from .xf import LayerNorm
+
+
+class TimestepBlock(nn.Module):
+    """
+    Any module where forward() takes timestep embeddings as a second argument.
+    """
+
+    @abstractmethod
+    def forward(self, x, emb):
+        """
+        Apply the module to `x` given `emb` timestep embeddings.
+        """
+
+
+class TimestepEmbedSequential(nn.Sequential, TimestepBlock):
+    """
+    A sequential module that passes timestep embeddings to the children that
+    support it as an extra input.
+    """
+
+    def forward(self, x, emb, encoder_out=None, mask=None):
+        for layer in self:
+            if isinstance(layer, TimestepBlock):
+                x = layer(x, emb)
+            elif isinstance(layer, AttentionBlock):
+                x = layer(x, encoder_out, mask=mask)
+            else:
+                x = layer(x)
+        return x
+
+
+class Upsample(nn.Module):
+    """
+    An upsampling layer with an optional convolution.
+
+    :param channels: channels in the inputs and outputs.
+    :param use_conv: a bool determining if a convolution is applied.
+    :param dims: determines if the signal is 1D, 2D, or 3D. If 3D, then
+                 upsampling occurs in the inner-two dimensions.
+    """
+
+    def __init__(self, channels, use_conv, dims=2, out_channels=None):
+        super().__init__()
+        self.channels = channels
+        self.out_channels = out_channels or channels
+        self.use_conv = use_conv
+        self.dims = dims
+        if use_conv:
+            self.conv = conv_nd(dims, self.channels, self.out_channels, 3, padding=1)
+
+    def forward(self, x):
+        assert x.shape[1] == self.channels
+        if self.dims == 3:
+            x = F.interpolate(
+                x, (x.shape[2], x.shape[3] * 2, x.shape[4] * 2), mode="nearest"
+            )
+        else:
+            x = F.interpolate(x, scale_factor=2, mode="nearest")
+        if self.use_conv:
+            x = self.conv(x)
+        return x
+
+
+class Downsample(nn.Module):
+    """
+    A downsampling layer with an optional convolution.
+
+    :param channels: channels in the inputs and outputs.
+    :param use_conv: a bool determining if a convolution is applied.
+    :param dims: determines if the signal is 1D, 2D, or 3D. If 3D, then
+                 downsampling occurs in the inner-two dimensions.
+    """
+
+    def __init__(self, channels, use_conv, dims=2, out_channels=None):
+        super().__init__()
+        self.channels = channels
+        self.out_channels = out_channels or channels
+        self.use_conv = use_conv
+        self.dims = dims
+        stride = 2 if dims != 3 else (1, 2, 2)
+        if use_conv:
+            self.op = conv_nd(
+                dims, self.channels, self.out_channels, 3, stride=stride, padding=1
+            )
+        else:
+            assert self.channels == self.out_channels
+            self.op = avg_pool_nd(dims, kernel_size=stride, stride=stride)
+
+    def forward(self, x):
+        assert x.shape[1] == self.channels
+        return self.op(x)
+
+
+class ResBlock(TimestepBlock):
+    """
+    A residual block that can optionally change the number of channels.
+
+    :param channels: the number of input channels.
+    :param emb_channels: the number of timestep embedding channels.
+    :param dropout: the rate of dropout.
+    :param out_channels: if specified, the number of out channels.
+    :param use_conv: if True and out_channels is specified, use a spatial
+        convolution instead of a smaller 1x1 convolution to change the
+        channels in the skip connection.
+    :param dims: determines if the signal is 1D, 2D, or 3D.
+    :param use_checkpoint: if True, use gradient checkpointing on this module.
+    :param up: if True, use this block for upsampling.
+    :param down: if True, use this block for downsampling.
+    """
+
+    def __init__(
+        self,
+        channels,
+        emb_channels,
+        dropout,
+        out_channels=None,
+        use_conv=False,
+        use_scale_shift_norm=False,
+        dims=2,
+        use_checkpoint=False,
+        up=False,
+        down=False,
+    ):
+        super().__init__()
+        self.channels = channels
+        self.emb_channels = emb_channels
+        self.dropout = dropout
+        self.out_channels = out_channels or channels
+        self.use_conv = use_conv
+        self.use_checkpoint = use_checkpoint
+        self.use_scale_shift_norm = use_scale_shift_norm
+
+        self.in_layers = nn.Sequential(
+            normalization(channels, swish=1.0),
+            nn.Identity(),
+            conv_nd(dims, channels, self.out_channels, 3, padding=1),
+        )
+
+        self.updown = up or down
+
+        if up:
+            self.h_upd = Upsample(channels, False, dims)
+            self.x_upd = Upsample(channels, False, dims)
+        elif down:
+            self.h_upd = Downsample(channels, False, dims)
+            self.x_upd = Downsample(channels, False, dims)
+        else:
+            self.h_upd = self.x_upd = nn.Identity()
+
+        self.emb_layers = nn.Sequential(
+            nn.SiLU(),
+            linear(
+                emb_channels,
+                2 * self.out_channels if use_scale_shift_norm else self.out_channels,
+            ),
+        )
+        self.out_layers = nn.Sequential(
+            normalization(
+                self.out_channels, swish=0.0 if use_scale_shift_norm else 1.0
+            ),
+            nn.SiLU() if use_scale_shift_norm else nn.Identity(),
+            nn.Dropout(p=dropout),
+            zero_module(
+                conv_nd(dims, self.out_channels, self.out_channels, 3, padding=1)
+            ),
+        )
+
+        if self.out_channels == channels:
+            self.skip_connection = nn.Identity()
+        elif use_conv:
+            self.skip_connection = conv_nd(
+                dims, channels, self.out_channels, 3, padding=1
+            )
+        else:
+            self.skip_connection = conv_nd(dims, channels, self.out_channels, 1)
+
+    def forward(self, x, emb):
+        """
+        Apply the block to a Tensor, conditioned on a timestep embedding.
+
+        :param x: an [N x C x ...] Tensor of features.
+        :param emb: an [N x emb_channels] Tensor of timestep embeddings.
+        :return: an [N x C x ...] Tensor of outputs.
+        """
+        if self.updown:
+            in_rest, in_conv = self.in_layers[:-1], self.in_layers[-1]
+            h = in_rest(x)
+            h = self.h_upd(h)
+            x = self.x_upd(x)
+            h = in_conv(h)
+        else:
+            h = self.in_layers(x)
+        emb_out = self.emb_layers(emb)
+        while len(emb_out.shape) < len(h.shape):
+            emb_out = emb_out[..., None]
+        if self.use_scale_shift_norm:
+            out_norm, out_rest = self.out_layers[0], self.out_layers[1:]
+            scale, shift = th.chunk(emb_out, 2, dim=1)
+            h = out_norm(h) * (1 + scale) + shift
+            h = out_rest(h)
+        else:
+            h = h + emb_out
+            h = self.out_layers(h)
+        return self.skip_connection(x) + h
+
+
+class ResBlockNoTimeEmbedding(nn.Module):
+    """
+    A residual block without time embedding
+
+    :param channels: the number of input channels.
+    :param emb_channels: the number of timestep embedding channels.
+    :param dropout: the rate of dropout.
+    :param out_channels: if specified, the number of out channels.
+    :param use_conv: if True and out_channels is specified, use a spatial
+        convolution instead of a smaller 1x1 convolution to change the
+        channels in the skip connection.
+    :param dims: determines if the signal is 1D, 2D, or 3D.
+    :param use_checkpoint: if True, use gradient checkpointing on this module.
+    :param up: if True, use this block for upsampling.
+    :param down: if True, use this block for downsampling.
+    """
+
+    def __init__(
+        self,
+        channels,
+        emb_channels,
+        dropout,
+        out_channels=None,
+        use_conv=False,
+        dims=2,
+        use_checkpoint=False,
+        up=False,
+        down=False,
+        **kwargs,
+    ):
+        super().__init__()
+        self.channels = channels
+        self.emb_channels = emb_channels
+        self.dropout = dropout
+        self.out_channels = out_channels or channels
+        self.use_conv = use_conv
+        self.use_checkpoint = use_checkpoint
+
+        self.in_layers = nn.Sequential(
+            normalization(channels, swish=1.0),
+            nn.Identity(),
+            conv_nd(dims, channels, self.out_channels, 3, padding=1),
+        )
+
+        self.updown = up or down
+
+        if up:
+            self.h_upd = Upsample(channels, False, dims)
+            self.x_upd = Upsample(channels, False, dims)
+        elif down:
+            self.h_upd = Downsample(channels, False, dims)
+            self.x_upd = Downsample(channels, False, dims)
+        else:
+            self.h_upd = self.x_upd = nn.Identity()
+
+        self.out_layers = nn.Sequential(
+            normalization(self.out_channels, swish=1.0),
+            nn.Dropout(p=dropout),
+            zero_module(
+                conv_nd(dims, self.out_channels, self.out_channels, 3, padding=1)
+            ),
+        )
+
+        if self.out_channels == channels:
+            self.skip_connection = nn.Identity()
+        elif use_conv:
+            self.skip_connection = conv_nd(
+                dims, channels, self.out_channels, 3, padding=1
+            )
+        else:
+            self.skip_connection = conv_nd(dims, channels, self.out_channels, 1)
+
+    def forward(self, x, emb=None):
+        """
+        Apply the block to a Tensor, NOT conditioned on a timestep embedding.
+
+        :param x: an [N x C x ...] Tensor of features.
+        :param emb: an [N x emb_channels] Tensor of timestep embeddings.
+        :return: an [N x C x ...] Tensor of outputs.
+        """
+        assert emb is None
+
+        if self.updown:
+            in_rest, in_conv = self.in_layers[:-1], self.in_layers[-1]
+            h = in_rest(x)
+            h = self.h_upd(h)
+            x = self.x_upd(x)
+            h = in_conv(h)
+        else:
+            h = self.in_layers(x)
+        h = self.out_layers(h)
+        return self.skip_connection(x) + h
+
+
+class AttentionBlock(nn.Module):
+    """
+    An attention block that allows spatial positions to attend to each other.
+
+    Originally ported from here, but adapted to the N-d case.
+    https://github.com/hojonathanho/diffusion/blob/1e0dceb3b3495bbe19116a5e1b3596cd0706c543/diffusion_tf/models/unet.py#L66.
+    """
+
+    def __init__(
+        self,
+        channels,
+        num_heads=1,
+        num_head_channels=-1,
+        use_checkpoint=False,
+        encoder_channels=None,
+    ):
+        super().__init__()
+        self.channels = channels
+        if num_head_channels == -1:
+            self.num_heads = num_heads
+        else:
+            assert (
+                channels % num_head_channels == 0
+            ), f"q,k,v channels {channels} is not divisible by num_head_channels {num_head_channels}"
+            self.num_heads = channels // num_head_channels
+        self.use_checkpoint = use_checkpoint
+        self.norm = normalization(channels, swish=0.0)
+        self.qkv = conv_nd(1, channels, channels * 3, 1)
+        self.attention = QKVAttention(self.num_heads)
+
+        if encoder_channels is not None:
+            self.encoder_kv = conv_nd(1, encoder_channels, channels * 2, 1)
+        self.proj_out = zero_module(conv_nd(1, channels, channels, 1))
+
+    def forward(self, x, encoder_out=None, mask=None):
+        b, c, *spatial = x.shape
+        qkv = self.qkv(self.norm(x).view(b, c, -1))
+        if encoder_out is not None:
+            encoder_out = self.encoder_kv(encoder_out)
+            h = self.attention(qkv, encoder_out, mask=mask)
+        else:
+            h = self.attention(qkv)
+        h = self.proj_out(h)
+        return x + h.reshape(b, c, *spatial)
+
+
+class QKVAttention(nn.Module):
+    """
+    A module which performs QKV attention. Matches legacy QKVAttention + input/ouput heads shaping
+    """
+
+    def __init__(self, n_heads):
+        super().__init__()
+        self.n_heads = n_heads
+
+    def forward(self, qkv, encoder_kv=None, mask=None):
+        """
+        Apply QKV attention.
+
+        :param qkv: an [N x (H * 3 * C) x T] tensor of Qs, Ks, and Vs.
+        :return: an [N x (H * C) x T] tensor after attention.
+        """
+        bs, width, length = qkv.shape
+        assert width % (3 * self.n_heads) == 0
+        ch = width // (3 * self.n_heads)
+        q, k, v = qkv.reshape(bs * self.n_heads, ch * 3, length).split(ch, dim=1)
+        if encoder_kv is not None:
+            assert encoder_kv.shape[1] == self.n_heads * ch * 2
+            ek, ev = encoder_kv.reshape(bs * self.n_heads, ch * 2, -1).split(ch, dim=1)
+            k = th.cat([ek, k], dim=-1)
+            v = th.cat([ev, v], dim=-1)
+        scale = 1 / math.sqrt(math.sqrt(ch))
+        weight = th.einsum("bct,bcs->bts", q * scale, k * scale)
+        if mask is not None:
+            mask = F.pad(mask, (0, length), value=0.0)
+            mask = (
+                mask.unsqueeze(1)
+                .expand(-1, self.n_heads, -1)
+                .reshape(bs * self.n_heads, 1, -1)
+            )
+            weight = weight + mask
+        weight = th.softmax(weight, dim=-1)
+        a = th.einsum("bts,bcs->bct", weight, v)
+        return a.reshape(bs, -1, length)
+
+
+class UNetModel(nn.Module):
+    """
+    The full UNet model with attention and timestep embedding.
+
+    :param in_channels: channels in the input Tensor.
+    :param model_channels: base channel count for the model.
+    :param out_channels: channels in the output Tensor.
+    :param num_res_blocks: number of residual blocks per downsample.
+    :param attention_resolutions: a collection of downsample rates at which
+        attention will take place. May be a set, list, or tuple.
+        For example, if this contains 4, then at 4x downsampling, attention
+        will be used.
+    :param dropout: the dropout probability.
+    :param channel_mult: channel multiplier for each level of the UNet.
+    :param conv_resample: if True, use learned convolutions for upsampling and
+        downsampling.
+    :param dims: determines if the signal is 1D, 2D, or 3D.
+    :param clip_dim: dimension of clip feature.
+    :param num_classes: if specified (as an int), then this model will be
+        class-conditional with `num_classes` classes.
+    :param use_checkpoint: use gradient checkpointing to reduce memory usage.
+    :param num_heads: the number of attention heads in each attention layer.
+    :param num_heads_channels: if specified, ignore num_heads and instead use
+                               a fixed channel width per attention head.
+    :param num_heads_upsample: works with num_heads to set a different number
+                               of heads for upsampling. Deprecated.
+    :param use_scale_shift_norm: use a FiLM-like conditioning mechanism.
+    :param resblock_updown: use residual blocks for up/downsampling.
+    :param encoder_channels: use to make the dimension of query and kv same in AttentionBlock.
+    :param use_time_embedding: use time embedding for condition.
+    """
+
+    def __init__(
+        self,
+        in_channels,
+        model_channels,
+        out_channels,
+        num_res_blocks,
+        attention_resolutions,
+        dropout=0,
+        channel_mult=(1, 2, 4, 8),
+        conv_resample=True,
+        dims=2,
+        clip_dim=None,
+        use_checkpoint=False,
+        num_heads=1,
+        num_head_channels=-1,
+        num_heads_upsample=-1,
+        use_scale_shift_norm=False,
+        use_middle_attention=True,
+        resblock_updown=False,
+        encoder_channels=None,
+        use_time_embedding=True,
+    ):
+        super().__init__()
+
+        if num_heads_upsample == -1:
+            num_heads_upsample = num_heads
+
+        self.in_channels = in_channels
+        self.model_channels = model_channels
+        self.out_channels = out_channels
+        self.num_res_blocks = num_res_blocks
+        self.attention_resolutions = attention_resolutions
+        self.dropout = dropout
+        self.channel_mult = channel_mult
+        self.conv_resample = conv_resample
+        self.clip_dim = clip_dim
+        self.use_checkpoint = use_checkpoint
+        self.num_heads = num_heads
+        self.num_head_channels = num_head_channels
+        self.num_heads_upsample = num_heads_upsample
+        self.use_middle_attention = use_middle_attention
+        self.use_time_embedding = use_time_embedding
+
+        if self.use_time_embedding:
+            time_embed_dim = model_channels * 4
+            self.time_embed = nn.Sequential(
+                linear(model_channels, time_embed_dim),
+                nn.SiLU(),
+                linear(time_embed_dim, time_embed_dim),
+            )
+
+            if self.clip_dim is not None:
+                self.clip_emb = nn.Linear(clip_dim, time_embed_dim)
+        else:
+            time_embed_dim = None
+
+        CustomResidualBlock = (
+            ResBlock if self.use_time_embedding else ResBlockNoTimeEmbedding
+        )
+        ch = input_ch = int(channel_mult[0] * model_channels)
+        self.input_blocks = nn.ModuleList(
+            [TimestepEmbedSequential(conv_nd(dims, in_channels, ch, 3, padding=1))]
+        )
+        self._feature_size = ch
+        input_block_chans = [ch]
+        ds = 1
+        for level, mult in enumerate(channel_mult):
+            for _ in range(num_res_blocks):
+                layers = [
+                    CustomResidualBlock(
+                        ch,
+                        time_embed_dim,
+                        dropout,
+                        out_channels=int(mult * model_channels),
+                        dims=dims,
+                        use_checkpoint=use_checkpoint,
+                        use_scale_shift_norm=use_scale_shift_norm,
+                    )
+                ]
+                ch = int(mult * model_channels)
+                if ds in attention_resolutions:
+                    layers.append(
+                        AttentionBlock(
+                            ch,
+                            use_checkpoint=use_checkpoint,
+                            num_heads=num_heads,
+                            num_head_channels=num_head_channels,
+                            encoder_channels=encoder_channels,
+                        )
+                    )
+                self.input_blocks.append(TimestepEmbedSequential(*layers))
+                self._feature_size += ch
+                input_block_chans.append(ch)
+            if level != len(channel_mult) - 1:
+                out_ch = ch
+                self.input_blocks.append(
+                    TimestepEmbedSequential(
+                        CustomResidualBlock(
+                            ch,
+                            time_embed_dim,
+                            dropout,
+                            out_channels=out_ch,
+                            dims=dims,
+                            use_checkpoint=use_checkpoint,
+                            use_scale_shift_norm=use_scale_shift_norm,
+                            down=True,
+                        )
+                        if resblock_updown
+                        else Downsample(
+                            ch, conv_resample, dims=dims, out_channels=out_ch
+                        )
+                    )
+                )
+                ch = out_ch
+                input_block_chans.append(ch)
+                ds *= 2
+                self._feature_size += ch
+
+        self.middle_block = TimestepEmbedSequential(
+            CustomResidualBlock(
+                ch,
+                time_embed_dim,
+                dropout,
+                dims=dims,
+                use_checkpoint=use_checkpoint,
+                use_scale_shift_norm=use_scale_shift_norm,
+            ),
+            *(
+                AttentionBlock(
+                    ch,
+                    use_checkpoint=use_checkpoint,
+                    num_heads=num_heads,
+                    num_head_channels=num_head_channels,
+                    encoder_channels=encoder_channels,
+                ),
+            )
+            if self.use_middle_attention
+            else tuple(),  # add AttentionBlock or not
+            CustomResidualBlock(
+                ch,
+                time_embed_dim,
+                dropout,
+                dims=dims,
+                use_checkpoint=use_checkpoint,
+                use_scale_shift_norm=use_scale_shift_norm,
+            ),
+        )
+        self._feature_size += ch
+
+        self.output_blocks = nn.ModuleList([])
+        for level, mult in list(enumerate(channel_mult))[::-1]:
+            for i in range(num_res_blocks + 1):
+                ich = input_block_chans.pop()
+                layers = [
+                    CustomResidualBlock(
+                        ch + ich,
+                        time_embed_dim,
+                        dropout,
+                        out_channels=int(model_channels * mult),
+                        dims=dims,
+                        use_checkpoint=use_checkpoint,
+                        use_scale_shift_norm=use_scale_shift_norm,
+                    )
+                ]
+                ch = int(model_channels * mult)
+                if ds in attention_resolutions:
+                    layers.append(
+                        AttentionBlock(
+                            ch,
+                            use_checkpoint=use_checkpoint,
+                            num_heads=num_heads_upsample,
+                            num_head_channels=num_head_channels,
+                            encoder_channels=encoder_channels,
+                        )
+                    )
+                if level and i == num_res_blocks:
+                    out_ch = ch
+                    layers.append(
+                        CustomResidualBlock(
+                            ch,
+                            time_embed_dim,
+                            dropout,
+                            out_channels=out_ch,
+                            dims=dims,
+                            use_checkpoint=use_checkpoint,
+                            use_scale_shift_norm=use_scale_shift_norm,
+                            up=True,
+                        )
+                        if resblock_updown
+                        else Upsample(ch, conv_resample, dims=dims, out_channels=out_ch)
+                    )
+                    ds //= 2
+                self.output_blocks.append(TimestepEmbedSequential(*layers))
+                self._feature_size += ch
+
+        self.out = nn.Sequential(
+            normalization(ch, swish=1.0),
+            nn.Identity(),
+            zero_module(conv_nd(dims, input_ch, out_channels, 3, padding=1)),
+        )
+
+    def forward(self, x, timesteps, y=None):
+        """
+        Apply the model to an input batch.
+
+        :param x: an [N x C x ...] Tensor of inputs.
+        :param timesteps: a 1-D batch of timesteps.
+        :param y: an [N] Tensor of labels, if class-conditional.
+        :return: an [N x C x ...] Tensor of outputs.
+        """
+        assert (y is not None) == (
+            self.clip_dim is not None
+        ), "must specify y if and only if the model is clip-rep-conditional"
+
+        hs = []
+        if self.use_time_embedding:
+            emb = self.time_embed(timestep_embedding(timesteps, self.model_channels))
+            if self.clip_dim is not None:
+                emb = emb + self.clip_emb(y)
+        else:
+            emb = None
+
+        h = x
+        for module in self.input_blocks:
+            h = module(h, emb)
+            hs.append(h)
+        h = self.middle_block(h, emb)
+        for module in self.output_blocks:
+            h = th.cat([h, hs.pop()], dim=1)
+            h = module(h, emb)
+
+        return self.out(h)
+
+
+class SuperResUNetModel(UNetModel):
+    """
+    A UNetModel that performs super-resolution.
+
+    Expects an extra kwarg `low_res` to condition on a low-resolution image.
+    Assumes that the shape of low-resolution and the input should be the same.
+    """
+
+    def __init__(self, *args, **kwargs):
+        if "in_channels" in kwargs:
+            kwargs = dict(kwargs)
+            kwargs["in_channels"] = kwargs["in_channels"] * 2
+        else:
+            # Curse you, Python. Or really, just curse positional arguments :|.
+            args = list(args)
+            args[1] = args[1] * 2
+        super().__init__(*args, **kwargs)
+
+    def forward(self, x, timesteps, low_res=None, **kwargs):
+        _, _, new_height, new_width = x.shape
+        assert new_height == low_res.shape[2] and new_width == low_res.shape[3]
+
+        x = th.cat([x, low_res], dim=1)
+        return super().forward(x, timesteps, **kwargs)
+
+
+class PLMImUNet(UNetModel):
+    """
+    A UNetModel that conditions on text with a pretrained text encoder in CLIP.
+
+    :param text_ctx: number of text tokens to expect.
+    :param xf_width: width of the transformer.
+    :param clip_emb_mult: #extra tokens by projecting clip text feature.
+    :param clip_emb_type: type of condition (here, we fix clip image feature).
+    :param clip_emb_drop: dropout rato of clip image feature for cfg.
+    """
+
+    def __init__(
+        self,
+        text_ctx,
+        xf_width,
+        *args,
+        clip_emb_mult=None,
+        clip_emb_type="image",
+        clip_emb_drop=0.0,
+        **kwargs,
+    ):
+        self.text_ctx = text_ctx
+        self.xf_width = xf_width
+        self.clip_emb_mult = clip_emb_mult
+        self.clip_emb_type = clip_emb_type
+        self.clip_emb_drop = clip_emb_drop
+
+        if not xf_width:
+            super().__init__(*args, **kwargs, encoder_channels=None)
+        else:
+            super().__init__(*args, **kwargs, encoder_channels=xf_width)
+
+        # Project text encoded feat seq from pre-trained text encoder in CLIP
+        self.text_seq_proj = nn.Sequential(
+            nn.Linear(self.clip_dim, xf_width),
+            LayerNorm(xf_width),
+        )
+        # Project CLIP text feat
+        self.text_feat_proj = nn.Linear(self.clip_dim, self.model_channels * 4)
+
+        assert clip_emb_mult is not None
+        assert clip_emb_type == "image"
+        assert self.clip_dim is not None, "CLIP representation dim should be specified"
+
+        self.clip_tok_proj = nn.Linear(
+            self.clip_dim, self.xf_width * self.clip_emb_mult
+        )
+        if self.clip_emb_drop > 0:
+            self.cf_param = nn.Parameter(th.empty(self.clip_dim, dtype=th.float32))
+
+    def proc_clip_emb_drop(self, feat):
+        if self.clip_emb_drop > 0:
+            bsz, feat_dim = feat.shape
+            assert (
+                feat_dim == self.clip_dim
+            ), f"CLIP input dim: {feat_dim}, model CLIP dim: {self.clip_dim}"
+            drop_idx = th.rand((bsz,), device=feat.device) < self.clip_emb_drop
+            feat = th.where(
+                drop_idx[..., None], self.cf_param[None].type_as(feat), feat
+            )
+        return feat
+
+    def forward(
+        self, x, timesteps, txt_feat=None, txt_feat_seq=None, mask=None, y=None
+    ):
+        bsz = x.shape[0]
+        hs = []
+        emb = self.time_embed(timestep_embedding(timesteps, self.model_channels))
+        emb = emb + self.clip_emb(y)
+
+        xf_out = self.text_seq_proj(txt_feat_seq)
+        xf_out = xf_out.permute(0, 2, 1)
+        emb = emb + self.text_feat_proj(txt_feat)
+        xf_out = th.cat(
+            [
+                self.clip_tok_proj(y).reshape(bsz, -1, self.clip_emb_mult),
+                xf_out,
+            ],
+            dim=2,
+        )
+        mask = F.pad(mask, (self.clip_emb_mult, 0), value=True)
+        mask = th.where(mask, 0.0, float("-inf"))
+
+        h = x
+        for module in self.input_blocks:
+            h = module(h, emb, xf_out, mask=mask)
+            hs.append(h)
+        h = self.middle_block(h, emb, xf_out, mask=mask)
+        for module in self.output_blocks:
+            h = th.cat([h, hs.pop()], dim=1)
+            h = module(h, emb, xf_out, mask=mask)
+        h = self.out(h)
+
+        return h
diff --git a/ldm/modules/karlo/kakao/modules/xf.py b/ldm/modules/karlo/kakao/modules/xf.py
new file mode 100644
index 0000000..66d7d4a
--- /dev/null
+++ b/ldm/modules/karlo/kakao/modules/xf.py
@@ -0,0 +1,231 @@
+# ------------------------------------------------------------------------------------
+# Adapted from the repos below:
+# (a) Guided-Diffusion (https://github.com/openai/guided-diffusion)
+# (b) CLIP ViT (https://github.com/openai/CLIP/)
+# ------------------------------------------------------------------------------------
+
+import math
+
+import torch as th
+import torch.nn as nn
+import torch.nn.functional as F
+
+from .nn import timestep_embedding
+
+
+def convert_module_to_f16(param):
+    """
+    Convert primitive modules to float16.
+    """
+    if isinstance(param, (nn.Linear, nn.Conv2d, nn.ConvTranspose2d)):
+        param.weight.data = param.weight.data.half()
+        if param.bias is not None:
+            param.bias.data = param.bias.data.half()
+
+
+class LayerNorm(nn.LayerNorm):
+    """
+    Implementation that supports fp16 inputs but fp32 gains/biases.
+    """
+
+    def forward(self, x: th.Tensor):
+        return super().forward(x.float()).to(x.dtype)
+
+
+class MultiheadAttention(nn.Module):
+    def __init__(self, n_ctx, width, heads):
+        super().__init__()
+        self.n_ctx = n_ctx
+        self.width = width
+        self.heads = heads
+        self.c_qkv = nn.Linear(width, width * 3)
+        self.c_proj = nn.Linear(width, width)
+        self.attention = QKVMultiheadAttention(heads, n_ctx)
+
+    def forward(self, x, mask=None):
+        x = self.c_qkv(x)
+        x = self.attention(x, mask=mask)
+        x = self.c_proj(x)
+        return x
+
+
+class MLP(nn.Module):
+    def __init__(self, width):
+        super().__init__()
+        self.width = width
+        self.c_fc = nn.Linear(width, width * 4)
+        self.c_proj = nn.Linear(width * 4, width)
+        self.gelu = nn.GELU()
+
+    def forward(self, x):
+        return self.c_proj(self.gelu(self.c_fc(x)))
+
+
+class QKVMultiheadAttention(nn.Module):
+    def __init__(self, n_heads: int, n_ctx: int):
+        super().__init__()
+        self.n_heads = n_heads
+        self.n_ctx = n_ctx
+
+    def forward(self, qkv, mask=None):
+        bs, n_ctx, width = qkv.shape
+        attn_ch = width // self.n_heads // 3
+        scale = 1 / math.sqrt(math.sqrt(attn_ch))
+        qkv = qkv.view(bs, n_ctx, self.n_heads, -1)
+        q, k, v = th.split(qkv, attn_ch, dim=-1)
+        weight = th.einsum("bthc,bshc->bhts", q * scale, k * scale)
+        wdtype = weight.dtype
+        if mask is not None:
+            weight = weight + mask[:, None, ...]
+        weight = th.softmax(weight, dim=-1).type(wdtype)
+        return th.einsum("bhts,bshc->bthc", weight, v).reshape(bs, n_ctx, -1)
+
+
+class ResidualAttentionBlock(nn.Module):
+    def __init__(
+        self,
+        n_ctx: int,
+        width: int,
+        heads: int,
+    ):
+        super().__init__()
+
+        self.attn = MultiheadAttention(
+            n_ctx,
+            width,
+            heads,
+        )
+        self.ln_1 = LayerNorm(width)
+        self.mlp = MLP(width)
+        self.ln_2 = LayerNorm(width)
+
+    def forward(self, x, mask=None):
+        x = x + self.attn(self.ln_1(x), mask=mask)
+        x = x + self.mlp(self.ln_2(x))
+        return x
+
+
+class Transformer(nn.Module):
+    def __init__(
+        self,
+        n_ctx: int,
+        width: int,
+        layers: int,
+        heads: int,
+    ):
+        super().__init__()
+        self.n_ctx = n_ctx
+        self.width = width
+        self.layers = layers
+        self.resblocks = nn.ModuleList(
+            [
+                ResidualAttentionBlock(
+                    n_ctx,
+                    width,
+                    heads,
+                )
+                for _ in range(layers)
+            ]
+        )
+
+    def forward(self, x, mask=None):
+        for block in self.resblocks:
+            x = block(x, mask=mask)
+        return x
+
+
+class PriorTransformer(nn.Module):
+    """
+    A Causal Transformer that conditions on CLIP text embedding, text.
+
+    :param text_ctx: number of text tokens to expect.
+    :param xf_width: width of the transformer.
+    :param xf_layers: depth of the transformer.
+    :param xf_heads: heads in the transformer.
+    :param xf_final_ln: use a LayerNorm after the output layer.
+    :param clip_dim: dimension of clip feature.
+    """
+
+    def __init__(
+        self,
+        text_ctx,
+        xf_width,
+        xf_layers,
+        xf_heads,
+        xf_final_ln,
+        clip_dim,
+    ):
+        super().__init__()
+
+        self.text_ctx = text_ctx
+        self.xf_width = xf_width
+        self.xf_layers = xf_layers
+        self.xf_heads = xf_heads
+        self.clip_dim = clip_dim
+        self.ext_len = 4
+
+        self.time_embed = nn.Sequential(
+            nn.Linear(xf_width, xf_width),
+            nn.SiLU(),
+            nn.Linear(xf_width, xf_width),
+        )
+        self.text_enc_proj = nn.Linear(clip_dim, xf_width)
+        self.text_emb_proj = nn.Linear(clip_dim, xf_width)
+        self.clip_img_proj = nn.Linear(clip_dim, xf_width)
+        self.out_proj = nn.Linear(xf_width, clip_dim)
+        self.transformer = Transformer(
+            text_ctx + self.ext_len,
+            xf_width,
+            xf_layers,
+            xf_heads,
+        )
+        if xf_final_ln:
+            self.final_ln = LayerNorm(xf_width)
+        else:
+            self.final_ln = None
+
+        self.positional_embedding = nn.Parameter(
+            th.empty(1, text_ctx + self.ext_len, xf_width)
+        )
+        self.prd_emb = nn.Parameter(th.randn((1, 1, xf_width)))
+
+        nn.init.normal_(self.prd_emb, std=0.01)
+        nn.init.normal_(self.positional_embedding, std=0.01)
+
+    def forward(
+        self,
+        x,
+        timesteps,
+        text_emb=None,
+        text_enc=None,
+        mask=None,
+        causal_mask=None,
+    ):
+        bsz = x.shape[0]
+        mask = F.pad(mask, (0, self.ext_len), value=True)
+
+        t_emb = self.time_embed(timestep_embedding(timesteps, self.xf_width))
+        text_enc = self.text_enc_proj(text_enc)
+        text_emb = self.text_emb_proj(text_emb)
+        x = self.clip_img_proj(x)
+
+        input_seq = [
+            text_enc,
+            text_emb[:, None, :],
+            t_emb[:, None, :],
+            x[:, None, :],
+            self.prd_emb.to(x.dtype).expand(bsz, -1, -1),
+        ]
+        input = th.cat(input_seq, dim=1)
+        input = input + self.positional_embedding.to(input.dtype)
+
+        mask = th.where(mask, 0.0, float("-inf"))
+        mask = (mask[:, None, :] + causal_mask).to(input.dtype)
+
+        out = self.transformer(input, mask=mask)
+        if self.final_ln is not None:
+            out = self.final_ln(out)
+
+        out = self.out_proj(out[:, -1])
+
+        return out
diff --git a/ldm/modules/karlo/kakao/sampler.py b/ldm/modules/karlo/kakao/sampler.py
new file mode 100644
index 0000000..b56bf2f
--- /dev/null
+++ b/ldm/modules/karlo/kakao/sampler.py
@@ -0,0 +1,272 @@
+# ------------------------------------------------------------------------------------
+# Karlo-v1.0.alpha
+# Copyright (c) 2022 KakaoBrain. All Rights Reserved.
+
+# source: https://github.com/kakaobrain/karlo/blob/3c68a50a16d76b48a15c181d1c5a5e0879a90f85/karlo/sampler/t2i.py#L15
+# ------------------------------------------------------------------------------------
+
+from typing import Iterator
+
+import torch
+import torchvision.transforms.functional as TVF
+from torchvision.transforms import InterpolationMode
+
+from .template import BaseSampler, CKPT_PATH
+
+
+class T2ISampler(BaseSampler):
+    """
+    A sampler for text-to-image generation.
+    :param root_dir: directory for model checkpoints.
+    :param sampling_type: ["default", "fast"]
+    """
+
+    def __init__(
+            self,
+            root_dir: str,
+            sampling_type: str = "default",
+    ):
+        super().__init__(root_dir, sampling_type)
+
+    @classmethod
+    def from_pretrained(
+            cls,
+            root_dir: str,
+            clip_model_path: str,
+            clip_stat_path: str,
+            sampling_type: str = "default",
+    ):
+
+        model = cls(
+            root_dir=root_dir,
+            sampling_type=sampling_type,
+        )
+        model.load_clip(clip_model_path)
+        model.load_prior(
+            f"{CKPT_PATH['prior']}",
+            clip_stat_path=clip_stat_path,
+            prior_config="configs/karlo/prior_1B_vit_l.yaml"
+        )
+        model.load_decoder(f"{CKPT_PATH['decoder']}", decoder_config="configs/karlo/decoder_900M_vit_l.yaml")
+        model.load_sr_64_256(CKPT_PATH["sr_256"], sr_config="configs/karlo/improved_sr_64_256_1.4B.yaml")
+        return model
+
+    def preprocess(
+            self,
+            prompt: str,
+            bsz: int,
+    ):
+        """Setup prompts & cfg scales"""
+        prompts_batch = [prompt for _ in range(bsz)]
+
+        prior_cf_scales_batch = [self._prior_cf_scale] * len(prompts_batch)
+        prior_cf_scales_batch = torch.tensor(prior_cf_scales_batch, device="cuda")
+
+        decoder_cf_scales_batch = [self._decoder_cf_scale] * len(prompts_batch)
+        decoder_cf_scales_batch = torch.tensor(decoder_cf_scales_batch, device="cuda")
+
+        """ Get CLIP text feature """
+        clip_model = self._clip
+        tokenizer = self._tokenizer
+        max_txt_length = self._prior.model.text_ctx
+
+        tok, mask = tokenizer.padded_tokens_and_mask(prompts_batch, max_txt_length)
+        cf_token, cf_mask = tokenizer.padded_tokens_and_mask([""], max_txt_length)
+        if not (cf_token.shape == tok.shape):
+            cf_token = cf_token.expand(tok.shape[0], -1)
+            cf_mask = cf_mask.expand(tok.shape[0], -1)
+
+        tok = torch.cat([tok, cf_token], dim=0)
+        mask = torch.cat([mask, cf_mask], dim=0)
+
+        tok, mask = tok.to(device="cuda"), mask.to(device="cuda")
+        txt_feat, txt_feat_seq = clip_model.encode_text(tok)
+
+        return (
+            prompts_batch,
+            prior_cf_scales_batch,
+            decoder_cf_scales_batch,
+            txt_feat,
+            txt_feat_seq,
+            tok,
+            mask,
+        )
+
+    def __call__(
+            self,
+            prompt: str,
+            bsz: int,
+            progressive_mode=None,
+    ) -> Iterator[torch.Tensor]:
+        assert progressive_mode in ("loop", "stage", "final")
+        with torch.no_grad(), torch.cuda.amp.autocast():
+            (
+                prompts_batch,
+                prior_cf_scales_batch,
+                decoder_cf_scales_batch,
+                txt_feat,
+                txt_feat_seq,
+                tok,
+                mask,
+            ) = self.preprocess(
+                prompt,
+                bsz,
+            )
+
+            """ Transform CLIP text feature into image feature """
+            img_feat = self._prior(
+                txt_feat,
+                txt_feat_seq,
+                mask,
+                prior_cf_scales_batch,
+                timestep_respacing=self._prior_sm,
+            )
+
+            """ Generate 64x64px images """
+            images_64_outputs = self._decoder(
+                txt_feat,
+                txt_feat_seq,
+                tok,
+                mask,
+                img_feat,
+                cf_guidance_scales=decoder_cf_scales_batch,
+                timestep_respacing=self._decoder_sm,
+            )
+
+            images_64 = None
+            for k, out in enumerate(images_64_outputs):
+                images_64 = out
+                if progressive_mode == "loop":
+                    yield torch.clamp(out * 0.5 + 0.5, 0.0, 1.0)
+            if progressive_mode == "stage":
+                yield torch.clamp(out * 0.5 + 0.5, 0.0, 1.0)
+
+            images_64 = torch.clamp(images_64, -1, 1)
+
+            """ Upsample 64x64 to 256x256 """
+            images_256 = TVF.resize(
+                images_64,
+                [256, 256],
+                interpolation=InterpolationMode.BICUBIC,
+                antialias=True,
+            )
+            images_256_outputs = self._sr_64_256(
+                images_256, timestep_respacing=self._sr_sm
+            )
+
+            for k, out in enumerate(images_256_outputs):
+                images_256 = out
+                if progressive_mode == "loop":
+                    yield torch.clamp(out * 0.5 + 0.5, 0.0, 1.0)
+            if progressive_mode == "stage":
+                yield torch.clamp(out * 0.5 + 0.5, 0.0, 1.0)
+
+        yield torch.clamp(images_256 * 0.5 + 0.5, 0.0, 1.0)
+
+
+class PriorSampler(BaseSampler):
+    """
+    A sampler for text-to-image generation, but only the prior.
+    :param root_dir: directory for model checkpoints.
+    :param sampling_type: ["default", "fast"]
+    """
+
+    def __init__(
+            self,
+            root_dir: str,
+            sampling_type: str = "default",
+    ):
+        super().__init__(root_dir, sampling_type)
+
+    @classmethod
+    def from_pretrained(
+            cls,
+            root_dir: str,
+            clip_model_path: str,
+            clip_stat_path: str,
+            sampling_type: str = "default",
+    ):
+        model = cls(
+            root_dir=root_dir,
+            sampling_type=sampling_type,
+        )
+        model.load_clip(clip_model_path)
+        model.load_prior(
+            f"{CKPT_PATH['prior']}",
+            clip_stat_path=clip_stat_path,
+            prior_config="configs/karlo/prior_1B_vit_l.yaml"
+        )
+        return model
+
+    def preprocess(
+            self,
+            prompt: str,
+            bsz: int,
+    ):
+        """Setup prompts & cfg scales"""
+        prompts_batch = [prompt for _ in range(bsz)]
+
+        prior_cf_scales_batch = [self._prior_cf_scale] * len(prompts_batch)
+        prior_cf_scales_batch = torch.tensor(prior_cf_scales_batch, device="cuda")
+
+        decoder_cf_scales_batch = [self._decoder_cf_scale] * len(prompts_batch)
+        decoder_cf_scales_batch = torch.tensor(decoder_cf_scales_batch, device="cuda")
+
+        """ Get CLIP text feature """
+        clip_model = self._clip
+        tokenizer = self._tokenizer
+        max_txt_length = self._prior.model.text_ctx
+
+        tok, mask = tokenizer.padded_tokens_and_mask(prompts_batch, max_txt_length)
+        cf_token, cf_mask = tokenizer.padded_tokens_and_mask([""], max_txt_length)
+        if not (cf_token.shape == tok.shape):
+            cf_token = cf_token.expand(tok.shape[0], -1)
+            cf_mask = cf_mask.expand(tok.shape[0], -1)
+
+        tok = torch.cat([tok, cf_token], dim=0)
+        mask = torch.cat([mask, cf_mask], dim=0)
+
+        tok, mask = tok.to(device="cuda"), mask.to(device="cuda")
+        txt_feat, txt_feat_seq = clip_model.encode_text(tok)
+
+        return (
+            prompts_batch,
+            prior_cf_scales_batch,
+            decoder_cf_scales_batch,
+            txt_feat,
+            txt_feat_seq,
+            tok,
+            mask,
+        )
+
+    def __call__(
+            self,
+            prompt: str,
+            bsz: int,
+            progressive_mode=None,
+    ) -> Iterator[torch.Tensor]:
+        assert progressive_mode in ("loop", "stage", "final")
+        with torch.no_grad(), torch.cuda.amp.autocast():
+            (
+                prompts_batch,
+                prior_cf_scales_batch,
+                decoder_cf_scales_batch,
+                txt_feat,
+                txt_feat_seq,
+                tok,
+                mask,
+            ) = self.preprocess(
+                prompt,
+                bsz,
+            )
+
+            """ Transform CLIP text feature into image feature """
+            img_feat = self._prior(
+                txt_feat,
+                txt_feat_seq,
+                mask,
+                prior_cf_scales_batch,
+                timestep_respacing=self._prior_sm,
+            )
+
+            yield img_feat
diff --git a/ldm/modules/karlo/kakao/template.py b/ldm/modules/karlo/kakao/template.py
new file mode 100644
index 0000000..949e80e
--- /dev/null
+++ b/ldm/modules/karlo/kakao/template.py
@@ -0,0 +1,141 @@
+# ------------------------------------------------------------------------------------
+# Karlo-v1.0.alpha
+# Copyright (c) 2022 KakaoBrain. All Rights Reserved.
+# ------------------------------------------------------------------------------------
+
+import os
+import logging
+import torch
+
+from omegaconf import OmegaConf
+
+from ldm.modules.karlo.kakao.models.clip import CustomizedCLIP, CustomizedTokenizer
+from ldm.modules.karlo.kakao.models.prior_model import PriorDiffusionModel
+from ldm.modules.karlo.kakao.models.decoder_model import Text2ImProgressiveModel
+from ldm.modules.karlo.kakao.models.sr_64_256 import ImprovedSupRes64to256ProgressiveModel
+
+
+SAMPLING_CONF = {
+    "default": {
+        "prior_sm": "25",
+        "prior_n_samples": 1,
+        "prior_cf_scale": 4.0,
+        "decoder_sm": "50",
+        "decoder_cf_scale": 8.0,
+        "sr_sm": "7",
+    },
+    "fast": {
+        "prior_sm": "25",
+        "prior_n_samples": 1,
+        "prior_cf_scale": 4.0,
+        "decoder_sm": "25",
+        "decoder_cf_scale": 8.0,
+        "sr_sm": "7",
+    },
+}
+
+CKPT_PATH = {
+    "prior": "prior-ckpt-step=01000000-of-01000000.ckpt",
+    "decoder": "decoder-ckpt-step=01000000-of-01000000.ckpt",
+    "sr_256": "improved-sr-ckpt-step=1.2M.ckpt",
+}
+
+
+class BaseSampler:
+    _PRIOR_CLASS = PriorDiffusionModel
+    _DECODER_CLASS = Text2ImProgressiveModel
+    _SR256_CLASS = ImprovedSupRes64to256ProgressiveModel
+
+    def __init__(
+        self,
+        root_dir: str,
+        sampling_type: str = "fast",
+    ):
+        self._root_dir = root_dir
+
+        sampling_type = SAMPLING_CONF[sampling_type]
+        self._prior_sm = sampling_type["prior_sm"]
+        self._prior_n_samples = sampling_type["prior_n_samples"]
+        self._prior_cf_scale = sampling_type["prior_cf_scale"]
+
+        assert self._prior_n_samples == 1
+
+        self._decoder_sm = sampling_type["decoder_sm"]
+        self._decoder_cf_scale = sampling_type["decoder_cf_scale"]
+
+        self._sr_sm = sampling_type["sr_sm"]
+
+    def __repr__(self):
+        line = ""
+        line += f"Prior, sampling method: {self._prior_sm}, cf_scale: {self._prior_cf_scale}\n"
+        line += f"Decoder, sampling method: {self._decoder_sm}, cf_scale: {self._decoder_cf_scale}\n"
+        line += f"SR(64->256), sampling method: {self._sr_sm}"
+
+        return line
+
+    def load_clip(self, clip_path: str):
+        clip = CustomizedCLIP.load_from_checkpoint(
+            os.path.join(self._root_dir, clip_path)
+        )
+        clip = torch.jit.script(clip)
+        clip.cuda()
+        clip.eval()
+
+        self._clip = clip
+        self._tokenizer = CustomizedTokenizer()
+
+    def load_prior(
+        self,
+        ckpt_path: str,
+        clip_stat_path: str,
+        prior_config: str = "configs/prior_1B_vit_l.yaml"
+    ):
+        logging.info(f"Loading prior: {ckpt_path}")
+
+        config = OmegaConf.load(prior_config)
+        clip_mean, clip_std = torch.load(
+            os.path.join(self._root_dir, clip_stat_path), map_location="cpu"
+        )
+
+        prior = self._PRIOR_CLASS.load_from_checkpoint(
+            config,
+            self._tokenizer,
+            clip_mean,
+            clip_std,
+            os.path.join(self._root_dir, ckpt_path),
+            strict=True,
+        )
+        prior.cuda()
+        prior.eval()
+        logging.info("done.")
+
+        self._prior = prior
+
+    def load_decoder(self, ckpt_path: str, decoder_config: str = "configs/decoder_900M_vit_l.yaml"):
+        logging.info(f"Loading decoder: {ckpt_path}")
+
+        config = OmegaConf.load(decoder_config)
+        decoder = self._DECODER_CLASS.load_from_checkpoint(
+            config,
+            self._tokenizer,
+            os.path.join(self._root_dir, ckpt_path),
+            strict=True,
+        )
+        decoder.cuda()
+        decoder.eval()
+        logging.info("done.")
+
+        self._decoder = decoder
+
+    def load_sr_64_256(self, ckpt_path: str, sr_config: str = "configs/improved_sr_64_256_1.4B.yaml"):
+        logging.info(f"Loading SR(64->256): {ckpt_path}")
+
+        config = OmegaConf.load(sr_config)
+        sr = self._SR256_CLASS.load_from_checkpoint(
+            config, os.path.join(self._root_dir, ckpt_path), strict=True
+        )
+        sr.cuda()
+        sr.eval()
+        logging.info("done.")
+
+        self._sr_64_256 = sr
\ No newline at end of file
diff --git a/scripts/streamlit/stablekarlo.py b/scripts/streamlit/stablekarlo.py
new file mode 100644
index 0000000..e57500c
--- /dev/null
+++ b/scripts/streamlit/stablekarlo.py
@@ -0,0 +1,381 @@
+import importlib
+import streamlit as st
+import torch
+import cv2
+import numpy as np
+import PIL
+from omegaconf import OmegaConf
+from PIL import Image
+from tqdm import trange
+import io, os
+from torch import autocast
+from einops import rearrange, repeat
+from torchvision.utils import make_grid
+from pytorch_lightning import seed_everything
+from contextlib import nullcontext
+
+from ldm.models.diffusion.ddim import DDIMSampler
+from ldm.models.diffusion.plms import PLMSSampler
+from ldm.models.diffusion.dpm_solver import DPMSolverSampler
+
+
+torch.set_grad_enabled(False)
+
+PROMPTS_ROOT = "scripts/prompts/"
+SAVE_PATH = "outputs/demo/stable-karlo/"
+
+VERSION2SPECS = {
+    "Stable Karlo": {"H": 768, "W": 768, "C": 4, "f": 8},
+    "Full Karlo": {}
+}
+
+
+def get_obj_from_str(string, reload=False):
+    module, cls = string.rsplit(".", 1)
+    importlib.invalidate_caches()
+    if reload:
+        module_imp = importlib.import_module(module)
+        importlib.reload(module_imp)
+    return getattr(importlib.import_module(module, package=None), cls)
+
+
+def instantiate_from_config(config):
+    if not "target" in config:
+        raise KeyError("Expected key `target` to instantiate.")
+    return get_obj_from_str(config["target"])(**config.get("params", dict()))
+
+
+def get_interactive_image():
+    image = st.file_uploader("Input", type=["jpg", "JPEG", "png"])
+    if image is not None:
+        image = Image.open(image)
+        if not image.mode == "RGB":
+            image = image.convert("RGB")
+        return image
+
+
+def load_img(display=True):
+    image = get_interactive_image()
+    if display:
+        st.image(image)
+    w, h = image.size
+    print(f"loaded input image of size ({w}, {h})")
+    w, h = map(lambda x: x - x % 64, (w, h))
+    image = image.resize((w, h), resample=PIL.Image.LANCZOS)
+    image = np.array(image).astype(np.float32) / 255.0
+    image = image[None].transpose(0, 3, 1, 2)
+    image = torch.from_numpy(image)
+    return 2. * image - 1.
+
+
+def get_init_img(batch_size=1):
+    init_image = load_img().cuda()
+    init_image = repeat(init_image, '1 ... -> b ...', b=batch_size)
+    return init_image
+
+
+def sample(
+        model,
+        prompt,
+        n_runs=3,
+        n_samples=2,
+        H=512,
+        W=512,
+        C=4,
+        f=8,
+        scale=10.0,
+        ddim_steps=50,
+        ddim_eta=0.0,
+        callback=None,
+        skip_single_save=False,
+        save_grid=True,
+        ucg_schedule=None,
+        negative_prompt="",
+        adm_cond=None,
+        adm_uc=None,
+        use_full_precision=False,
+        only_adm_cond=False
+):
+    batch_size = n_samples
+    precision_scope = autocast if not use_full_precision else nullcontext
+    if use_full_precision: st.warning(f"Sampling {model.__class__.__name__} at full precision.")
+    if isinstance(prompt, str):
+        prompt = [prompt]
+    prompts = batch_size * prompt
+
+    outputs = st.empty()
+
+    with precision_scope("cuda"):
+        with model.ema_scope():
+            all_samples = list()
+            for n in trange(n_runs, desc="Sampling"):
+                shape = [C, H // f, W // f]
+                if not only_adm_cond:
+                    uc = None
+                    if scale != 1.0:
+                        uc = model.get_learned_conditioning(batch_size * [negative_prompt])
+                    if isinstance(prompts, tuple):
+                        prompts = list(prompts)
+                    c = model.get_learned_conditioning(prompts)
+                    if isinstance(model, Txt2ImgDiffusionWithPooledInput):
+                        c, uc = c[0], uc[0]
+
+                if adm_cond is not None:
+                    if adm_cond.shape[0] == 1:
+                        adm_cond = repeat(adm_cond, '1 ... -> b ...', b=batch_size)
+                    if adm_uc is None:
+                        st.warning("Not guiding via c_adm")
+                        adm_uc = adm_cond
+                    else:
+                        if adm_uc.shape[0] == 1:
+                            adm_uc = repeat(adm_uc, '1 ... -> b ...', b=batch_size)
+                    if not only_adm_cond:
+                        c = {"c_crossattn": [c], "c_adm": adm_cond}
+                        uc = {"c_crossattn": [uc], "c_adm": adm_uc}
+                    else:
+                        c = adm_cond
+                        uc = adm_uc
+                samples_ddim, _ = sampler.sample(S=ddim_steps,
+                                                 conditioning=c,
+                                                 batch_size=batch_size,
+                                                 shape=shape,
+                                                 verbose=False,
+                                                 unconditional_guidance_scale=scale,
+                                                 unconditional_conditioning=uc,
+                                                 eta=ddim_eta,
+                                                 x_T=None,
+                                                 callback=callback,
+                                                 ucg_schedule=ucg_schedule
+                                                 )
+
+                x_samples = model.decode_first_stage(samples_ddim)
+                x_samples = torch.clamp((x_samples + 1.0) / 2.0, min=0.0, max=1.0)
+
+                if not skip_single_save:
+                    base_count = len(os.listdir(os.path.join(SAVE_PATH, "samples")))
+                    for x_sample in x_samples:
+                        x_sample = 255. * rearrange(x_sample.cpu().numpy(), 'c h w -> h w c')
+                        Image.fromarray(x_sample.astype(np.uint8)).save(
+                            os.path.join(SAVE_PATH, "samples", f"{base_count:09}.png"))
+                        base_count += 1
+
+                all_samples.append(x_samples)
+
+                # get grid of all samples
+                grid = torch.stack(all_samples, 0)
+                grid = rearrange(grid, 'n b c h w -> (n h) (b w) c')
+                outputs.image(grid.cpu().numpy())
+
+            # additionally, save grid
+            grid = Image.fromarray((255. * grid.cpu().numpy()).astype(np.uint8))
+            if save_grid:
+                grid_count = len(os.listdir(SAVE_PATH)) - 1
+                grid.save(os.path.join(SAVE_PATH, f'grid-{grid_count:06}.png'))
+
+    return x_samples
+
+
+def make_oscillating_guidance_schedule(num_steps, max_weight=15., min_weight=1.):
+    schedule = list()
+    for i in range(num_steps):
+        if float(i / num_steps) < 0.1:
+            schedule.append(max_weight)
+        elif i % 2 == 0:
+            schedule.append(min_weight)
+        else:
+            schedule.append(max_weight)
+    print(f"OSCILLATING GUIDANCE SCHEDULE: \n {schedule}")
+    return schedule
+
+
+def torch2np(x):
+    x = ((x + 1.0) * 127.5).clamp(0, 255).to(dtype=torch.uint8)
+    x = x.permute(0, 2, 3, 1).detach().cpu().numpy()
+    return x
+
+
+@st.cache(allow_output_mutation=True, suppress_st_warning=True)
+def init(version="Stable Karlo", load_karlo_prior=False):
+    state = dict()
+    if not "model" in state:
+        if version == "Stable Karlo":
+            config = "configs/stable-diffusion/v2-1-stable-karlo-inference.yaml"
+            ckpt = "checkpoints/v2-1-stable-unclip-ft.ckpt"
+
+        elif version == "Full Karlo":
+            from ldm.modules.karlo.kakao.sampler import T2ISampler
+            st.info("Loading full KARLO..")
+            karlo = T2ISampler.from_pretrained(
+                root_dir="checkpoints/karlo_models",
+                clip_model_path="ViT-L-14.pt",
+                clip_stat_path="ViT-L-14_stats.th",
+                sampling_type="default",
+            )
+            state["karlo_prior"] = karlo
+            state["msg"] = "loaded full Karlo"
+            return state
+        else:
+            raise ValueError(f"version {version} unknown!")
+
+        config = OmegaConf.load(config)
+        model, msg = load_model_from_config(config, ckpt, vae_sd=None)
+        state["msg"] = msg
+
+        if load_karlo_prior:
+            from ldm.modules.karlo.kakao.sampler import PriorSampler
+            st.info("Loading KARLO CLIP prior...")
+            karlo_prior = PriorSampler.from_pretrained(
+                root_dir="/fsx/robin/checkpoints/karlo_models",
+                clip_model_path="ViT-L-14.pt",
+                clip_stat_path="ViT-L-14_stats.th",
+                sampling_type="default",
+            )
+            state["karlo_prior"] = karlo_prior
+        state["model"] = model
+        state["ckpt"] = ckpt
+        state["config"] = config
+    return state
+
+
+def load_model_from_config(config, ckpt, verbose=False, vae_sd=None):
+    print(f"Loading model from {ckpt}")
+    pl_sd = torch.load(ckpt, map_location="cpu")
+    msg = None
+    if "global_step" in pl_sd:
+        msg = f"This is global step {pl_sd['global_step']}. "
+    if "model_ema.num_updates" in pl_sd["state_dict"]:
+        msg += f"And we got {pl_sd['state_dict']['model_ema.num_updates']} EMA updates."
+    global_step = pl_sd.get("global_step", "?")
+    sd = pl_sd["state_dict"]
+    if vae_sd is not None:
+        for k in sd.keys():
+            if "first_stage" in k:
+                sd[k] = vae_sd[k[len("first_stage_model."):]]
+
+    model = instantiate_from_config(config.model)
+    m, u = model.load_state_dict(sd, strict=False)
+    if len(m) > 0 and verbose:
+        print("missing keys:")
+        print(m)
+    if len(u) > 0 and verbose:
+        print("unexpected keys:")
+        print(u)
+
+    model.cuda()
+    model.eval()
+    print(f"Loaded global step {global_step}")
+    return model, msg
+
+
+if __name__ == "__main__":
+    st.title("Stable Karlo")
+    mode = "txt2img"
+    version = st.selectbox("Model Version", list(VERSION2SPECS.keys()), 0)
+    use_karlo = st.checkbox("Use KARLO prior", False)
+    state = init(version=version, vae_version=vae_version, load_karlo_prior=use_karlo)
+    st.info(state["msg"])
+    prompt = st.text_input("Prompt", "a professional photograph of an astronaut riding a horse")
+    negative_prompt = st.text_input("Negative Prompt", "")
+    scale = st.number_input("cfg-scale", value=10., min_value=-100., max_value=100.)
+    number_rows = st.number_input("num rows", value=2, min_value=1, max_value=10)
+    number_cols = st.number_input("num cols", value=2, min_value=1, max_value=10)
+    default_steps = 25 
+    steps = st.sidebar.number_input("steps", value=default_steps, min_value=1, max_value=1000)
+    eta = st.sidebar.number_input("eta (DDIM)", value=0., min_value=0., max_value=1.)
+    force_full_precision = st.sidebar.checkbox("Force FP32", False)
+    if version != "Full Karlo":
+        H = st.sidebar.number_input("H", value=VERSION2SPECS[version]["H"], min_value=64, max_value=2048)
+        W = st.sidebar.number_input("W", value=VERSION2SPECS[version]["W"], min_value=64, max_value=2048)
+        C = VERSION2SPECS[version]["C"]
+        f = VERSION2SPECS[version]["f"]
+
+    SAVE_PATH = os.path.join(SAVE_PATH, version + "_" + vae_version + "-decoder")
+    os.makedirs(os.path.join(SAVE_PATH, "samples"), exist_ok=True)
+
+    seed = st.sidebar.number_input("seed", value=42, min_value=0, max_value=int(1e9))
+    seed_everything(seed)
+
+    ucg_schedule = None
+    sampler = st.sidebar.selectbox("Sampler", ["DDIM", "PLMS", "DPM"], 2)
+    if version == "Full Karlo":
+        pass
+    else:
+        if sampler == "PLMS":
+            st.warning("NOTE: Some models (such as v-pred) currently only support DDIM/DPM sampling here")
+            sampler = PLMSSampler(state["model"])
+        elif sampler == "DPM":
+            st.warning("NOTE: Using DPM sampler with default sampling parameters (DPM-2)")
+            sampler = DPMSolverSampler(state["model"])
+        elif sampler == "DDIM":
+            sampler = DDIMSampler(state["model"])
+            if st.checkbox("Try oscillating guidance?", False):
+                ucg_schedule = make_oscillating_guidance_schedule(num_steps=steps, max_weight=scale, min_weight=1.)
+        else:
+            raise ValueError(f"unknown sampler {sampler}!")
+
+    adm_cond, adm_uc = None, None
+    if use_karlo:
+        # uses the prior
+        karlo_sampler = state["karlo_prior"]
+        with torch.no_grad():
+            karlo_prediction = iter(
+                karlo_sampler(
+                    prompt=prompt,
+                    bsz=number_cols,
+                    progressive_mode="final",
+                )
+            ).__next__()
+        adm_cond = karlo_prediction
+        adm_uc = torch.zeros_like(karlo_prediction)
+
+    else:
+        init_img = get_init_img(batch_size=number_cols)
+        with torch.no_grad():
+            adm_cond = state["model"].embedder(init_img)
+            adm_uc = torch.zeros_like(adm_cond)
+
+    if st.button("Sample"):
+        print("running prompt:", prompt)
+        st.text("Sampling")
+        t_progress = st.progress(0)
+        result = st.empty()
+
+        def t_callback(t):
+            t_progress.progress(min((t + 1) / steps, 1.))
+
+        if version == "KARLO":
+            outputs = st.empty()
+            karlo_sampler = state["karlo_prior"]
+            all_samples = list()
+            with torch.no_grad():
+                for _ in range(number_rows):
+                    karlo_prediction = iter(
+                        karlo_sampler(
+                            prompt=prompt,
+                            bsz=number_cols,
+                            progressive_mode="final",
+                        )
+                    ).__next__()
+                    all_samples.append(karlo_prediction)
+            grid = torch.stack(all_samples, 0)
+            grid = rearrange(grid, 'n b c h w -> (n h) (b w) c')
+            outputs.image(grid.cpu().numpy())
+
+        else:
+            samples = sample(
+                state["model"],
+                prompt,
+                n_runs=number_rows,
+                n_samples=number_cols,
+                H=H, W=W, C=C, f=f,
+                scale=scale,
+                ddim_steps=steps,
+                ddim_eta=eta,
+                callback=t_callback,
+                ucg_schedule=ucg_schedule,
+                negative_prompt=negative_prompt,
+                adm_cond=adm_cond, adm_uc=adm_uc,
+                use_full_precision=force_full_precision,
+                only_adm_cond=False
+            )
+            
\ No newline at end of file