From 14e5568890f8153d3508833a507d5150b5efe6bd Mon Sep 17 00:00:00 2001
From: zzhou <zhouzhenghao1998@163.com>
Date: Sat, 11 Oct 2025 14:18:27 +0800
Subject: [PATCH] =?UTF-8?q?2025/10/11=20=E5=8A=A0=E4=BA=86=E4=B8=80?=
 =?UTF-8?q?=E4=B8=AA=E7=BD=91=E7=BB=9C=E6=9C=8D=E5=8A=A1=E9=83=A8=E7=BD=B2?=
 =?UTF-8?q?=E7=9A=84=E5=8A=9F=E8=83=BD?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 audio.json                                    |    1 +
 audio.wav                                     |  Bin 0 -> 67530 bytes
 v2w_service.py                                |   95 +
 voice2word.py                                 |   40 +
 whisper/audio.py                              |   13 +-
 whisper/tokenizer.py                          |   13 +-
 whisper_service_deploy/v2w_service.py         |   94 +
 whisper_service_deploy/whisper/__init__.py    |  161 ++
 whisper_service_deploy/whisper/__main__.py    |    3 +
 whisper_service_deploy/whisper/audio.py       |  157 ++
 whisper_service_deploy/whisper/decoding.py    |  826 ++++++++
 whisper_service_deploy/whisper/model.py       |  345 ++++
 .../whisper/normalizers/__init__.py           |    2 +
 .../whisper/normalizers/basic.py              |   80 +
 .../whisper/normalizers/english.json          | 1741 +++++++++++++++++
 .../whisper/normalizers/english.py            |  550 ++++++
 whisper_service_deploy/whisper/version.py     |    1 +
 17 files changed, 4120 insertions(+), 2 deletions(-)
 create mode 100644 audio.json
 create mode 100644 audio.wav
 create mode 100644 v2w_service.py
 create mode 100644 voice2word.py
 create mode 100644 whisper_service_deploy/v2w_service.py
 create mode 100644 whisper_service_deploy/whisper/__init__.py
 create mode 100644 whisper_service_deploy/whisper/__main__.py
 create mode 100644 whisper_service_deploy/whisper/audio.py
 create mode 100644 whisper_service_deploy/whisper/decoding.py
 create mode 100644 whisper_service_deploy/whisper/model.py
 create mode 100644 whisper_service_deploy/whisper/normalizers/__init__.py
 create mode 100644 whisper_service_deploy/whisper/normalizers/basic.py
 create mode 100644 whisper_service_deploy/whisper/normalizers/english.json
 create mode 100644 whisper_service_deploy/whisper/normalizers/english.py
 create mode 100644 whisper_service_deploy/whisper/version.py

diff --git a/audio.json b/audio.json
new file mode 100644
index 0000000..258fdce
--- /dev/null
+++ b/audio.json
@@ -0,0 +1 @@
+{"text": "\u4eca\u5929\u4e2d\u5348\u5403\u4ec0\u4e48", "segments": [{"id": 0, "seek": 0, "start": 0.0, "end": 2.0, "text": "\u4eca\u5929\u4e2d\u5348\u5403\u4ec0\u4e48", "tokens": [50364, 12074, 5975, 44237, 10123, 10440, 50464], "temperature": 0.0, "avg_logprob": -0.5544378757476807, "compression_ratio": 0.65625, "no_speech_prob": 0.1877238005399704}], "language": "zh"}
\ No newline at end of file
diff --git a/audio.wav b/audio.wav
new file mode 100644
index 0000000000000000000000000000000000000000..24d93142f223243357f6626ad789bbd95b93d773
GIT binary patch
literal 67530
zcmd>m1y@{4vo;Xif@>hSI}A<|Tn8WA8JvOO4haE*ySux)OK=G8?(V@Y`AE+D-uK+I
z?!ABDtF>mTo~r7u-rckJ%-YZHfr5e}Fm-hM2>Hmx0res;n%%;|7V=`!IXPK?bqWw4
zIl@KY3Jp!rhmhmqU;L=-Y%FXn9Ng?|FZu=Jd<0wAyyzFvtNf#&p=BXpL&q1}72<3T
zhW_J(g4(W{Y5Q&drCUDz%MAtflJM`GKfeD4UOfN6AO5T1|7g5=X;9~AZ)o)r)p4{k
z2ft|O-!=bLALf5*^{2!?>ipkzp_L$D3&WSb$ROZ9HU1pe-;WK{^QRKCIoR6%-!&a<
zo&Ggfe|c11Fj&ga25fEeC+7)bVPpCd!f=56)#<+)QoYRaZ~q&xy~%I=TYe80)ydwP
z>|b`LgQJo4iw$;gbohPHzvXw1>dT`39_VlRjsFs8D2(66zm|c`!~yK^8^8MW>C@X6
z4f~>TKHx&ZLqWlu{&~T@cwjIC;h}#|AcljZ<DdK&`?tJc2jjoYKkqm%*!^$#xA~G|
zZ2LFg?>2w$gAU?s;`$ONbF}}Xf2aPg_4hIVh94ldwtwabV)|cm^pXQj_p<1J=E(6c
z|L=DHTI^T%FZPQ-!#dhK*+4_R{vCV$#{~uZqW=DwfJJ-3^nb;EzyFDo{s+JM@zT$~
z`uler`qzKrfAaq}7qBn2U|#zCxBQ7C{u}+(KlDqV|CZlz=(B&Lf5jjEfp7nVU%k|Z
zUi}aL)Bd01`&&M#{-2Z6J*(38Z=?QC`~PpPwZ{K%4juXcs<AD6ajRj(>OSldaSh;@
zWQk8vo4Vw|D@(o>@vGtXa%XR}S?2l^6jT5!yhtt?)a!3C6i`s_ecmG6ZIkFik!Jkz
ziialgc?0G3OUxRIwqz%x;rLa_kpRZ~y{!(YBO%eZmr$=2u02TvU(q1hvn=N}Gni^}
z<szCE$8~o$>n)?bZQ?`q-z;C)TDiVupIH8S5tzNL0;vxj4d@x2a5PwUDi*J0KJ6OZ
ziN_`kD@9ptEHac!A`E_>7kmFj@<i_I?%6rI=#E)6qMJ%rkHC9v^6@4l^zMlP2R$j#
zkdMm5y25CtfdN|gNG4=Fkx_JCMn#w=vB)F8V3YD_E7Bem=RNaU7p^L`M_u3Qk`x&d
zeC2XV4%;FZHm3Jo$f4@mPI=JS_)#Cgra45URvRv?aqaifwC;mMqfcj;MvoznvqTzH
zu|zKnXHH9h_#;y|E26{+pPVMK{oyxGs;GeF3G`GWu(bjuTCTi3En}7hk&k4th1OkY
zQIAO<B@r|(e8zVyy^JIGGrLmV;vG%8;qY=ZQi-XlY6s-m@bp>B){UiL?C>i3kT{>)
zpt~LuDsBFpQbBxeNgd;iCfs~N!~PUmmJ^rL;1kP&;O81va7yp&{9H0)b7|?;)_eWw
z#f#@DxB2#oK2oHir|TrgHs=#1o{bs@qoj?*{L!^E`milnDnGw*j2l|HXrlK-jBW+h
zovDSYPGZTT$W)1$GVn?5$xP7-&_wg4Ge3m)$$gOYK)H;*M>^PndDFzm&E}XzsG%m9
zcHZ5$Yn%J2Q^{tjzUbO_=_-_SR`-$}@<Kn9mSNY<kKS0sYuDfS7#!)Q6rb@V=cBmS
zZXdIy(TY}zTJw$RhJCuXvdvy(VrgP6tGm%_v4{@4wr-B^saVPx+TusGZ)KV6oL=Ff
zm1y_=BEn8_irACh1u=a@AyNJ&+G3sEB~3~8C@>njc!{v%lpsimo%bG7b%Xv`>#e>8
z7sB2C{`J|#%xWGSX#%;RrHmXW`>oS3*Oy4?sinD?PuQNtg%ooXODI6fTGq)CYG&L3
zXwQaO+B!QO60oifHxShNO!6_mM30QPUrK}Sm?>WRgR*AE(5L+GCBbXV0ht*&Y9s+}
z8m~k-D6EyLuAG~xEY%s}lIagG;YXV|@!XXiV#icewe@DO;bl_&^|1TQRFpay`s%fQ
zE{e#A63ENxx>x!~U0B}rRTv74+GAW3CUkSyTg=MwCzg+wimc4#mS3#?yhYWz0k0zG
za!v)ig@H^r65Z>M%gqNEtTdsYOwp?gW<#3@P7bF76G=ebQ5O8QD#%V)wtlV!9EKqj
zp_K|E&Qt^bPS%|Dd=!PO@=Dl63jiXyB*4L-4b3|u1D!Ux5-wBlzFlPBNX_WCA(i3$
zE+T;}dZ2omFWa^BJ2d|zlwuw)brj>U+P3^E!&*npxGRPuMBXG6TDg5_bb>O~!W}qs
z50(<;T1@*5due4waG|QNr-U5te6K}p3UQ$+OXR#ZSW@&NV$tYHKf&elC`Qo${Z7|r
zBrm^4cP}+uzC&$bgR#~f-tcgiQQrS_2;xL4WbC2}BG)vlo4wha9BSJ(r@jP>ja^?n
zc<{Xn(%m=`N^MH+dmq!+Rz{J{oIV!ezfAg0na~m$xc`KFv^d>JZk$uTJuL?Rbn^qC
z5vQF7WY)?4XvN9%BqZsX#(SMivnsvV+?^Xk_r*I;rRo~8UrIVB_x7vIb67q}t<AeW
zT|MU~K6Zj>JKb|j-jcvZC-{@XBMPVNSx~$$9A!!geWiLwd1D97dGHeRsaRdm4;!;t
z1KgFBVyT7{CGuw#wWyFK`SOX(ud;<f1+$&kt*-M5Rs+R(sqmAFN~ZnF$R1MHsxFDB
zWwBk#X_WHf67A(Q6WK~Z5DOQVj=}+OICD9ZO#Rf!9U4Afc(6^sbyK|K5<>8n&9e=K
z)Af@gsWwV#qNOU5a7(TdW(J(H9ObLY$}>fP7#mryS2yBnIB5*oE3-ZwX(yn%%p#$b
zWa7-H&Q+q^JuUizH;IVZ(;u~JI_zvIF0Cz?u`5$em)cO!wWl^c$$E(?VihxAC12}H
zySzt5ARL9W^{mCe-x-G~{H#)9X^#;q9x2m;01fL}Iy2zvB<DDS$+CQnmtYk;NH0|g
z4KGJaMs=JSIHw^U3$*ang7xWbL_VWmsp&d|Sf;;T<_UvkF6WL9=xE*vwYnBa3+kLH
ztq^GwEu%nYUuk<)9j4YZXJC?AI+ETDubETM7T#r^`?@6dVd&W_8rgt-UI2yB-z{Xv
zrv1TRPikpU9g>hV%oxScj<|>udlq1MWF$K{RbRh}HB-WCaTS39!juza;V)FRp-&s%
zBC&!#%nf*R6Pe9XjhEi6sO4&YJXqKmxb(%v#ga<j>h}8TauKYL#$_7E>{S}^9&z=f
z<2fu>34O18=n!{1VUmkKB<rY3l-Zb<()1!OIMf=~X?UFt*bnaAzJHJnG8`Tl%a>sJ
zMvT~BD5T5mo0G)87TaU<%uIsE0#Od{bOG)90tA2(T5FpXWmQQgU)1<HJf?$#lmrO`
zO!B{ngHyd!b(xW|SQt4T$JiK-Tbc`F_a>YA%pk<;?J8Ofa^yIfodf+s)kN^+-12m;
zH-3oumMTH<iO0#2xw~<^$%!!s28qz)-!cfTLX#Q}VrWix^SG`Iaf5oEmemF?uuX`o
zQ+wWVBPGiq(dwh?SP7(Q;Tluo+59Y`DKHtcaAnYpDXJNmfmkJ(f#O<8(j-8ogYBDb
z!O^`dp7HirVGxy%*V2?5z`=E>=~kP&(!L$@gc6Q6%}DhsC-t=Wv&GtaQ=jgLaNcBp
z+l+~RqC;8LJ!qD^6zb>zsi87yiP29zmkGG1uf<wBI9nXQMkteBNK`@CD}+9DP+R0{
zXH#zydMr#uD?JjW-M$L)vx66IZW)@B3suK5O+bp?@%X`cK{X&kVBvRaF1<b#iG6bS
z^M2o-e+5wDDC-8Zyh->&0H*5U#<Vfttqn-JY1@OY_Pn-T9@{cg_-(rY8RCp2S3oMH
zIM~z!i6D<&t}vyLC!{*x6bRX_0uA`mq+D=lc+%r#hVH*LeU#LU=~kqL<G#`s{57br
z8AglBZHR=yzr+BQtney|Tkjor!IZAk&6^9!Ci|ZuSN(S5HMXM1fKuQXaY4<>xs7V6
zhlxJ>4mRpilI>(ptH(nN*EVB1+gSH78P`l{Hm)4ez+{mYZ+|rPGy@+(9Pa(%yKI`i
zETJtpwATGgE>aNm>-=obrpr@0if+lSjcsxqot(2mms9C0e4DMM+m@Yzrz8RElR8#(
z$1E7Qgc;nJw#@vfUl#}gy7Tjyl-y+!Z9CV2!?7H{f-vdI>^gfb;=8|uObY5J^Xqqi
z{HW;z{v>?WuX{gS*0C0rrjp-A%xd=KSI$zii2TT562!{jh6HNt>*qqzOP2?uS~nqi
zPrjao6EnZ=wpPNP+}%XAd%4=gH}gSr#W78{i9i)<VLh&%VlgHdS=cY!VsVn+pOZzU
zf0)WiU<_G0iT2H1?xi*`CnyBuM%#rp%-iIK9xHDuy1TyLP*@uA5IKC@C_X{)<V=Dw
zZh2^)v@?7C9%?r&`I<n^kfVh#ZOD#rDLjc%YqR`N?Wi$Kr<J24r^1FSa%e3ex#^>=
zLClHbbWrPcY?Aq#4h{G$rUw|XLio?uSMNp|%i4M2*YKyBHp=HU0=kR$wgoCYh-Ua3
zx8+f5hSa~s;_J*E+jmhSz^x?Kd|Mg&92V-4zc7cF@uo`2KZ3R|5#Yex7c7NLzt;n9
z+<Yh)V{^r*2h1Zm0lt*Rf=mN}yP{5doifGA$s5FWdh9!+NZ4wRyQt23p~{6s1fSb&
zoyy0=$m=+_nwLcSXt9z*q%Iq&Zrs0VtYD%`_YGK;EkwE>y~*DsNXB^%Hxn+VsdGk-
z78TxVb{LHlu67wrzE+YDI@vCX+0{ZCBq6og3O#}GLPFesaKc*YkJ~q1n!0Q1rrY<T
zgeb26Xv1l4K(R8kZt~>nZDS3*vqd2}_O=uIex%Zi_S4C*W>2$s$}+D)WXLl1)=4)!
zd~~kJHfMb0Eh9^LEaFIDoCOarNlc&u;%p$l=@**oWsSIIb5VUF-LJZb27`S|!Y+A}
zlYs~yH@Xsm=EsNWLnddhav)!)&@&+Fk_ov#SH&w(i9AWQIipJ`g`BIZ;~=tfaPt+?
zgl&$sS_*^#)Zvc#%^277C(|pB$>n>D_$P;eg%0EJMv;<w*z<WVDJ#9YeE%!Wi^)c4
zrX3UEuzYx|TQGEpt^xwoYU<)OU7@Xtm3^$5l#NQ@_x+Gub)T7cw=*-LV|NvEi;dB!
zYx@s&TD7LuSmO@g+=-jC?oo#Zw*lb1b;L<vsQ&togiT0sfnn7Oo$Iva(P&%nBs2LZ
zhVS$TWVw$ewBpi)^8sCNPQ;t(4)Go3VZfcrG~~P=M*YYU*KRXv!FF6hiUq=)nNNW+
zhI<(cqXav_-sqGg{AQ}(dBcNC9`aUK_V6zKdrcg}sL>pF4+8xo2|*Rk^N`ckx>ThP
z->Ie;jwtET5M`_I;vY$W5l4FlxH%zs_u!{gw3IhH{4n-40&3mGR#ms}RScJV5WDtA
z$TnO9wW`O~E!H5j$KSoAvy4$|vRC@8mk@0%r@j%bq5<Jz3#68m^ybze>vTY{tQ<zE
zm?l~-lnt)Xb*T5k9y#tcO5niT)2!T0y@%SnG12YDGLf|B#rc}xlZ3=q0$l@=HzU;C
zKp(R4#_o602euPS=f*oDDH#f4kNG=zrkwNP_tK;q^!@P^q8JSo+1FQ!(wg$QA!FiM
ze7F2u&Sr=osJo?{iLz<pSiUw02}R$X$D)sSD4s5yXt^^FZ{xd=9BXl+t(Z6y*F{|=
z`6upeO}ATzS34ZICdkiPW{dxHd%CJs&}9Y_B&}gE2VCV;re97moKR0N!DjchYD6_I
zO*2P@O(;t#`*K2Qs4+T0U8QmF@IqJvHg~u;k3R<9%oZT#o6X<sYw_qGUOtr927mvh
zZ__+E%r{?jybu{E!Mejy$B&z|=oJ197i0mIG~CBqt3I14+Z|ww_Mu#Q_A#5s(6)jJ
zet2MmR^0H7Ay+315<&#g1j~7_Ab4G=o|hDq7KXFMJj+$L6u5tqJW5#8hh)#p35``U
zR>v$ne{C8dE{#(FPh~B}s~M=-6Zy?F@PivK<j98c;PtFt`JHNfUO~@i!AHwHFE1U6
zdejbYV!IU8I-s4)#yhRzcu%bLc)tvz378(_pbr4bN55pCG<Y+r7oRlP1Rt*{P*C8p
z1|T2gI4iE1{Z-Y`Q@?%QYDS4kKT|_itL8Fng<5ZAk&dx6D%yXd>(f>5Q4@Lnuh#Pu
zA?}4yq4dGK__|EX*{k7`nKAhap;T}D*0aNU&xnq}D2tszi@Aj5q93cQ-SwkS2Q`@A
z+>{wk`?wBM?8s1Idm#=`Yaa&%CWH%6v~GdsUtnb!vY<<_N(hvI6O&7-?=Ew|)7U|@
zgPGOS+H78;?6W4htZkI=^*k*PG)F;oAvFu%(Pk8aWf;5)X;Kg1_^JEc5fsU)yS)mp
zs1;*NwQS+P0iq;qNqA9;7?5xnFc>t}wjvjB9Iw0b1#GWxjRZCyfr|T?DF`ns9=;+m
z00G&sOhN2Zc+G(x`Y6Dp_34>CoTLo1Z-*aMoSt9z9L{nY#1?2*BEsnD)Um%0NOSTF
zk=OjHGSwW$9P(EFptTgmGGXSO-cw>(c2S%?1{eDw05!W?fG=qw*ZGWTV24X;eDRHl
zqDTqQUiPM}a4y~?!u2Qbz^uZn#a^ZYFPutaO6}r+%(BO~{*NY;AMt)zUdnX)8`O!%
zmYHK$3U;Lh+kba%l?8pFu&hYERc#pLq?W>3kx$h1Q??5%)Nl+jPrV=splB>I`f?Cv
z*Risl%P?Pp)AQ(Hxo9QA>!QNz0G--`H~x;8If1dNgMqs5Yp6$OsXD%;)Ugh^;BI0%
zju_1v`Pls+I$BT__Veg-_He`%sefz%M7!z}H!b8&O53WqC1OgGDba0j?JT*3!3Gr(
z3@e3;D(6R3ys|yXs+3hav`%B4L{e($KgF<nqAFf-niPUvoubmj=khSA5Yql}4JEa?
zhl({0Gfr!o=!zR)z0(1gCrZ>gk5nvD$&(Wnue@i|kj8s!<uRvl){Vx)bw2?}J8IGr
zpv?}Ovb^agIcG2eoo!b4VpOK|eT&up9tlbz+Sjtt6A6+*reEda*!~s?^MRZq95>gu
z`@=T3CL&Dv>!X$<!7>MgTH9pG<2$J1H0~9N%`1t79ClIr6fpHQFPGX;^p$St*uz^d
zPpuiA4Ab(5k~qRov2L)D9Ll|MY2}uRw@lx)A}dFQc7bDxcW0uX3`5mY1n_DUCo3T5
zjIg$%C8*)P`QHITu7i^8={s%q6DxknV+LClsfiU0gROacCU@5d?JKU^%-Y5hwoK8B
zJ)I?NIu1&%vdSogdOVbX2oxoBFK?Iq?I(|2dvC$h?Yq(FkP>CCqzxt8v2*KRUd9k1
z*Y9VxxeC_ds^1BDMv9A4>x-XQbF2}rX0ED-Y-^BBYS&|ROGm>p!uLjdlQN8<K0(P!
zVnkyzCi8e>Y5sg})iR9{p>}GOcJ;C8$jc79e}5@CBTK;uhl;YwD0s|_HPmVFNpgn$
z^gd1B!%CFrqQMKZTrPzf6(LwDT>QXFRzW~k`m-{uuOY0FCDf8}UY*<SY+@}=XIznA
z8^`U4bOH_c%dHGV1f0-jY~^;bGE1h`)E_)~`4-54DslxSnj>6?cKBq(%@0kEL!bRH
zs|6+_Z7`x##fmOSc224qP-~<wNC|g;DJ-?Ak1mfj(CZ(@IVbU#+O>L=>|}4I!!d?=
zSX%V0v}a7L3Alf3*}^OPJY%>+OOA)s-~hpsa?;09a<T?Jac+~yW*O19AFCzA<;z{5
zWop>$xl-sJAwb=vRiDsop74g2cSD?s&C-*n15JH=!Yh=hIK9%CW+Va6D!2SHYMqH+
zy)b$P(xp}sF8CDnpO@M$mg_QUE1G^KgQ-(ZUTb-3<tGW}5_{`QJ<O5Q7bq56q`LRH
z8<kaV^P`*Fxrm^U<WTSI)ai4j4|^0XT~@PxJ2qzioY?HiOgu+eu)_Z!{75(I26%Y9
zidCj>F=A#Pd2Wl*4sd-ld%qhoQ>WW<%X3=6oO;@yUu)Ju{XIRP-P_u(^T(w;E&r&Y
z`5?wT`wX7CuFZZ|$NK!U_T*O||N6&<-s(ar-pY3+CySgl=aR$FM@jugKp}6zCs{Yp
z8tbbt{?Ig$5K1C-!C&!y1k=Oh$3YtUHDA7Z^KKt!*mRS-N2SJijk7iC;^7&}!-v5m
zDwEvh3O)`+=kE_huL6|mxp>m^L)ptz5~(LTi!6px2#|G3THH+HV?S8DDO8tR<S*>w
zXb36y-b8SV|4{=?o-cDGRkopp5$0DYMmwAn`SlypRr-zS=Nl|87qNl{Qh?17yhae~
zqv^%c9A8i8)|1!y{b7q5^}6~-264sVFEbUX_%BGMTqC1?dB8$BdKxi`(uFrNkagZ&
zQ6-W6ooLQl0=GBZva6APPTp&18pEsC={loDUa|F|C!<%N^E80|OBh6qFW2{n`Eqt=
zK|~Z{IG`mhMjHcGjD40eDAYZKdn{>vlcI8Xr$6%IpjdH&vSvf0e=*Ih&#}STNy15*
zKqBRG+aZUf1G~%zyT1i*(N>3KCMWKo<guP%CEzDGU_K(3b}nw~3!EDk@m*Ec*VndI
z@vEe>3*c!PNvCLy&XfC<-4Ajl0z@)HGX?Bn+{;bZkvT~qn*)<5mS-aPwT>8$F`sZi
z1Btk_#$7Bn8@K;fp8WTcT;F@S8f{9Xz{bok-292P(9#8!Z6bYKv5v-4k#=$+?jaN@
zo^+kuQ%SG{-5s_<DZw{6vhid>rtUEL(wOCjS;gtWmtWqyIvV>wum=gdCOWDp$|zPU
z2Zn-jbaAm5ic124QIir{;1WwMWeU4Kz9HZV?>74NbUS@WO&%Sc>av|w7jGVTcy2-%
zjzbqK!ifvAmdAgMRjC===W0k}I#4w{_2bTnc7NepV8=y;&D3=oi(o<iabB`6CsR=0
zE^LPon)7=t^II$bo%bndtFn9D=v%3UZCtKOb#y*sdcg<COM(Ye0Nyp3YdzN;K;7!k
z+^^7N64ncWaflSPvk(uI00{oDmfZodnXNXF+G3jy6yBus{=$WjyqK3<4@kd}jAZ}9
zrdxdx>L)?j^n4`~OE|}%Rsu6X8U}={%cMv_>Dl8)Va+<tR&5M#e37;;Rq^O!4CJ9&
znQQ3=0jK6v&-$G}^2hs<!MW*7@AAVIx2Uz1PwOU6(auXfU!~yVMpd*9n6k|Zrha|5
zCk~m5E~m&L!L@`>uWB0X&6%h%4poE$eW!$Fx^XgPEn!2P3+Z{Sx%8PlxE?7elSURn
z$lg1W2Nj5^MmL8-ly9h7HLwrhTlD2gt46CmdjkQ}v*hUe29&7i%8XKSGN;Q(?4Jg4
z`LYw31<0C-CmiB)VMr9N8FrB5nNpe=64w{X`o#81^vj9lY)93!6v`{%xlk$0jTZHa
z^UjwULn_v3!jN+YV6o_9f}A4SM7r^DbS0wd$dOH_rl~z!r*uUlkHAGzQ-?<sS!)V!
zNy7*e-XM)%{EVU05fi~TEMgyOopx_+CfQr^2rsXXfHf$j$?v0ASJGrPMqm2uEk<Ca
zY<%5d3Lq_GXaXomx3Ce2)73^L6CY|>g|W2wczYV@0r42_cB94fIwjeek6vzr({yIr
ztO`j>`W941Gpf-5TL!9Ys;s0I<i<!AoLp&_3E75&48{58Jxq;TQ;QgJWgPSRo5nq?
z&z%6uOZy6UbsGMQ3DixP<>|&~UKwvrqlmzl!E?yJWT-roRoE##p`KjJR2$b}Wt!t-
zGI*5disSP1fP+%1!gs|newe)N1)uQY5g0)C({IVoQvy;C5G7*Q*&#DN<?9vIZw)_g
zoV;-zN^Ki+DV6;A(UlQ6@+EXYkDaQ1WF*IcjJ25m1kPyCGM7ZKU=w$5t`2_b0oBfI
zSLzGGTpVyA=#vu~E4F-G2Ijri@H9gk3QDu`&qwN;l@ZgSmadV=ByT*t7NTHYrx_zR
zj>GKzJf9Ha07tz3Qd^Mv{AeRu7&9O+2aFj&ZHoZya#CU7#2wWT`3&JtW>z!~|FlNM
z#uL^swcHlGq_7noZ_itYq-yPPLvS@5@i4Y%xdc3LO${iR=5*NVpBN4%DZpG^yg{1Q
zSXuH*XW*@Qz#7;58dan4*>8!x_()cz*@Lh!kJ>jNS)gQ&UilDpWhW)bsT$iwdO)^U
zb(?2LmI<2>nZ5Al=6OQXT&gE@MJ$_YUKij)O7S+2XoePp-!W)IKQWErfjhLX95h`7
zR*~6EAjQc1yk{gQb3H%R&4`NE;DM)y+dk@WH<amSy;{D?+S)gmDzJ#PXGjdnlLjJd
zXo5p<RECEjDfK3z3c>9F|JsyZ7~#7w9F2jN^!f$;DWR*6qMY|J8x%AEw0a3XG$*YB
zhtplVz7lbmNWGsd36;kWF4lU*m#kMWRP_Hu&%f~N{}(+^nfMp;`oGfiWdC7i|Ipjp
z|G+>0JN`SJ|A!vG{zshhZ}|741G|62FDCRq^RfPW9Ohq-4*d518PEBDk3-M@1OJ1*
zhe7^F{AJ$$EhMXd==tNe<P&1Hg{*dp>UNA?QBNKu)-bI;+fEOfsMenXc~5ShsB7nM
zNk^DdVC~{dl`sOsD3uHNp1)DYAv~|Twq)tREp{2iqU^*BtCd_-Ll=mO+i(ZfMK@Z5
zZGYytgG&x8eh~cRA<QOa`>;g3%fzBn{YWZY0;)l*)C-m_AW{)Oru_nCUhhwv=leyB
zD&NTJt2=^nQ=o&aO&b&%Jr|HC!#n6EuI{@RBzosAV@D0(ucptvg(MR`vIdMZNu+kx
z!PEB9vxk5F=7Uk*|MYz5Q~<Az(vETvuunA1{oxy#QB0$Da<y0<ULY<;y38&b;e9hd
zp+r&LG!`#n2XYFzfjgd}0a#l8%Gy{p&;Oa_vWP;*9@n+Kz`9iGrk4g?+m)B`2P_O#
zk#LUR+rcQUljAOe`S0ae$m%1qwNkHHzcerwe-<0l(%?hoA6(W~^IQoXJs{5lnD|rD
z_r!5U4)zf+9N5BBa0pZ?V|Gp_tcyK72WvN4PH9(agUPSHQ0^<={ZOXh<Dx5T$f7W$
zZAYFdEPKpnUalVxj=N26b<Z}Jmm!`u0B*+w)^1p+JAB`YeaaOEoB(yiGJ`lgE`y4;
zy6v#y@zG2agb`&_Y-KUNV&I&m^RJ(ffyL4MMvRWC#6s{t#z2%!S%megZTus~Iya(m
z@qj4QJbSOkWZs(|h?JJp#ON<@8B4NYZa!<kwq8m5Tjm*=Zt6na&6i!7ug_auPFbN>
zO=))s%CliJ$jB5K_=2P2iB82@7z~%H@3&dIy0_AQw)ud%-Y(+n+LR1_(Py_iGuTeR
zK6Kk%6pwmD$4>UV#e6Suy4qi3a4(5Md?e(7V0@aAZj}ONOxCCM?Z4b5t*EQMS3CHH
zPb*MV*_^*xOC>&qX+c{JAh8Z9=O@Q{^&aE3N?i^y59FO+>%$H4ZHKAz(Lu)`>%{F3
z3e+g9Pm}?imwn4XO#4YgKUf)5tnNexLCiUJ4MioS{tpGvggFdc<YDT}QXR#HM8Q%J
z8-sv!I74Cl&7LJTH41962!cRd`(SaSWH%ej$##P#wxA&RL^*<v(m7-DViOZuVY#{0
zt~K*ksUqana>pr9j{|sFBc7-&%4*z%U70aBYrs{uG{Z_I;98G{`myDri!(^0f`gxv
z!MsxiCV<FS6sxe(wte_ACh}7z4%KuJJJvb1esN5Z0&2HhSvPwd9qTu`%k7=F1k6c1
zEMhp4gjp;fBQu~}qG3ctLI<e3BbAl6?}7!4WYMpbkx+gxZ?_vx1?m)9Ap>(+IX$TP
zTsFk17I6VUOT#Fq2M(|~QDZ^00KsL@o5EumABtwAk9!~H^D$Ft_Tu;g8ZOw+Ij&aT
z=yF9sL-CeP-(^x!)cGWx#C^MCKj|HBR?S4@$o{#tn3m&`@^1gE-sElWergPr!-+sR
z2P=a1-Z>kv;>XyiC@i1SE0=3A73Qk|#mS_d*p(qSlNzg;OPjKh>XIHa#qup%P-;Hh
ziF|4o$(^_m{C?ZJIyinjeUwBk=9uJ6q^<b<b7vtIra@(BOykm(4-^Z4V7YfPwKZ=>
zZtcbaWyPUhL4qfb`(r#)6J5S1A0jj5N55$%Fl#i)e#I(;(w5+U^LZZ`%$WEvaAr_7
zDDLiT?mPY=WYO|4M1HpEiWxgMeGR*>v0=QhpJ*wAzhcrNc}Nown|wmw+UJFeJ6cv;
zEd3;5xhj9=y3;q_Mj5<bDU+p3@o=?KG~IWh+YC{ro1t`T4d(}BZ;#|y`sdZP`jwb*
zg^qr_QteiNjUTFMV>>3on0YUj<mCAt@2&<?Qd>jl9=;O>-Fz;*n(<j^zf7@FcP0jd
z*0?|NN)#-2cDArNff!sO&h?B^94hsFagnbq3r=NTCL~|hAX>JtMB@2SP0qG&8GbZ;
z8rNgCGdCJ-5EfW<{J?8NaWdf@w3J0&Zyme6+TgtxOcU-~49rgx6fUZthr~q(Ff&@-
zJHi<Of0oCvg8G7v&>SY_yL44TbX7py#(_3dvO!i9YEi+Du8Rx>Lp9HKHtxSyAB5N#
zJHwR)SJz=!CYDRdKqIY)!4x7$H9OLqZ|5yl{cB|0O@{M}jdnNsu<O@EzK0vQie1l%
zLo>@=FKyV8VIO^euN&LqXKtr^?RK7QkFh5~xPw~S!`32iyeXAJ#rA%Q$I=^0Ez({>
zu>x8R-+_17;6WX~%D`I_JD4Z&=YV9==N!f48f2$5{uKKi0`0@rlbSrpg7%(KucS7F
zZ#7;m!4P8GX1|K~h-y(UcY?5$<(5?soJ03pLQz+6;UXYog*`0$zSKxYq-Bdff>r9a
z)MacguG2KvkSY#AD9#K=YA#7!*>~r(ESKpnY>m_9i?UsUAB0ngt)@}U;aB&(AdIP~
z&vWT@-$Qb@@<wq|F^w`Iy%0xF|BTHR{}U-iQ+uO0F5;CFN*|D$!8vB<(mL9y^!`!u
zXw%?3<EI5yxrB(?g4A*)9zu=K#e+!XQfX_4p>@-QIB%c007vG<P53mYsp)-eFF}d=
zUaROh7w)WO=NgUCQX8T^XQ&9HRoynU%?~SLzWdm@NmgOCRb`W!GvIGeeAY63aQBfz
zF$%(9>JRNg&a+v7x~#$iWI~*#Cskabd~jk(5Wd~C+K~2KNRug(6_Gi9bFFl+Y0W$$
zNTv2z*Zl2LtYBEl>yvT{JF|Vo=b|~;HQ2;}K^fg~JHFbmJ*c0q<v59*L^_0}MLN{W
zYs#A(J90R|8f1?gGupN8vT&R+N0Gl`BRrMiLO>o4sNQ2bU3{hD`PG82=iFzRQ#dAX
zxg$Xkrb>?!>eo@;390nX$B-};<`Fp8JE9cNzVDtWy>P|7zI8Ko9t*)W*7{84B9e4Y
zu-Ou8Dss%JsnkFGlBU!fztuQ@(32})$nPXhL-78bP}R63PK7(7bA5}^l}7N6vEba?
z;%6;Ol~;?A(Sn9-)qG!1@99uwI~mLpyxOqoIIB{M(i;Xp&f~f!9X+|pP15r4M{5H*
zkx9l7Agp_r_OOgTj0ko~zJ;0M+I*S=&AjZJ*K%e;e%ijH`KX8Vq3}jzWlpUnK|FXT
z$FVwxE%fu`i^>flk$onj_Ou_()g;OiA@bsa&-s1Pt6}}x%HLa)PytEkdzAFNnYGIb
zmgkD!fb`em)@8?nX%LumX?O-+3I+L3#q@Hx6?0*eY06PU6DPc@B(epd#OA#9BGIZ0
z`C^$BEU&xg_1o}NM1Z^~+U&b1b!Hj5lMX!hS4O>r%7=^_G&dAy2`u0R=^+_51%eY1
z5&T3xX{yB4P^C(u9_<iJ%`;uB?Xx08WXUM+Wj+t>AgpssOj}7o%o^`dFA7A1kR$EK
z(B=>TS7{L^CtO5Afka7&o<A$pdh6nqnoEzGdXE~Nt(dvKs3@aPJ&E0k3w_G-SFM|A
zMxF<C+`Zv1)YjPFwI~LD=G^Ec19#{%o^dA*p=MGzaea5s_Y4y8kC^)T#jC-!jDUP`
zAP8%wsJT%Uu*KdBR>$puItF3kphlRXoa3-<cREF3`qnA2hg`tMV!6I?gE|cgh=S5@
zPSYGA@7js^CeeRYD)5Gb5EaB{Lz04bz`QEdD2l`36fQy$l(n)q+}d)}zA%M;+7p}O
z^|h{A!uj?0vjH0ets9~|dJdi=TVK*g+PBzBa0IroVb;o%;>>qvpd7Kr-Vd_TTn6k=
zQaLJ=XpPx{g_C*xw4{ep5lV7$WEal(*(PQ8dEM4Mb==L6;fv#7rrO2LZ;kc$&pG&?
zvLX@9k}wwVU~<Cp*-iX#2=D!j#j1pz<5eZd<CH2(9c8yRUd35OQ6R#KoHNx*g%5lf
z%C>i~A8tNCLszAigOz(TRZO#R;NL~aADl{j^piH)bl>L2U;5)-PiU8-FVoMgs=$+n
zA0KhmhOsQQ^nS3Vq0FemDZ{W+N3*vXU?&J!$BM6_Fb}JMVlXFjm5h>Y;-Bi(fR6;0
z<L2tQneawuqVk+ru0kp}{f4G!@uaT-r@ddg{Z8))N=|L++v@`gXpHV73N(D0<E3Xw
zJ4#A0_GanOOTLtZi}w=MJM=))h;PNb4*_!2Gf%C&y9n2!s!^RZ_Fbl#0wNv0+%GG0
zylX@R1efcm9!Yn7gO1Nu`~r=JAU{<yK128mL6<hReI(<m2U+Bvcq+st_siwt!f>pI
z^o)A`Wjbi6jl{-q?mXS3(wXU9#=!#RFXKjN*))bLeqc6H+(12=4|TAIFkeU^-*m!!
z(a{fmI#u=LL0a_k)$IgmtDgPyGBqv=c;{;KS>Bu>f~n8Rg@rLMKS=N2@gQ!PvQhvh
z5(taS;w-C)(wfQ!^RVJMWGW#iun&ho%_{MDX&AXYHbpVr{n>dVh+V^R(FZEUa6+>T
z3*k&!EV<3BhzVJl*u#r?{Ay%eMOv`P=8!I0-70dx1mY=6u2CF#8W(XwCUfMwB#MCR
z^-Rx9CRbT!T9BG5XW%R57?9a&9!!zl1-4RI)mj_rP@<?Z&b70QFuz?P;>B!nDkOGl
zdqRRQ4--$$x5PBuagl2`InzmXOGZK`w8aC{a6(#S4u<Ai>g7n0dwJTqyya><$`<Mn
z+%&QjwuIFK*Jrqg0F~S&eyA#Eq=WqHP&#>$5d|4C-*bF}Ap#W<t`IG${v7J)R}S*x
z7vsX@%*@A#n1CYb5e5=N*m}vp-ZvxiVd5!NjI^R*_F}vV=H||Jo3$9Z@Xoj8*PLCB
zDk(EH-UZ}pc##Su%3e@+yyc3Ps0L_wrFtr?`8B~U(T8eU91OT*^M1$CE%fX;TY-UK
zLt268ukfPyL%$}T$L_~;7P}_A8}33?*d+vr9*4aAX4Gd~!-fY=k|p0AJyamgDoN}z
zO1`U&vFH9)<pIxQ%cN-xVz>6}T;YONuUB?{+hHwa9~gjYTC{n4YW>LkgRCs=L8vmt
z>JuF7w?r$HjQ}~0`tN5;N*hb^;VU}>y*Wgq>9h>6D^nteY0^a|1HK;G@X`@>c4;oL
zAWWes7w%qx0NOT*Q&pvX%a0aiIK?R-!8F7MpuGHb+o45x_>@%$z(3D|o!oX~Ad;bl
zAANd3Hci_uUUPpnoE>s%rapW?z&xnhiYjWN0ninq5n<2Kt7JnRP3BN%)#*<G95s?9
zq!`+=IP(ecFEmC{J{FGTBHd2W??T;}zQ)O{WqdRUUN5=p(N#n<g`S?9X-JL}CY%fr
z?scs;EJTZc#8=yF2=GHyv%&zTSTPT?&*Sup#qa|-pc%h((3Kr*&XlxdsVhYh#M=U@
z;G8C0CS@ZcXyUv`;6TOpvO1e|YLp_bE^k24-m3MQDa^UYg90cHu@gy^LZ#J5%1M@K
zR&#O+8Ir-}%;%#rehj;+GjlP%yF?nZ0IVB=9}0AsRZ2F`R44oOb{?QNa>hTQo_Ncr
zH7enGY365>n<r0on&pacV1(Gvqx*ZH-j87;_d4Np3`}1-&K(K9-)%L1<1B9#DsV|p
z=Qu2(GPQhVO2YB=@O{q+PZ+UDp%2A(53F^ZWS}_6llFB$U3m)NM~W}I$XBf5{MC%>
zmc`Zsx|T~<jwWrvodL&b5k}GlOw=Iqo8%ea`$)|`R|K5`iiCGX93rlkc+dd!3le3|
zvE)gqZ$FaIJ{J_8rLlB*b?v&Ul3J1A*Iw11eYZY;jYE@gY0-3Y-?okX@@!y!t$m0|
zS~xxQ_8ng<ll!37f%`Z*<Zz$nXNOat{LcmC=)jG|6nd4yHWQYzmYW1VzimJY$blr^
z8DL>4GYJrx+bYmP8W+$mw#Q5SA)$TGJ6#|fs~qBEq6FcdZj+zQ7Wg(HK5NiEe1Q;^
zS!9qqVSzDI$1p`(o1&?XCFUMarw?#moh15kZev$6!V2UI=et{1TNV0{gerr{qTE$T
zlfqu9J+D!gg3KFF5_8PHnpY=<GDO}~c4hq)iu*p+8mTX)nv%tfnc5PQpSk#1$l{Pn
zIl)}D@zMNA$DUwsy)qC53s4vIs^3uCN{M>L$~4VNlfzwo-H7`8!EU#RZ$^IxS^&8W
zGHYBLu97H4yI8NExo+s`p2X64<xIlH`1%hNE7u?vY}77FP7hrrUHuep3giTrjwFT>
zEz^d&CH^&NM1qPaB|A2lAb_V_Lcb!JYLRlJ3F-P-INQKngpO1=y6r9}Ygo&~v{~hS
zpH;kQL`+e;2c_Cj5pXKq__o=P1aR}Ku_C2#Z1<-whg3h+wk$=ZBDeiBYsd_dLRI_Z
z#GscpO{dTuYlU4x;tj6mAYnFUjIab{efAH1L&x;$!OU#v{LGo{kYWnNEX6rlZ|R&0
zE2}SsB1R65EO5M0_C0+S@d8$@4xCrNUX6v~n%xVOk@{?pVD2Xku|=C;SgAiJfK37|
z<Bi&H(TBF%a{4AYo1PflNa9hv;kzx3k}!SA7;&k36t^(-y}2ncalQe}sl&Ld5vk_-
zvRH#~-ZIa`6!dio-Ku?q4M+>Y${ox>O}KV%1#lr5{cuR~ow>T=uO%-cKbkK@PLQ{O
zK<7bNvABw6VyzCQLG5ce+5s$S`4N@$&GnfrX|1Vp^MrMu8J-$1tHN05qPfMIRBT(*
zznv=P?lWY?V278d?OHBlB@(Jj;r6P|P!A@_w0^ZdrFsvBt|6a3s8M(`ybts&p;IZ+
z+Avrk3Xzs9@JAk|u|_Jl=HZ<?Ux#+oQ+uK|w~(YrOT)BqNM|xkDh7X?56}rVu9kR`
zV5LT;iP|(ImMxx0cD(xjNNr{%j%ClZlKEiL@Znkjyh(+2^|2q`#K}mceMIz@j%~T@
z@c~r?B?y=X#w)5P5r6laO}TQN=Y5+J{ALe(<ihJH%r0N7v{IE)1?o`q*C%0B&U_7(
zroli!CaWe`u<NG=>w1ftDi8WX*_lJ8SVOYi@>`n^k6Wx~a6%t9)9dhgNYtX6et{G^
zM}8fJ&3U~CZMDQGap}E_epJl*9=j|D#r)N)RPf?dY*BLt_w~UeQlNSQdb{?5!r;)v
zYgk_mpLbv`G_^8MSu0B{K7Ok<$^5C7_P5J*(=J@J&TXUe9r6moyt7T_PDhp^*TsZ*
zc72L8*4ofaSbzkIES_D`E%Ojaog4Q+Nr|0vo+vv26gk47Z>180kzX>|G*nmXnzGO1
zUYleY_u)W#m`S;saVKoN9(;He<GMM;%*!%V(R!B`yFyD>gMw6PWQR@24eR@Q&4i!#
z>g=66NKq=)9#Qp5ajj=cB<|iKif@<JUq-3EufihwOJtT=ugpU1X6V$RvfANBV2D8|
zM;e^oo9V?68P9sPsI|&8M1T_>E4uwBLGG4=af?_mpjx7`c4vqs2K)}%pqhDrUlmqe
zK3i4#oV(a@BJ~TvJX2AWs2d|+4tPhGj@e1Bl|AE&H$sAJtmO)>?0rmT*IPvaj!(<y
zk7!|S770ohj8@c^cs1^5=KJ;E0-_tJt*ggpd)hY{g&lHx(*>I#mTfPme_3qeX7{2)
zRMT+w)_hT4BSQ=l9D_!eC==2AFUj&->ISftOul4#Q^oyNn^z7;d~Y86^3oee`wY`-
zIc%ewK-Z@T;bHBf-6l1sfc4V84qC;ZHYy~FJ5tLj>g_7WR^#-URWlNG3EDL@#j*Q!
z3$w{EEc$0tU;QB)E<hVV!=>duHM$xSJ_mbR0Pc=d=lJ606lc@q4Yyrl(y{|-F^Qb&
z_FNYTz?{Q6S0&~fNnoUbiA#oo)Ne<oi^qmTEgk$tuKq!c{BCe&*wsbF){kjdCv#Y#
zPECBb%uk%)Iw($crKc!ZcU|s-yt{wfRZzIFf-=jr|A-1fFV2Xl5_yJ{?=eLgy>KyQ
zKM^hf^9^zUv#gC5Vbm%Nme>%xg>Si@5VvkXH51R#hB*b_fT^}`6ry;5!zuA$A)IeR
z@q8-cWJSK#m`vQmUi_|!WPHdR+UMM;IPV@rthD!UiC&BNH;K(JapIKZ9Z#;-CJ`Q@
zQj(pE>jX37M1wiD!=1?W^aMlL!pKHe4O#At;YC)5-CF5pvYX1&NZ^KG$U3zQ0K+mM
z_#;ysr`Ar58gajvl-$gkQE(tL0aGw|qR7%d=2L3I+t(D1AvJ@Dhw8PoNP;iBCQ3mX
z6LvcZ95FU1aXWd3mf4XEAS49EaU_W5#;+e4miIC~c}B8pZ8@%`J}jaW4J0mfKVSFG
za;8q+Ht*6m%^_8^d-qo#UBA_zDQ?s_3!39Z^{)T5E90Hgw3@ux2@4|+gFwH<$~{Es
zzCHYvP!sOMi2p-h9a<VZ^YQdg^d2{mc0V`B`>VV7YXYf-x}nZ);l4u;{iGw^@Pej1
zkBr>CjXj0w*B0ZZr#augtrV!R2!VQMWpajRjDh6zqJhii*U?TU30YCb<q7gmu!17;
zWys)KZ!%RAy{_5_i<;u4;7<z!979+v3|OX|ewVe41n<6iJ1z3EfBgVm@1BLq&1ZKs
zu__7VfeYn&Og@P*=&F{KV0rR8HsZwN&JC)ZI+TvnSM^dIo89L5K|6(77M~p($|F)O
zWA#yHcDoV7d|U0RJh(8?)3GLa_hQ1DxN*!45kKFQ2w-lwSpX3UcyIz)N<6K^N6;5L
z@$t~edT^Me@t5k*K10(G8<9s4>2a`c8g=q-ZZ1L-Oxz4r@M~<6G)ZyNZz9P1<zx$$
z&CEPmUC=ELF&}fTb7s_|>5&91cIFDoQ5Uf^GRfDj=Fo~na~v+s)HP(zJ7Zc5s3xa!
zvZlCtl3<r{*HW8quUE%|Os{~^;+qna)3%mj;sR+jzo21WuHXZej@#w9`Pd-$1nL!k
zZHA#>GAi<@bPJ{2P9$@C^MZ{_$NcYa7P)W3=re*pNiGj#<o-&jEGl5u*IRAMXjGg^
z2oE2-K@eD^va_4GPr!~lg2YV=lzkO#GA(KSF#|c%T!W-&EuB@cvG3l7jOlLWtd5Mg
z?ZZ`$m5*AsJ!D^~`%TcJAfaC~9Nx$Esca<4r>LZb51Cg&HmEcYDyHa2P=eSS9Kvj(
zu~mGHpI^V1t<70|BF_l1k5~x=!qNbADKoLQiJbiV-ZYeeHQ%G_#kN2+%C~mF?BI9d
z?zyXQj=?DetP)SQ_#Gg-+NPR}$X<Cat%YQZYb52y(l|KzS;Q%@vC=f?drm)qk>Ptv
zwQ=!<3$&D&+V<Yj5oZ5d%-~v5fW!>ik^0GMeNkf5{Nj8ayP3AqnFZN$g9@a*WU|%X
z>^*7q#8IuRMRRqER?T>(%JJPrO=Di3?bcdPEgkO-C$V4KteNv{{Ar~^SMuVi<>10S
zA5MmJZsTKqz6k;JvT9*l2b%;9Qtd(Xj5h;F;POtBws0h6^T#AJnzWcaq$z6qHN*sZ
z3P<zKXnCLCE?)zb7twk;8NR#1QwiU0skVTh$iZn%69eJFN(<{zWxm5}M>;G?2xCh>
zZxPDIwP&!|Q(oL~uMWB9%J-#ki;H)(FW<I}jw=fY!A5xbE-$UqXe39{1n7?D&c?Tq
zG3=|-8|{aR1hW)VP15!I{4&j>k$bD_xb}IGQ=!0Xlv{3Y=AZy@c2|-{ldx$~?c?Mh
zrP^}xJc984;0@0)V};C(^-}0(in}z3a*rr7p1g6Q&@#}pjt2WSh{3lhURz7X-JgVL
z5M=yeC&%K42HgY0W<p8e0Y*&=T|lnA9IZhdfJbk<keJ6E#Y{g9b##;b1|A(+HgEET
zxw87JGYfL@d|y`%VC76tZ~-^;?U$Yq0z)E)o0uMt3DdF@!Oc>e&ni*s%Dkq!pxX21
zx-X3Z+|4Z!6z6-0KX^6bc7|I546N*1?e3l&WGklu$zdQ})WUddDygbH0ArM@T%2vD
zxEpWj=5=iiW@LYOI$^DR2tmPCZ{iAy;E|(cBE_hPhRLR+h#fvzwn9iy^UKa4fOh=%
zRDl`bIQMeP=T<E4>RqXj;bh%MTmX48p&bo=9FQO}_rQ@#enoX^=5ryMV+_m}5sO-L
zX$I<rb*Tn+mKeO64=etHQj9%uSix&4TA<mVLC@do=V_%y*@L6|O&~1tP4|N{q|mLK
za=DhG{f1G*jCEM*SkSnw9zQfoeMPJ&sWj&ZPSokO?R0H&j%COF<Dn%;@>y-nTgAz=
zyi$)-L#UePb31hsD=?q_7k2#jKNkESo7Y9^|Kf)KH+CHQrQd(cZ#Ej{@0-hiv)j-A
zj{b|S{u}<yj{lu4|I_|I=}><*umAa@fWMr7;&lH>_`|OMbMyNCf8u}G@xRskga5&X
z|F`k~-Yk#wkM@`U!K4Vk*>Pypm#gT!2|R|jF5F~M31f&3P*aJKjN!m&p9DGW7YCyn
zy~5^Wf1}dsXiBE{K}S{Z;ikC5yvnSnqscvGShjXIMtWL0F^;*zcDmXV`qySLpy1`b
zs%Y&+hK=BJ?OXP+b;tM!QLf?TfwwF?8D1YRw^)ICSsjI}O|2q|IU=t{D)4o}aU}hA
z20)n@<sJG=lUoIP>YDr3%|}*rEZU2pC}ZDIYc}#I`;WcNd4x<535g(0;G0&x4tN@n
z|Ehd17R`GfvchQn6PHqli&FdISCsF=eDzhA8Z8ek^DF^s=5t1i2J4z+%e~SG#;mwS
z`&N7A;a<0P!ER`rdSgfO{wh?94i^oZI!pQ+$c4cDGR{px-;C3OrRju2<r+?W=gULt
z*(&tMD;DRIob1<vZaibUPFBPhd9iLIiX^HZuw`l$snPRVIg<^L!U@hyL?6!dH-I+e
z_9w!Qd(SFzK==#ftUG6!<kAJgX|PMJfO=2Ax4=Y*j~tSTsBbJgghEA@!7IwGX3`6G
zr(7a!cNac8whdESWlt_g(?|T%zC?VZg_ySUCx%)7X(o-#o>-d!_m4}76K*Y*@fD*5
z)K6|7Z&d`9sf?)b4@6DUBo4s}?p}lo+Gx^7&fHpPRN-5Oy)K`qXI74$%CPOy-g7oZ
zFozx`0WYEMet&~9!@m`VYE*i8e6>7*S0=`VjR6xH%6Q;ANE%yjy&!A+1410yoJZ{E
z;e;x*%kbOuj&(QN=En3dz~d5Mjwk&A1(Jt-qxyGa!A`VRM~|aR0CUf<;0|v;8;dpL
zH@nvZx{rp<JX&4W00wJx!lR6(QiX8<*UbU9?KR`;S3<!Zt~Kv<Zi<c+)g<r=aR&nj
zCPlxX{tC42WpHfH%nETgf_#O!<7GImkkN)<$Xv{Hh#>!pW_;tKO)F87h3?IaY8hgw
zY{~RW4+U>i(zL*Yvp+E&1x4cr)^wg3#g@DC>7&|AbjrK%p?5+23pqJk5+URP@M(N8
z7L(%*V|v)bsNbD|tiu_jx)sP4xaJE7$B6+?hjv@?S#-DtT&h&sUF6YwXFprXSw7qp
zi{z|s=9JeX1mR(UW>zxbVy$`b3V>m#NvUd!^?AZ3JnzB8ECLf7>D^IF{i_f8H`~ja
zdzvalY$nZ3paE^;g|A6mp&@~HuUb<GFk(rwIXQwBjpu6ch%Mj{&BC=DX<hqWN+Yjv
zyGY|>zT$wcL5mn)1#^z*)i7C@8a7gO)jvDxE<t~@)>sq-&{vPhQzPxDO@;d2Wt*wX
zbyp`ES(_iyFD8&MsInJ_;qs4)yq9B^zf)HTYv2zGTyhFP<BEniwpQlH<wcDPhDF2P
zMZlnP^0z8~`6X4(Wj^kKvl5lr0!RRWRa@!fF)(mhTQQOTA7O7D6xZ{udk1&-!QI_G
z5Oi<`cXtWy1PJc#?hNkk?(XguTmmE{m*4lCQ^)Rm-~MB&dUnt5soJ}`daciTp6-QN
z@?HH|qw>yd)j6y8a+-I0HX$%j*a6CO&`lZ7dRAOkqTHo+zM*995asm$0Z%OY`^n7S
zoZEA&e>(jV4?pK@-!-87gLVlV_(%oaDNtw2%xL9Oo+?f=P2b?!o*>_`a@`W8Ovk;-
zi#Ha7&t0G1L-kw?Z5K-x_Yf1F!*1fIrwk4qRRGc~mru>4SnLe77vC21QfK(-uat%{
zdBJVXQ>%G0Nl%NJxexq3RD-*druu!=UzGNjW=GT!=Lbn0#@*PJTsTo?8YLr4SVw3S
z<wLD+ZXFJ1=I5g&Wcz6uhC^ALY-&o!lLS~2gvk}UyAcbmC6cpTu<7#T?E!99mWAPT
zuWLLLEImx)g5_~N!$o8X45Q?s71dN=7C#<v=46Hw7IY~?(@T@tR=tbfs5Q*;GnHAt
zVs>N~5Wj^@R?%Nr??-Aoaap~+70aX5RS<qHXHaTz5Lsi2)R^gBK;_uycigehvb%)#
zx^lU`U^+xOfJIo3t#&gPgW{TPNP>y}u$OIe{A`>Hu6oG<C01?(C*LBMK=R&H_)vjl
z2?(S?2p^FgoC!ZW@RW&BA(Ozu74stC&ZKo&qQq$JaMcu{-s*)-uW<X=^v+dZZSdxF
z{zor7;svv8$qfP)zd8jTELHw63!JI+a(8WNg`msNK*E7W_2?KzI%7<V^e-we_*Iiw
z=Akfieq!T6l=pL!m2JACckypvTiZ9eRi-v2>LITaw}gvrtRvr^OklnDVs<A?w*sbw
z7ITD-7L>XI(r6{4*SukMt@$hy_vs_p5cu`cpR+VW$90ky+Gvt`+&Vzp!Nt2VR^OrT
z)x#A;%<aJAlZ1)bLy<hHp0%LKuyR_>kRWIaj6Bp3#1bgsIIFOfP2S$F=UoN>jIw@R
z&`QK1pFJUbb-6n|pt2Sr<h&ZK$1tgOV1;6k^>QVJvPk9{v(+|C&0&YpT8@(y?e?Bt
zq*^nj6uAR<&pg@IA03R>M@bdR2ob#HlpEJBIBC?YS57F-pox%UM9qi5RXSBZHgDjN
zd8VA7xquN&HKp~76a+KgXF$lvPfhG{a5xH!@-MYN3A7s@2omoDh&-!P$*f1$CW%8z
zm3VNEe^?j3n=e`En6x~;4j-u~U_qM6>Cj2hK(@ZjM7aM19TX_G=r^+0ZR1K%<Ec{^
zPq2ucQyMG4hk#bbB)}5Z!+D*WbKfzqc1v@<NG>a`(b28SrxoSF*k?y_d$Lzu1F|Z!
z2S19`>}7JCcWhPB1|9u{4IP0y27fND_&GL)Zhrl7W5^KIplE0Bl=66Tu4)#HGlYZq
zgkRmP<tZFUk0_^qu9FBspF(DZNiAnkjqIrR3vJH^!ar^Nes;{Kxkhhse?0FS_34w@
zVD$mXZiLVDvgUiMglLhCVQ1wpg)q|WZ)svVrBkm;QlZ?!c46Ea`ErCib{Exp7N%!v
zgVpNR(@vTQ(xbC7M>!ZsW<#01vH5{2JpNefr0A9iQahPFD4-;q<4(Ix3Tr`ItbL^m
zRXD*}jgiFefdpzhEe9^dtqG|<bIGnIg6wOtN%kLz+5>#BoxwU*49A}<1=`6|zC<e!
zAR=-tGjx36*>Vc;7QnI4ENMJUt@Q~|111AxdJeLu{0ijJ5iH2nEv&+y7+v;v_d-*J
zV{Y)@VZmpaA;n#h!Xa7AT1E}c^<jG53?uw0^me6AB~2SIudEEhUmhvfZOUJDIsw@A
zKD@eJBbBn87^e-_<r0fVtN=}u$LI1t9F-&hKYWh7PF?M#CRb2`8f-joq+h+jj2veo
zgF7yOQC5ndW7ac?!mH7d2n%QEU@lUkXfC*{u+dx(Ap-;#7v}{NA~H`bW8X=a^=4*O
zf&Dc<VBpRuYqy%=M~4mn0kE#tY|FV8$0P@Ho*D(WB8wf!L%GrR#N~JiB@qD|x0`|5
z+Yyv&?iQj%B235dpe^3laTN<LMV5~o0Gn8Gn5^IY(a}XqdV?&3je~<g<*VTeAF%ee
zRNz&t#926rT8@s8Bv+eT%duD5;o4k9o9`JWxAFRpmT=Cwifg)iwSkWPsIa=so;&_f
zF+84q6g@dCOE>18?sqVy7W(bBL3)i!_z(D-j#<r>#=2RaZnh>dVXlvPfsm-q4s~3F
zw|p6KU*#&{_&W8tkh!~^+K60HJO@eZM_}rWkJ8gt>c6Rl=Bo=(Y122L8~pM3SA_cd
z?vsVPelwfTikdA?D%y#m)amfo9fj0M1>2>?nJ(Fc1mKRuA5oYJU4VN0wVioNZ@dQh
z(KGcbWT(~KGLV29_kOHl$>AbK_L(uwVCaN_*}*+_Q;k!F<%WXf-3VRnJ1zt7(`Y=+
zoY)bQLP{r{YnR)yM}WN1Qg16CNrsR98n89OyVBB)rnJ_OmeKzvgF%f4r^-Na=e=s2
zP++0W`FRy3sbF=@is?YqvDevbQ6>z`?ZH=ptqD_6t&l?sTBzM=A0DnsD_KUpkmyMm
zXO7Z`&uGzupQ|*X#b8=BY_=s8N!ih1!ZMEY(BuT{kYftYVij(%+%P+hP2`!<$iwWH
zewj-JoN$<``n5baHPI=j89mpCNgPr_fzMTe=JXR4kf^MmWEH)XCma-FMzKfW-ikmC
zO*!W#g0hL$jTI&!`4AsF>*;2-F2yp|u%Ohifs=m6nu||=_^Oxb!iiw4%&YbEOSUlL
zA;E__5XY9kyMu1}Ce|QXeIi?-yR_8+f+VOvXh-BiTVaadzd=#$;!X;XMzsyU%er4w
z^Ct2)S&4SsI=(5o^gw}^-6HM3#|f)DDc$>89N1qL{CGp@Ug*^~Irt$xYCcIiv5EW!
zSYF{~)3()&32hA;RT5s0X7)d^{}O=P!=}UihO{tQ*5xPOQ-kt!?)nG(no<(vrbRs5
zXp5ctwPB%2rr_1wCBPz2@%*B#VhWjGtebHF8V`zLV+)B&iE29;OKZKAmp~Pp_v$pA
z${*i?DJ2j#mgOK<A9{^@<)=%iij{(K4`0tHt<|o89Ixz`5vk^^h$PmonaxP>extBZ
zu70g|QLHUIltBO5i`0%fnx&7}A%;s4_V9<7jbaY{owKvuTmKlOb(Sv7aMH#WF#fW9
zetBK(;?C2H8DYu!*Vyj%MB3Kh$!usm5eMQoub&TRTfAS4L5&6GsJv3m+cR-XMRm2g
zT6=jekeCBf8@c*B8O#-!JlHYaA@4{6m5k$SCm@T+$Pzip0S)*EI7Fs2W-mdn^>#{L
zoppjILLYw!V7&bdgW9N}hMHJ_^3b^0KFx@PHHU1BLl$>%<Z4ImXK+or{&03^Ljw<|
z3!k9h4OvG=j$7-X@tReUsLJ|SfqHTK_ur~!3NFu^Z`%0ZE%@1=bhL#aAo*&s8L`!%
z0@RYDGE*~`%Yi8z=1VhxbT}2WkKh(^qOdbtLb;^qSrcq_CQ;Co424P7{_9Fxixeir
zn$_l{k`Em<G5+kyH2Xp58czaC<F9nIeP)^Pz)_4SVi|q*@VLKZJC|z7Qy~ww`AF6v
z@{Z2i=P5(w7A7%uj41yx(2$A<nB9+*-i7suuW|T2u7RlS(1bAe*Ulnx4S?hBp49l=
zjr$zrr#-Tj_ESfZBa>~^6Q-Ez_{QOhv8z|?kbO%vW)vyYg*Rru>63{r)aQBjKlHOs
z(Ue!dog*?D+jS8Y2U%PV)%KZ5QUgPq%&v5M17=2PAdt@L^u9`e49_<O;wGH}rO++-
ztuo7)=7Kod$P|PnU)TP1tu*{@h`Yo)Lj&UuS|&R3CT5byA9mzoP~xG?W_8LI#|NzD
zn_MINYu0#7ZTEE=j%GvAYJLc=ZRVA@IGbm-qz`sF(aB@?F55i9=I=0Ff*skq60!1~
zbv%YVt9moNB1yqneCbf3t|e+ld2x%S3cL?;<Iy3a1a8CXC2O37-SeGr)&$Un8aLhY
zqfzAiZi$4@G67MMf827K*v%k193{OOOlkvpOyy}LkMc}T^nl^3+6o2!(Pb_H^HKV5
z9UMTBFQa8i(B1)|u~msIjG>MD1zpLRNU<_1(aT&H?C=q|)HO9WxVnHPmNVHSU4{Lh
zek*a`gInv7GLsY>9Ifnjw%%hJsSW=?!1L9r(IR2Ojs*p{F4Jc>jMQq6bl2Q5tlO`V
zb*Rsn!S0qM(N`SU;;bV^g_-32*4w!eL8DmEKYwuFMD_n>GtH}hg&&>TXuRx{75gq-
z8jV-qdj*zmV5*|69bhFo1v<ypfVZ38wkKS62`9Qsl;bE=w_~rh5B&c?3D4(ZZNJcd
zek&8OWa1z{Z7}rtshFu(y{?y0LyzRx;F6ODa4`e1uN@~m>)FcpkZvnwaWz6!J8U|;
z?J{1wwr25oD(wzz8{Or7vGX2B%D>6+CTz;3YPZhdeAai=?oYXNjNs^&t%;z@G!ooY
zP<Z^I6PWtWqjx6vIN(xRDQG}9cCFi3=8pYpQ<Sp;umV!S@D7z)LBJCiu&DMGp3T;f
zn4zU<OeBl}4^_M~86&@c%_!N{QZMXub>+u1ciecLeR$JcpWScORvi!k{Q7mCC0P>3
zJ=ZB)*Wl2)Tk7X^Q497{-Tpbmt#VuWxczWQyeju6Uy>6CcatM{qiV*F$HOBoRM)Ss
zaAkbqk4~D}UlyOQbN8ND7K=D+qF;yUtBk?n7L*ms_H(7`FJO+(<RzLoAT#$OE-Cn8
zkT$|(nuyO32VthTzaEuiEr_JZcO{w%3A?CiZ7$yvD9d~zDV51XU!*5d@u@>WWsyQ0
z*B1htuWD(Mm{=0tSX_K@4YPCGz}9=EmhZBLYK6o6cmW5F{i3p9h+>0WwE`iQ3_YGR
zTrpWpa@)<@jwW+H2{iKSo6s-q0YUbRtx%4Re#h}?k2K*URst!NR%;2hc8k%p8aFkG
zNuqV^xs_$=VLUvxZw$Oc8eN|s!Vfg6-YjK<Q(F06Y*TQ>m8|4@SJ9tQlIu_0pWP60
zPSAqw&*ax_+4v{M@(Y3x)5Kx@ANJ~3&WLz|AAdqvs{V`t7O>cLVa-OseR}_<q(2`O
zLMa&A*6*2ZeP<-rh0vs9%t?nm*Drb50*SjO!l)5~_?RSQn;vJw(ID|hg}*d`OhG48
zz@_G%O74!=eYp&E6TuZT57-1>AnrrY;dgXAmM`n7-Bjj6rd>6F$(dNFA$%LOUusV#
zuI1zjT@sSPOkHA#$%~eC0c2D%0d8*A^Ez4{0=j$zd+O;7z<DIu>=~Fy%&;;f!&)B2
z)C`$f<j_+pW(=*|Z0w@`2i8XZfjAT}_uJ@Wl+g6rxe2i8!y!b!<Q}M#>BU*?=8w`u
z0Ayytm-S~MVHI=32;s;*Pa!eiZ_Nda!}QAgs){9%3!9AX?svj0R$`;I+vb_?hXFOJ
z%CZUcYSxQU<}94>pRJ8-STV2Dz>g3le6?&m*ur5DkiT0;VzqGidIfp}mel22E{=g*
z{5awlr3y<89jwkoSXwghWReL564}are`gv)8(zSDyTvCterX&$=ZVgq-Y#EUN(I+*
zo#sd+r+OfsaB9O=m(tdSlwR6?<4FLo4Ck6g%?-BKIS}jM`WH7clb;Mi`Uc~1ritfY
zNhxprn07N|9$noc#kcA`%=Ot$1ed4Z6!829Ps-}WkKntEF0`Y{mwh;_(-^m5X2v83
zLL8_eRQMIe;@J6X`(cs``Zj#>&|uxcqmQ%Jrh4H*is0+QvGU}SkQ+n`#{*geFn6xm
z2K8daV1OnQ+q4b0@$#mLE`+%#3E3BK9XqWr_SO+D8W|j=usM{~!uH+?Y8fkbI$fa`
z9WE<f?#L}56)ZKVrjcDBbINBE^x;wb`9wxNS_KZLilK^pemTN?tS=7H?^uiFY!w{w
zteCd6OL&Te3kn*=5XqF7#YLlJX2}~6^*2_lPFxIHDs35X$Qic^WT~uj0z8nyy75$K
zkX3Xz0<IwfI(X$OEW;nxg7Hu-sn`di<e^H#%p6DCD#^NB@TJ+++s#%F+JiAX8+4Bw
z#2bt~A)q?T1)pD=(Vf<5nLC)$4bi?{7+F%(Ff~+<KeY2S1QC9y2G)1D88hrOQW$7(
z*;dGY2E3%GAhE(hknz<dYq`N<LQUo|Uguo0hIp;&D|V!*GpJo_t7hV%eI6fZyd%)o
zq8ZE9Hk-tcs!S{*ec1)}4dwQ%JT`sNnfjKH50UriV6oHw>tIfztOE?iHh7+H;)eXx
zC~(>OTl1-{P&j5q^YV{&#6?DE9GK68!n7x0OOShE*C0vYoreJmGmY>O%qE1>xz+Ut
zYQvkdWp>D3hX@;$2Nu`ZS)=PP;}zh8dH+S|Ycf`hmJD4<w}rUb1+`-7sLk``FMU}>
z#qYR$-1vk!-lJ3@WNMcE7+3;a;<OC>r=&yX2q7Wx+zWcP8piy{ccY1^B8dGe6Kd9p
zZ$)J(O$qiJbrFCGP(i8&b+%Q?G#+!P)71~XS<dESnBKxHjC5v<iSo$kJ!-C;d`t&X
z&qQh4gVrqJWo(ZN2dV;l-%9UnZqG$J1ek=zih^*J@FpDJb7md;aAxPy-OB3D_z4LB
z-IU@mdO;8}cT_=_P!r=8)tof-YR@^BtF(;U5`35E@qA5o<N6p2b17D$4q3iS(|7qy
ztW7;d>%&4!-4DG`Wj727P0~lnZ-C8b^lHc=<NVjmzQtQ^1)0M5y||H6Y3PCDBxu%9
z2>PEBMF2ZIbB=Mi`qBYcF-MV1{(@?U8lyv4i|XXaej(lWaTW7K00^G=4I5sG6b=HP
zuf{dp9_B+SvlyJAr6kyEsnb&vRYDMjI2ZWm<UGvw=Aj8w>MZyfkH!uIAKw{6*Ca4f
zH&20GiVeBMOR}j^-cRGD`g_5*`fMT>-FaKqS&bDuW2;jZhA50>UK~4=>_21Im9zuL
zX;R*^q)Pq<s#yG`;yBx|jI2Y>l^D&fszOTYM{P{tu;?rHQnd`<Ko_);7(5Z9S*EEw
zZI}5k0T>dHAGRt?)iy|RH0Y}_;_Z!xTM?-(nbp;#8RgTxE+_)w#ilql>XN~+(V1(<
z^<~f#IC~4`3PI}?p6zapfgs#%2SwK_1%%cnjz}{?E#5P=uW;yVRazw4bW`<<OE!QN
z<9?9348W7XVxYhq^1Eg8ReH@JwOtyii*bD`g;e55Jz>Jc@>m`vEsHs#6XU}HyKoC5
zXN^6Vz)Ms8xxPjmD#uBY90fRA-ollvB?#Q;A)efol@YCVBSp@yI%h69O>1@Ga=d1C
z<NNnB(>g4C0g}6mgMrqhUf53|$85)KdyZ|PjpLUnEveM14kgi2f017Cns>p5X+%f7
ztrmaF=f%V>oj;h`skJ1pmPrObiUm+(*GZiBR-WD>vzva*#1ePw)s%QUXVgxW3|w(h
z07EFOLW1MGgf-mp6ouimibDB^|MJz8>?U6>UK2)NZY^AB_sl}zek}PH$W(Btkqe$t
z7f#><n>`@mXD-fj3_$x#;vR>rv^K@r_>1%>MK7Tsy`_!foQ2OyiCGPGSag2}HBETF
zI-&fX0^;#=rj_fzqdA|7rmWBY#dxhm>SJpkC9BQbxP0)qj!IETE6(L<$*eMayn#+t
zh%IZ`6RM&qyM}Yg841)B9m9^*^^DR8=rU{Tc{T|4v;|2vY<DZoem#?lJ*GdAn2WRs
zCyN#(_ESMW4&TTYWLs;z^xGM{Qi*Rro>+U!7X8)VI+1rhC`Da~?_^I$tyY2k%`ZPf
zwZ<mAC^crP?O2!Ej&VwLc!@H0&eo1%hmxhk3eGmvPr;_#x~9BnVO3MPv)^y-=QZCA
z8S`&Cr@?Mp?}&p-!LwughW&Pf&<1N~L76a1`q^ADRXXzK$<BUR&4;2VEr{|0;`+?0
zsLF_SZzjDOaOs|+1E+yt%I3}n!%oN2N+TV4VN-4_B4uj=HaAFam^mFXKX{^v9k$c_
z<dG~~IrHWVT3f>9QO)*zpE^$!2QCOD8T7Co!>Y@)$`O>9n&4^}#Lmo2guI60Lf~>q
z{AO&HTm-U>E>&q%*VB?y<d$R0f;qQlbDfbGC|~YLoHu|&;V_nTUKw;IlsI2cpo#9#
z(e@@jJ=L_%U<~1jS#xVH*3c<#_+`tuKzXPz7+PzC-DGcf%js7OLBv<}BvS|1eDM3o
zG%9UXWLkmaEOG-u0fJ3g_T>Py@7d$gCYQ9Gkwz~`G0q{RG3-*zjl|GhBO=;V3QEA8
z{9}-9mED`Wg*fs3dS%w^tv(>PCSrKSCig(4M{K#~?|S+bpU<>PUz&>hE)r!rWHI7p
ziw=XIEzDjA6CX=)(v*r^+PWR*!}Hie$trkHc|<`d`BN<Z@~b;yi6{9*<}(J@%5<IC
z4E&v(`6DdBgUk0sRpj4xb6Mh7VKql*L3fXJ{DtG&m|$*!n}8_mtH~VTj(SPYfr_Rc
zjbUA$rhu6<-$^-mtbWuKzyzk2q6zZ}0P)y+Qi>O0jTXgWN~|Ab_#b3qTZ3{37i>rz
zb#o@rJMgYyN_&-KQ2865mo}KAHRjhvh6Qd5fV<2Q$XYu`W-+waE!wQx9?PR)6sjIR
zFRz%RxTUhiTA9wxx-w%O3~r8;rs{Osx86H3=BOwq5s-rHq*uejc$2AC_lV=(bRY)b
zt_ZY5nu{u2ZNy^eP`3~}<Ih~ixfSq6g^`jn%egeSHiSf6;xDHuZl-EkT=YxQ(ta?>
z(ec8XC`1^v!E|e6{440hYbN^Tjnd>C)R+R`aJM^l(>zo(Vhroc;bER4X@da}X<nU_
zwkjXA&5wdid>=naMXF0A*v|hCjQ+uv|6ds0_}?+~{|%%6gRy`8CuaT!bN>(Q{l7fk
zG5-7b|ApEA2Ns8f`OoqH%=ceJ-~Ynw5HJ6E{h#^&_3`iRpFZAy^!^RwLt_8;{(lv*
z|8xJl|9<_y&QF5+4@Rfpt9tB2p<&2>Ud6H2R79ag#T%|*Py>HRBB1uUQ7B+h6Z-Ar
zeSHr9BxCb!QhEr*QJKt#^t;Jb2EXZ<N{Nbnq?{Vf*rTQ|UO9Q*5kymdzWrq^LkmAt
z_5u0s<eu6kvHx`Vd*&Bo#aO^j7^4_-9F5NxPvS|q!+0hFRpk`I-OlwwJv4~8^GCQ?
zm2h*&1#uWAVPcK8L0(vc%WM&PhW|G5s&D#4AJ!?iU(hkE)wp?(7j7_Tnkj*Yoz6*;
zKz)PDP4hbOeBuV6D(?WKjGuXD@@l`6*$88~OThD=F=YB`7-faXH+12-xn@ripaxh|
z+OjX@s6e{eE>gjb(6vnyGovk4&;4@}%RH;L%*%9RJ{H3fx^kknmLcG4^A2%n7+-^|
zp=#85dw$iz<I~`zvmvTEs!vVOXZ$&#RXlIu&dXiP+@6pcyz^wsd(XYMs>FYzz0WRq
zI0L1iPX3qAbRstprGsV<`B%t!O}y}eE1S{E1lZdKxy1b`vRqw4mM2L>$F5wSw?4fq
z@UGt-2pcOd)feL1dhmsui*B$ys}xqAn=HXCB$XAYxW#OHYGh%IwG$4}lrJ>=d&Ja~
zN<kq1NUDIwk*9#5S>d}kJ*g^{!GcP=&*#SX_-Qg6jWkQ#Eq&MvnUc=FnVm8jYprUO
zxQ|t{n#-0OKqwac>7fGL&5xR~{91VcOe|G{EJU-rHxLX@WB4XDpDT^Kz@`#Zmk>g(
znJxQXpY?7?Qv7Yy%=22>Ygu8fzI&3l$xaWqsCHIr@e$4x&gPrcAq~d1>4BlIZksod
zYliJOQZwrzxAISSEYUUEcKd5D_z$7N>F?v@U&j@LCvr&DE9ESPA%uN;_{954oI%}L
zK}FQb1mSiJ;5e#et)b;gkZSVjOGIeKN>HKdM5maIXA2t3&ELr^toDs0bo-#`%2$&l
z%l4Fw5nzqLY^1$vLN5mFA~@Xpx5n{8g)CRr%A)bh){rQ2wk#Hn`SM=5gL5N;lRZVF
zbz@Gi{jxDVL8C26`nJKUQ7{2K=4exVjfp9&edAM#nPF38E{KmNVV|C#c`H_v*e&b#
z26ok8FE&)&G)1>-SV>(k0nsl&D*3WrZg4rUoYra#_U*<&KSaAaisr-v!(yg{q;O>|
zz1;rVD9~$${3ug8nsB5N$!2!OTcA{3!61jw5EI&LF%i8KA!$>pBD_oq_9&tp7*F7;
zm&yfJr(P)6!NI0$Q0E^?@kc<Wh&u!px_4|B?%IPcn7L3R-gbO%5A4n;mDLsO)c`bR
z<K?t~o{J8l`*X$6Ic**0(4QJCrRU7=k~z-@dUiE#X>-B3*h4L=<e08u8;*9bEdd$E
zE<)4?<9ix+5a@h0k1};=Hg0Xb80=R$HrLn7ZZZ|0xu{{2h)k@Gf*Du~%zhK#GyP$<
zi9}cja^2bU|7vZm37JyPB?|2?vo<H<!5Ffc&@JYsn05GVz+XLeTp#o3nRnVFk!d-{
zIzHQ<cW2PQ={M?pziIj01l7YE<1_y%&OW_qU@mLP@-Y3P#=~j~oBKP)&lJ@hO-TP_
zE)OETDs9BRQrWIH`JTpLqp7?gS%hz$ERw5G{h@6yreCe0?OHURI<|~vL!oi<J!SC>
zk)eLrJjS}vpv6ztTF<r|1H?-vaW;^y<d^kBQ@TJPYDw_0VG!^`V*=P1Jp;CIaw}Hl
zoVS#$Og0kpY#-!;i`U}}#0V;|SaSKH5Scn6jC;FQ;OvcgkgO+lise5di|A@`O%k9a
z(^3L~M6)F`*~cjx`$?``XpCF7OWohZG*<wbLyI7#=QvIC00yzGu)`Emqp_$I51r~T
z7EU?YO)pGLF3OH#lk_vQm^e%9*H!-r@pP-R>iq#YmXbd5sG%H=NcuB)v&X6r4IvO}
zt)+2vq}2Pwj;Vla;4-&1hE)EQOWJ$Apye@Z9?X^(Zgm6kT}Fa23K4dUsl9d%8aaNu
znZIHjT9SCiDq<C5K>XT#XgKPdZQGn{(X{nOG*^g1wko4d<*Jp<v-2#}Ej#V|-gfgc
z1mNTGd!~Pxwo!FVS+Mzisk71qsV@1Mn+??h^P8(QXu@l&Tyea~L|-Cw^E`i-;6^lV
z&cq;W^`1%^sqQh`zkt1WYr8h>{f1!EucXRl%NPI>M47BTzqE9eKU#$|)m(LbNSU4n
zR1)(GLK3bE7>+zPpPN4uMIGr}d9Y6L6V@yk8z#9tH%k$hWOg<%7hvn5OJ6m-amIYj
za(2{9IFw`K)-uTnXn)G!!BtbO+y0p*R25RUCd>%MFmX#?O4m%!D9=P$&RCagTS`WF
z6jH*%4FuLxk_uf7*+i(Seo6Q=P1TWO4z$={jve=Saw$#BP_etbO)+{_V60}6a^S|y
z@waLiny|E<f2pRJNRIZg+rG4;8;wlfiOz=a&?OuktBroY*u>wqCT`$yMj8V)(W?xb
zqg3z2lI<^JOcg1kA!&Bg*tkqDPsQ=dsw-UN<P3SWqyNEqC|1JP>d3{fj7ees1Fi6D
z@#J_LeUQL3xXKrzH$P0WtPFFZxWd^6QiUMJMjUH+;&Ut5mFcH5L3c0%!}Q%%&Zqo5
zT5%z}w;C=gXvwkWiac%;GCHfG-f#(Vo5JA&lwfgjf4WS$PewM`&!+V-2jDw1SN8-j
z3L>i}IAapinz6W$;IHQOn8vu{#fplNJNU<+g*qKhcv`907s9CIX&zRKj$Q?-2xXni
zb>?gVo1*+4Z6D!PjhZbAA`={hbY$a4PQm*!>c9j+T@}X4{U}$PSLzECR2q#k`pR%K
zf@@bur0XZHe&kTBOvW<w!k%t6UupB3S)abq^xv^^v0f8tcSP^_uZsSOEIv?jN9><n
zOl$QO7SV+4Mq<YEXs+MtESzg2%D-8fH(q)<=w~=|tiphgiaNGjMic=-ah*?gyE9$R
z&QHdoHyTL#R((SID6rdGWkgL;_DFs3MzH2=SW1gWyx;leYC&@;Bk|}MA?Tm`Um4vR
z1?P{|wigQpUs4<>V;`jkock)42$7W~W#_rEf3CNskbsL|@T+)A3>@+3JCiE|F+P1*
zQUY6Z??J>c1YYU)N^E7N_~;vf5!@0=8M;M5>8Jv)zi1r#*d+;|nky?T9^dmh&c^%$
zAMF~lR$`X>8CJkm_>M<CQ79P5TeyX3LjZgE!0%<LZ|6W&%kr|43e33}vLt!QplMeH
zMBToIkNr<=A&P4&N@+>XHODQFoSwD1m}k>j(BAMavswegWMP0S*r9_S6wk82#LK2^
z!#Fj^W+~r`7t;hvF;<tE)WD^}z(Axv8SWIP?8D&Es8E!UV%gzRFVhOyfi{xgQj|J3
zgZ@k$`6JK+raMM&oq7RW_FcDsNg^MN1udVk@~Ym-?L?1FBNrDy<kVI^+P71rsQIte
zz%|_k00|+uj}H*YysV+kU`S$?l%zt*bxb-toDz*hheeNtm3sQ11>1qkya@tqk4gk9
z&L@PYVFEb9^Sbc{IL`L9v1>ktEKYI#ZiTyh!ZZrlnJgI0X^PXicBp_8`{6@upACld
z{=n7(?g4`uk#V!-PPq0<ZXZ0nwQqE`qIPRwV(gh}>&u@rO&uzFDOnL=-sMTuz<b4w
zI-Bu;zz8g<Q8h>(*q>+iv7ch%eAZz6$X>`a@wMZs%r<y=U7OXetd~xA-YIbKsp(GX
z&KMp>?G_1`#q%><x1ejWLsX?CBEwEMBYOB^(i(##B_%k|_N(%bGJ*OGI2@(yU6FJb
zatW5=zai2n@CQDjx67Uu^Ey_S#$i_1f9jB#O7J3^?i(KY-TCt>C|+tDQzoK_<-(ps
zh)d18E4eHI6L@_`2b)}epUM2nz3Ov!aX^EBuD3ENr2{K&78Q|X*rV1fjNsiT1Lb6&
znK5MwDOr$)TSBCE^*vLSQUmKag-)oh{a5vVhaAQcO!w{2=lNcb#xK8`yxv4td!s_h
zH&-}iHt%{Zwej#W>61D&Lpc^lSyl^}+ZSG423NM2+eHchZbd@&m?)~|J137LA*ulJ
z43o(xvrpI*%42l)yq^K|BR;JaYoFIX&5J+$E1SZ_SJ`AK^-`Bua?)J+P*$H>q7PXJ
zKY4KLI*BC$c{b=MR8OmUqn;TIohkBtqVCv8Y8n4H!HVEQ8|pcqszzp4W?pVloq$>F
zFt#Sm$T_-%FA~KB(gJ6Llz0V%vBUOTllCW?y;<$A`qSuLiKt_@WlV7z;u0j1qz>m$
z|M@?ZAV~1JzpSrWPjjx{10kx>vz3R{-?4Z}fy5D-*o8~b$|Tr?FX_9%CZaIq1M0R*
zF)8vYrI58abfbN(l>nku`fzf^z}GcG_&Nu+ei#^Wa>ZsTHr86+)r?CVTU=Lb?DsqR
z5u~62g#!&IVVDx8LRQbEwx26;80xN>l4QM(6ly_J%`FW$_AZiynJ@<7gSi9Tb-{zU
zOf1VZ(Q{ncx6*2uhc~?0AdZ*v$Pw=IC2gAbKq#6Lwb&?S+*kx5+az{C`JEVyGMQJ}
z8U22mvdeC}JsF=BcTA`*f|`Uv0(yTWc_fb00SZ5pI4lE|s$u+l2;X7sf=3;plr@e^
z7)F^I%@m(QMc`Z=;nn&Q4btNMO)L-;9{KJ@M<yjhJ}4s{-%+eXvVuqAsS`tXz7%|?
zvCqu}<7(RU3>Eo30}?Ag1pmjD%&2QVUwc$$$6m`dWAX*0%LsLAtx<kB#sEsMzx-7*
z8F{!_BIjPb3PcEgR6^6^rR`Eu^+m?%2SxdyWjGWJMS(iDoO^(-RX}@9DK$`DJw*4T
zyaGKKA^5;kh~Mo_T*dOR-vvV8_Q#SfsY$*EZISk>H{kew(Zh$AX{|CX4{%$%Yt%Lo
zTi@*x8zvUyIfyg*UgQh@_FW}=$(2NmV$Jg=s<YQR3*H&3ruR3iBcFiYw=m*Q6GL^%
zSEa4Rojar4C+?FGn`V}lj%iTjBYC@oEW+z2@+)7yjlC?Srm$C2+u|_$s2~|-R(zw+
zM&IytNFRR)T^tvlZt0h><I#g~q9yo9fhFq=EsiSWuQ(Z1#z!M7hYC7bdfsducbQxy
zH6y5Fn)X%>MoF#hw`V62f1O34$35p!tB4qFNR)_lA>Z}#@s<KE*C@vqL9x~?FVmD!
z<9X}QYk`HCg-+N@5iZ`=-KD@a=Ihw2fs9xw`XB^XP>Nrht2YC7vV0J_#pCG(p=5E!
zc)E9RfP{_*r9wL0+Z`r2`%&Rlb<qL*Ldt@U4%`iBL;yl>1XUbtDJ1^pv|VpxV4TxF
zSI#T4TTmTNzr*9qEF4tL!AW(*jww_dV<>)@r=~t!v6^WZ&)O}~8ta=uxGM$=ao`i(
zed<dEbBIA1A(c%P<42y2-1;oy&mWGgU1S(>o);-LC|~9)^FWWJ^p~}U16GYfv(F^8
zHTVCOJzJ_I2C%`_R%+&I5F5PLG#yaWT(e<l(7(sUUri!Bs5(D#p)d|bQ2;MBCi0g}
z1<hM_H8mUlK5i-*6W8j;IHbW;I3<toXB*CZI4CQfn|}6RT7|m;JJqW2JpQ@Y33=h1
z>E9jKHZimhbOTjn%@{sqzRnUdNVXFYL!wt0AsR%;8buz1vp76rz#<8etCzoDpJdGR
z`~#YO3!XQX_g9o)Z`P|6v7h?I=p!hv#!Fru>Q-c=qJL6`M%^maSEbO>Lpt(7ORgYF
z{pm4wxkR&W??%0IU1$9wFGE{Mw}OSVY>eb&dcSFm*Fa~PVgWn3K5{HY3~!^dAWPG5
zPn=@FDLkc@jfYj4JCeiAc~R!&0p?~Kj}b9|E*r7>4h+Q9swYnlDNcYdmbo+vvKwz(
zD7tcK;pOJmRE@|(s%Z}N?<;-voB?|#hl%V(NbW7KGIU6+INH&RzK91mpSuDZZ)?si
zi(={JRFqN8lXUH@Ia#722Vbl%0G$b%(vZ)Znx6C;!I9XSdeG_EX+H>?Gu2MqtIx3$
z(M!asyQ4>FXO+lPrYaY#iwCts2za#S6dk5J+P{{=kTsrY)oa>2tI0O=CFE#_{Pu%W
zNM*WfQn(oFCltBqnPS!<UxPKcQu0?6ev!*jT~U{~0Dw1I+ds&7b@G`xES4DLww3w<
z-eOdmGZ2{KS)p27NlK{`9`5F@fr!bEip4rTbTrky<rgdS{j^Fh>JeG$RzaR+29fa(
z4YRCSsDBz%)YJZia4&W91ItULlXuO*t4VHBlNJZZqDSVlxz}Xd?kRWawPqUAea0vA
zgT(p_Wwvh6sb*Im&v#66T}@DMN9@Rh8z=Pd32XU$Q**xP@2Y7^PTkByOS5UFP9*20
zUialLS6nZ(o=w(Ie^GKuHX>LmH^lWd+N$AAr2^e97ZH03!Ph*(>Wj*&dR4qW%7YF}
zAF`Xav!7dpdeS%Di&2(%VoU+L<}=(IY{HmTcl`C|k$kTArYZV{#`N^=PtP3G62}_i
zQ`!r~9q3+~wiOhUX2Y$VdQ+|zwAMA;2PHpTu`VdoS+&15NT#;H@n_)IH+!}q!?k>)
zxhmzA!(ZHprcM2X`s0C{LxPvo5Cj5U9H<{75sZk4#2zRgIL)>fs}TDcWsN9VtBUy|
za&VZA%TiSo=Mrh#W^aX^lnPT;*RGqX#&{`%SU_$OQTAd>?)Z#@-euwOyjIKKYnFtt
zD4Xjtm+2cd663Rn=t+gLyWW#8iz~P>U`;a*{UBvi8o6w}@B?^-UzM$m4^PGyG#0Cz
zATr57Ow=c@8df=h<%o6i{c9PD()UV;h73N%R7CbZaQ}J-)HYEt86dZ9Ju<=b9^Ol7
z&d|w)%9S=o#CDYT@Gxt?j0vVw3z`#Q*^&chXDDf;$4;}=H-W|urA%du8`5vz5+YW9
z2(tpuk~dXiDZ?#LSZp$9@<8Lu;?0LlPc>g$2BgmQT*}YAp?r(Z2>=emRVJqu?2p;C
z$8U!<l%+quEgcER3UrPi`#N-leo6Ll9=y}(QA(hKr{F&EZ7BVfihfFvCVOtdwCsXM
z<a#Wn<X6sjSZn6%pDBmvh(y_S$j$+}wr>G^IBOuO)?0;^{wR8uyb}?QwaZ!-Uw3G{
zvcPfS%V{Yy#S4#7muOGH`>|ReE^NzQ&OS_v?_?tC(r{HP2UhybWjD_=*adl)ImuNv
znPh6O+dGuZa4lOo)Rt}w5KACLbs!Lp^tC@_@2CO$$Z?cMWN>e*4>uVN%<oWY_b)<T
zl*eIDI%)4!>~KGy-c#^KdOdON$p$&x&0R))A^ufoB1v;Vb^6MexMD~WlBF6Xd30aN
z@@ok!>9f||-YoVcg^ok+!dGQHRwmJz;|*rNXW;t9_)sRyP4k>+f6SrIuSRV$QQ@CG
zz5qu57L|-CUHfSal{MA}-;R1XeGT2i2Q~@)jnLk+SgCZ?B6lSL1QDh{*z0AUY`wbI
zD;BF~?n&fSI_98m`ip+PwQltyuHm#SN`{wo=g^erq*bkt<1bFGv|QZj;!GGE1meeQ
zQ7K1(JQN7cDl@!Nr*^2B*)5xRk66h-mOS$LlnQwo+g?wywOuA@l96syo;A?=S*~b~
zCyvB>&?8UX?QhR^*fEZ-{gdTtxUkp51yfIw_0(afVA}Sn$Li`Ky{A&&H@R&5zk=YC
zb2+iIJSPh@p`X_5k>X@fDln~B0TfW78f3H#uL3Hv18&$0*cb76=vj_fslYS9P$uz)
z#3*)l#?cnPuh(40r;Ml4GJZCmd@(3AvZSgWtTEq@f-+`oNGPdKc0}63qAe?V?BPl5
zvcZq53tZnqUeM0BS`I`lDfg_pE+E9}eWo2e9n7KI&VKuLH@#0i5a}a6gii^Pf2Lms
zF~wECxc52@jZtn>!V22YX-oQRlPcVuE3UWhZaM;OTH~u>-fql_x}hl3ocFhB)2Dy&
z%25md<-WozYjl6@W0-F8;X6q6SAR5XSnp(w!fvYHKHn0-`}OABP+CV{ZOtW*U}pfo
zmPT;tZW^C}LVj4wi5(w(0v8k$<Am`=wxmx03!mjyIf0C0AHhk9I2&ES{t=$uDhcuA
z0zm2A&3{Zw)`-3K#FY(vSV?)&?1ajYI8EHxqEF`yG58j)s;(WPayr9t(<vn)ZCP^x
zm$P8Gziy<04d&SClOcdkoLk0G@#oh%zu6K9Fqcg8J6^Y`FFS1U>{R3$J%~?v(MikD
zKmc99u%8=0{<BTU;zmUr2w~eRQEJyGid<QrVcjg#^XddKC1a-&wii{DLl5&xJ6-zb
z@(bbcwf->#t_~zNJt}HVD}Eaol}OZf#m!sZWcSWH`z@t1J8#g9vg_&g{e1K1(1BuJ
z?#=ooZDJGi!^gj4&UU-Z+TUCiPVD2%OBH~_+IVYuopV<Rd-aX0H2J~G1Apo6rgJ9u
zqY`r@<7-xl8mC^PDrSYxaw><)HNU49cMB&+J=mNQ40G$ZM_AwZ4w$HqP^?1-0F<7^
zKU*it=yB37BC;(!PjM(CXV2WrQJf<w;QU;?5J-OHMUeT9Msy{Vc^>JAREoviK*#Hj
zcwyEg^Mvi(Pjx)@6;OuYIy>|wR-9SE+bn*UkpwWD(+6a(U-=%5S)`JJalWFye8G~;
zfq*`4vxSa=@MuG?)fys2MlD+vIyENTsD)<~C{ee`7oQJOgRa74LdY<d0xhv(s>n_R
z3N+0M+io{%#VkCCk~k5gc`$Ew9Ig8!m33rt8-RIFax>R_W9H)c77}ZC?hb|I*Ji7*
zb({MOS&`6~lHeOw0|}rfu=D9lJ`a~bTdQU=S1!mIh3wi@I3i@Zld(`aL&;Zz5`<37
z<O;lwY=H@KD9a@R)ezdv1kk+5){3-nA5mcSk}e7+2lUrt1J#EPO_GfZr{(D}_0Qxm
zl+Oc@6_<^0^zl&|g~o6OA5NcVG!pvPPTvy0!AXCc3sRz7qv}H*7l96fxOD4?7>n52
zT|5SV>Fp`nRv{XsI=PW7owr8g_qE6z{E~9%vbH+Z;xAg~MEdFcQ0SDvm>{I6<kVRU
zaklE1{>S`46$^=cgtr9<>JR>=j8pTO*)V1^N->w4@*rk%VM{wt1Esnpcl4hmI<{Na
zdc#H4t6z6w)p2oz*ujw8sw7CMb6wcv6Y(WG{tv22LbXJvMW`{MPR?>oFcjt5nZ(Np
zGCbEzx*P$UGW-6SXb_KK32x;yE6AKw+-X){BD?zWvs%HJVz+@&p*G@rhnsg34ns&e
z!0ez>#2)K?2CSQ5<Du`qmcwBtYt)n>&m{)-@Sj|ZR9P@>SxJ(Oi+lq|GBpH$d$|8r
zqXtr7Fff`q>DEa}UM^hD2em-J!oimd3jTP>PYYeHYYV@TLtJa2Sst%L$Nk+pQ2KSk
zn0AiqQ7X`r7*ksw$D#US-M$68jv}*63~01*w?UOIQ#e%GU>|uX3h}b?Kx%BYO6f?<
zV~mg3)|f8lO51ZHTfe@vG6kG5*N7Af#^-uXnHA<LOiYtR86W=KTBm~pK}C)m!D*tw
zEyW9g`A^m#nx|KYb69^jmo<jZ3^D{5g&_{f|H(G_t4mTzGkG^XoH#jNHr800U=sP+
zapiA&EA4d0qYqWONW3`ols}|(v=U|HBl35m_!>-fqz-BP7JD*zB#$kf)1L}U=g)Td
zS&LVAI5YC5S!FR^*$;X>Shcz<oVIBbi&^)dcJ<7seh;d|#o}}a(F`jx%1jI-<`cSJ
zTI%v<=c!a$jBNY_&eE$=f6C@4Ci=oPcnfT3JD1x(Kp}_h!j<S~eqj7Ym&b=wxkHk1
zEwk(_1cQm_!nI;Dpx3$bYsC?ZoNfvdw_cfL8q)@H7?#RRaU?xNeyfmPnIh)dk>hjF
zGv}i!QK$r<d3S*9P`{Kxa#!Um<r_Z@gVflL-Bg*KWY31Z&WTDEjj=QGIBftG3xG2$
zXwe1#w{0d4vt+D)#95l>7n2UYAq6?#P_irmX3UD7y=YC3N9Z86BMbBn0h)_3zlyMt
z7#9!tT;-m&VlzigT|=ayr9DbeixSLKV~r#|oP^p!Y%U^X6^8P7t+;OwJ-=E>N+4WW
zBu?QIOn5$SQK!Acvm|PHHjuz<REelV5-Z(8fz&U72IG@0U2&;pqd$KU2cXm41YH*{
z+bW7`hA0fnW=72yQ!C#qiQTPS4G<w!(!<Q>j`PBIJTab-=PGhQ7oTou?FPM}YVbjw
z|F4$qAGG@aMbiW2{*5mG-)Q<jH2c4tE&eZh{lDn=|DOo_<KX%KX#f2Ff71C6V*VfP
zA3FX&E!F?0B|O}JjQ)3e{-51{JK-JqKfm~AzW+hv|JRA{|K<K9u>a6>{Ev_-3L7;9
z*A6D~Wto(v<DsUstwC*QBE_g4Xch5Rjn4mwo2xO6kC~wuPF*L?W>vQQ>}8Ei7EE?#
zeU{za-hgr-?l2L@vbfC9L%;>wu!pl+x(T0_eJL<Fvn*_KLw)T);=T8T<}&j3?VQPN
zHm7aXFS6?L49fHV=hx$%7LogRwXY<3`<H&YWaaY#^*sms5k#)b@<{s@iraeF;}tYI
zk}Zgf*K=3~n#NV(vQZ5AMZEe7ufohfY&JWR6T+t}g<ls;YN%s_pk~NB)lHYk8|veK
zr5BOZJgxkWSdx{d!@l5P)lsR72D;&Btm-(GnIR-6?cyN1^Q39yyj2!A;W%^-_d1qM
zhY)p~uV^j9s2(_2-EJ(5#wl((f*p9uqaInD(L$Hue@qodI9Jh(wFmJ$pl#_E_c4q2
zA0#D|;|4|vxX@KH<Xr!mi4H~2%~zifgSQtsZh)ed<FcMiQ4A_Ir7kPLG(#4NMAK?5
z=&uW?%3PRJD2Oc_H)<D;T8(fvcEcDc&xvtgI?2kShEs0jhZ>nd&PIR;0*Ob1QT7-n
zi(_Q^)w&qV7`=s3cdZji?aXm8lR=YG_gj^biFPJZK$HU6^Xkj*G`Qwxwd%i`=iI#>
zwHxJF9UJ?jf~$GvPmG`~aRP#P7MxnyA1+?XO7=)7`3?U?)Lo;95JOP%2lReuE5)&V
zZ=5A|&DQ5mOhOC+&ILP_6jzw5#JZ`&rVBfd6HkAqHDkHa`jboJlz*i|m!ysRXgur-
z@cDL9IrN|fQm8gSK|C$eD)<Y^Lps-h>n%Jf6{w2Wb#o|MwybfCjBZcsjl2S&JJle|
zpRR0H8uy$m!B#VQJ}^1ki?5eJ)APfm#~09ixd1TaNrv!@0&J!s<&e~<FIjVH{<4np
zJ{9k<D4KfURn@#f`q-+lqcT`UZsB=`7WezQ9F>Wumm37Q%0;O*lyP(-$kg+k-Zd?v
z>qBPcPG%$|LCA~&Wt|KgCrsYTXr-7v{3r6;>IK0YD%SdJ_rKlccM|lqKSW-;#6`>4
z>=fJV+dSYAmK6;G$uc}@dNqi7h1NUjltT(9jn83BYRBB=hs~WZ`PRiT85HxYw&C<h
zF>Wcz3Ia8yjs7ha^oU&^;rQ+jh}2f{^fh7Y<s&(*$!lLt2Sx%?0Uu##v1NS{_oh2Y
z4no;v2~xb@0@KzYm%pvGFujMbjrLeXSK8dfTixP!yEwy9YVl#SN?1<-)hdf5B=6A#
z{io7NZ0six?)@=FS$GT;pe{l=>7*H(Z8na6qR-$kKyS*-;@~<y!beHFfC2JPjDO7+
z>I2w0auN$Y1&^B@>r&VF>F$91l)H~g)PFsE`#qRne9!#0124}9t4n5)k;Q80=J0r%
z<yu-|up%012I$J>CIaIDjdX1HQkb>G(8c;W@RJFhE~$q`{*i#$b@&RK+iIMW!QYAt
zGi<BsjaVnQkgtCl9K0ZZonRJd`&mK&4Sy`N4@NhmMS2yzcAyyUCY6OwZ#V6uAErs-
zN)b-&s7s=cX4q(a+k+)ZQNGuG)y`z_KF9Us*$j3#Sm(|hnNH8am5n!?@+H8u-udtb
zWmLAvk%x}&0;+gu``J}`wIuO{**1v@UA%LjqAN;q5iDd5O8MrN_CR<Z$zZx!C9fxZ
zXK{~!ldSkSc6X^;6pzCzBf%3N`$DJqC+``z5o#>;MK}ZVi;wn0MZR}>0({Hgm?hPP
zVuXisF^!!0bc;!vOQtZ}{fZ6Hsp5D7h=p=$xID0pon~1V3DR+_s-#(fM^bM-iaD7E
zU8bfysn-P#6;U`4)e7FoM|OL(O0O$-FAKSSB}NTj*F~Zs6Z8vF5V-g`dEL_-yS;Ou
z6_~iWo18vTo;F)lX&4z|mm;@bpJqcj&sLO5&u0SZCMkvRyja((B$Csi*eg_xnhG$Z
z=7}M?exIrl_&Suq?3GEXI#WcqmQ1bra1}SWBHnmlmK;l!{F2)A<9fZK0bIH4P~^&S
zX>-g8E6Je$uRXKtOZ&$6wFiNLi_17>mMePX?rnrLQ$1o@Sg!Gu-zDv-KkdzAIIRK)
zW#aWMZly4qJJCuH8gXE$tjsKfL>jnJze}qw6^x)}8CSI+*lckiHllcmlHpg;o%8+C
zcrCG16~cv5dL?{N3<-+d&M-30<9{a*MB>HQ`HChKsPSd#ylKuI4~)kR_8gyR)DL=Q
zV&YLjGE5o4(I`sewiYiVS~;5{Tz~26L#*g69m+-Jp04-Ced#baIqD-h??(9|xupix
z*JL5?%6n_$Kux)1QKaA-(<N8W^fN$X(2%DW3FE8Y+RJ6JqPj}{@@Sc(V-l&?MPZfd
z=Fjv33W`J31V*Q>6iJ)9f`g5=<%Mh3RsCA~O4Ka+EMeYi_lqiPvKuT4XWhBup)zqs
zUfi0DrcG<O*b91;v=TIL%8KEHZ~p#4@Q<#~rj6xAy>24ec2cFL_;nXX!Apr_;;lzX
zqBgs@kDUgOy!dyylX)yZ$$IFUO7{~Z$;j#Baat5%Z<j+(ey~Qx@Y%Q}AF{*5;xtWc
z=DC@@(9On6jTM0+^07jODdCS-7#T{QhoVgaKaDUKYW5}Aqg)iJZQwHK^x*;QwJZbE
zpVrnIJZ`1-5&*S=Uq)Y)RwdeLy!5%8y9mJ((~b#BJp_r(Nk3<aYFC8?IwYLF{ZmBs
zkvu053uYr9aa(C_I>cFNt5(pLZ&3fF33;^bh00ZdmxsXdjsyfK(5F!hF_yQr+*2Wn
z&|Tk09-dAhQ->0*o(;5m^QZbBt9Wt!AI8oysI6~d_l1_?v`BC$F2Ra>i@OGbLvaaK
z+@Yn#U4pw42=4Cg5Zv7<S}3%={Lee@+;h*Bx%swc@2q69Cws41zemz}=NS#Hv+5?8
z+MQE4yd(AP*ty>~&pD_3#M#nQH27*u2WIG})zm(h59x4uB|}T_t-KSDRc4i4pqNcx
zDwj*>+di9pb!M<cNzHm9v63+VepFWS_m&5v-jK|SG(fH=$!qnH8z0i)#8aU07i||x
zc|Cl+v7?nm=4ocFp?NJqY;Y<m`}YZk(QE=OpLuVak1&1G_qa9XidKzv-TpBVvW*kD
znIDR1N6`uK=Z<}bv~%{Lk!cOHQ8g%fYT%&~m8!rncTuL*S9L5U3+7sdRg9Su`(Nx2
ztZOW`p0(lPl0m03zMTB09w{KYz>!X6mAoe}CDG@?I#yQX*!|>Jy3eA7Mb8Smib}a>
zwvI`5_g;Ulm@?TCW7#B6^$lHLA7x$hQR^{_Edv$f=<G&%FdN5P_JEzsu1rp36Hf6^
zbRWf=)L!z%M?yM+xnF-o0$~D$$}tsgn+yF6qU+1%(lQj7{MLK&JJLTs^y>_j(xz#8
zrs@)b-TFyzD{*U<M#?VJOb$HiM7Yd#e#qkUa#_7KXntcyk^#JLZ`Ey+ob!W2bRUJh
zfse}_2Mh-pQF`WA&$JxQD^P-z<-Jornm8$RT!0t}ECeB;a2~j49)VtwXWY{{eY+zk
zB0J&8v*p;*8tw|y$^U4)pKH?FXQ(^j2n949Sstz$RPC4Vt$aGs1T}95jU<`yH!ohU
zwtL05!d1><Q0O@R=r)$H#jg^jEb&<G*3O*`r~PKmX7rT9$~AexW3UubVcsqusg-6z
zgx<p$-S6KzpTdnQ9biz%O?itc?lPq(Dnaj<S9+{jClgujNH_9YyIi>_x30W2Myc;;
zyX+DbmT@Xp-IT}v+uX=R<lX_%1>q$@@`EEQD^X;=au2H+z&>&o*ereR?56clFfBnY
zx8NwfrU}6rNCpa)&|We&(0l_8De%ZoH)KU-xU=&u+3A@ZsIb_4q%kN+?M`<X&phHm
zYRA?xlwi|&y5gIiwb2upDD2Nk1sOf-7BExd5-5Ew8w~2gvn95W)KKP+2snc<fvTG#
z>t>T}Mx>_-mSxF#_*jg};-w7WQEUU3px)sY`34PY40l{d&wbToy5@=~m8wz!PTN=N
z4pU2SrTZlv$B`r~?Y~Oz0pgz(*)Qg52mU09h>4qgR%*EV0xx0TNS}q(W#{^9YFWi~
zS}F3U3?%6~(_YnUSjCXwi1=0PqCdU5SiRvT(WQ>bt%J!}+v_}9kFemmee=@NH-Gr&
z(z_9tiW&ZROr>=@o2y?Bm!VQfth%hcg5ym7z@zQ`TPzX#m8`jfw2RfHsdG#8N@T1Y
zD7LAp8gk_zv0OSz6g@lqV3bgO|3SPMzWm9~F;ZCnLTXDF-&mqij3W7k1qr)XI@^e)
z%h$3Tkd9cbAu3+39+&pR7-oCkfbUJ@_oAt)sY<Bawa#YsBE{>UXkp>EKzX3cPLj?;
zzOSe#WXiB#x9D?~l`yO*AP>ID^!YP*1hC*A$N*gz5)P?Y;vmFf;x|WFfN)8@_Qo1Y
zX07e_yN9WwRHS%4U@43m3vZBX)4hUEt<P+WkdKh(-DLaoHy|*wSRe`J)gg#LVeWER
zVWwe#yV%z1WJwyWHMn3+cSE`2$?yR&sl)$)XN2$0aJgVuYR48Vvq4^@47{>9Hhf`Q
zE>ye%v@1|+>8$JbUT3vgmNF^0m7xI>$BW1$<L%$RIuTG44_1n4llCW$;c6E}kGI|3
ztuZoB_s$5k5!d|sxW_@}Z@nNSVp(w!+{yP{aXPFtc*6asqdqWd)<s6teqhIZUJ$v;
zN+~)`%dpKRqgCdV5j^9s%UP&80OepVwyWzA#?4dVc;4hkn4A?&(dov`;t|i5LTgs$
zS9SGfH^Dt?Mk)*Z31nKQ%2J*@h;gosE&aGcYD7Ao|2bd4?<kS*jb~&IuZzGYWaW#{
z-hK+NC*8h>8$ua~;y~e5N%W?8Mub^nm)AeD<TqVg8p76Px4EcahLorLKA&kpv5@fb
zsye&x?E5`ajYC`CXtu|PSC1(gjdEhuY~gbdBOI~E+rw<{8WlbP)f7CT|9I@Vb{j!L
zk=d!eOx@3zetIjVTgCkdi{o6S5M@&UD*)p|^OPa{XvEKYqaw0TmRh}Nmi~aV)_zG~
zYRSOIh*#Z2$ZJpV!}9X8mO0MTijjPNmoO1FYDr~R!@A?R0fxby)zve|`7+&;l#H~y
z$3ard9to~w>2$?Ljb{<};m$-&oO8<_Kx4$eNEvB=6h5@o(s9uzvsvJFPR~Mohf#}<
z$FOLya(AIgk8jV$(kc{zP2bM4730=T>s{KMGQ)3Hj4y3Q3n(k%{xOwYW|!Kb$(GEN
z22fHh__7QBQF)-Fl`oN3<y7=-cqTZ=-PK-QXu7dxJ=s{AaCDwJRn~a?5##{rcls0+
zJP~<S&F5OH5T$L{>4P_mtBZB$kO~w6FPA)!NsW~3glp~!k^0#q^ZD5O5bZ(kxH(%x
zXK2rg2)4pRQkty|!kV3IV^b}gW~N{~#IA*kQ4$&`_;_&%37=d@s70b<eBFNw)EuxF
zLgAKiE|%^Ijlw93yTkp16{E)KS%qxspZ_@YMg11O2Rd1v97(Nwe9jPkrY7m+_7W(!
zz94#>mXmCLNLvs_nNlpMwJk2^+B^j|&zJhVC26f^iz;I>&n`=M->jobIn14Wo{rgl
z>fOndw?-Ug?=ye1{B)bs)05J6<f}!<@9DkRW~rG+<mV~jcK-3(lFxu1>T0E;8J|W>
zR0zqR3z_=Qc`n@(;dE~qL$txFt!Ngom*xEvSqgUGMQ<MwfYRq(=iAmOkvw}e@^(?n
zKR(qn-TW$c>*Nv0OBD^;JKE}I6erk2`T_lkUf!ssOSHM@&jP5QvW?SpD)bml%(})c
z66Z<7>_6&Ha=vFqxhhFyuILQ}=bPHr-7Yffk%XU>5$T-(-$1p9H_hNQvZxq7pp-aV
zV3dE&pY{IZ$2(GD2z#khzm~y;`^e=fErGH@O|uTmF-(%vhk3@vR))<c&$#GLIz7Cj
zWO;NJA{n5cFo+VB?qW9j!@*9Dy-}&Tp77V)d7aGMOhsmtn@6@rcw(X_8MQufX}1Kp
z4yl&#6t;(VjfV7k2%x+Qf)9J;)H|AZG~k3At)Ezl?8S?taw*UF&>u8jNvehMSKOXa
zRw{+Pufc<bGr;EE1m%xyHcK%0T<LStPc79-UT2Zb>Oi?xUu1K!4xp$7CEb?YOX=gb
z!?thS*+7%^#f63A4a?d9)R$e4w|LBL%0Wj(&pYSO8>w@0%|q?4D`6CkmxJlaGwQCT
z(qi9&2MMEbH-rg|wCEFxntUCR->#0-2iyG~j*ZIvA|FZjU_Nr9t|h|tLRv1$>2Bmj
z>z_N5#fPDyKTPB5(u+Tb{7gnIdIdUR7gFp%2`^Hy-GKHz_@nAy{uH;0gzqgjzgomn
zE^N&T$Sw+LgTt89tln70HSVq2MiRTq?F2h}lByMcl%L+R@V~a}Y(}y2;wsV~DOYXS
z$B40~raV6Yne30T#PGB8;KmjPkZ~nP`tlS<d-=RB8ccf9P4!6wAU?r{<{5!w(iSw|
zG+mR6!VY6!;r*D4$n<w>V~SHX*3)>i4l7%?c6R02iV+gTy>R4k|4H(flzEcsO$@zQ
zZZ!*2q)hfe0!MX#JqLR#KC**(N|H;%sVqDJLt{Rj1d;D`bB`6$hQ5jT6#X<zsO_w=
zFpOn4B<WTMw_JIB&T5_95L8H$g%PE3Ro6c!-dO`T2}>&p3TDF&a@QAH^A~n^i_cr^
zPdo<KFYIfttVMwfl#m2)2(`6=u4=N#`MXPVF?j5-WAo(o)Z)THg?Xb|6eXUjGr)?e
zqz&nMl%tg=VP8ka2{9XkbnwA*8qaZ_UYNdfmX<ttA?Lg33Ln@VBM@thvy2IeYOu~$
zgwNTd1`e~-wd;jfMoJqRQCvk}zR%r~9|eXy9zPTHibT258n2aNh7Rnl3XbTsxnQ0C
znL<J;&#0+(zy|a^H_E&tcEWHg?;meiIcD1K`d@J5HKxZ|Yjv+9hBl^gs3O}p&0}~(
zU*1~|>N-sD*=enlocHac!qsbaAm3fe#e~m{^dS3vZ}{Jbo91(d^*{d8u1mZwjtf4S
z5<W-=_qN@k<1V}zrEQY$Z@hG7u$s_{ITyuq)=4SlKNkB7U7jX`Wcs-xlhcCWnb<Oo
zy<u(_y!{-u&HYNW3mRtDvh3*df?SD+DAxrPCMT-_%jG2&&fBFGeHqWdvs2FFx`Wc@
zLc-`fQUbT5(JXkwx$m<9bJF3rzIw_UVt7_cZZy~Ye)T~v(r5M2?=784d3Y(DpWT{-
zjd=2^hE9e?3Gyq9`$h}jdjG<pX5R1kk(8v&2VuEGMT9<P*vn;m!bv9wn?2>cpjpwn
zVNQ|?V{koGFJTDedt;MVlYU!H)vm(3GZ>?d&lLV@S8syhIX+F=>zBZ#hzDbW@TU;#
zF8wgMjPI2B>_Md6RZ7AL)-UaA_FD$h>V|g^TvMsfG<uSfl%+l}y+n+1;8V*`tx<{<
z;8O@n`gp&Mh0<VwX1?aiinvD+Q?o<icRa;F8mWUsu$J@770!IsfaNgD{!>eu!}Kg+
z?K|V909xR--ALYkt%Df*%{1o1FQnH36+HSBI6!0pk5b|RA-6rDrcnhDFqf)DK9GNU
zDAsUE>>CR~B2$vnFg8WK=3yEi%PmK<oW4=@D`?6_;#DLg-uC&%Vr5ozR0Dfd^JDn_
zZknR=erVlw*iK~oAqhqCd(Q9-EBZ){eb0975#psSr&SO54T|5pZ#RT6`E}55_(0cj
z2jf;dBJfSu@0g8;Kd%E4)_>(ofjFHjeZR1MeQ0I%@Lp-|+L-<6sFS@W8t^GfWv=of
zO{IfdQbwiW(#$wW_TtyQ0w<YyU~t51#!noLU3IF?goHMq#qko0w~ZAZ`?jT#4^c1^
zZ;4J~OpG(aTI%Z9zp-h$dz`LG7#0%2=X`!q<+W>!O?DGLN>E^0=5L`eiI<kLD<=i#
z2Qm1`VAD8N!y9{);3bkDItf=WE_M-oU(v@y3oQMR6ZE^9aNQQnOHvc@^KG;p>B@<3
z-FUm=O+TvuK;EqTGyPCTvKVP$6twbz#(WJ??aJ{6OM3_b{sx$55zpoA?4oMqV}&xF
zNj)3b2}MXr+4x#d$eCezPhKJkt=zuvn@d;*?1dwm<A8{Qg`v0;N(C{%+;$E)^9qX-
zux)mgeo6xX77`twAH%wX>HhG;ivUyc*jwFzhOqbqc{>ladVw(LHn0H(0c-KRAWb@V
zV%uJ(rH^J_CA<p)6Pd*5sTBErHX2xPL&Ey6rC~3?x)lLB58p2Y1APom#&guwCB&yF
zr$6AAB?@>fJ+c)CI-!=w>o>#}Gy9tif8<`((NGauWLJvY;5~t5j(9|~ESZ$FOFpc3
z#14*1x^w_IlU2b6`?Fn(*n2s>>iaNmeGZb@143Zq7w<kNVpa{AOUEU<jD<uIb@@&2
zv23DEc7y$TgWh{bFU*ac1?S<{!})9A&`QOo&4hKzTgK3bjDbm(_40?W;Yfpj{w&Ns
zS)DjN{v~G_LNzeW67Ua@SpOP%sNgvwyen4_p_wY4j3=@7{Ssa2y7+7<jo_VsN}?y<
zOsY4zLXh5eeGm#l8)kRIcvghOsp@^Fv|GzDKT?(%@xlKO{Rz<tg`^jjE&5#vEn_}b
zBwAF4LNu#Tcz7N;R20CS;F_X09p{4A7#cQBe=-FNS0S+p<CL^TkukGorV9dxAag8(
zfy)9k2vu_ro3qp<ry1Wgb1>U?kEBxjx)$M?`~)Vfm4PTPH2T-UM1>wn1bxk;qk3*)
z5NeGhBNJ3u8Vfoztz5k*-b|(>@BP{3kW_KhYf0+5#KgPQSnE)PKx{D)USL(Kb?4ja
zJc`KU>S%uaIl5YMpFvfSEb)0U<4z>fXk4Cq2OBWz4fzosVK<{40N0^ryr`buTji3-
zYMVoC@Wn5IN&+czIdl!4HjTORp-$l4iC#4`sE})(QAUmX2emu)Yns;_CUGRc8L!sA
zQdyMH$H(w5@=Z{XYH)Q?G{S%8-dL`mJ@u>T_#{_FH<@yDRE@)%D!!keb12?5zSKH6
z<W8}g6YdT0>&e768Ne`Xrh1|jTJ!nUhpF<^<6*GYBqXUgV<sgqDL8Y?erv@_Di)Ue
zRV8P;WZulx+JLXD?8v}0TNh8+?BWgVROj7$8_m)tRIm9Qk)QaWY-NietBSKmB;00C
zlW4I^ninD*glFClCo4JmZ+F5;xoe<@nDgSI3m>`nTIv%QO-*0P%J^uYh}W6R@P42h
zvSqim_*N~IPS8{$N&TvrQ3-eNfVMF0m$10vfmRRGtX{rN_Mp+7A%zmO$M4MWiJP`R
ztznqP)gM?j9w{~&179-^JiJzbXJKQ%Oz2p5bKvn3B#a8dwARIyR%h`d2K8D(F*!%8
zG9LA<LMr{+E-=VS(2J`Va#5myE*q?xR<rfFR88&TtDHE5p*gR=(;{AL>G3&9%8+s$
zQjw}NMy8f<s#EVXE%dV4GnMs^1e>dAWcO-9HKXLSAjENu>wWSxKVom?X=0%DZ1m^+
zHDzx4Lh-P?gec*N1Zdzyu+co7U8>9es%#R1g1nNvl0^Gk;b3F*Crs3Tf{A&`q|iff
zktWjTEi<oRq{p@u^<^L>jPHR{age=cx_E^68ze(C2_?}qn^?cce6wPGgO`fhFJikb
z-Jgs+su%t$iZz_54?L|cSVUfh{=SI2Zh4xJweP-BIrx^Ef8$F|pFltfN$8fQ(OVGt
z{kLGhs$=!22p-kugIQJg@l|o-8D%g3F9t^A-s46d7RN~?Z$GAug}DaYp!vY5z8Cr>
z<W^%htyLR|g*?6OjxbtE9(!tMvE3FuA^d#L39+KH3=VH2w^weK_bvhc;F_@`o7)Q0
zdO&*37y70xEdsMRF)`&ZWSWQFK9EUA5ASC3Jy*<V##H4p(b?`wvecQxYD<_qsObA#
z#42x0Yr%05oVMd;N(sw11st8GQI;Zhy0hn`y`{`l5V;ww(0E~U7rnLi+CunHz6wRb
z^V}=#QD3NQZYt9n@~hx%M=F28HcZq9*OtvQE;3Q?B_FMG)(|7sA#O0xGs6CdAc_bN
z!?;6#c)1J}dSWOOhpRr8T7_{Tw<3nluXLs(D5VbwP39}sJT6I#234{-8S*q!wCSEK
zBvIFcgM^%b7Ki3ZI%#B1oy7hCLX=MFf!17Yr7tqihGGH{8M!j!o0U^vQKxYWQUO!(
zf0kMXF<*?5KNYYvO1=9~zMx#Nn;zz2tIx;Hc!HWd^nan~&lLOrMbUHO{);;Qe^c~-
zi+caZ<Nd!xy+2Qd{2V3!->-jB@qhE2|BGt>M^QO4_J5!KOvV54p#Nv_InMt){x47Y
z&;L38*L?pm{?B4_<o~4j|L%Ezo(k#TDf-KQQFILA@;M)6k{<D}rouUk@eNkt7U|>r
zor+3@5-N4nJ^F;d9}M;h3qFRy3Lt*5KS|-d-X9It4k>Rxc^9q<C7mqxT<7_NS^#ei
zE*~}&7c!4V4b(T^^{vGvf4g9P(aC6HNBnWUQTE`z_3a3Lil(;R$<!!E(>8EneKM@U
z>#vCEW9Pz+WoF#jM{E|`zB7LQ3+bJA;qDsCe?|l5Y=M8c(Kqk9uT!xx3mh7%*49p;
zMI6!&FrwPU-h0xHk4w&(!3p?fC<PMjX0@V)mtCQpkN%X=4MnGQb|s>F6YEB1`Dn_V
zwIoFA3V;{Z)TFk>d5F{;4CP7xK`067QdjaPxAtecz_lLdinv3Y0R(GS%?%D^zgk9@
z>kb1)8`wp3m8I(m!(J#dYBHA0MTMcS@kmQ-&n=#Q9{q|za2hV?#5(Ly2?|CCzi;nL
zQF9%sSQAoSj*0QLAIj%??vI|3c(6>=3_CPMtisc3(NLyXZfo;2TauXR-F+c*0ws-<
zk*RQ(w`9cA7E}hl?gfL~172SWRvh{3+DCES0CyE-49pUYy2#qti$c&Sp`TeXC|SS7
z3PuUYRyu&v<2Y;kyiO+J{U~^Aq*~lOUJ+HouFlAULN-udJ$1O6t{D7{he8%FqRp1?
z=enfOHP%o=p-{UU1|-j=6!jzOb4FjqjTX5)0ivh$uovAB;x<^fJ<<Z;x?(j)+H3nF
zR|J|?QBGNl%sh|w*els~3V?|y@DD5`M!qfyY_b{GO4%#Ijx&4r`_i+r=wgK?^CW}i
z{C!^lXa^X}GpS{l6%2E?Q~x{W@;YhX7X3o`=ZH{EvfC#;a=3ks3#0ANqMu8C;+r)F
zpp4J6(Rx3tf9@N8%XIN`6#wbN*Q^S;?nwA7$~TZ7H-A<qWXt^AOD7i=fd!2~`p}!}
zO&0l-$awEkoT*`QQuC}|<3~GNuWo5%q<W8JD&sW<VB%J^QXFFCWXjWNSNi5h8n0$_
z2Oy%9Kh@$JM+2VU%uGPbeR5TD20<Ij%v|oa%Y{e)`A+*HXqm(;DHLj6)=YfTPHjS$
zH3RmGqMNRPR4h?Q7mEJcrGC$DHvMo0(Jm}mpig}_AYOKEx;M>DLe<!WT?q7Ak~)ry
z=s04UYAMw+7_&81XSdXJ8cm1})T*saT(1a1x2EJ|=}U8;s=*YjOD2rtvehk>Yf+~;
zK0CXnf8in3lIti%U?vcXs#ox-0uR3eQ&kW{>sRqzxhAO7dY=SYV+;x}*&(_JB1MYU
z8|+{!H6X?7l~%?I$Cc}wpMyLyS?EJ@A})<+_Jd)vyZ}ZOhqA9Oj*5_PL-oOd#`(Pr
zU}4YLWRd9ST-VBILIOEtO*{AW_LiNZ;<~v$B(<_vxWe-VL#f~dq0Y<H^r^cl_D!B6
ztUaQWPx8ei*L-!R#cXXJ1slFZ{hT<qbjBkQzp_&H?1-J40-X3y*F#c-UdG?gi}UCE
zTAjALYzkFG-u@lF+gxAjiC}7d3Q27!*|4xk^*KpiY}<#o*YwlLp&S8O6$PXv`1+s%
zWxFbmap%46YknrHt(6P>*^6F`DzieBri1FfTll8|Rc)UB%-aSyiT-)*4_bKBwAzub
zOz+Y$o<pO@ltm&;EP$Qqd2^>WxCTSB0Ukv+pH+k_r8SsZSvLtVSD?iR=V{%(vTYAx
zBccju6gi&0qh@4C9&aVXs=yK>O<xi<)LiMfI6Lm^_VhaB@Rfn#N`7s|6tT=LU6)kZ
zo`{|cW=78jV!pUBNs1uEmge{2O&me7QL??WK3W!CnES4AoBirkQIuh;doRS1R9S^l
z-BzAQO`1f+O_jlqY+e~Y+-ssUFrt?EF_bacy$)RJeamxPUa0#~<3kG}#N#sV6G3=O
z^>~VO9$|MNE}kT#-5E#{rTh@i5XDaz%3q{y6-P*#(m&+q<a`QIQ3t)Jr-Ui8AiZSx
zu8@#Rh{@Fp`tF$&Hxw?8z9L-=K(^v5peMw<o1Pj7P)C13X}C9E9v(~*_fj<#WPo;c
z8MOaZ=UVtVi{Kv_Ds>P(CU@k~d24%oS%ttX{BwZ<7lXv0ub=hT42k-y_rWT(o?{s-
z1-GOi&t=f1>)+w&2g&P9k{KULpJ=Lh#e^$Nx<2xAg^I2jib8dwV~lTVlKi!2lYrMP
z!Ac^pEsAsoGb@YFK5%0tQ7^ZDz`@kXPPqNhl-{Ko?LTpF>~9MB26KzaIahm44LgQ6
zpO+{6(sH-3<0Ge&6gTXO$N0<#SG6ghiCxBVBS^9jJNMw#?<&Bq{SBKZT(jI$F(Do3
z7<EU(p&gx>8MHvli*K9JR(;#h91*7WHJWEx>!#%OZ%I6X;bV*u@}X<A1=Xyh(Bo1;
z53~_W-_DfK__xViYC#O?<y2JZ^JlAwYQp)zio3A>x68$M;iXB#^kh=E7<k8n1!9Lg
zJDW2F8L5|M$|el=i;Jt>X{;6ONUx-!cx;v^mnUJ=tq7BH)2D+p$}ec@rdoliM3sQ}
zBxyw(<?;#E#jl^O)D$KI{41HvA(gygAn7a6jIdNzXqwOR_R&~#6(oyiF22~wR7FNo
zL?aqS@U{KAeTzc(ck50H!AH%0g8W=Zb0Ohh51@L@cZQgZW$3`)$ka69SlT)l1^Tm2
z95^C-fpl`&uXM!JSRRa<M?B%%xP*SI%Hdr5E60L%v$lDkfG;g*Q=I1!bAj0l7S_xL
z-%$S*N1<a&iTuLPXXck_a7#zie2p3~6p$NgEKl!dE+Bp=Psi<zXTU>Og+$x`o<JMI
zn7{{)Ria+?BtS-2eKZVp`5hs&zz7rvBqr|sTo8?wu#k-B{F9!r^EQOCu`7WmSsSn^
zwo&(U%;BQT^)HWmwv_|r#--B1a;iseGYx;6e~i3K0FhBvX(keMPJ%3y&rNVRS+4OX
zerxwwoujm@W7y~4x8I0EUnYHUdE@yRJSkd%Tshch;2f`1&lTIn86)3zVCs1KCZr7Q
z{WlS9EPSC(HOszsLs|L|c~nuYLY>=YhB#+2R;lt#b<dGdxv&OqWre(Wr?;JQ!0i~V
zqASQuDrL^g(5R<ZGd*wXi8UBI-Np83#gr(dYVz_`EX^XA?Gdvn0&iuE^EIORmy)vJ
zuTR+Dx^i^PgBcw^1d(!oX^y8GH))|4M+T@)Ri(wTalW<k%W#!=gh=yAX_M_?$=anc
zUmTm6Hr&e$u`u`Kwjz%asEW<{f0LUki1ZKx&6JtaZzJu0eMwMwu{PFtX6c6dVnY7l
z6h-B0;Yd3Nq3C5N2Ovs4w^sLd<96cjn+Z0azihr58FiLg4nbZe93_|aYk6bEC05gy
zYvb$&MGI&)s|!OLO<}?|Y<mbhM&++XpNY!4GPl4BPzlT`S_L3z>H@ez0@xB0P}Tj3
zMDwrv0HGwVso0m`65azA=lhe$yH3lJ^thI)Qd7XmZZ}Pzp*5FEp4iTjgf?reRC&3V
zoyxdv6Dt}Hbp<#01Q6FW$koiHZPG-a;})QY7x_rRs#5N2yUOd1LIbZ1iYYn~G*609
z>vHVb=2G*!6S1H_oEW+Ll#|CB8XNK)Xj|g;2?c&8dUTf%qZ8Stxb2N2SJIVCNi5w=
zenrT5sqad(b-&yz05fx5ZN5H|O0@B5`gP<bAk)e;90&GzS-5?6FmEs(`aYh9)uM<6
zL~D<&_T}9M$X#whE)_7he|jM*v=D7ZwmV=-o=y^LyBnEmpn);qW+<Uqtp3v5CyVKN
zr!S0VPy-#{`hK{K;7W&k+)2$rx&d>+)ta7Nh89b<4)Yg=T;kFs#CW6lRp5a!uE(|%
zA$wMUo9I@<>)i-P5Jkfe_fSS}9$CDOHro+Qp{`z_p~2KdkJ|3gjV}2hXE2()suYG>
zV0LUbq&G2)DZ$`#todkbT|>s~9HSanulNt`Djn^lN10;}-b;+%qPWuJPpxi0u_w#=
z{}gDByJvzFlvcU8W8?JLYW<%o7OVs{5|RQ_WNf9@TV@;RNTv(fmxsbV^V%(-RS8uF
z1HMfN<Rv7qoLu0tT6yK#)Z%{E7rYj#XMq`kz_SVcpFh0@;Ihy;%dPCX3?Z%nRhO`~
zpRpDv7VdEhc>*qY{0&2rq8-1Xb1)5^t-iO0OW10X|AtPw`2O6S|CN61dRsv1(3>46
za}y;c%)|Tq(7`cV*I5W7VaG2|^gXaOpcBJxJNUyPe8|hhN$+TJ1h=1nR=K#Jt!W6f
z&;+BH8jZ|7$<jhZR>~TFbMG%zlJ9KAF;z)!LQc@{{rHk|Ta^c&EurW;Iw)+&yynCU
zGrzxa-kkL(o`xpA=&BVNEvT=@4My<NSD1!ATFzX52l(OZP@SX0>>(g?X3nlT0<9nS
z+e@#IIRZp@)4f717LyKZOQ;4k1xpinLIw%*d2g2#WVI5VahozFwyC!YpMTEPdv+`S
z;Epz)Qu{)LZ60q^_{~T@0Tk2Yl-|Nn`m5*gs<Tw8q=G!&JAyZfCw7G5-9Yx}c~gyu
zu4^H3u2k1xmGPf>TTz-fu*}vMT_%P@WnP%~@{JNTN7D6_(M6i+E~t!*Jel_)M_LH^
z<WkV#c-<{TL0=_fZ-*gdiGlA0QYV{!y{_T`$v5lk3ZeVaq=yDwEm7?R>^)}D8y|RM
ztq(ujt+rMuPC<N9C(T0AL7jm@+#VU<uY!?1T^^SX^JeRa2B)1wU=nW#id3dU;ja^>
zR2C%WJ7mQ|zJr=_zDhvltVQCu4FI3#VP%;1>K~puS2DyJk?R!DSbxw+xwMd;_*GL{
zqW4?okTL*8a{2^97i)Ls3ccj3u@$6t{BS@>K6U1KUH(bei8iHyURC*CW#JD8aV)u)
zK-0ajqIXu1j-jO_S;d-q)g=+d8_Nb!Re7hIwf9s~uUDqblC{Q$35+O9?dtXkahjdS
zDor<{-xfQO$arA;4U>B4d?LDX>{<Wqr}@(##P)zM)LKU2OPKyaSV_2pm$9Ju9ZLD^
z3H{BsC%jPngTwt#o=gD?>AZUew$k*7C|Nwb6%2*aNbz8qFY#A!FejxDKVMkyM8Oqu
zZFM)_8MB&FgF@jW^Pq9c`MRvwoQ%=)`d_}hY!o}rf+sSia%U!+{%F6u{f*tY^;M3F
z@;a}b!m)>jg?3cRo=$yHi+szxsf|tp%3B^WXFPuuYfyK9H8=|S`gqGab4@M6eVq-R
z54zx5V5GIgV{fT7ZKe@!G`~qwny-~gH!UNF(LthivHo&?6Z?jy&vZHgEUp#TP2HIM
zKF!l|TCS=7g!NP<Wr4hK*!4j(TBFt~P|+?dobJN_>g9&5R)iKFtxz_eC;yausdi3d
z6DzOzp0x5kRoNG-Y@Ib<*vYT3rX^|J_VQtbq7|E~THRM491n}I&q+y!vb+soGmd_q
z$l@^39-X>#JI{Zje0gjsFo3Q}h#&4P4#o?VmMz;0x~?hlrNHhM;#gd2VV}Kc^&qzy
zyq(HsbVt;1<?{6B1j$3EXPw1UTxZE$`)rl}yc`?U*n?glb;;S#FQUWwZhHHvajdmx
z`&0D+CB9=<56g|JC*mmk?gPi>PguwB=Ppr~sEd)R<$=F{g>6G7_h5UeON4cMm8(|m
zCN>F(j)VK?#;saT^G_4>Cn3pvxWVc-Iv<=bLw4Gv*+-`_qR$5Bfo%rCPi(e|PaPQx
z#$zIzO~b#M!r$AUgt8`&-8TobPL*z!zXX1x|5#HM*f&z}<UTNNiq%2#i@8LZhN(Kx
zlf@2Lm!bhn99_6jhIBltZ5BP$36Ygtku{`PUGkz(6V%x8<eR4}KH%D@G_9L&Ez!Vh
zX)0=p-T1u3mS%>;(MkiIZRk5mZX)VwFrc!j8YT^2!yl4*;SoYochCfop>QC>W;W?r
z?NxSVzH2=CJT6_D#wpqy^(O^-{z;Qyl6xn!Bdhn=^b7|TwqA6D*`gTJXZj50sc33-
z1xfVM7?(<T9vGY<L|}s;O22R2+3ZpPicmd?2@Bj?2dbUySe7e+a#IR%n8QbW9vDjj
zLC*)CWL;ihdmh?M%B$P!8jX=e;rCODx{+~?W5b&(7%jLQqgI~)1k+<tM$Vv$Lb((o
zg&_NMa_<+KdQ#O5TReaNY5ALrzy<{2)3JeUs~e85gpPmh!0BnJNUYgryl*?n&IAl<
zJ~~g;5sS9)6gcXEOgi1$|CCPoI3sVyW+oT<)bmJviVMii%+sv(_K~q^azusP`rB8!
znmoAsfM(r&g0=^0r|E+)s^r?ms!kryiXuLo8U)mwwSUC->WDlUGjj0X5A8Gv+!{#+
zCQlvMwk;!2XpvqDPcrr(C>Xo7sytqB0^FfS7%er`io<ra8mqR8#4Jw@jDi62rSk@x
zgweD2mIlJV?|o^!Hze8eCaZ)KS{`(26H9l+zVqS_)9qC)HX)K}BzVvRErdZU5nRx@
zR6;!)jc>8?Y;w6@Nvg}13xzZve!8H^H`lSsNsk!%P;q$c$<0p!oaDZ`kzl<q!B37g
z8Q$u%X;OD<c+$H3)PG2aI72!Izb}p=2oC~hW9#3Rt4|MA8!U6NMtFr)wyU8OI}9w3
z>DXKr(&VvY#iP9lT&tkvy{T}76sf|O`iZ~a=R~F{=6P60_tx#{S<IS8*((oK%5dO}
zvU|=ng;O=Ue)nGH$y18K3aRY2n})WfqoBDONh5<sTAIqzb}MFsnRXD2%&3JcV~KgT
zoasAWO#^FL4t{<1#sT}L#1=Y4zG!0R<;<krt(fTxVhSTgWAmkzuZ-i&e6UnFosn>K
zDhNuN>JoGvC)(42aSPysiP+JSJhfOo^UwETD4}2at>3(msOZ@7%f}h?syzN3(Y;Uf
z$L}=ttSQ^j(r&4CxpLT-LE|`?Yy2IVD}}aF6IFZOOz`z&r|stz$K5T0o!Xed&G^v@
z-V5K-9n-O#D@7PlwI|t!S^35}|Kf^~-JEKrCN|9FcIFG%yt1k+Czpa<NinTFsWPLJ
zpYh)?n(^X+qx>fQl3Ug#53L!t@8}!RY6~cJkX^FpU@u@%NgJBX7z_D)#b(^Hx+kti
z3e-0|^-QDWm99Feh-wfw#P~>iyfq1fiXRU&N-8&LOC9fQ+4o(UI(Q4Mk;fB$%qtXp
zOvei}z}G!)#PGDl8B3<rQ04a#1D&&srU|E00vo5JK5KeEW*(7?N0|Zh+=!EnXhOm=
zu?1f#vCw1@R-P~<{SImi1;tU`<qF*&mZ#@1W&$VJ@?^n(D1aU=8$jtkD=2WsH*}5|
zf9Sih!<9KWwcxd(icw4VAmk5veY3Vp8NGUhmAcYxe#{nFbAB9-=@zT%zgmOm>u9J_
z!fpq#lY@axkLmU4GG91;3awVlnii>xHFK1q7ve557|MOSpv!5RC$nxxzq~8yf`afy
zUx~vs5=C{AWML6}Q(g^fibq|7Ln%#*R9jQThS#6{GxzWtMQln|9P}QI(Uu*+0{$f4
zO@SsS6N@C%$;fdoX#s*S6Az5^AtOVufN6!1pnz&5hsgslpO(`8nEVvz|EA`Xdy3P&
z%N-h7mBFvtjR!1E19Xx3s$9tORTIVA*MC`Xq}J%4^|e1HiGNDJ%y!qFe^{^e+KD!7
zYZ5=lp6|eZlhgBdAjIscecgARc2(3&Aa3Y+m0KSt4#|rXyX_^y4^Qk7-m$C=A;KBL
zWzk})n!H`fU}<TILytC*v{_`WCHi51*Qj;JW@%jVq<(r<Zr=AQi&__7#0EX%i8QPA
zPHqHlx_(~6={~118<9#U<QMADFyId%Kna?TYp5T&I4aO99x9)2WS{hg6}``5Sk8l$
zB=@JjFRD-lJ6bLOSdi(OH-CGx4f!-&tg2aO1-9_hJ_3F8ulOPoxqH$5tiaWa@?xmj
z--t{=RQ2xEBIL3=D24iLoKC7TZ$G*Z(b3ozSzuxlqp;?moHO3(5th!R#(%Z>iLF#Q
zOgm?>O_eNwoGK-MNz22|#W5;Qs6`>dZJTHpHUy^SyVbKnTY!5NB#R$4kB8@`{>*CN
z@Yj>ORPB%AR>wnfl9#Z-dy_aWuA~|6>dI(zA*G4BlV6a_j^D!To5&dKkakc_Q}{Gc
z!S*=zxl8;cmgmUK1UWquU3_4;>pNkV`hn@W`5tRJp~ydN4xB-DWg@%2Yzoi?(rZS-
zl0>kyyc+K6Ya`8(1(R2A_UbT_-qo$AR}Q+_m6JMdO>J)02EqzQ65OT;FZkRYRQO!c
z--{<;D3FgzL&E|>`szMe+?`LvLk07_&rvb5z6X>DPmRNidwi%LZcAnt&|V04esi^~
z86%p|s4{a&pY*^~?bX@7IhCPE-#%5hO2H!N-}$nWWIUXZDdvU|3<??;D->@TvRO4v
z&o77)8VTMoRtAoeim+z@({KnS{en1b-?mFl9E)4<2|;(aN)$+iA6Rv#a_Cm(tQJKK
z4GFbXZF3s=?f~IZX}{8P(45ntBIAe|7RreMk}xNPMt&R+>O`(HyA+u5D+7?(4px4>
zCzG0HM<f-+<?uZ*s=Wa&Ocoz*>&XeRZ4T_N3L(&>Q|S%%8z=KZ1Zf#MM*C-k=N);Y
zjWgu6)YNl|y61J5E~*!M_mkc~>fzAv)SIdTEs-}=2}98LuN13Ie`~)AbimsT3T2rp
z0d`}ZvoeH;;9C&z5YJv})aaq9`yw@ze5TbW(4SL)D89&<aYc1z5%wW%MXTD~{1T@e
z{|d-n1)f)}W#t>^)3K@bqp{O~So!k`lef3^ahC(h+uyn=o8lP0d&k!^D6#Z%)7E}(
zsSsEu<R+=DwfAzEgs#>#h=h<;U80Jmw6}B&V!Qsl7KG0Q18`XTzN6Nb@1z+%dCfU_
zYE```42i-FtX5G$dLL*{N2A7hymp#DwHt~<f47lV*2+sG(n|Bxe$*%$nu4B?^M{mT
zV0oG6{=}9B(I~B9S$A5A(4(-?7OTfj+>(ilGqU9!DgRXR$D?LaO#kUQyf(8e-*VOK
zcOweh8|;n2Z{?!0UY~G*y}hdq5MpleBk3bX=F-)g@>pB1GrclE(qb~_Q7qz7an`j%
z(Ysu~rF5ArqUi@8SN6)J%sL5HwGLXtXmUeqs#jC{!`!7kjM%!t!MNJ0pxcRMBX%*j
z)MlF_Ma)jii;gy#g+_?qiyLk+PQdx86HCvYwh*JAMsFs?#ZCutdBV7&5;`tgpQ(GS
zTZ}>vq+MKCXeX28vj@H(SWhuDIE);uednuLJh0O%rY#AtW9R{i%U57t=Rqr0aG~1T
z3S|B7G!r~Mc6Gk0K)&iF?MS|jBp`+wRyR+pCEOQv+la;mp0=p3Ai$Z-d&i?R5T>@L
z^j#%@EpC2Id;r|b?T)+fW6*E^E4c~32i7U3;2*mhP@!nF?JkS8wE=))Vaf}9Le8`}
z4G#?Kb&-bPxk72?A!d?ssy`wG#S3dAORruv2qHjrF>4jm<+x7!7ww!6)D^%!L}oMd
z^w!iHuIY={&spyOq!Kkv(G;0J9VO3p)DP+w6BgnT0D=AeKLj;UzalI7DKoOXjl>cs
zk@*uvcd)PL66>eGRFw;+dA}juX|UmZO5eFYlB<#NYfmH7=8;iS(u5<rM|h#!FZ^oV
zB#2LbrF@cidEv@)mFU<%B=XjQSEH%zkWZIn-*JS*b!<RfUY3EsI4HF|*xa4Y3T#Cv
z`G>!_PeCNj^_b&xj^$WE%dN1TkLB*Y3ukRr?(&1Wc(EnvlZ2FKN}581gUCbTbXl2(
zeWSVO>2%;)q>(-e)vB()2*GH874U)@mePPbvpm&I*pS(yabm0ImvYJND_1C<fvrEb
zhZf(Ueof(o@;!rIx4h4l-H2(oS-^fSG`&zyn>?}xeg4>vXHf&HBy5jDJg}^_dJ?-0
zRa?tfMHq^A!n3PNUc<sUJ)3d{8^v{eImZWP)rzUv20<MJHhUk<HW=DcuScL*o`~u4
z&B*5QLL9sETXq)aUjM9$ZJmT7vLo8~@2db^PG^Y3q)8v*!~Ju7Xn+%sMjHAaSUx6G
zJTXXS;M;iSuf84h(%L?}D17B|Sq<h2{4wPIA_xnq$fIVVsiYA9VQLM0n=@F%KoN6T
za|TvVO8Leo`bTP0KD=E6A~-}yc4q^jYS^^Lg>tCoGI(@H=uzHCu@#I%-GVD+NB93P
z2>sc({r^Jfopb+%r2mfy{rM2`ztx-n3yJ@yar{5WX#XqX{<pXP%P5ZYAI9~6k5&H{
zLjN~J|1U)TALsu&^8P=5pZ`Gk|6A<*e<J;V`#eg_|2Cr6`Cw!4n73X@@N)9%`2cU-
zQx((e!3aaD7|vnt?kAPn!Re1OzVBi_7NBVt@1_=pc=Z+u{3ZD9#@NW3OoQcecdX0P
zn`6ZU|9-pjJ6{Y$Ig`z^eR*#CNC6@v98_&QB&*b0W8rB%GdLuFGxN4JC-7@1zvevM
z;!#WFQ=jGAcwQkr6;`T4Nsf2LR*t%=9#a676~hZ$H|Cpn*=VDT-{W0b1l@%D9n<{R
z{cqo>>Gu6tSp^i7FN1%0HUb1nV#^wa>y|Y0g42~XSMl0Dg9^!t$iWKmrL%OzvYe_a
zIHn^Up#6T$0K-D-0F}@mJaSqHh^c@XyDj!<1FU)2!UD`>lBCXQvwSz_+v*hhJR1P`
zs{P~Hj=!{C6PwLORHcot5Nx{~SPmznle^D15^F4p=EVA9j#!1Z5kh-wTcKY{ex0Nn
zR@kd#^kaarvpO8_<*6~}lJlDbbk((;K=T{5uF&0LESG~$6<8+OaYeNp8~10vNG%Bw
zBQCV@M%aqs32+xYQx*v!2#`<fU7wR4N>`2kc5Io!tzO?^7aui9dh)C^hZ&2%F};<%
zl*<?TQRyL!ESL5h8qS@}6RE;RL6ki@B_89|@lu4DOP2k%f2}XAXkX9V@j;(<!pf#L
zrn^&KD^Jk<JlCo9D67~S?jVJtR)ght@hnGwwj*OxqKk*dLia&dnnh6}N48U`YEg@<
z{{48QG=A&h=bUJ9&c<i2Sg#EkPK)*55Ar4l8GN$dT7QERx*xm2w;?*s&7Ei!W-prp
zJF~g!v(Gi)n<(>vW5P7@Mtf#m+p*@E5mU=;C*hSyHq`t5{KPnuF^U7CR$mae!c8s{
z;T7k+xdl@)+Vb-YaRu4~v(vP)Iyx1Vh(#gx-Gyv(M_FkFhFMd66p~5y<nl$y%ljyk
zX)uWIQ%ISfz%8#tK^Q9sl4`4Sg1h7V+{Q?zJ<j@ab>e5W3*7PPTq<zN4}|QfdVZUv
zu-|+M{TdrG59{R^hfoZtHe0=cfT)IeBg}hN=qRbFAiN=1bo|%-{jJZkHA8oOr#T=Y
zhN9UMFHS`)Wh@a^v-8$+Em<QXPGxKSWLpS!qO15t=KeHoy+z1r9BF5!76p?4!aG^?
zosSrjMbR0)Ko~rjZtrfJu?7Xp&}&VMk`@~el<SBGPOpVsJJ{f%z75|~V|+h9c)6#o
zI-464no^V~JUH0G0bF{K_Cp$l+!n}G1Qu*?CtOW|u1ufvP&@D@CYuGrr{OQieBahH
z_$6SK5OV!2S_wg6Oo^VD7ZkX$9-Lt`OQ^{R6lfXPxt|efi^<>~bV+E0Z9xY4eQ6iu
zVS{}r3`JgHuE$_C6?u^NbB0%WJsLd*i#YNj+zYH#(pVSo!~rN|W&(;kxlJ?c1acO{
z#-K&hG}GsuewB)EVVn1uHS_vVxlps7JNCX^A^IHO?iWQ%aQn2ySoQ={HSQhe7tUil
zH-+}vspXdPB;AS+3(lF|N%?KOH7*%6(4MdhE@Hor{8?CMlfNGX{TCjr@SXXE0ik$n
z7+?g8$X}<n81=Js92)1fsxJ9cv$oF8ARQ(r91WvuT6B^Ktp(p_qwkvDd9g-GZ4^ES
z9+`Ci54KbjE`nbt4)Ckpk;Vi^-oqykeSNm(R}VK+@O0_pW(0Lsdvylgc3+(JwP?5z
zE~6%A^+~B>0jqu81RK-dVqb&>*KdPkUXv(AtO6?{m)7QzfzEA(P=<MVi&o4t`6a1U
z+2%qYaAB0lSzO~>Vu5>I#LRGCg&aO<y1Ymjcr1^NkW)&rvjuiarxgb)8<X|N<&NnR
z7i1r4(#M*qZfz>GAcXO?1hWp`SRRXio|PNwDflP>l<DU8Fi^KPgP4xi$@<_B=zNVJ
zFe#B{WyMP+mW@N$+7gxp^fTd^?r1hrF|P+2M6?|uJj{^IRK64Ung^HX7@YiI7`u@t
zt_HmF-0NCA5HDvw{3W0(Cz{@*Iv8O?kaMC6tbE%NPSb3LVcno;fo(sU)KGe)Dx)<l
z@XdKtWoU%u?>~2d=Vt(r&V;U{lauSV?xd`=)HzqG8@aE$CM;MPZ@F7=nPLutVkAoT
zfDy#GX@K0vD~3J3o<vPg<3Lv3SwSVH`Q?YIS~coO#nOq{Q9VLSh|IsZeViJg?6v}y
zHJri{>#Eqzw<F%)>y&gh3OGFPZ;M&t{CROi_VDM=1ck1L;LUy}lZpc*^ezpcM}M4Z
z1cEiQN466tDgo5LZ_2aV%`()kL+8Q0y_0F$qM=dICG3sd7S^>qAvJr<<EaS&6{$(U
zMM0jtQm$!cO$7=@ws!9!*3Zdf<`%CqH4My>V$aW)F-2Yq!d5g=Om*zHu`8(wC^T?^
z>w@|se46^L%Z3*9MGADkdzjA76KabkL&L?#^51!P@WQ+m5?O6T@n7gqluqr~rG$@r
z9we_rHG*ry9WK!u{+vg5)d0Km9cdQi9gT`|Vm^D+FyJcm)WPt-#x84HK;cTbYU1cT
z)ZY#ciy{Yl)!o?XtZtRjre`80gz<tw(y(>DUsY0JMl7$-a14feI$eDl>}MVP@;!E!
z6DS|K2EO6_tXXXL9t5%+sw(v7ZO;>)yWZ&p)7?+$?Ajb6<cpj}-_mN#Oi%v=D1o_-
zQ05RjKE^7wiaKUEm3Xq0oi@_sy6$wfVE8PAY+3EB=9C+F`Ej2wTCmy6M@1y-nsiCe
zu&C0i-EiZd74@7CmL0hWD!G%}Y>(5KTQS&8aQ+q;TOc0Im5TfE1`%4Nr81bS%1-0<
zMoh5WSnu^`wXK$@CDNFh3hY(+;Nh`q>^wK_1(}UL%|2ML3U}50@2gUQ=c~cA-+CzU
z(U`DX;&<UkbJ~YXElnf)k3Q46SKrH`3!SqI)dM>>#^MVbx@fS<RHk2@k>!G<Z<1@T
z*zgzE-Msnn-Rear%TRWVB6PoXTkgDx_~cX>Jd%%%mgZY*qXy$!+jc84tucxS?}Iyt
z3V9A)Ui5FbNScSjEi?>|8#gqMUHVul^hHL?OIwbk=RsUQ3p#V#Rbj26X-|4n;JMvB
ztJ<(;Y??Pt(zfwwCSqmuN}}kCUStLmcq8us>QC*pggob@H7;p^K4+9k(cggn7Z5`P
zC7bfv${s<RkWX`rV7e~Ar$+)NqMWrC9zmgkDCOr~$?ht}FodGH<2#6rpnRB1U}I6t
z-va2W*4pc}<T;tyFXy&1Z_Ucc6&~yy+e5ZLb?V$~W|i7TQC;Q@w66jBkn!`zAH%VC
z#}c9SprR=#4Uh+{=CYz-ezX7+!Gn~%t~hzAmbR@_tZa+o5IPnc6DF@O3Cz77981+L
z27~=2Ultg(q8@G-WBlRU_OGYdR<jRZZ&-qzpwAwf*)0WVOm+o*dOdM|ILFP}@=rWP
zq7OFVdo)ziE4Ye=(M-JuIv71Z2OOkFu2V1;KDLqA@>MLlb4Rb7n}b{Z&-CV14!f-s
z_C@zn!JoRSp_L4yfZ!RSXpyJ9af|yuhV#yl6aI3)69b&|@2Du$TH5v->&Fvu;SHAF
z7w`*Q&kC5VzvoW<I(M$JCKy36peKfAlNa<x@oNrE%g;}me_tayTtph|kGW4z+Z&=8
z__+%`o2EvdGwBG)pP~xc4Oy4<`;U*;{l8pt?`ayExNujTuEgcEE-CweB#W!;1CO8!
zd->J7twokgP)^cl$)F2@H3WY;YBaoMR;W~9mh3Fa+$nbSwVt;qG?UXy50yt4q#o~|
zmTJ3K%__k{Z@%$~Y^x9OIQ+I-ru_sEi8^W1)R`WrPpyQ@UD-M+4^v$0=O);DMt3|H
zu#nr#M1f<tyalU^k|p5nKaU%NkYUP1+nVnY*xD4=Vy!|}1NinMdTB|2p3<+o+vIdA
zb&cl+1Ny=2^^F(vgTdXEq}C@r?L0r`Pa*n{D#B`tK2)PX1$<YL=4<Z-heo_(T98cx
z64c`(Z<dq;5RIMn8qwGJ^$*iLT7OL{XKR>;I|}IY-dht`#pT=Rs<-1s<DJFReHTt6
zUOe`dM^3Y|cnq_7eruwKtxZebLF()9MS2UtvWq8Q>4#_k$+IP~B|=lVz2$T1@PS$@
z)Fw1N9=60IL<Fm`@Jea;Y+?QBhC$98g945KiH_bn{GlWCm>{~-IhKuJ$VPRG>32Vk
zTlDQuttx*Q^DJua-+#cAb`GJz9+{5nrH0sBU^U&cuE^|YJ}{T;8wT(}@ACU<m|fL9
z*hF2V${B@UrJ)u=<Wc;tS?VI3sOxa}Tx8$ml2!GJHCeb2lK+p!z5*(aWeF4r5F~`)
zmH@$7APF8UA-MbE?y$IPgaE<aHMlOYxP~OS>*7vucTfH-&+dEwo%esA)2F3ss;X;d
zYxZ>YR6U*i<m1<>g;RzLGukAFuPr69z|MOcNwP}XYD4VXb|AaM3=J>XXxyCZ_2Oqq
zII@|D#p|e#?T=z1cr9l`EwPFol}o!xL0s8JIq0Oj`P(LoAbD&l`?Z&y3uocAchLkR
zwc$Kd8{AN_R0`uY@=pCS-iYww!?}F6+_*B!M7gbw_s>||*igmaiW!!0T~m1G{v6%M
zNc|R6V?pxLH2qvMw}?@TvK$pgnJ*sBOf0e}qb@4om8wf2f&t`JS9N0A3@qkVzrv-y
zvXG%GxIl4@Z>$%qd2pqiBJp&0r=t#5R^Q6sq`{EifFn-*BugTl-MIchgN())^yZ9K
zL<O7mc_zPaSbUV5q;mp#37fSm<~wEb-*59Ml=Z`PF%2>FC@&!g%%~0>G^{<^@h?Ps
z@w%(>C?~382J})zZYnAw=SDPrYM)Z>pk19cYw?ZI!Qp$y`QlrLRy!UJQf0NJp4jh^
zQAnZHg~LPwsw-FUa>%N7@OOHBq$fUXkx%0GviLAIRak<=kgEqJ2+Sl*=;X$3t8-cO
zO!bIwM<ayveyMnvpVbSIbvu*rgeK~7N_8Pk)X#gy;yTZpG?62prwBafRH{&6;hE!f
z=FrP1{+6dGzcjzaQW~$>q2Or$ObqQ&oP61l6TSRM&$CXQIhGftEu^ma-0^rY^2zgz
zH&0tIC}W|e{p;}*u%8o+1@XM%wuSNmx;bd3(q}V5X4oD_$qMPZSDG&8)6dMZr$+;@
zs!N|$g@z4CkGU54xYpLV46QJS=g?NoSnJm$Vcw}@L)&UBx0^FOo`yi^F)I>3X^hw>
z!mP#(Mi1nic`rFDo_}CX#j=tcEdIeD?WUt~#I}BR49PalaS-Cbl?;_+80~rc#-0+s
zm3;2ff3EEQ+SEX#j3-M_2uBNJZ)4t|dRr;FRGMvPOZ(T?VvPX51hJkbYhq55qx}`(
z!Pr+~tg3YO_VlWqBw`ih6sIUWg5cBb`U}oB5ZJ$(-oF8Vw;@3n;>}Z-YPpi`wN3|D
zpX_P+^iY>$z!$zGfB0@O_DP}AyPFb7T=qhqA%98GMJ1C$1<SKwp>_g}2dHRS<;!xc
znUZqb8I<4*_jpC|ywu92L=(b7simQ$c%F+b`EqO9{8vtRBRJT|?~yS`fp~PlHb8LR
zrM*@cxjuS*thA+4Tp<Q$ad1%FL9nb}?4z6S%JIQ>LZ5}Ey&6V>`M)i}vyXILQRhf9
z6(v$2M+f*zoPFRDrIRa|b*ddPY%Z5@*HftJ(I2dd)y&lv$8h3`pFjMkK(*x>d9?|<
z8P(+3VAp1anmRP08a8=*HLGCMQN^%`ve6~6?YFX2GHOz_ZX=r~XJcODf0{41K1+mI
ztgkM8CUDj&J=z$u58=rN&f#qc{`QR!t31btUF@+@Gc|hjj*G?kZ(mk%;UA1d`)q0H
zc$}8=Vhu-k;N)<yyGgYWhd+w3(tgn#R>nTgsockUUehBPJmXRSX*AO|d(pZf_vExn
zG)A`3$tL;7dyH|*9zDv}GYt4;fvAt$NcIgn65s`n-Ws8P+mm6;60nT<z=gOm*%WTZ
zn4mr6WLmQ1IJT&&sTv^uCo-IuT#M*7)R$W+v^6h}xF}Az(Ci_h>2TZg3i);ryZASh
z^A7pcccVG-J+#VsLMWq68Cr5(QF+8uNbvR#QekdPQFn)2p{~~%tU6{|XOJ^>7}6Ys
z_+28;ZS+nm`GZfd7!=Wj#7?`tjbxojRWM|cpm^S0j^Sj<6l2h_I4-+~ISoyj(Z7Ub
z%*7NF-2AV$h$GEtfEc5CseH~gk3(1DIC|T*k$wYrzJ34|+i>|M%2|$zu9}*?np`|}
zwF7Wqy)=oJ_HL%uO%(i%&2{K{Z*uMx^MTAT(zLD%IivY6yBBoAGSunF`Zv{Xs(JK#
z)A3ETorZ3L8tY(7dmXg7$L^8gp34p#?zrIYwjY|N%7sh%KE+kCsAg4r*22VRk`tU2
zY+Hr7qwqgQx`U3IS$!kx&aF@HjJ+vf%D=`3m$b#^IO|+%8;3(}m3cgt5A9gt;O_+=
zQd`mVo1+g{Tn(nbj;*IE1@Pwge@}Dm2v%ycg=H5dv1uET5$^ojgPlzk#q#&y<joez
zl~fT;r20$ek93vyjVwrb5=Yd-yMr@WB0zmwDmi{_*7Ie73y#XN-Sy0co4iC`yEdWq
zQ0-n>rh^F@rF^2YjuF}pJ#DUGxm4?lNVm@uL(!431%!frJ3G^>qXYF7iJfS|PFqCG
z^<&ss1E9>8zE`B8l>V|<m4{W%lIVni{xSsb^4x(8@F$~~T&b%z-&mzj_qgJ?4JTw$
zRe##~LO=Iz^&Cy(PsoZwV;#DwEqsbuKtYNeJx{b_v3ZJRAAO=_p?XeXy{OWKKRpU;
zdoL82X{ga;MM|VN9(=M<%wf@^#i0n%XE!o5H$Fi(kzmfr-K|==yjtmD34EhU@9x0O
zSu~yMrIaicrTFM!p+-EHIlK!Q6*x9A1v66sX$BEbgzGk%X_}JJheX(Otf~A_R8r)U
z_QpQa`GXbIL8rXVMj+#jZ(me<{l=qjM98U0d(VzZFVTH;pnWDob|#}x&g;7Wq61nj
zHFd)BSTD<Yp@#1pv~6jT&-_Pt{SJpcmVuSCKrOFk(XV8AXK7`tFf8-WQJ>i*raU(6
zW3=W-Yqef`l5tZ6hOcz!rwtx#nVzr9J2z6;+<-t~0qJ84%dIrMd&N5uVjZ}|>jh?s
zTwW0Q_>723`Pg};TX*dureD;8sUTub2-mgVBT|re3^r<tnv#gN=3T~HUfS6Amz<ur
zCGUBzQI#83qEnQ4Ilqt=t2xn(L>4g>aZB~AGHdg_NFPI|Rg}S2wHlMZX^l{w20Li4
zB{Hk+(^#l?j^icU+dP$Jb#iN&Hv{|>q3@Uv<0*w!Al8D;n$hyR9{7XI^{E*+&@FJu
zG-x&=e*NpHPvwd2c@S;hE@m&p50lJ(UTdJBi!!a4my-@A;;QCrn2Rrp@1>G~cDLTq
zqQK6{!1i2C-zeKg(jGV!Ijw<bPV3eB!mR_HNg9Z{w`IufEk<W*h8G}KE8r9h!NL9<
za`c|JKH7wY1k9YmVP~Q-!c@%j`idO$S$4^R$d0pWY+c}0{=4cOUF%&|sBCr7v}#se
z-ll51jIpJ+&7RvI!~;;h3Ck)@hKs9#%hmO^@jt1F<@qtii9oE8y|<K}naGSOQysRa
zY^D3QJL(?)*(9sdd3}CrdbKX!{dsdvFuJr=w(qM`{`<v=qw4XhRd@--V}^vf?<#BO
zmKJ=1on5)w?Kb*UU<vkJpM5i$o4{Gdp{`h39ItOyXp9m$%mD>~;oO^O(aic2m-=lQ
zKPU6EzfWYWNkspqM4bwe3ixTqf>RwK*QI!Lj?sJ0Fj%lXW|R_|aKvae@{*VSu8E1h
zBN>**zdmuG5v%p1wmymi-9$@WcQVk;Wfq<w^vED5B5b!+_Rud@Z_zyjlGF#W26@sJ
z5YIC<Ubd8lK!Z{&Hw9i1zpi>q(O@{5asJ2Fdg0g`YiYa7Ah$GUYLY!+QHP~wWu6mU
z+Fr?`m7G&0ti%)Uc`GJU^(Vg#y6Zs6mRkgW7T04JWPquo-OOEMra8^&r)j0WeHbYJ
z*;BUc<B_C&XEX8Qw*SrN1o)PJXNva%j~I>SiVs{-x3lE4-NR3>8?nD&wl9)1>F6l9
zhV8oG7nz;Mcg|lWmx_L-2TLyWmX?BU$i%)Wl9z%A3K($W!q`pd1t{UK{fTSzE(#Kz
z9Ia1n3dr@QbFxWKI@xr6wVk~+7sGv>_;{TJi%@rKe0<ad8n{!fZ?EV_^Exiwl0c?{
z`@I|6i^=Us*Q8J|MkC5L8pOIPTxuMyLN9~m_v6V*sctYgU6{M_+@SE%_khy1CE0g*
zW>F;@w;_vsthAd>v1%E(f8J6KrEK4Iwd@%zTMTWdc4gQJz7<~cUG%}Z$sT>a$=~UJ
zhv!@hgO@}cnYWw9pwRj*jvrrb!>p-u=m>w@HErdKyw4Bs42OPW8IWvs04WvdtZ)#H
zMJoK<7>#IV@i`$tvbn2(u+msjwSQ4hFOX^7KbW8WVl&XcM2UnwV#n5D;(AtPlb@ww
zTU%XImtNyX;R+`!3HQ|DT~s$AVb-j?8U{6{*O5HUR-q|)tox)d=Etjo#c917dk&&l
z4a0%w5youv-_^D*u^MvQZ#`s^<B5N{b!+F0{h-oU=BjyRvrTWnRcu*Bo{9#SX_XG^
zjw_o^Fb1*X(Eur4>#rNm@N}|Dr+N-cv*Ceq@0o4;4pZvC{&=0pQv~+Qy^`)jr+||?
z@p#Fb;7UWEShG{6zFCamZo#COy$cxVm{4u|d5(@PQci4AQbwMsk@ty%<~3vbXBu6J
zqqW92f^&P992=|5p+$KqEey;?#g%j=%tVgECCucUDPBsORT>JuEn)GZOrhk-L=F?W
zLV<LV`QnNh>f(75+fA^V&RW&q4@C^!&t1VJ+^KT%P``81Hd)tVg|{36WF$;KGP43B
zk&#HtxC$#CWBE<J40T9}36@e($^Fa_(gBU_SmcheP@OQje*7n&>m_c^UL%{rWar_e
z`-@??b5Ni6ramt1x$}n<M~hSMfizily8%yA$nT00BTCrVZ|S3W!EJ?z(1eNO0dEJ!
z{-w&%7Nq#Xu9)*f;que2bcZjvQ0m$e{(>-Lxg(Li-=8Vh$;Ot0F?pRHW4;H>VMg(l
zQH(pbvn9+GSIH}aEvHw~l|mU$MwCG1;zMFi&Q~ulNYU(4v>x3tnVLPhH*2SxfVxyB
za_Sdp@#1agFv%~~aip*^)@;?iJPq<RuIYvJmp>attesyBR7OLzK5<FU4rTRz@AbmH
z;qTN7Dy^7!pr7*T!6X(IVaCv#(?20QuF78d!oqEwI&^MFb-7wsVD(|$$C2vXblddN
zQz~T=7w^L~&&xR%t#7njt}4v>VAV6bRR`kT77r6Ypg%1N^h=1aN!hy6=Cd{5#v}g_
zP&-ByYvznkcU&P3%AxFBG~x(UTJISbBT~k3UmFSY^yG?#6`A+0GNO~1X4^wF)YO`<
z-HO!upz>Tr`T-d%1Ji?pX(03B($Y*O7uNF~%CipC735O&z<K_HQ%w|53W{&O4)w=_
zwopg2Drk?L`On_j*2RI&8Cs`^G85t8<&8^<U!v8FC*0)a_Q?mMyQSwO6fOjbitOKz
zQC~v|Lx~tzUywI0V!cdNEn|iM#EeH`)ynOvrYEuBZ-w0m6daK&Y=l_recn#%SX}ic
zZd!g*+w3(OPo8l%A!v}{_<pO<sIiX6<CeYJe<6k;%E!)T@=&dBriJxl+>Tq$S3Njs
zWNMeK+Lh?SkSIo?afutHuCZ@{kx|@);qok9a#&hno+*vD2ntgx=&7w+*kBQ<&&m6#
zu}sGT8w;bNCK2q=jFc9z*2O(CAKsFv-p}C}oLL05x%)R;$7O~+v9)HJ)*Vicd|N<A
zjUHECW<mNj?#BTvT7(&r%X1{19YbdLYxSe{oD{=~>#9`k(-h0KRoktQ_(z^9>^X}D
ztA3`OAq$=;F0b$S;*)s-?d%B-Oj`5xZ5VE?_ZMxy))@G*cIVZe?(V!;OG&OPF{}sK
z+sJKlHat>@7co{pCRlyyO30-wL&1jn^HObZ;h_3_bYUVN<7?r}3;}bMZ1IUs;0$!%
zY_wqafu%Or7gaye^X|18J-kQA4P}%I@#cQ=I{kPielO?(8E1bAq-;Xvy+k2pg?eK1
zO~)pLXRFiNF#O|36>x@~*KPPlln-79V5#e<WD~jkd=(GYz<*QDXl0M*H2KuB`(UU?
zVG%9U9NpYK@3Ps&qP5|rQ=764XY}~oJi}sS{r4jFv-j&7M3j@a5Wpq;Ut_uhuI>Nj
zN?)Y<->&rk%N6~fUw5DX%XNPLwfKL;fd4mFJm7Z!zrGIti!1xyHU2Mhe~Z!n7hn86
zoa<i*_v8P}*xvK+aOD4>kNlrsqyGg?iG1%$$9N4rE&TpG+j6i;FJs2Mo+em0F6hlN
z5T%Z|`Febo+BKgdk1N7@ppuGT^g1T8AwT|{*3nWb9`@k2l`)Ex`{2=TOETI<+498X
z$7_iL!2-Qy@4KcT=c4gf>ZDkCx+&H>CvnFa$MTiH&W5C}_Pk{LGOhXyJN)x^q7f7B
z+6AJ~;=GUPF+az{%yi>HGXWX~SB{Q04#S~_Hj_<B@Sfq4fyg(~HY8qE6^*T}$P(S*
znzLTy=>lY5;mQoY(AT#g0>M{aKe<sY*@EAos4(&OrmIWhGPXGUF-asAoxk3z=`ZbM
zVF;gLgNjEozK4Zi)TlZpIz!VP204AlczZ$u>q+w&o%D+|gSwh34Do_M>Siwqgm)b<
za5VD5tweGBLmC_HM@8(Eg=pF7OHzuro=y*BCj^GIyDD3L9}E=#K;)n8;t^dxTS`+@
zLVgLOLf%tEZKftJO<>*+s&mUWO)vH>RbLt}6FkewGOoc9yHY;w*QF~in#E|S@i?y7
zl}6h46Mp`5PaiwQa;jRC(3RNC*%p%_hT^hsG5N9UXAGWOX%8DMansE)zw?=O(|71K
zHiM75lTWwt=i{GWKW#2f;G(YT&5PjUeo(qEU@*a4rX_2;kkd3eK6GTc=hl!qGip-c
zx|+|@kRTI%85_`OYC7%VXzR^*3A{(4)l-NpX%GA{ecC6T2eUf9f|zoNF^S12KGCUo
zS${d?{P|)100)xs?ei-;>^6y!H1hP8U%rdovjSNykU?IOE{QXjmlwq9$D&@0i{H<~
z9)enXa1Oc16DR9G&VEE`iGZHT^pU>VN!6^<?S1N;5kDw;hz@c00}svuN#UAH#8*q9
zBWtrn$)+7?_S{{Rf<}f65#<6goE|>Zs;+}ySoTjI(!g5<S&mNfj4n37tyB4N^Qc1t
zqTsh$n{$H%{4D`;*B*hh?GGQk|3i$kbwu;PM5?%@bTdfUU_N{(UT7xA@tAjhJmhQ(
zp9&mSCSzVye#n9kD`_W!mh(!@+uq7=UP0<j7&&m5D!O4r5jZ&H>)`qq@A7~3gPb&W
z>{g@(hIY$&^~4|6YP_6`6>gOH;Tk8-q`d6;o24jxd^+B19sYne%SU5<vKOmdsU8zj
zDHqVD%;R_Dq6}@Q94%KV!E79un3y!EzF3P%-+OHNN+Fb{4RSr>buq`H)3h#LXdl3$
z^Kr<EZt>j+@6S=3kS1X@W`-Ci7*m!l<-WQOUcs={txU;?*-0)fT~ZZf+M3?UB2S9z
zkdH*sPeRqlQizFn)Uh`jnq#5hGtU<K8kW1%m2yXgqO{c3`|SaBK8Z!Mb*brip3t!o
zcW8v=z&iIS8n^rh0V&y&v_~7rScs>ff#8eIvd*ZAs0P0Evg|96jIi+I;}2q13%{ur
zr$R$JUR}QpWuNEk9I+jPRav0TYjxr_=nmVoHhPWVC-%AxMUh>6TKo1O<Iwx9YPBtC
z-y2-%9d(VK&XG#xnS5dqo>^IqFEef(Ytd*95jjiLuo`99>r@<^+Y(`v$KvwTE;N=N
z&kWZi6>(Lbzl)jvs)A#KJ0tf@PF90ao)*`*Gr|KN&Es26D4Ck3c841M4$XB&88jT~
zRU7IA8OACzY41$wrwI>l3|(T!!TssvspC~RFjNim$!H>>)9C}<-kAj#9Pi}Fv<xR|
z-VHv?U!I~uA5_I{?hkE};d@kO^tqvq@$pmo0lk_?uZpX9%YKh!P)%JqRGOK1%3<=1
zAU1X$t^@;zxXy9tOTLfDw9@=KZ89v8Z!Y$$x4lp9N+ZbFm)DJ;dNkYu{XGXwxH5z<
zwu5WNd}#Y!&9Zow%*fe<xLoa(+C@$L1dl(8I$x^#2xjxM2~aGJ`E7{u*NP~)$*&6|
zdCenM&0*v8Y+l}_7<$0IJeZ6yD%K8*m&bdQVccHd&zN(g>I?bSCMXu{D;L2|O`;Gr
zcAgaTNr_D4$#l>z)`kw&N}pogA%1YwZBtCqj0p2wAulJr+gU_#`6JD{*Kt3iUs^3S
z(0oDs_Tx#vKF*To^-g>#RLt!L8t0fDQO9^z#Zl~fY>=gxdpY!QH$AaEj+=@>7jO2W
zxPc~I^t(M1@8e<(Ba)Ft-`xF0&baIHD!Zge`{NdS8eQ(H#<J)io9^he3TT%39K|gr
z2g##4Pe~s5xh6rElRC|JHfbx135Q?@JWmu`p8N^5*Nj#bo=($H|8D=iEs3k<BWc}O
z_Y8~DR^B8FzN%Smq&6%br_P!yze?4Go?^$^5_P=*EXt<ybfR|7R!FRY)PZ28mGgs0
zyRc>Hni<K;QJ3`?$eeS~*HhDIb$p+(-7Xiy403)6*A24b*3h$h`$eSpJbGJV^3x+I
zJy;8@YbKo?pX#<!9e-BW%0Cr~o>Us1S=$xxOG=f{h5?*o*qK^q8Stdrc^fJq<K^1%
z&W(7&ICyP<Re*HFyJ34<Z?`9?-D?iQGHXL-9<L#zuS|SAMvT6}ZX(hI>&lP_I3t}&
zzT(Voap71odX!Q$U?H1_Or_x!&X==bzLk<eRIWdvOHnimGD%A4dMYhrX-vb^m-S{M
zY+jD{1jG9o<P*^jUA2~5{HYCQWjm{?xu%T&L{(Dc8_!3BF$*(SoDT7N)YjXh`6ZP;
zsSR9S#IfQU@xs5UtgBt$a<_@iteDI7pzqj53S9J`*7X>wCwhewv%c@=Ir0VeP)pBV
z*Q3>!toP~!rJIfi=*z?tfz&?67|ovrJ*UhNL+;mSJo`Oa;H87Lo7vzK>3-LC;nez&
zx$;JT?9;(Ie}UG)W!?U4laS&~D>!mfYoo}nt;do)uVxJ=ul*ByOY>8*&zr8&85B1W
za_z}C`zUb5hj#Us#~coi@B}CN3#{exMY&@?hWT()wWTsY3=GAg)oim$)zE-Urt&F_
zV>3=SK;u(21VH(e1;h(IJZQ>fsr6-<ndTE!vXQp!y=vPdV!CR-q-VcMF2rdv4RP!)
zR;RW0i}guWG3{@ks%cJIuYU+4hj)5$<d3U(TTH}_;N^|A*xJ6*6>PjZX4aV41qD!}
zg(&coTj^XuQfq9=QC1k_sVKjtH=xSLv8|eN23_%;+yAoejsW7VTa*NiK~zN&P+s<m
z8aR)2OjXgo<a;QV_Z<6?6$h=dA$@1lQcH7hu_U#+Ng6|KeA=@)-@X1^=wjuMmmzj!
z{7q0jX)dx6#q{9X!M<wFvsd!2*4;%o&y|}pE#xUqyCYLFjYL^Q6)a&aHc2aurCoLQ
zD#N%f44wM?azFJF_36Tj3&WJ8ggeR@#N%~kymj$9)e2$dL<&+o0g^L1vuCxHm9N%T
zPqhou8y1tv@OQVqdV0``2lJco*blB|vN-quh=m<NSWNZCD)#>D-x(m?l0xAaNq!MY
zMdbM@!S3{MB3cyW1=Lf9IB3+z4}-DPT<yQpNgR$hIBMUTgj58HcOkp&UK@CO**O$=
zDj%g-SzjlaRl3+!cONygQI55*KswnP>{M?5{Pfs=xi=j<;3XaC3ZFHp$*~lU3>&t*
zSy8Ex5ZAp!ry`OJ5<g)FtBf#1cusnB?L(nWA?|1r$9C$^<_nX^+tRtuU%NL}pGgoc
zFW|d`l|9?n{k;B)i}cMjWUVZ}xE^JLK*~Cnu9pSA6*Bq&<yYY^)g3OW#{_5Vje65g
zZ#-4jo?uI;%uq;w1eenCYL&NFW=7NGKodmDPRIQ8+E|^KZIdT$*No+%=`A(o?JJ4I
zp}L_NdTP^dk2bN;KwRcT)VIH(5-tKV0&DQD#rXH)9r~Zd)RfD<$Lm3FL)#L$G1<Ap
z#f4x;=9`X9adt-e3knkpVK*}UC9_i@K?yG83Ve|&?VUNRUDab<&#*zHYg^KU3M*48
zmP1j2vKP-|t(s@~re`E;@F_YdF*R=_wziD9TAF=vu0F0-(xqzPq7Vk#v09f%oD?(4
zhYbuzMPuqd9w48HeF4UZ-607rwmjqHC#Jad9+rf8`9Grkl{97>mEfluE1{@|ZOX=C
zwFK^Blx1&J+BPGt?I@cHsVW{+$gitwt&_>AtqbH0sj|rV`Dv*Kh(v)zT3SOoC9Z4m
z?1)z;T?3P?2*ivomdC2hQkAQAxGNQ(?d0)hNR*(l&VaSIk*BVyb9bKh^Tj~cs@5OI
z@*N=-yso`Ql89Mh<y;@V$QmE#2?5N7-J3`B>s?V>`3o~$9?}FNO4TKG&kgb5rFKWj
zn${#(v+&3KrI3(>Cdsbg1Iaj@m)qBZm@$=C0_R$4rQ9Povc?^tz<e{!(IuzT`ah90
z^voWuR%F$K<hRXtZvq{?8gYeDh81haZ1;<#I{M|jH4b%1Fp?*_M6GyBn#@OTSZvwF
z)pBCg!kfn3fE<mYIrEpb;A)fQNM(^gBf^xazQx!|^;Gf*hCBTl_rinTXx%`5rngnk
zaD*>baj>jJrv^iuj&$mb-Fzz_dSpL`$Ql+7ztrq=-`aImVdB*By~aNlWlK|7;I4Hk
zO1myXwJOq_nbrdLiZq0-M7qIW5I?185Rj$8;MYWbl!30QxMOn;{*7NONuxf6H{eY{
zzGhft7kXk_$gb{}q&D%HEcd8}LA2&$C@6$4f$QArorZs~p^VLhiHYz$X5|dFI?<C!
z+rSe{WKuTmpZp{^idAt|HiS`6@~Fy*2YI{?P{GB>a!<#Y#51r{F^TJRmIqSg58xD5
z>TsEZ&=zUcG;yJ4(0RYtGNFOtd5!$-wcbj`Vp&6ojr7xQm>r8<VDa2Uj7zYWaACYy
zQ4CgCEPqheht{E7!cjBg!zo9N-@n4wcdw}Tzs0WdZAf>`C$>5CW%9H)j2X05W}2z0
z4lpKRU+aE=2Q2)tio}Jv%C21vcoRL}I#n!Z6(Z(at63N`g?FmvssEaS%-mEhq47F~
ztK~u`ro$T}wK)0G15ol93#>lPq`(tWJuM`9(Ytihl7!3yeQ6*#wIfAZT4A=!^&ykx
zIg~Ftuy&Xy32Z;iFEVP$K8y&7vc8@kzap^qD7mWcGxd)Zsj_Ev^+EB-oz^a@*4sDH
zUAG#knu8lHG*d8K+Pu~4&`roNc$X?bsbU0j1x$W+wQWI?$Rt&@V#oU*LFXr>y&hx>
z0lLQ*dds$kSy6XSMa)*AfsQ3C3MMwM*ZVCU*O#?1HrJXmzqN?oJo7OjmyZ)GX0|b2
zO)WhWB~r7r+#Ok4Qvyxz$88O5)P5GYQP1p|<g8q=1$~<7>&OlzM_;Q^r{=H}^l<@6
z5{Ohj|23aP_g#))3RNCmPxMB|IJ&`60QAv`I`hRfdwFmJw7NburprFDi*h5C2B&U7
zYKvXBaxQ)h50A4bR3_Y6(K;QZ>TEI0$hX7TKHu-;XxY=UG^;gv7mjf@6?8dN*YnmV
zSYxCmbw)WUu88LojR3DJm!+Fpsq!{5k_YtEV+s~}QsZe8UqDpkERGcyM=yNoBqg3q
zxnNK`daki#Y1seO@idN;j#bMe73Y(Dn}C9m<2#Pu0tN2IaE)BL`t%poTU)=qKE202
zAj;Bmr*!6X;%6XBAUSrgN$$yrFF)!SiRwJv3t!azX_51izCuvZoP5UU&gyh2v774M
z8!R%N$CFJ>3k9~p41SK7IPn5@vR-`__UsxRh4njD)hFvZ9D^_R&vkFu@*Ha74)<tx
z+%z^8S6<|$L|M6P^;eV@R}7Zc#OV`gN)3f8YWsw%6YoN7)^GTeIgVRI7#;FE3wI0U
ziodseHlG*5qw7X4fAyr>x_PJSNp!2MERLtB$GPvL!ilSP$mTrGj(YOG^87F?N?r}J
z93y+E#!E4NU4;?kf5Rl9;ti3BgTAobIltyjYg4OOLGuifI2aid=N|W-n4CT6Ev;>f
zTnQ2<YuB}A$VyZ#LtUV;`WU4FUf-w4Q*l2Ra|0X4mc@dTix=9sD{m$XG!%;|jNv+E
za~;*ml>A>?!bP~s;0$8!+f}Ylre77*x<KW&7Q$fkAM*>hbKBvZB0nn5nr*l8$3J;I
zul{!O+-uvdsNmd9dGXD9u^JGY;^7*sG`YI(^Yk94R@Vo`HAS$FlXQLNW={K@zE8+p
zX3OYH<5lq-E!CP{nsh)zmh(uPEfHr-VM+@jgQun%9i(BACVAOaNyN*Eb$nRMaB=$&
zw-(y1<$~5vccn|Om9e7s#G=QB26JmZmBN;^negpZWEj^9@g}KLx0;<Rc}w%L9#!Lc
zQ8K@Lz<It9)R)$vW<5wgUjAyw<|Td%OU;9V3u_t!?mxcd#IO0r?Bh0-c+>o2&8nZr
z(DXHhKFbg|X`}b%<r&e+70f>z|JBzpt8;0&TS_ODCb#xPtaFI1WoO5yT2doSx8FyP
zWmbpPvCYH`9GFGSTpy|D%9D~HXI}DzpQ|<N!8K*$f<bEU@{isHsiUV1xT?=_sWR8M
z9rP^9q$5My8yW2xw1;PZL*7(ie`5)$uMv0>bQrUr?n`T?Ei%vGDXxkVHg?eSohGVm
zd%Fl#{<#EqNFSe?w&Su-EJ+DJE{~m1U7pr^Uph8B(!gZbkJ)nKT@T27)JeclJ)?!}
zQ<Az@Zy>-fm*0W<We+jEgP+61RG;ZfeHts&;9&Z2|J(g+FTrQ197gJGy<Jt<q96T(
z{jrRn$iHMrueMM+&MWk&t3E$^REDznN$Wuf#&_Q9yzJX9TfTg#JMSz5<x2))lVu;<
z-xNG5u&}EN|HblPg5fGUkcF(ODI6Qsn9jPEijX!rae8u;CUCBPWW&ACXuOB<(H`U6
z%65G4tm)6&+X9I!lgI-I2lQf#V9g_Q%t1i!R>}@S!6m)SSii;_Wt}j$J7Qq?bCUb@
z4<^_mC<*#Qn&N?B{9B~o=_zN#-F6eE5?9uNXA-6g>D~_h19x9kr(Ie2e%Ii3rNUD0
zm`=+1A5Fn%ec!{}UYdipsTR!De(yUUI~4_s(0!${p+@#q?|6*CLC0v{XEKsGd;Mm#
zW?)Nqv}5pDBizjCK&z=b;?Lxs4Z)ehC-WY20kfgklf{~E7lOZJkiz%Mj<wjilfx?o
zh}OHI5<B!|7SgeN#uD`wVQA_la+_G6EJL(y*!IjP53$KTv-if{ezT4AdL7ke-o+=$
z9Kf7j&QvyCA!a5#vhBmu6ROzXo7Z(NWzNU`_+5RG9YLx``nA|;(kKwUdMovv4=?iX
zfiDuB(ywE+-jrIf2iKu6;zLaHQuWkWw4YtIh`f|yRNkd~s77Xx7)PY~<RhwxA_0BB
z6Ka{!S|j%>C#2wJI>RrFDP2TA>>r?_yGm$0C&KBf0!dID3Mz4^yoeObY`3i+OzBcA
z<7r71ZGuw2v&Z;y;89C)*n4FDGKPqGb1PV$@LkM1$i^FP)z<lXgXBLPOIhMrqtkt5
zd*ro88$9@Fvo>obB-XoI1TtefQmff4sZ}~-@nGn@o9ts*H#Mfg!@!tDP~WR~DHZPF
z*e)ND&o9XFklv%>(>P$r2s$BuRN&%}Lahj%dy{$f(rm{SY@5ixS@(V{)2tev0N1Vo
z0|~FB)>wcn5qWQhm*8Ww^5HlE`2Bl{z(GvH%QC_s0SOdX^H?Tq1NxxkxM?*U%fCGv
znuenRe|qC}!gbN|YO;a;V&Ki69&_RMgm-(4ezw{nu3)yIu|9rxxsgMai%0w%0q;}K
z`B*~g&geTEh4ucd30+@2!-Zk?x)sbASj-rxoGUyu%4lq}fFC?IIqMOgdWuRDT0b?P
zsQ5sQPJI+_hgh9%rR|-&9z)KPzywJ4a7}VT;Q5O}Pt1#YqmgRNAbz^Wl+1eT;-8a(
z;`I7@O?WI1JN;{)ro|{Ck!L5{q<oyrH)9N8E>a}IdA0U6)+}DQp>AtqQ+$B&*<IVL
z8p~dT#esAl?`SG$yN;qhJk3}J62^O>6Q5eN%C$fsrAhbp3lPbT8**6YKhKC+68nh!
zYHPSYzw}Q=)~cn&?#<5F`=*yeLw3`U;#mXjy~Uey#q?QFen;=9eg|5d!xj7^q@0QM
zaPWGguk&l#U||HT)U(R0cd$Hx?2L4xc9`y?MvTRKS6R6kr>z>Ee3IN!Hz%0xPI-U0
zlY6_V$ud`D;Zc*su^*5;%5`>19|h}_bb#;mCsa`^vDq#)SCmS02cB4(AAUt*?B6B=
z2Q27{l{1V&duPsu7?@*I*O5s9ud3yk|3{2adF~PZrI3$(d14S@!R#kr8ZSWoB(<<1
z9z~aJON%6zWI{{{O(1zyI!k415Z1FW<5xjqU%P1rruLUl|M189FC6jJ6|cVM*ulv`
z8Rq`Qdy*LxIEF=pTj)3MbjjZlo93#^QMGyl`of_sEW5CMwWy|A(8Jp2M25=({rO6N
z=KDgV4wdSycgF=y^x&B-{Tw4td(Y-T@7k~PrqY5x&=RWIwiCAzSivhb>pcUs%&K52
z`EY5GXjSVlUY_pV@@)U2ykyzhQb}=DR%W_b%@K^!LIy9XQR^Su+_9$Agep~WGlWi6
zOd{jwh6ReWqAC#}S5;&B%BmXK2)xe4lHMM0kr|F4rrBtlE^}7&$aAcV=c-D^c}9wI
zgO6ZxU#F5+U47BiC*Y{YD?<I(rg|uXY-?8AEAy5Pia#SXOQ{*yVlS93g+(UFSw|R@
zI%+EO)Zlz=^)NeP8m+{d-M%dTZoef92A|Gg>N75yw^rogXqF}Wej5iuJkohR<mw?$
zq(e{%N7Y5*ELo!tNXW_8$tOQcT6}ML>LJ1#Cg0-2TmP9bt(_`2zbJ~f&ZZ;7RP3Oq
z-<lIYxs|YBfxse6WSU{YWh%P-g^j?YJ%)+LHN@i0KH9&3yZbL9hcQ{lpw1EDIcHYe
zhg^A1I2djrqXUPk3cEQp>y`7%PG#W}mg30DXm-~0O%WZ$MmGt@4Og7|oI~X6McBs5
zW0baY&C$EM-qhQgi*l=h7;mlIHHf6#=Y_^5-fZ8D{i(_U1$M4;>pKr%znF8w&T{g8
z(sQX-JdYetk<)HQ)>``2IMdDcnQ_&$>Xg$ITg0-o01N#CP32rrOunkpiKKJU)IgU+
zY7G1`s-QB=9(xvtpCYJbACv@aW-r8FFAe^B9{pA)t;4#~8{TX>4<4lr8O^>@%(JMF
z$6sXiNZiB|d486PLiYkJf66FE6LRnz)pM(i_%rqwLw-$DtkM^5PYZjvGJ{Gm4Bb4@
zUW&bW8ggP76EPfjH8u2IFvwp^IZq985tJJ;(=fg*o1-IQ^@zPP_p*M|LewYDI(3nU
z1qWXB$7)C<PJ8=P+2Pn{TPzpM(!%28P@1q;j*|>=R<5TN;q`4qs!?q<z45dld99Bx
z^tP5)AE!)I2d(}*;dLT)o=EPN;wG1pmgt<xt(&+|@^|JcwVW~!JFbX>(<$qK!RGY7
zyUoci?l5>Oh+@(&gy>|UsM5cpt1b`co7yh#*TT_uqwgZ%E6gKaNI0rQWtpArC$sD<
z{Um=%(6ySGZB(tdY#N`pUHCAGXxrZU&Th?|vp`A%hd0)YJ4HwW!hsGL?rv?1gaMrw
zX)HPN;>*lTYDG#4BQ}%2Ms`}Iex5Q0on>ad=38~z4f)nwdpsG^lU3`cPv4_iqe^(A
zU^(Rod|2@{UX(}2#a_+*;rmOUlLdQTuQFwj><dVe)Fh?+(A38A7ubVbLA7R8P7ZGD
z&IZk*<Y<*YHSc0VinhbHh<r<)J#(xUMjDaDJ-e9FU|3;f(X(%XRrSrE`@w9@OEX>j
zSyYuBXS-THlkB1d>_yPmnbm>ibcsltro%keHq8@m8=36Yf0O17ya<qtU~`Y~9(|$F
zuCFji)>}y<R-&u2U3^L&velwxJ0zXU*!nuyeOPxsht43`Z5I5b90i<Wlji%b)J3W{
zppX^+eRi7W7G2}sXB5IKa$U~y4}Km=1Xjpz1qJw6>zq+|w|QY+@Ilb6)Q=$Z-PhE0
zsal1x32^d_QxD@EJK3gT<;`5R!M9)2*w3r!L9boU)RSs_xO7Iv4@rlN^#ns+lQUEM
z6;Nn2()d&|@KW;(tO&Qy?<9-!rg9M-?HZbwkx_@5f;M-RUg+*Yt+X8K`JD5*)CVBU
zlKv^9RJpylQCFU7<(u?)U6kJ|l^UPl)Q%SBN)xZ9XLbePsZZFe-D1d<Wv!xuTlB+|
z%T)_Ptp&T8;8x{!ih@cl?d%I#7~9)HVn+~=AmBK5u(h{02ihmL=9Z>^cO+v1t)`2U
z5n}$S%irb^-;e*@F?|>HcRVf^Cs#XUB;3Ei{;PxfuMTpOdxtY{XyX*{I6kT4m)ewc
zgTrgG=Tt;052-_~<iF!f(;$76Q)Ua6qIkxc$n7TfDPYiQz=BNLjf&%ahXDBe+&a`z
z;~l?u!1SIFBw!~^GAaAEA$Ya)RD&L;Soca-T_nfK$Yq<OVWn7xY!$NpQp<D=``O*i
z6LD{8MWdq$O9#A8(kCXqzB!IrYmu0T>eR8_&06|%3eK}LmK`qJElAU0INwXEz%wO+
zu|5xE!Rf-zxp8M?gEJTe!8}->9Byg(z^?^Xe7Ff>QB{f-S%B{|<2ye~=C$eJi(iq}
zGU=J5qs@0uD6Yw>(8qK2|BeHk5+?bpMa6oymd6H}S-r6&?sLwTtNCI(cY{enSY`Xq
zxlB6f3vNS_`HnUxRDyXeLLHqj>DF}mvyAxl2Ff4*>`s~4fVw$7)DG}j!Ih+0_q8hP
zE;0%}pi{dlvSB=pfmosMm0H0Jp^6bBY2KknO+(#!E;IS8v&~!LDEz_Np06)#)dX3U
z_O_w(ya%wtemtB+4f~&E-AC>*yc9kjhuq2&BL3-V-d1NCxVpLq67Rj%T5|iTQGdaA
z&RyuNnIH_ZSHGr%?<7`e*6S=@)`dsokA4lQJ=|Lz*GzSs;q0o5b!QF5`H`p4#I46J
z1KBJz$7)HVBv(>{u#hQ~zW~m3MLV3YoWIpSuqBhyRlmiwcDFp<m~cza#j8hJA%${d
z_|l>K>oiQF*!_yirAZ5rCLcw`rJ^F}1bM_gMP{W(F~phCz~$IbiI!m!j=?-Sa=IXM
zc8l-H`HF|0Zyd`H<BJYAN&9J>m99tu;h-?84<l44h#!oiWFVH|i?t3n3z_=J#R9I`
zRXhHTK1wNmLL~Hbs)^Mm%&Na2CzZ%-20~Y(vQ?QQpD4JZ*4`e<9U{OWG~`T_Ut|jl
z?b&y&cr*Bzm|ewMd1_y(coxhxyt(IS-r_oMqVSA7CB1Z7OA{_ng!{0eMpa95EQL~m
zbtbXM1PqqQ3FXPRAuKK+bg?$ClmV-zSOT4x{1mI~NJ)rwet=A|<L0lGm<^fUPG-wr
zd-f{DR{0U3tm6ZI@nBWa990$Rm^?L=pW)<*Da;_SIIXIxN-tPmrC1?9-|9ISl$}5%
zs;VLftzLd+GeD1TX2LZxZ5AW9tTHPRt7;VPz;zC4UklsawW8&*6`abg<g6-Dl~$Al
z;Q(`%j|vJ>3HM5OcT|(LV2XoNgM*zhTydwz1vaJ@Cc>sOUY@EAPvy3wC}HaKgHzpu
zogaIu%sW4u2=LdStB!O=5{s9pm(c-0DOlx-i5L)-%;X&`Jw4pv%)CRAk64nxC>(zv
z1O|Fmy1Q!~M^{yX{|gWRDorlS-|>Q2lCRfDOHt7m2??DAFhg#DS+TKo`<og11%tOB
zDiTuDM<nEp-$;OwLD-ppmVX-nv;Q#`{wLyp?Ev8?(4+wq2LYhEE|7bk$p7N`cT<Pp
z`<HUh;y?KOZ@#E&wx(7_010Ya)BnCMaQi(2HcL|*r+=_=ws*DrM<3Cu0%KDtBRf+Y
zGXxH5l&zKBz0G#E{in@F;L(}>ZT_jLlNkbzC<t*MxjNZU|E-Uz=Iml@1N2RuU7Y`6
z?-4j^whmkfK@kNp{#!vtx<70L+yX>-W#(+^jOaZ0`0*nZ(4qn@p)e5=I`D4TzprS(
z3<_QlIx_ICK#p>Dak<9_P+tKRpo@4v+_#979Df-<I0nYRHUESoWB_7-{9}wbX!Tco
zL<|xH@PERw{sq6s|0_QPJ>Vhux9k5x@4w;$%ltQd1ik+sIsU@u9{<sQ(D{Gy{R{s6
z^?zB{f7Q!h_}#;S*fRHWGXeM91XwqsAl~H&*>V5gBHsCbpF_R{@VtQWKMK%CfKZr$
zN(o$c1yn$PC;%-)Y86C!7G&W5As+x0hy#NRq)S0w04f3x=%Y{r?JuALKc;&YK>8IF
zT41~mR74*isDK>Ehd`YIDnJ8~Vg+$<4RM(5p8g%s2kN~X2w8yV5`_tvL-5`F8|<Dp
z0v}il@(ECH{)TriE8gF60DMF&n|s=qKp#*H3JFjNfjNXcM}LO{>qGGSCr#kq-_rwh
zb}t)ZEr4F{<pJ&wGC=Eoy}&$z-#_F=q^h~+2doXDtNZl=>1OVAgwPGJPQ(GhdwCEv
z0fHz99*7F)^<G~HeIR58)^;xs0uKQPXy32xA27r{LC^vYDBjN_Xd%W3ogj3E7z2k1
z5xPU}`P=+Ib&Qw;u1Df>b}_LBT0{Z#_yD+vv`I)vBDF|JMu0v($s!>k()l3b5g`sz
z-v1Z0b2W7_0`R$Q&F)*owa@=*&;TRogqZwm!LYJ%b^#{X5&txS7I?Uja5dzCG&3gl
cfE}|?v$8b<rhs)KJw_thQ5XY<qRb)xA4`bRt^fc4

literal 0
HcmV?d00001

diff --git a/v2w_service.py b/v2w_service.py
new file mode 100644
index 0000000..18af6f3
--- /dev/null
+++ b/v2w_service.py
@@ -0,0 +1,95 @@
+import whisper
+import time
+import datetime
+from flask import Flask, request, jsonify
+import threading
+
+app = Flask(__name__)
+app.config['JSON_AS_ASCII'] = False  # 确保中文正常显示
+
+# 全局变量存储模型
+model = None
+
+
+def load_model():
+    global model
+    print("开始加载 Whisper 模型...")
+    start_time = time.time()
+    
+    # 手动指定模型存储路径
+    model_path = "./models"  # 您可以修改为任意路径
+    # 根据实际情况，选择使用CPU还是GPU
+    model = whisper.load_model("medium", download_root=model_path, device="cpu")
+    
+    load_time = time.time() - start_time
+    print(f"模型加载完成，耗时: {str(datetime.timedelta(seconds=load_time))}")
+    print(f"模型存储路径: {model_path}")
+
+
+# 在应用启动时加载模型
+@app.before_request
+def before_first_request():
+    global model
+    if model is None:
+        print("首次请求，加载模型中...")
+        load_model()
+
+
+@app.route('/transcribe', methods=['POST'])
+def transcribe():
+    if model is None:
+        return jsonify({"error": "模型尚未加载完成"}), 503
+
+    if 'audio' not in request.files:
+        return jsonify({"error": "未提供音频文件"}), 400
+
+    audio_file = request.files['audio']
+    audio_path = f"/{audio_file.filename}"
+    audio_file.save(audio_path)
+
+    # 开始转录 - 使用 FP32 避免 NaN
+    start_time = time.time()
+    result = model.transcribe(audio_path, language="zh", fp16=False) # 主动降低精度，使用 FP32 避免 NaN
+    transcription_time = time.time() - start_time
+
+    return jsonify({
+        "text": result["text"],
+        "processing_time": transcription_time
+    })
+
+
+@app.route('/transcribe_text', methods=['POST'])
+def transcribe_text():
+    """返回纯文本格式的转录结果，方便命令行查看"""
+    if model is None:
+        return "模型尚未加载完成", 503
+
+    if 'audio' not in request.files:
+        return "未提供音频文件", 400
+
+    audio_file = request.files['audio']
+    audio_path = f"/{audio_file.filename}"
+    audio_file.save(audio_path)
+
+    # 开始转录 - 使用 FP32 避免 NaN
+    start_time = time.time()
+    result = model.transcribe(audio_path, language="zh", fp16=False)
+    transcription_time = time.time() - start_time
+
+    # 返回纯文本格式
+    return f"{result['text']}\r\n处理时间: {transcription_time:.2f}秒"
+
+
+@app.route('/health', methods=['GET'])
+def health_check():
+    return jsonify({
+        "status": "ok",
+        "model_loaded": model is not None
+    })
+
+
+if __name__ == '__main__':
+    # 在启动应用前预先加载模型
+    print("启动服务前预先加载模型...")
+    load_model()
+    app.run(host='0.0.0.0', port=5000, threaded=True)
\ No newline at end of file
diff --git a/voice2word.py b/voice2word.py
new file mode 100644
index 0000000..13c6bb9
--- /dev/null
+++ b/voice2word.py
@@ -0,0 +1,40 @@
+import whisper
+import time
+import datetime
+
+
+def format_time(seconds):
+    """将秒数格式化为易读的时间字符串"""
+    return str(datetime.timedelta(seconds=seconds))
+
+
+def transcribe_with_timing():
+    # 记录开始时间
+    start_time = time.time()
+
+    print("开始加载 Whisper 模型...")
+    model_load_start = time.time()
+    model = whisper.load_model("medium") #
+    model_load_time = time.time() - model_load_start
+    print(f"模型加载完成，耗时: {format_time(model_load_time)}")
+
+    print("开始语音识别...")
+    transcription_start = time.time()
+    result = model.transcribe("dingzhen.wav", language="zh")
+    transcription_time = time.time() - transcription_start
+    print(f"语音识别完成，耗时: {format_time(transcription_time)}")
+
+    # 输出结果
+    print("\n识别结果:")
+    print(result["text"])
+
+    # 计算总时间
+    total_time = time.time() - start_time
+    print(f"\n总运行时间: {format_time(total_time)}")
+    print(f"详细时间:")
+    print(f"- 模型加载: {format_time(model_load_time)} ({model_load_time / total_time:.1%})")
+    print(f"- 语音识别: {format_time(transcription_time)} ({transcription_time / total_time:.1%})")
+
+
+if __name__ == "__main__":
+    transcribe_with_timing()
\ No newline at end of file
diff --git a/whisper/audio.py b/whisper/audio.py
index 826250f..8e1f594 100644
--- a/whisper/audio.py
+++ b/whisper/audio.py
@@ -1,5 +1,7 @@
 import os
+import sys
 from functools import lru_cache
+from pathlib import Path
 from subprocess import CalledProcessError, run
 from typing import Optional, Union
 
@@ -102,7 +104,16 @@ def mel_filters(device, n_mels: int) -> torch.Tensor:
     """
     assert n_mels in {80, 128}, f"Unsupported n_mels: {n_mels}"
 
-    filters_path = os.path.join(os.path.dirname(__file__), "assets", "mel_filters.npz")
+    # 使用 pathlib 处理路径，支持开发环境和打包环境
+    if getattr(sys, 'frozen', False):
+        # 打包后的 exe 环境
+        exe_dir = Path(sys.executable).parent
+        filters_path = exe_dir / "whisper" / "assets" / "mel_filters.npz"
+    else:
+        # 开发环境
+        filters_path = Path(__file__).parent / "assets" / "mel_filters.npz"
+    
+    print(f"filters_path: {filters_path}")
     with np.load(filters_path, allow_pickle=False) as f:
         return torch.from_numpy(f[f"mel_{n_mels}"]).to(device)
 
diff --git a/whisper/tokenizer.py b/whisper/tokenizer.py
index 2af8375..5842f1c 100644
--- a/whisper/tokenizer.py
+++ b/whisper/tokenizer.py
@@ -1,8 +1,10 @@
 import base64
 import os
 import string
+import sys
 from dataclasses import dataclass, field
 from functools import cached_property, lru_cache
+from pathlib import Path
 from typing import Dict, List, Optional, Tuple
 
 import tiktoken
@@ -329,7 +331,16 @@ class Tokenizer:
 
 @lru_cache(maxsize=None)
 def get_encoding(name: str = "gpt2", num_languages: int = 99):
-    vocab_path = os.path.join(os.path.dirname(__file__), "assets", f"{name}.tiktoken")
+    # 使用 pathlib 处理路径，支持开发环境和打包环境
+    if getattr(sys, 'frozen', False):
+        # 打包后的 exe 环境
+        exe_dir = Path(sys.executable).parent
+        vocab_path = exe_dir / "whisper" / "assets" / f"{name}.tiktoken"
+    else:
+        # 开发环境
+        vocab_path = Path(__file__).parent / "assets" / f"{name}.tiktoken"
+    
+    print(f"vocab_path: {vocab_path}")
     ranks = {
         base64.b64decode(token): int(rank)
         for token, rank in (line.split() for line in open(vocab_path) if line)
diff --git a/whisper_service_deploy/v2w_service.py b/whisper_service_deploy/v2w_service.py
new file mode 100644
index 0000000..372a460
--- /dev/null
+++ b/whisper_service_deploy/v2w_service.py
@@ -0,0 +1,94 @@
+import whisper
+import time
+import datetime
+from flask import Flask, request, jsonify
+import threading
+
+app = Flask(__name__)
+app.config['JSON_AS_ASCII'] = False  # 确保中文正常显示
+
+# 全局变量存储模型
+model = None
+
+
+def load_model():
+    global model
+    print("开始加载 Whisper 模型...")
+    start_time = time.time()
+    
+    # 手动指定模型存储路径
+    model_path = "./models"  # 您可以修改为任意路径
+    model = whisper.load_model("medium", download_root=model_path)
+    
+    load_time = time.time() - start_time
+    print(f"模型加载完成，耗时: {str(datetime.timedelta(seconds=load_time))}")
+    print(f"模型存储路径: {model_path}")
+
+
+# 在应用启动时加载模型
+@app.before_request
+def before_first_request():
+    global model
+    if model is None:
+        print("首次请求，加载模型中...")
+        load_model()
+
+
+@app.route('/transcribe', methods=['POST'])
+def transcribe():
+    if model is None:
+        return jsonify({"error": "模型尚未加载完成"}), 503
+
+    if 'audio' not in request.files:
+        return jsonify({"error": "未提供音频文件"}), 400
+
+    audio_file = request.files['audio']
+    audio_path = f"/{audio_file.filename}"
+    audio_file.save(audio_path)
+
+    # 开始转录
+    start_time = time.time()
+    result = model.transcribe(audio_path, language="zh")
+    transcription_time = time.time() - start_time
+
+    return jsonify({
+        "text": result["text"],
+        "processing_time": transcription_time
+    })
+
+
+@app.route('/transcribe_text', methods=['POST'])
+def transcribe_text():
+    """返回纯文本格式的转录结果，方便命令行查看"""
+    if model is None:
+        return "模型尚未加载完成", 503
+
+    if 'audio' not in request.files:
+        return "未提供音频文件", 400
+
+    audio_file = request.files['audio']
+    audio_path = f"/{audio_file.filename}"
+    audio_file.save(audio_path)
+
+    # 开始转录
+    start_time = time.time()
+    result = model.transcribe(audio_path, language="zh")
+    transcription_time = time.time() - start_time
+
+    # 返回纯文本格式
+    return f"{result['text']}\r\n处理时间: {transcription_time:.2f}秒"
+
+
+@app.route('/health', methods=['GET'])
+def health_check():
+    return jsonify({
+        "status": "ok",
+        "model_loaded": model is not None
+    })
+
+
+if __name__ == '__main__':
+    # 在启动应用前预先加载模型
+    print("启动服务前预先加载模型...")
+    load_model()
+    app.run(host='0.0.0.0', port=5000, threaded=True)
\ No newline at end of file
diff --git a/whisper_service_deploy/whisper/__init__.py b/whisper_service_deploy/whisper/__init__.py
new file mode 100644
index 0000000..f284ec0
--- /dev/null
+++ b/whisper_service_deploy/whisper/__init__.py
@@ -0,0 +1,161 @@
+import hashlib
+import io
+import os
+import urllib
+import warnings
+from typing import List, Optional, Union
+
+import torch
+from tqdm import tqdm
+
+from .audio import load_audio, log_mel_spectrogram, pad_or_trim
+from .decoding import DecodingOptions, DecodingResult, decode, detect_language
+from .model import ModelDimensions, Whisper
+from .transcribe import transcribe
+from .version import __version__
+
+_MODELS = {
+    "tiny.en": "https://openaipublic.azureedge.net/main/whisper/models/d3dd57d32accea0b295c96e26691aa14d8822fac7d9d27d5dc00b4ca2826dd03/tiny.en.pt",
+    "tiny": "https://openaipublic.azureedge.net/main/whisper/models/65147644a518d12f04e32d6f3b26facc3f8dd46e5390956a9424a650c0ce22b9/tiny.pt",
+    "base.en": "https://openaipublic.azureedge.net/main/whisper/models/25a8566e1d0c1e2231d1c762132cd20e0f96a85d16145c3a00adf5d1ac670ead/base.en.pt",
+    "base": "https://openaipublic.azureedge.net/main/whisper/models/ed3a0b6b1c0edf879ad9b11b1af5a0e6ab5db9205f891f668f8b0e6c6326e34e/base.pt",
+    "small.en": "https://openaipublic.azureedge.net/main/whisper/models/f953ad0fd29cacd07d5a9eda5624af0f6bcf2258be67c92b79389873d91e0872/small.en.pt",
+    "small": "https://openaipublic.azureedge.net/main/whisper/models/9ecf779972d90ba49c06d968637d720dd632c55bbf19d441fb42bf17a411e794/small.pt",
+    "medium.en": "https://openaipublic.azureedge.net/main/whisper/models/d7440d1dc186f76616474e0ff0b3b6b879abc9d1a4926b7adfa41db2d497ab4f/medium.en.pt",
+    "medium": "https://openaipublic.azureedge.net/main/whisper/models/345ae4da62f9b3d59415adc60127b97c714f32e89e936602e85993674d08dcb1/medium.pt",
+    "large-v1": "https://openaipublic.azureedge.net/main/whisper/models/e4b87e7e0bf463eb8e6956e646f1e277e901512310def2c24bf0e11bd3c28e9a/large-v1.pt",
+    "large-v2": "https://openaipublic.azureedge.net/main/whisper/models/81f7c96c852ee8fc832187b0132e569d6c3065a3252ed18e56effd0b6a73e524/large-v2.pt",
+    "large-v3": "https://openaipublic.azureedge.net/main/whisper/models/e5b1a55b89c1367dacf97e3e19bfd829a01529dbfdeefa8caeb59b3f1b81dadb/large-v3.pt",
+    "large": "https://openaipublic.azureedge.net/main/whisper/models/e5b1a55b89c1367dacf97e3e19bfd829a01529dbfdeefa8caeb59b3f1b81dadb/large-v3.pt",
+    "large-v3-turbo": "https://openaipublic.azureedge.net/main/whisper/models/aff26ae408abcba5fbf8813c21e62b0941638c5f6eebfb145be0c9839262a19a/large-v3-turbo.pt",
+    "turbo": "https://openaipublic.azureedge.net/main/whisper/models/aff26ae408abcba5fbf8813c21e62b0941638c5f6eebfb145be0c9839262a19a/large-v3-turbo.pt",
+}
+
+# base85-encoded (n_layers, n_heads) boolean arrays indicating the cross-attention heads that are
+# highly correlated to the word-level timing, i.e. the alignment between audio and text tokens.
+_ALIGNMENT_HEADS = {
+    "tiny.en": b"ABzY8J1N>@0{>%R00Bk>$p{7v037`oCl~+#00",
+    "tiny": b"ABzY8bu8Lr0{>%RKn9Fp%m@SkK7Kt=7ytkO",
+    "base.en": b"ABzY8;40c<0{>%RzzG;p*o+Vo09|#PsxSZm00",
+    "base": b"ABzY8KQ!870{>%RzyTQH3`Q^yNP!>##QT-<FaQ7m",
+    "small.en": b"ABzY8>?_)10{>%RpeA61k&I|OI3I$65C{;;pbCHh0B{qLQ;+}v00",
+    "small": b"ABzY8DmU6=0{>%Rpa?J`kvJ6qF(V^F86#Xh7JUGMK}P<N0000",
+    "medium.en": b"ABzY8usPae0{>%R7<zz_OvQ{)4kMa0BMw6u5rT}kRKX;$NfYBv00*Hl@qhsU00",
+    "medium": b"ABzY8B0Jh+0{>%R7}kK1fFL7w6%<-Pf*t^=N)Qr&0RR9",
+    "large-v1": b"ABzY8r9j$a0{>%R7#4sLmoOs{s)o3~84-RPdcFk!JR<kSfC2yj",
+    "large-v2": b"ABzY8zd+h!0{>%R7=D0pU<_bnWW*tkYAhobTNnu$jnkEkXqp)j;w1Tzk)UH3X%SZd&fFZ2fC2yj",
+    "large-v3": b"ABzY8gWO1E0{>%R7(9S+Kn!D~%ngiGaR?*L!iJG9p-nab0JQ=-{D1-g00",
+    "large": b"ABzY8gWO1E0{>%R7(9S+Kn!D~%ngiGaR?*L!iJG9p-nab0JQ=-{D1-g00",
+    "large-v3-turbo": b"ABzY8j^C+e0{>%RARaKHP%t(lGR*)0g!tONPyhe`",
+    "turbo": b"ABzY8j^C+e0{>%RARaKHP%t(lGR*)0g!tONPyhe`",
+}
+
+
+def _download(url: str, root: str, in_memory: bool) -> Union[bytes, str]:
+    os.makedirs(root, exist_ok=True)
+
+    expected_sha256 = url.split("/")[-2]
+    download_target = os.path.join(root, os.path.basename(url))
+
+    if os.path.exists(download_target) and not os.path.isfile(download_target):
+        raise RuntimeError(f"{download_target} exists and is not a regular file")
+
+    if os.path.isfile(download_target):
+        with open(download_target, "rb") as f:
+            model_bytes = f.read()
+        if hashlib.sha256(model_bytes).hexdigest() == expected_sha256:
+            return model_bytes if in_memory else download_target
+        else:
+            warnings.warn(
+                f"{download_target} exists, but the SHA256 checksum does not match; re-downloading the file"
+            )
+
+    with urllib.request.urlopen(url) as source, open(download_target, "wb") as output:
+        with tqdm(
+            total=int(source.info().get("Content-Length")),
+            ncols=80,
+            unit="iB",
+            unit_scale=True,
+            unit_divisor=1024,
+        ) as loop:
+            while True:
+                buffer = source.read(8192)
+                if not buffer:
+                    break
+
+                output.write(buffer)
+                loop.update(len(buffer))
+
+    model_bytes = open(download_target, "rb").read()
+    if hashlib.sha256(model_bytes).hexdigest() != expected_sha256:
+        raise RuntimeError(
+            "Model has been downloaded but the SHA256 checksum does not not match. Please retry loading the model."
+        )
+
+    return model_bytes if in_memory else download_target
+
+
+def available_models() -> List[str]:
+    """Returns the names of available models"""
+    return list(_MODELS.keys())
+
+
+def load_model(
+    name: str,
+    device: Optional[Union[str, torch.device]] = None,
+    download_root: str = None,
+    in_memory: bool = False,
+) -> Whisper:
+    """
+    Load a Whisper ASR model
+
+    Parameters
+    ----------
+    name : str
+        one of the official model names listed by `whisper.available_models()`, or
+        path to a model checkpoint containing the model dimensions and the model state_dict.
+    device : Union[str, torch.device]
+        the PyTorch device to put the model into
+    download_root: str
+        path to download the model files; by default, it uses "~/.cache/whisper"
+    in_memory: bool
+        whether to preload the model weights into host memory
+
+    Returns
+    -------
+    model : Whisper
+        The Whisper ASR model instance
+    """
+
+    if device is None:
+        device = "cuda" if torch.cuda.is_available() else "cpu"
+    if download_root is None:
+        default = os.path.join(os.path.expanduser("~"), ".cache")
+        download_root = os.path.join(os.getenv("XDG_CACHE_HOME", default), "whisper")
+
+    if name in _MODELS:
+        checkpoint_file = _download(_MODELS[name], download_root, in_memory)
+        alignment_heads = _ALIGNMENT_HEADS[name]
+    elif os.path.isfile(name):
+        checkpoint_file = open(name, "rb").read() if in_memory else name
+        alignment_heads = None
+    else:
+        raise RuntimeError(
+            f"Model {name} not found; available models = {available_models()}"
+        )
+
+    with (
+        io.BytesIO(checkpoint_file) if in_memory else open(checkpoint_file, "rb")
+    ) as fp:
+        kwargs = {"weights_only": True} if torch.__version__ >= "1.13" else {}
+        checkpoint = torch.load(fp, map_location=device, **kwargs)
+    del checkpoint_file
+
+    dims = ModelDimensions(**checkpoint["dims"])
+    model = Whisper(dims)
+    model.load_state_dict(checkpoint["model_state_dict"])
+
+    if alignment_heads is not None:
+        model.set_alignment_heads(alignment_heads)
+
+    return model.to(device)
diff --git a/whisper_service_deploy/whisper/__main__.py b/whisper_service_deploy/whisper/__main__.py
new file mode 100644
index 0000000..d14f205
--- /dev/null
+++ b/whisper_service_deploy/whisper/__main__.py
@@ -0,0 +1,3 @@
+from .transcribe import cli
+
+cli()
diff --git a/whisper_service_deploy/whisper/audio.py b/whisper_service_deploy/whisper/audio.py
new file mode 100644
index 0000000..826250f
--- /dev/null
+++ b/whisper_service_deploy/whisper/audio.py
@@ -0,0 +1,157 @@
+import os
+from functools import lru_cache
+from subprocess import CalledProcessError, run
+from typing import Optional, Union
+
+import numpy as np
+import torch
+import torch.nn.functional as F
+
+from .utils import exact_div
+
+# hard-coded audio hyperparameters
+SAMPLE_RATE = 16000
+N_FFT = 400
+HOP_LENGTH = 160
+CHUNK_LENGTH = 30
+N_SAMPLES = CHUNK_LENGTH * SAMPLE_RATE  # 480000 samples in a 30-second chunk
+N_FRAMES = exact_div(N_SAMPLES, HOP_LENGTH)  # 3000 frames in a mel spectrogram input
+
+N_SAMPLES_PER_TOKEN = HOP_LENGTH * 2  # the initial convolutions has stride 2
+FRAMES_PER_SECOND = exact_div(SAMPLE_RATE, HOP_LENGTH)  # 10ms per audio frame
+TOKENS_PER_SECOND = exact_div(SAMPLE_RATE, N_SAMPLES_PER_TOKEN)  # 20ms per audio token
+
+
+def load_audio(file: str, sr: int = SAMPLE_RATE):
+    """
+    Open an audio file and read as mono waveform, resampling as necessary
+
+    Parameters
+    ----------
+    file: str
+        The audio file to open
+
+    sr: int
+        The sample rate to resample the audio if necessary
+
+    Returns
+    -------
+    A NumPy array containing the audio waveform, in float32 dtype.
+    """
+
+    # This launches a subprocess to decode audio while down-mixing
+    # and resampling as necessary.  Requires the ffmpeg CLI in PATH.
+    # fmt: off
+    cmd = [
+        "ffmpeg",
+        "-nostdin",
+        "-threads", "0",
+        "-i", file,
+        "-f", "s16le",
+        "-ac", "1",
+        "-acodec", "pcm_s16le",
+        "-ar", str(sr),
+        "-"
+    ]
+    # fmt: on
+    try:
+        out = run(cmd, capture_output=True, check=True).stdout
+    except CalledProcessError as e:
+        raise RuntimeError(f"Failed to load audio: {e.stderr.decode()}") from e
+
+    return np.frombuffer(out, np.int16).flatten().astype(np.float32) / 32768.0
+
+
+def pad_or_trim(array, length: int = N_SAMPLES, *, axis: int = -1):
+    """
+    Pad or trim the audio array to N_SAMPLES, as expected by the encoder.
+    """
+    if torch.is_tensor(array):
+        if array.shape[axis] > length:
+            array = array.index_select(
+                dim=axis, index=torch.arange(length, device=array.device)
+            )
+
+        if array.shape[axis] < length:
+            pad_widths = [(0, 0)] * array.ndim
+            pad_widths[axis] = (0, length - array.shape[axis])
+            array = F.pad(array, [pad for sizes in pad_widths[::-1] for pad in sizes])
+    else:
+        if array.shape[axis] > length:
+            array = array.take(indices=range(length), axis=axis)
+
+        if array.shape[axis] < length:
+            pad_widths = [(0, 0)] * array.ndim
+            pad_widths[axis] = (0, length - array.shape[axis])
+            array = np.pad(array, pad_widths)
+
+    return array
+
+
+@lru_cache(maxsize=None)
+def mel_filters(device, n_mels: int) -> torch.Tensor:
+    """
+    load the mel filterbank matrix for projecting STFT into a Mel spectrogram.
+    Allows decoupling librosa dependency; saved using:
+
+        np.savez_compressed(
+            "mel_filters.npz",
+            mel_80=librosa.filters.mel(sr=16000, n_fft=400, n_mels=80),
+            mel_128=librosa.filters.mel(sr=16000, n_fft=400, n_mels=128),
+        )
+    """
+    assert n_mels in {80, 128}, f"Unsupported n_mels: {n_mels}"
+
+    filters_path = os.path.join(os.path.dirname(__file__), "assets", "mel_filters.npz")
+    with np.load(filters_path, allow_pickle=False) as f:
+        return torch.from_numpy(f[f"mel_{n_mels}"]).to(device)
+
+
+def log_mel_spectrogram(
+    audio: Union[str, np.ndarray, torch.Tensor],
+    n_mels: int = 80,
+    padding: int = 0,
+    device: Optional[Union[str, torch.device]] = None,
+):
+    """
+    Compute the log-Mel spectrogram of
+
+    Parameters
+    ----------
+    audio: Union[str, np.ndarray, torch.Tensor], shape = (*)
+        The path to audio or either a NumPy array or Tensor containing the audio waveform in 16 kHz
+
+    n_mels: int
+        The number of Mel-frequency filters, only 80 and 128 are supported
+
+    padding: int
+        Number of zero samples to pad to the right
+
+    device: Optional[Union[str, torch.device]]
+        If given, the audio tensor is moved to this device before STFT
+
+    Returns
+    -------
+    torch.Tensor, shape = (n_mels, n_frames)
+        A Tensor that contains the Mel spectrogram
+    """
+    if not torch.is_tensor(audio):
+        if isinstance(audio, str):
+            audio = load_audio(audio)
+        audio = torch.from_numpy(audio)
+
+    if device is not None:
+        audio = audio.to(device)
+    if padding > 0:
+        audio = F.pad(audio, (0, padding))
+    window = torch.hann_window(N_FFT).to(audio.device)
+    stft = torch.stft(audio, N_FFT, HOP_LENGTH, window=window, return_complex=True)
+    magnitudes = stft[..., :-1].abs() ** 2
+
+    filters = mel_filters(audio.device, n_mels)
+    mel_spec = filters @ magnitudes
+
+    log_spec = torch.clamp(mel_spec, min=1e-10).log10()
+    log_spec = torch.maximum(log_spec, log_spec.max() - 8.0)
+    log_spec = (log_spec + 4.0) / 4.0
+    return log_spec
diff --git a/whisper_service_deploy/whisper/decoding.py b/whisper_service_deploy/whisper/decoding.py
new file mode 100644
index 0000000..49485d0
--- /dev/null
+++ b/whisper_service_deploy/whisper/decoding.py
@@ -0,0 +1,826 @@
+from dataclasses import dataclass, field, replace
+from typing import TYPE_CHECKING, Dict, Iterable, List, Optional, Sequence, Tuple, Union
+
+import numpy as np
+import torch
+import torch.nn.functional as F
+from torch import Tensor
+from torch.distributions import Categorical
+
+from .audio import CHUNK_LENGTH
+from .tokenizer import Tokenizer, get_tokenizer
+from .utils import compression_ratio
+
+if TYPE_CHECKING:
+    from .model import Whisper
+
+
+@torch.no_grad()
+def detect_language(
+    model: "Whisper", mel: Tensor, tokenizer: Tokenizer = None
+) -> Tuple[Tensor, List[dict]]:
+    """
+    Detect the spoken language in the audio, and return them as list of strings, along with the ids
+    of the most probable language tokens and the probability distribution over all language tokens.
+    This is performed outside the main decode loop in order to not interfere with kv-caching.
+
+    Returns
+    -------
+    language_tokens : Tensor, shape = (n_audio,)
+        ids of the most probable language tokens, which appears after the startoftranscript token.
+    language_probs : List[Dict[str, float]], length = n_audio
+        list of dictionaries containing the probability distribution over all languages.
+    """
+    if tokenizer is None:
+        tokenizer = get_tokenizer(
+            model.is_multilingual, num_languages=model.num_languages
+        )
+    if (
+        tokenizer.language is None
+        or tokenizer.language_token not in tokenizer.sot_sequence
+    ):
+        raise ValueError(
+            "This model doesn't have language tokens so it can't perform lang id"
+        )
+
+    single = mel.ndim == 2
+    if single:
+        mel = mel.unsqueeze(0)
+
+    # skip encoder forward pass if already-encoded audio features were given
+    if mel.shape[-2:] != (model.dims.n_audio_ctx, model.dims.n_audio_state):
+        mel = model.encoder(mel)
+
+    # forward pass using a single token, startoftranscript
+    n_audio = mel.shape[0]
+    x = torch.tensor([[tokenizer.sot]] * n_audio).to(mel.device)  # [n_audio, 1]
+    logits = model.logits(x, mel)[:, 0]
+
+    # collect detected languages; suppress all non-language tokens
+    mask = torch.ones(logits.shape[-1], dtype=torch.bool)
+    mask[list(tokenizer.all_language_tokens)] = False
+    logits[:, mask] = -np.inf
+    language_tokens = logits.argmax(dim=-1)
+    language_token_probs = logits.softmax(dim=-1).cpu()
+    language_probs = [
+        {
+            c: language_token_probs[i, j].item()
+            for j, c in zip(tokenizer.all_language_tokens, tokenizer.all_language_codes)
+        }
+        for i in range(n_audio)
+    ]
+
+    if single:
+        language_tokens = language_tokens[0]
+        language_probs = language_probs[0]
+
+    return language_tokens, language_probs
+
+
+@dataclass(frozen=True)
+class DecodingOptions:
+    # whether to perform X->X "transcribe" or X->English "translate"
+    task: str = "transcribe"
+
+    # language that the audio is in; uses detected language if None
+    language: Optional[str] = None
+
+    # sampling-related options
+    temperature: float = 0.0
+    sample_len: Optional[int] = None  # maximum number of tokens to sample
+    best_of: Optional[int] = None  # number of independent sample trajectories, if t > 0
+    beam_size: Optional[int] = None  # number of beams in beam search, if t == 0
+    patience: Optional[float] = None  # patience in beam search (arxiv:2204.05424)
+
+    # "alpha" in Google NMT, or None for length norm, when ranking generations
+    # to select which to return among the beams or best-of-N samples
+    length_penalty: Optional[float] = None
+
+    # text or tokens to feed as the prompt or the prefix; for more info:
+    # https://github.com/openai/whisper/discussions/117#discussioncomment-3727051
+    prompt: Optional[Union[str, List[int]]] = None  # for the previous context
+    prefix: Optional[Union[str, List[int]]] = None  # to prefix the current context
+
+    # list of tokens ids (or comma-separated token ids) to suppress
+    # "-1" will suppress a set of symbols as defined in `tokenizer.non_speech_tokens()`
+    suppress_tokens: Optional[Union[str, Iterable[int]]] = "-1"
+    suppress_blank: bool = True  # this will suppress blank outputs
+
+    # timestamp sampling options
+    without_timestamps: bool = False  # use <|notimestamps|> to sample text tokens only
+    max_initial_timestamp: Optional[float] = 1.0
+
+    # implementation details
+    fp16: bool = True  # use fp16 for most of the calculation
+
+
+@dataclass(frozen=True)
+class DecodingResult:
+    audio_features: Tensor
+    language: str
+    language_probs: Optional[Dict[str, float]] = None
+    tokens: List[int] = field(default_factory=list)
+    text: str = ""
+    avg_logprob: float = np.nan
+    no_speech_prob: float = np.nan
+    temperature: float = np.nan
+    compression_ratio: float = np.nan
+
+
+class Inference:
+    def logits(self, tokens: Tensor, audio_features: Tensor) -> Tensor:
+        """Perform a forward pass on the decoder and return per-token logits"""
+        raise NotImplementedError
+
+    def rearrange_kv_cache(self, source_indices) -> None:
+        """Update the key-value cache according to the updated beams"""
+        raise NotImplementedError
+
+    def cleanup_caching(self) -> None:
+        """Clean up any resources or hooks after decoding is finished"""
+        pass
+
+
+class PyTorchInference(Inference):
+    def __init__(self, model: "Whisper", initial_token_length: int):
+        self.model: "Whisper" = model
+        self.initial_token_length = initial_token_length
+        self.kv_cache = {}
+        self.hooks = []
+
+        key_modules = [block.attn.key for block in self.model.decoder.blocks]
+        value_modules = [block.attn.value for block in self.model.decoder.blocks]
+        self.kv_modules = key_modules + value_modules
+
+    def logits(self, tokens: Tensor, audio_features: Tensor) -> Tensor:
+        if not self.kv_cache:
+            self.kv_cache, self.hooks = self.model.install_kv_cache_hooks()
+
+        if tokens.shape[-1] > self.initial_token_length:
+            # only need to use the last token except in the first forward pass
+            tokens = tokens[:, -1:]
+
+        return self.model.decoder(tokens, audio_features, kv_cache=self.kv_cache)
+
+    def cleanup_caching(self):
+        for hook in self.hooks:
+            hook.remove()
+
+        self.kv_cache = {}
+        self.hooks = []
+
+    def rearrange_kv_cache(self, source_indices):
+        if source_indices != list(range(len(source_indices))):
+            for module in self.kv_modules:
+                # update the key/value cache to contain the selected sequences
+                self.kv_cache[module] = self.kv_cache[module][source_indices].detach()
+
+
+class SequenceRanker:
+    def rank(
+        self, tokens: List[List[Tensor]], sum_logprobs: List[List[float]]
+    ) -> List[int]:
+        """
+        Given a list of groups of samples and their cumulative log probabilities,
+        return the indices of the samples in each group to select as the final result
+        """
+        raise NotImplementedError
+
+
+class MaximumLikelihoodRanker(SequenceRanker):
+    """
+    Select the sample with the highest log probabilities, penalized using either
+    a simple length normalization or Google NMT paper's length penalty
+    """
+
+    def __init__(self, length_penalty: Optional[float]):
+        self.length_penalty = length_penalty
+
+    def rank(self, tokens: List[List[Tensor]], sum_logprobs: List[List[float]]):
+        def scores(logprobs, lengths):
+            result = []
+            for logprob, length in zip(logprobs, lengths):
+                if self.length_penalty is None:
+                    penalty = length
+                else:
+                    # from the Google NMT paper
+                    penalty = ((5 + length) / 6) ** self.length_penalty
+                result.append(logprob / penalty)
+            return result
+
+        # get the sequence with the highest score
+        lengths = [[len(t) for t in s] for s in tokens]
+        return [np.argmax(scores(p, l)) for p, l in zip(sum_logprobs, lengths)]
+
+
+class TokenDecoder:
+    def reset(self):
+        """Initialize any stateful variables for decoding a new sequence"""
+
+    def update(
+        self, tokens: Tensor, logits: Tensor, sum_logprobs: Tensor
+    ) -> Tuple[Tensor, bool]:
+        """Specify how to select the next token, based on the current trace and logits
+
+        Parameters
+        ----------
+        tokens : Tensor, shape = (n_batch, current_sequence_length)
+            all tokens in the context so far, including the prefix and sot_sequence tokens
+
+        logits : Tensor, shape = (n_batch, vocab_size)
+            per-token logits of the probability distribution at the current step
+
+        sum_logprobs : Tensor, shape = (n_batch)
+            cumulative log probabilities for each sequence
+
+        Returns
+        -------
+        tokens : Tensor, shape = (n_batch, current_sequence_length + 1)
+            the tokens, appended with the selected next token
+
+        completed : bool
+            True if all sequences has reached the end of text
+
+        """
+        raise NotImplementedError
+
+    def finalize(
+        self, tokens: Tensor, sum_logprobs: Tensor
+    ) -> Tuple[Sequence[Sequence[Tensor]], List[List[float]]]:
+        """Finalize search and return the final candidate sequences
+
+        Parameters
+        ----------
+        tokens : Tensor, shape = (n_audio, n_group, current_sequence_length)
+            all tokens in the context so far, including the prefix and sot_sequence
+
+        sum_logprobs : Tensor, shape = (n_audio, n_group)
+            cumulative log probabilities for each sequence
+
+        Returns
+        -------
+        tokens : Sequence[Sequence[Tensor]], length = n_audio
+            sequence of Tensors containing candidate token sequences, for each audio input
+
+        sum_logprobs : List[List[float]], length = n_audio
+            sequence of cumulative log probabilities corresponding to the above
+
+        """
+        raise NotImplementedError
+
+
+class GreedyDecoder(TokenDecoder):
+    def __init__(self, temperature: float, eot: int):
+        self.temperature = temperature
+        self.eot = eot
+
+    def update(
+        self, tokens: Tensor, logits: Tensor, sum_logprobs: Tensor
+    ) -> Tuple[Tensor, bool]:
+        if self.temperature == 0:
+            next_tokens = logits.argmax(dim=-1)
+        else:
+            next_tokens = Categorical(logits=logits / self.temperature).sample()
+
+        logprobs = F.log_softmax(logits.float(), dim=-1)
+        current_logprobs = logprobs[torch.arange(logprobs.shape[0]), next_tokens]
+        sum_logprobs += current_logprobs * (tokens[:, -1] != self.eot)
+
+        next_tokens[tokens[:, -1] == self.eot] = self.eot
+        tokens = torch.cat([tokens, next_tokens[:, None]], dim=-1)
+
+        completed = (tokens[:, -1] == self.eot).all()
+        return tokens, completed
+
+    def finalize(self, tokens: Tensor, sum_logprobs: Tensor):
+        # make sure each sequence has at least one EOT token at the end
+        tokens = F.pad(tokens, (0, 1), value=self.eot)
+        return tokens, sum_logprobs.tolist()
+
+
+class BeamSearchDecoder(TokenDecoder):
+    def __init__(
+        self,
+        beam_size: int,
+        eot: int,
+        inference: Inference,
+        patience: Optional[float] = None,
+    ):
+        self.beam_size = beam_size
+        self.eot = eot
+        self.inference = inference
+        self.patience = patience or 1.0
+        self.max_candidates: int = round(beam_size * self.patience)
+        self.finished_sequences = None
+
+        assert (
+            self.max_candidates > 0
+        ), f"Invalid beam size ({beam_size}) or patience ({patience})"
+
+    def reset(self):
+        self.finished_sequences = None
+
+    def update(
+        self, tokens: Tensor, logits: Tensor, sum_logprobs: Tensor
+    ) -> Tuple[Tensor, bool]:
+        if tokens.shape[0] % self.beam_size != 0:
+            raise ValueError(f"{tokens.shape}[0] % {self.beam_size} != 0")
+
+        n_audio = tokens.shape[0] // self.beam_size
+        if self.finished_sequences is None:  # for the first update
+            self.finished_sequences = [{} for _ in range(n_audio)]
+
+        logprobs = F.log_softmax(logits.float(), dim=-1)
+        next_tokens, source_indices, finished_sequences = [], [], []
+        for i in range(n_audio):
+            scores, sources, finished = {}, {}, {}
+
+            # STEP 1: calculate the cumulative log probabilities for possible candidates
+            for j in range(self.beam_size):
+                idx = i * self.beam_size + j
+                prefix = tokens[idx].tolist()
+                for logprob, token in zip(*logprobs[idx].topk(self.beam_size + 1)):
+                    new_logprob = (sum_logprobs[idx] + logprob).item()
+                    sequence = tuple(prefix + [token.item()])
+                    scores[sequence] = new_logprob
+                    sources[sequence] = idx
+
+            # STEP 2: rank the candidates and keep the top beam_size sequences for each audio
+            saved = 0
+            for sequence in sorted(scores, key=scores.get, reverse=True):
+                if sequence[-1] == self.eot:
+                    finished[sequence] = scores[sequence]
+                else:
+                    sum_logprobs[len(next_tokens)] = scores[sequence]
+                    next_tokens.append(sequence)
+                    source_indices.append(sources[sequence])
+
+                    saved += 1
+                    if saved == self.beam_size:
+                        break
+
+            finished_sequences.append(finished)
+
+        tokens = torch.tensor(next_tokens, device=tokens.device)
+        self.inference.rearrange_kv_cache(source_indices)
+
+        # add newly finished sequences to self.finished_sequences
+        assert len(self.finished_sequences) == len(finished_sequences)
+        for previously_finished, newly_finished in zip(
+            self.finished_sequences, finished_sequences
+        ):
+            for seq in sorted(newly_finished, key=newly_finished.get, reverse=True):
+                if len(previously_finished) >= self.max_candidates:
+                    break  # the candidate list is full
+                previously_finished[seq] = newly_finished[seq]
+
+        # mark as completed if all audio has enough number of samples
+        completed = all(
+            len(sequences) >= self.max_candidates
+            for sequences in self.finished_sequences
+        )
+        return tokens, completed
+
+    def finalize(self, preceding_tokens: Tensor, sum_logprobs: Tensor):
+        # collect all finished sequences, including patience, and add unfinished ones if not enough
+        sum_logprobs = sum_logprobs.cpu()
+        for i, sequences in enumerate(self.finished_sequences):
+            if (
+                len(sequences) < self.beam_size
+            ):  # when not enough sequences are finished
+                for j in list(np.argsort(sum_logprobs[i]))[::-1]:
+                    sequence = preceding_tokens[i, j].tolist() + [self.eot]
+                    sequences[tuple(sequence)] = sum_logprobs[i][j].item()
+                    if len(sequences) >= self.beam_size:
+                        break
+
+        tokens: List[List[Tensor]] = [
+            [torch.tensor(seq) for seq in sequences.keys()]
+            for sequences in self.finished_sequences
+        ]
+        sum_logprobs: List[List[float]] = [
+            list(sequences.values()) for sequences in self.finished_sequences
+        ]
+        return tokens, sum_logprobs
+
+
+class LogitFilter:
+    def apply(self, logits: Tensor, tokens: Tensor) -> None:
+        """Apply any filtering or masking to logits in-place
+
+        Parameters
+        ----------
+        logits : Tensor, shape = (n_batch, vocab_size)
+            per-token logits of the probability distribution at the current step
+
+        tokens : Tensor, shape = (n_batch, current_sequence_length)
+            all tokens in the context so far, including the prefix and sot_sequence tokens
+
+        """
+        raise NotImplementedError
+
+
+class SuppressBlank(LogitFilter):
+    def __init__(self, tokenizer: Tokenizer, sample_begin: int):
+        self.tokenizer = tokenizer
+        self.sample_begin = sample_begin
+
+    def apply(self, logits: Tensor, tokens: Tensor):
+        if tokens.shape[1] == self.sample_begin:
+            logits[:, self.tokenizer.encode(" ") + [self.tokenizer.eot]] = -np.inf
+
+
+class SuppressTokens(LogitFilter):
+    def __init__(self, suppress_tokens: Sequence[int]):
+        self.suppress_tokens = list(suppress_tokens)
+
+    def apply(self, logits: Tensor, tokens: Tensor):
+        logits[:, self.suppress_tokens] = -np.inf
+
+
+class ApplyTimestampRules(LogitFilter):
+    def __init__(
+        self,
+        tokenizer: Tokenizer,
+        sample_begin: int,
+        max_initial_timestamp_index: Optional[int],
+    ):
+        self.tokenizer = tokenizer
+        self.sample_begin = sample_begin
+        self.max_initial_timestamp_index = max_initial_timestamp_index
+
+    def apply(self, logits: Tensor, tokens: Tensor):
+        # suppress <|notimestamps|> which is handled by without_timestamps
+        if self.tokenizer.no_timestamps is not None:
+            logits[:, self.tokenizer.no_timestamps] = -np.inf
+
+        # timestamps have to appear in pairs, except directly before EOT; mask logits accordingly
+        for k in range(tokens.shape[0]):
+            sampled_tokens = tokens[k, self.sample_begin :]
+            seq = [t for t in sampled_tokens.tolist()]
+            last_was_timestamp = (
+                len(seq) >= 1 and seq[-1] >= self.tokenizer.timestamp_begin
+            )
+            penultimate_was_timestamp = (
+                len(seq) < 2 or seq[-2] >= self.tokenizer.timestamp_begin
+            )
+
+            if last_was_timestamp:
+                if penultimate_was_timestamp:  # has to be non-timestamp
+                    logits[k, self.tokenizer.timestamp_begin :] = -np.inf
+                else:  # cannot be normal text tokens
+                    logits[k, : self.tokenizer.eot] = -np.inf
+
+            timestamps = sampled_tokens[
+                sampled_tokens.ge(self.tokenizer.timestamp_begin)
+            ]
+            if timestamps.numel() > 0:
+                # timestamps shouldn't decrease; forbid timestamp tokens smaller than the last
+                # also force each segment to have a nonzero length, to prevent infinite looping
+                if last_was_timestamp and not penultimate_was_timestamp:
+                    timestamp_last = timestamps[-1]
+                else:
+                    timestamp_last = timestamps[-1] + 1
+                logits[k, self.tokenizer.timestamp_begin : timestamp_last] = -np.inf
+
+        if tokens.shape[1] == self.sample_begin:
+            # suppress generating non-timestamp tokens at the beginning
+            logits[:, : self.tokenizer.timestamp_begin] = -np.inf
+
+            # apply the `max_initial_timestamp` option
+            if self.max_initial_timestamp_index is not None:
+                last_allowed = (
+                    self.tokenizer.timestamp_begin + self.max_initial_timestamp_index
+                )
+                logits[:, last_allowed + 1 :] = -np.inf
+
+        # if sum of probability over timestamps is above any other token, sample timestamp
+        logprobs = F.log_softmax(logits.float(), dim=-1)
+        for k in range(tokens.shape[0]):
+            timestamp_logprob = logprobs[k, self.tokenizer.timestamp_begin :].logsumexp(
+                dim=-1
+            )
+            max_text_token_logprob = logprobs[k, : self.tokenizer.timestamp_begin].max()
+            if timestamp_logprob > max_text_token_logprob:
+                logits[k, : self.tokenizer.timestamp_begin] = -np.inf
+
+
+class DecodingTask:
+    inference: Inference
+    sequence_ranker: SequenceRanker
+    decoder: TokenDecoder
+    logit_filters: List[LogitFilter]
+
+    def __init__(self, model: "Whisper", options: DecodingOptions):
+        self.model = model
+
+        language = options.language or "en"
+        tokenizer = get_tokenizer(
+            model.is_multilingual,
+            num_languages=model.num_languages,
+            language=language,
+            task=options.task,
+        )
+        self.tokenizer: Tokenizer = tokenizer
+        self.options: DecodingOptions = self._verify_options(options)
+
+        self.n_group: int = options.beam_size or options.best_of or 1
+        self.n_ctx: int = model.dims.n_text_ctx
+        self.sample_len: int = options.sample_len or model.dims.n_text_ctx // 2
+
+        self.sot_sequence: Tuple[int] = tokenizer.sot_sequence
+        if self.options.without_timestamps:
+            self.sot_sequence = tokenizer.sot_sequence_including_notimestamps
+
+        self.initial_tokens: Tuple[int] = self._get_initial_tokens()
+        self.sample_begin: int = len(self.initial_tokens)
+        self.sot_index: int = self.initial_tokens.index(tokenizer.sot)
+
+        # inference: implements the forward pass through the decoder, including kv caching
+        self.inference = PyTorchInference(model, len(self.initial_tokens))
+
+        # sequence ranker: implements how to rank a group of sampled sequences
+        self.sequence_ranker = MaximumLikelihoodRanker(options.length_penalty)
+
+        # decoder: implements how to select the next tokens, given the autoregressive distribution
+        if options.beam_size is not None:
+            self.decoder = BeamSearchDecoder(
+                options.beam_size, tokenizer.eot, self.inference, options.patience
+            )
+        else:
+            self.decoder = GreedyDecoder(options.temperature, tokenizer.eot)
+
+        # logit filters: applies various rules to suppress or penalize certain tokens
+        self.logit_filters = []
+        if self.options.suppress_blank:
+            self.logit_filters.append(SuppressBlank(self.tokenizer, self.sample_begin))
+        if self.options.suppress_tokens:
+            self.logit_filters.append(SuppressTokens(self._get_suppress_tokens()))
+        if not options.without_timestamps:
+            precision = CHUNK_LENGTH / model.dims.n_audio_ctx  # usually 0.02 seconds
+            max_initial_timestamp_index = None
+            if options.max_initial_timestamp:
+                max_initial_timestamp_index = round(
+                    self.options.max_initial_timestamp / precision
+                )
+            self.logit_filters.append(
+                ApplyTimestampRules(
+                    tokenizer, self.sample_begin, max_initial_timestamp_index
+                )
+            )
+
+    def _verify_options(self, options: DecodingOptions) -> DecodingOptions:
+        if options.beam_size is not None and options.best_of is not None:
+            raise ValueError("beam_size and best_of can't be given together")
+        if options.temperature == 0:
+            if options.best_of is not None:
+                raise ValueError("best_of with greedy sampling (T=0) is not compatible")
+        if options.patience is not None and options.beam_size is None:
+            raise ValueError("patience requires beam_size to be given")
+        if options.length_penalty is not None and not (
+            0 <= options.length_penalty <= 1
+        ):
+            raise ValueError("length_penalty (alpha) should be a value between 0 and 1")
+
+        return options
+
+    def _get_initial_tokens(self) -> Tuple[int]:
+        tokens = list(self.sot_sequence)
+
+        if prefix := self.options.prefix:
+            prefix_tokens = (
+                self.tokenizer.encode(" " + prefix.strip())
+                if isinstance(prefix, str)
+                else prefix
+            )
+            if self.sample_len is not None:
+                max_prefix_len = self.n_ctx // 2 - self.sample_len
+                prefix_tokens = prefix_tokens[-max_prefix_len:]
+            tokens = tokens + prefix_tokens
+
+        if prompt := self.options.prompt:
+            prompt_tokens = (
+                self.tokenizer.encode(" " + prompt.strip())
+                if isinstance(prompt, str)
+                else prompt
+            )
+            tokens = (
+                [self.tokenizer.sot_prev]
+                + prompt_tokens[-(self.n_ctx // 2 - 1) :]
+                + tokens
+            )
+
+        return tuple(tokens)
+
+    def _get_suppress_tokens(self) -> Tuple[int]:
+        suppress_tokens = self.options.suppress_tokens
+
+        if isinstance(suppress_tokens, str):
+            suppress_tokens = [int(t) for t in suppress_tokens.split(",")]
+
+        if -1 in suppress_tokens:
+            suppress_tokens = [t for t in suppress_tokens if t >= 0]
+            suppress_tokens.extend(self.tokenizer.non_speech_tokens)
+        elif suppress_tokens is None or len(suppress_tokens) == 0:
+            suppress_tokens = []  # interpret empty string as an empty list
+        else:
+            assert isinstance(suppress_tokens, list), "suppress_tokens must be a list"
+
+        suppress_tokens.extend(
+            [
+                self.tokenizer.transcribe,
+                self.tokenizer.translate,
+                self.tokenizer.sot,
+                self.tokenizer.sot_prev,
+                self.tokenizer.sot_lm,
+            ]
+        )
+        if self.tokenizer.no_speech is not None:
+            # no-speech probability is collected separately
+            suppress_tokens.append(self.tokenizer.no_speech)
+
+        return tuple(sorted(set(suppress_tokens)))
+
+    def _get_audio_features(self, mel: Tensor):
+        if self.options.fp16:
+            mel = mel.half()
+
+        if mel.shape[-2:] == (
+            self.model.dims.n_audio_ctx,
+            self.model.dims.n_audio_state,
+        ):
+            # encoded audio features are given; skip audio encoding
+            audio_features = mel
+        else:
+            audio_features = self.model.encoder(mel)
+
+        if audio_features.dtype != (
+            torch.float16 if self.options.fp16 else torch.float32
+        ):
+            return TypeError(
+                f"audio_features has an incorrect dtype: {audio_features.dtype}"
+            )
+
+        return audio_features
+
+    def _detect_language(self, audio_features: Tensor, tokens: Tensor):
+        languages = [self.options.language] * audio_features.shape[0]
+        lang_probs = None
+
+        if self.options.language is None or self.options.task == "lang_id":
+            lang_tokens, lang_probs = self.model.detect_language(
+                audio_features, self.tokenizer
+            )
+            languages = [max(probs, key=probs.get) for probs in lang_probs]
+            if self.options.language is None:
+                tokens[:, self.sot_index + 1] = lang_tokens  # write language tokens
+
+        return languages, lang_probs
+
+    def _main_loop(self, audio_features: Tensor, tokens: Tensor):
+        n_batch = tokens.shape[0]
+        sum_logprobs: Tensor = torch.zeros(n_batch, device=audio_features.device)
+        no_speech_probs = [np.nan] * n_batch
+
+        try:
+            for i in range(self.sample_len):
+                logits = self.inference.logits(tokens, audio_features)
+
+                if (
+                    i == 0 and self.tokenizer.no_speech is not None
+                ):  # save no_speech_probs
+                    probs_at_sot = logits[:, self.sot_index].float().softmax(dim=-1)
+                    no_speech_probs = probs_at_sot[:, self.tokenizer.no_speech].tolist()
+
+                # now we need to consider the logits at the last token only
+                logits = logits[:, -1]
+
+                # apply the logit filters, e.g. for suppressing or applying penalty to
+                for logit_filter in self.logit_filters:
+                    logit_filter.apply(logits, tokens)
+
+                # expand the tokens tensor with the selected next tokens
+                tokens, completed = self.decoder.update(tokens, logits, sum_logprobs)
+
+                if completed or tokens.shape[-1] > self.n_ctx:
+                    break
+        finally:
+            self.inference.cleanup_caching()
+
+        return tokens, sum_logprobs, no_speech_probs
+
+    @torch.no_grad()
+    def run(self, mel: Tensor) -> List[DecodingResult]:
+        self.decoder.reset()
+        tokenizer: Tokenizer = self.tokenizer
+        n_audio: int = mel.shape[0]
+
+        audio_features: Tensor = self._get_audio_features(mel)  # encoder forward pass
+        tokens: Tensor = torch.tensor([self.initial_tokens]).repeat(n_audio, 1)
+
+        # detect language if requested, overwriting the language token
+        languages, language_probs = self._detect_language(audio_features, tokens)
+        if self.options.task == "lang_id":
+            return [
+                DecodingResult(
+                    audio_features=features, language=language, language_probs=probs
+                )
+                for features, language, probs in zip(
+                    audio_features, languages, language_probs
+                )
+            ]
+
+        # repeat text tensors by the group size, for beam search or best-of-n sampling
+        tokens = tokens.repeat_interleave(self.n_group, dim=0).to(audio_features.device)
+
+        # call the main sampling loop
+        tokens, sum_logprobs, no_speech_probs = self._main_loop(audio_features, tokens)
+
+        # reshape the tensors to have (n_audio, n_group) as the first two dimensions
+        audio_features = audio_features[:: self.n_group]
+        no_speech_probs = no_speech_probs[:: self.n_group]
+        assert audio_features.shape[0] == len(no_speech_probs) == n_audio
+
+        tokens = tokens.reshape(n_audio, self.n_group, -1)
+        sum_logprobs = sum_logprobs.reshape(n_audio, self.n_group)
+
+        # get the final candidates for each group, and slice between the first sampled token and EOT
+        tokens, sum_logprobs = self.decoder.finalize(tokens, sum_logprobs)
+        tokens: List[List[Tensor]] = [
+            [t[self.sample_begin : (t == tokenizer.eot).nonzero()[0, 0]] for t in s]
+            for s in tokens
+        ]
+
+        # select the top-ranked sample in each group
+        selected = self.sequence_ranker.rank(tokens, sum_logprobs)
+        tokens: List[List[int]] = [t[i].tolist() for i, t in zip(selected, tokens)]
+        texts: List[str] = [tokenizer.decode(t).strip() for t in tokens]
+
+        sum_logprobs: List[float] = [lp[i] for i, lp in zip(selected, sum_logprobs)]
+        avg_logprobs: List[float] = [
+            lp / (len(t) + 1) for t, lp in zip(tokens, sum_logprobs)
+        ]
+
+        fields = (
+            texts,
+            languages,
+            tokens,
+            audio_features,
+            avg_logprobs,
+            no_speech_probs,
+        )
+        if len(set(map(len, fields))) != 1:
+            raise RuntimeError(f"inconsistent result lengths: {list(map(len, fields))}")
+
+        return [
+            DecodingResult(
+                audio_features=features,
+                language=language,
+                tokens=tokens,
+                text=text,
+                avg_logprob=avg_logprob,
+                no_speech_prob=no_speech_prob,
+                temperature=self.options.temperature,
+                compression_ratio=compression_ratio(text),
+            )
+            for text, language, tokens, features, avg_logprob, no_speech_prob in zip(
+                *fields
+            )
+        ]
+
+
+@torch.no_grad()
+def decode(
+    model: "Whisper",
+    mel: Tensor,
+    options: DecodingOptions = DecodingOptions(),
+    **kwargs,
+) -> Union[DecodingResult, List[DecodingResult]]:
+    """
+    Performs decoding of 30-second audio segment(s), provided as Mel spectrogram(s).
+
+    Parameters
+    ----------
+    model: Whisper
+        the Whisper model instance
+
+    mel: torch.Tensor, shape = (80, 3000) or (*, 80, 3000)
+        A tensor containing the Mel spectrogram(s)
+
+    options: DecodingOptions
+        A dataclass that contains all necessary options for decoding 30-second segments
+
+    Returns
+    -------
+    result: Union[DecodingResult, List[DecodingResult]]
+        The result(s) of decoding contained in `DecodingResult` dataclass instance(s)
+    """
+    if single := mel.ndim == 2:
+        mel = mel.unsqueeze(0)
+
+    if kwargs:
+        options = replace(options, **kwargs)
+
+    result = DecodingTask(model, options).run(mel)
+
+    return result[0] if single else result
diff --git a/whisper_service_deploy/whisper/model.py b/whisper_service_deploy/whisper/model.py
new file mode 100644
index 0000000..e537447
--- /dev/null
+++ b/whisper_service_deploy/whisper/model.py
@@ -0,0 +1,345 @@
+import base64
+import gzip
+from contextlib import contextmanager
+from dataclasses import dataclass
+from typing import Dict, Iterable, Optional, Tuple
+
+import numpy as np
+import torch
+import torch.nn.functional as F
+from torch import Tensor, nn
+
+from .decoding import decode as decode_function
+from .decoding import detect_language as detect_language_function
+from .transcribe import transcribe as transcribe_function
+
+try:
+    from torch.nn.functional import scaled_dot_product_attention
+
+    SDPA_AVAILABLE = True
+except (ImportError, RuntimeError, OSError):
+    scaled_dot_product_attention = None
+    SDPA_AVAILABLE = False
+
+
+@dataclass
+class ModelDimensions:
+    n_mels: int
+    n_audio_ctx: int
+    n_audio_state: int
+    n_audio_head: int
+    n_audio_layer: int
+    n_vocab: int
+    n_text_ctx: int
+    n_text_state: int
+    n_text_head: int
+    n_text_layer: int
+
+
+class LayerNorm(nn.LayerNorm):
+    def forward(self, x: Tensor) -> Tensor:
+        return super().forward(x.float()).type(x.dtype)
+
+
+class Linear(nn.Linear):
+    def forward(self, x: Tensor) -> Tensor:
+        return F.linear(
+            x,
+            self.weight.to(x.dtype),
+            None if self.bias is None else self.bias.to(x.dtype),
+        )
+
+
+class Conv1d(nn.Conv1d):
+    def _conv_forward(
+        self, x: Tensor, weight: Tensor, bias: Optional[Tensor]
+    ) -> Tensor:
+        return super()._conv_forward(
+            x, weight.to(x.dtype), None if bias is None else bias.to(x.dtype)
+        )
+
+
+def sinusoids(length, channels, max_timescale=10000):
+    """Returns sinusoids for positional embedding"""
+    assert channels % 2 == 0
+    log_timescale_increment = np.log(max_timescale) / (channels // 2 - 1)
+    inv_timescales = torch.exp(-log_timescale_increment * torch.arange(channels // 2))
+    scaled_time = torch.arange(length)[:, np.newaxis] * inv_timescales[np.newaxis, :]
+    return torch.cat([torch.sin(scaled_time), torch.cos(scaled_time)], dim=1)
+
+
+@contextmanager
+def disable_sdpa():
+    prev_state = MultiHeadAttention.use_sdpa
+    try:
+        MultiHeadAttention.use_sdpa = False
+        yield
+    finally:
+        MultiHeadAttention.use_sdpa = prev_state
+
+
+class MultiHeadAttention(nn.Module):
+    use_sdpa = True
+
+    def __init__(self, n_state: int, n_head: int):
+        super().__init__()
+        self.n_head = n_head
+        self.query = Linear(n_state, n_state)
+        self.key = Linear(n_state, n_state, bias=False)
+        self.value = Linear(n_state, n_state)
+        self.out = Linear(n_state, n_state)
+
+    def forward(
+        self,
+        x: Tensor,
+        xa: Optional[Tensor] = None,
+        mask: Optional[Tensor] = None,
+        kv_cache: Optional[dict] = None,
+    ):
+        q = self.query(x)
+
+        if kv_cache is None or xa is None or self.key not in kv_cache:
+            # hooks, if installed (i.e. kv_cache is not None), will prepend the cached kv tensors;
+            # otherwise, perform key/value projections for self- or cross-attention as usual.
+            k = self.key(x if xa is None else xa)
+            v = self.value(x if xa is None else xa)
+        else:
+            # for cross-attention, calculate keys and values once and reuse in subsequent calls.
+            k = kv_cache[self.key]
+            v = kv_cache[self.value]
+
+        wv, qk = self.qkv_attention(q, k, v, mask)
+        return self.out(wv), qk
+
+    def qkv_attention(
+        self, q: Tensor, k: Tensor, v: Tensor, mask: Optional[Tensor] = None
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor]]:
+        n_batch, n_ctx, n_state = q.shape
+        scale = (n_state // self.n_head) ** -0.25
+        q = q.view(*q.shape[:2], self.n_head, -1).permute(0, 2, 1, 3)
+        k = k.view(*k.shape[:2], self.n_head, -1).permute(0, 2, 1, 3)
+        v = v.view(*v.shape[:2], self.n_head, -1).permute(0, 2, 1, 3)
+
+        if SDPA_AVAILABLE and MultiHeadAttention.use_sdpa:
+            a = scaled_dot_product_attention(
+                q, k, v, is_causal=mask is not None and n_ctx > 1
+            )
+            out = a.permute(0, 2, 1, 3).flatten(start_dim=2)
+            qk = None
+        else:
+            qk = (q * scale) @ (k * scale).transpose(-1, -2)
+            if mask is not None:
+                qk = qk + mask[:n_ctx, :n_ctx]
+            qk = qk.float()
+
+            w = F.softmax(qk, dim=-1).to(q.dtype)
+            out = (w @ v).permute(0, 2, 1, 3).flatten(start_dim=2)
+            qk = qk.detach()
+
+        return out, qk
+
+
+class ResidualAttentionBlock(nn.Module):
+    def __init__(self, n_state: int, n_head: int, cross_attention: bool = False):
+        super().__init__()
+
+        self.attn = MultiHeadAttention(n_state, n_head)
+        self.attn_ln = LayerNorm(n_state)
+
+        self.cross_attn = (
+            MultiHeadAttention(n_state, n_head) if cross_attention else None
+        )
+        self.cross_attn_ln = LayerNorm(n_state) if cross_attention else None
+
+        n_mlp = n_state * 4
+        self.mlp = nn.Sequential(
+            Linear(n_state, n_mlp), nn.GELU(), Linear(n_mlp, n_state)
+        )
+        self.mlp_ln = LayerNorm(n_state)
+
+    def forward(
+        self,
+        x: Tensor,
+        xa: Optional[Tensor] = None,
+        mask: Optional[Tensor] = None,
+        kv_cache: Optional[dict] = None,
+    ):
+        x = x + self.attn(self.attn_ln(x), mask=mask, kv_cache=kv_cache)[0]
+        if self.cross_attn:
+            x = x + self.cross_attn(self.cross_attn_ln(x), xa, kv_cache=kv_cache)[0]
+        x = x + self.mlp(self.mlp_ln(x))
+        return x
+
+
+class AudioEncoder(nn.Module):
+    def __init__(
+        self, n_mels: int, n_ctx: int, n_state: int, n_head: int, n_layer: int
+    ):
+        super().__init__()
+        self.conv1 = Conv1d(n_mels, n_state, kernel_size=3, padding=1)
+        self.conv2 = Conv1d(n_state, n_state, kernel_size=3, stride=2, padding=1)
+        self.register_buffer("positional_embedding", sinusoids(n_ctx, n_state))
+
+        self.blocks: Iterable[ResidualAttentionBlock] = nn.ModuleList(
+            [ResidualAttentionBlock(n_state, n_head) for _ in range(n_layer)]
+        )
+        self.ln_post = LayerNorm(n_state)
+
+    def forward(self, x: Tensor):
+        """
+        x : torch.Tensor, shape = (batch_size, n_mels, n_ctx)
+            the mel spectrogram of the audio
+        """
+        x = F.gelu(self.conv1(x))
+        x = F.gelu(self.conv2(x))
+        x = x.permute(0, 2, 1)
+
+        assert x.shape[1:] == self.positional_embedding.shape, "incorrect audio shape"
+        x = (x + self.positional_embedding).to(x.dtype)
+
+        for block in self.blocks:
+            x = block(x)
+
+        x = self.ln_post(x)
+        return x
+
+
+class TextDecoder(nn.Module):
+    def __init__(
+        self, n_vocab: int, n_ctx: int, n_state: int, n_head: int, n_layer: int
+    ):
+        super().__init__()
+
+        self.token_embedding = nn.Embedding(n_vocab, n_state)
+        self.positional_embedding = nn.Parameter(torch.empty(n_ctx, n_state))
+
+        self.blocks: Iterable[ResidualAttentionBlock] = nn.ModuleList(
+            [
+                ResidualAttentionBlock(n_state, n_head, cross_attention=True)
+                for _ in range(n_layer)
+            ]
+        )
+        self.ln = LayerNorm(n_state)
+
+        mask = torch.empty(n_ctx, n_ctx).fill_(-np.inf).triu_(1)
+        self.register_buffer("mask", mask, persistent=False)
+
+    def forward(self, x: Tensor, xa: Tensor, kv_cache: Optional[dict] = None):
+        """
+        x : torch.LongTensor, shape = (batch_size, <= n_ctx)
+            the text tokens
+        xa : torch.Tensor, shape = (batch_size, n_audio_ctx, n_audio_state)
+            the encoded audio features to be attended on
+        """
+        offset = next(iter(kv_cache.values())).shape[1] if kv_cache else 0
+        x = (
+            self.token_embedding(x)
+            + self.positional_embedding[offset : offset + x.shape[-1]]
+        )
+        x = x.to(xa.dtype)
+
+        for block in self.blocks:
+            x = block(x, xa, mask=self.mask, kv_cache=kv_cache)
+
+        x = self.ln(x)
+        logits = (
+            x @ torch.transpose(self.token_embedding.weight.to(x.dtype), 0, 1)
+        ).float()
+
+        return logits
+
+
+class Whisper(nn.Module):
+    def __init__(self, dims: ModelDimensions):
+        super().__init__()
+        self.dims = dims
+        self.encoder = AudioEncoder(
+            self.dims.n_mels,
+            self.dims.n_audio_ctx,
+            self.dims.n_audio_state,
+            self.dims.n_audio_head,
+            self.dims.n_audio_layer,
+        )
+        self.decoder = TextDecoder(
+            self.dims.n_vocab,
+            self.dims.n_text_ctx,
+            self.dims.n_text_state,
+            self.dims.n_text_head,
+            self.dims.n_text_layer,
+        )
+        # use the last half among the decoder layers for time alignment by default;
+        # to use a specific set of heads, see `set_alignment_heads()` below.
+        all_heads = torch.zeros(
+            self.dims.n_text_layer, self.dims.n_text_head, dtype=torch.bool
+        )
+        all_heads[self.dims.n_text_layer // 2 :] = True
+        self.register_buffer("alignment_heads", all_heads.to_sparse(), persistent=False)
+
+    def set_alignment_heads(self, dump: bytes):
+        array = np.frombuffer(
+            gzip.decompress(base64.b85decode(dump)), dtype=bool
+        ).copy()
+        mask = torch.from_numpy(array).reshape(
+            self.dims.n_text_layer, self.dims.n_text_head
+        )
+        self.register_buffer("alignment_heads", mask.to_sparse(), persistent=False)
+
+    def embed_audio(self, mel: torch.Tensor):
+        return self.encoder(mel)
+
+    def logits(self, tokens: torch.Tensor, audio_features: torch.Tensor):
+        return self.decoder(tokens, audio_features)
+
+    def forward(
+        self, mel: torch.Tensor, tokens: torch.Tensor
+    ) -> Dict[str, torch.Tensor]:
+        return self.decoder(tokens, self.encoder(mel))
+
+    @property
+    def device(self):
+        return next(self.parameters()).device
+
+    @property
+    def is_multilingual(self):
+        return self.dims.n_vocab >= 51865
+
+    @property
+    def num_languages(self):
+        return self.dims.n_vocab - 51765 - int(self.is_multilingual)
+
+    def install_kv_cache_hooks(self, cache: Optional[dict] = None):
+        """
+        The `MultiHeadAttention` module optionally accepts `kv_cache` which stores the key and value
+        tensors calculated for the previous positions. This method returns a dictionary that stores
+        all caches, and the necessary hooks for the key and value projection modules that save the
+        intermediate tensors to be reused during later calculations.
+
+        Returns
+        -------
+        cache : Dict[nn.Module, torch.Tensor]
+            A dictionary object mapping the key/value projection modules to its cache
+        hooks : List[RemovableHandle]
+            List of PyTorch RemovableHandle objects to stop the hooks to be called
+        """
+        cache = {**cache} if cache is not None else {}
+        hooks = []
+
+        def save_to_cache(module, _, output):
+            if module not in cache or output.shape[1] > self.dims.n_text_ctx:
+                # save as-is, for the first token or cross attention
+                cache[module] = output
+            else:
+                cache[module] = torch.cat([cache[module], output], dim=1).detach()
+            return cache[module]
+
+        def install_hooks(layer: nn.Module):
+            if isinstance(layer, MultiHeadAttention):
+                hooks.append(layer.key.register_forward_hook(save_to_cache))
+                hooks.append(layer.value.register_forward_hook(save_to_cache))
+
+        self.decoder.apply(install_hooks)
+        return cache, hooks
+
+    detect_language = detect_language_function
+    transcribe = transcribe_function
+    decode = decode_function
diff --git a/whisper_service_deploy/whisper/normalizers/__init__.py b/whisper_service_deploy/whisper/normalizers/__init__.py
new file mode 100644
index 0000000..896d5e3
--- /dev/null
+++ b/whisper_service_deploy/whisper/normalizers/__init__.py
@@ -0,0 +1,2 @@
+from .basic import BasicTextNormalizer as BasicTextNormalizer
+from .english import EnglishTextNormalizer as EnglishTextNormalizer
diff --git a/whisper_service_deploy/whisper/normalizers/basic.py b/whisper_service_deploy/whisper/normalizers/basic.py
new file mode 100644
index 0000000..8690ae7
--- /dev/null
+++ b/whisper_service_deploy/whisper/normalizers/basic.py
@@ -0,0 +1,80 @@
+import re
+import unicodedata
+
+import regex
+
+# non-ASCII letters that are not separated by "NFKD" normalization
+ADDITIONAL_DIACRITICS = {
+    "œ": "oe",
+    "Œ": "OE",
+    "ø": "o",
+    "Ø": "O",
+    "æ": "ae",
+    "Æ": "AE",
+    "ß": "ss",
+    "ẞ": "SS",
+    "đ": "d",
+    "Đ": "D",
+    "ð": "d",
+    "Ð": "D",
+    "þ": "th",
+    "Þ": "th",
+    "ł": "l",
+    "Ł": "L",
+}
+
+
+def remove_symbols_and_diacritics(s: str, keep=""):
+    """
+    Replace any other markers, symbols, and punctuations with a space,
+    and drop any diacritics (category 'Mn' and some manual mappings)
+    """
+    return "".join(
+        (
+            c
+            if c in keep
+            else (
+                ADDITIONAL_DIACRITICS[c]
+                if c in ADDITIONAL_DIACRITICS
+                else (
+                    ""
+                    if unicodedata.category(c) == "Mn"
+                    else " " if unicodedata.category(c)[0] in "MSP" else c
+                )
+            )
+        )
+        for c in unicodedata.normalize("NFKD", s)
+    )
+
+
+def remove_symbols(s: str):
+    """
+    Replace any other markers, symbols, punctuations with a space, keeping diacritics
+    """
+    return "".join(
+        " " if unicodedata.category(c)[0] in "MSP" else c
+        for c in unicodedata.normalize("NFKC", s)
+    )
+
+
+class BasicTextNormalizer:
+    def __init__(self, remove_diacritics: bool = False, split_letters: bool = False):
+        self.clean = (
+            remove_symbols_and_diacritics if remove_diacritics else remove_symbols
+        )
+        self.split_letters = split_letters
+
+    def __call__(self, s: str):
+        s = s.lower()
+        s = re.sub(r"[<\[][^>\]]*[>\]]", "", s)  # remove words between brackets
+        s = re.sub(r"\(([^)]+?)\)", "", s)  # remove words between parenthesis
+        s = self.clean(s).lower()
+
+        if self.split_letters:
+            s = " ".join(regex.findall(r"\X", s, regex.U))
+
+        s = re.sub(
+            r"\s+", " ", s
+        )  # replace any successive whitespace characters with a space
+
+        return s
diff --git a/whisper_service_deploy/whisper/normalizers/english.json b/whisper_service_deploy/whisper/normalizers/english.json
new file mode 100644
index 0000000..74a1c35
--- /dev/null
+++ b/whisper_service_deploy/whisper/normalizers/english.json
@@ -0,0 +1,1741 @@
+{
+    "accessorise": "accessorize",
+    "accessorised": "accessorized",
+    "accessorises": "accessorizes",
+    "accessorising": "accessorizing",
+    "acclimatisation": "acclimatization",
+    "acclimatise": "acclimatize",
+    "acclimatised": "acclimatized",
+    "acclimatises": "acclimatizes",
+    "acclimatising": "acclimatizing",
+    "accoutrements": "accouterments",
+    "aeon": "eon",
+    "aeons": "eons",
+    "aerogramme": "aerogram",
+    "aerogrammes": "aerograms",
+    "aeroplane": "airplane",
+    "aeroplanes": "airplanes",
+    "aesthete": "esthete",
+    "aesthetes": "esthetes",
+    "aesthetic": "esthetic",
+    "aesthetically": "esthetically",
+    "aesthetics": "esthetics",
+    "aetiology": "etiology",
+    "ageing": "aging",
+    "aggrandisement": "aggrandizement",
+    "agonise": "agonize",
+    "agonised": "agonized",
+    "agonises": "agonizes",
+    "agonising": "agonizing",
+    "agonisingly": "agonizingly",
+    "almanack": "almanac",
+    "almanacks": "almanacs",
+    "aluminium": "aluminum",
+    "amortisable": "amortizable",
+    "amortisation": "amortization",
+    "amortisations": "amortizations",
+    "amortise": "amortize",
+    "amortised": "amortized",
+    "amortises": "amortizes",
+    "amortising": "amortizing",
+    "amphitheatre": "amphitheater",
+    "amphitheatres": "amphitheaters",
+    "anaemia": "anemia",
+    "anaemic": "anemic",
+    "anaesthesia": "anesthesia",
+    "anaesthetic": "anesthetic",
+    "anaesthetics": "anesthetics",
+    "anaesthetise": "anesthetize",
+    "anaesthetised": "anesthetized",
+    "anaesthetises": "anesthetizes",
+    "anaesthetising": "anesthetizing",
+    "anaesthetist": "anesthetist",
+    "anaesthetists": "anesthetists",
+    "anaesthetize": "anesthetize",
+    "anaesthetized": "anesthetized",
+    "anaesthetizes": "anesthetizes",
+    "anaesthetizing": "anesthetizing",
+    "analogue": "analog",
+    "analogues": "analogs",
+    "analyse": "analyze",
+    "analysed": "analyzed",
+    "analyses": "analyzes",
+    "analysing": "analyzing",
+    "anglicise": "anglicize",
+    "anglicised": "anglicized",
+    "anglicises": "anglicizes",
+    "anglicising": "anglicizing",
+    "annualised": "annualized",
+    "antagonise": "antagonize",
+    "antagonised": "antagonized",
+    "antagonises": "antagonizes",
+    "antagonising": "antagonizing",
+    "apologise": "apologize",
+    "apologised": "apologized",
+    "apologises": "apologizes",
+    "apologising": "apologizing",
+    "appal": "appall",
+    "appals": "appalls",
+    "appetiser": "appetizer",
+    "appetisers": "appetizers",
+    "appetising": "appetizing",
+    "appetisingly": "appetizingly",
+    "arbour": "arbor",
+    "arbours": "arbors",
+    "archeological": "archaeological",
+    "archaeologically": "archeologically",
+    "archaeologist": "archeologist",
+    "archaeologists": "archeologists",
+    "archaeology": "archeology</span>",
+    "ardour": "ardor",
+    "armour": "armor",
+    "armoured": "armored",
+    "armourer": "armorer",
+    "armourers": "armorers",
+    "armouries": "armories",
+    "armoury": "armory",
+    "artefact": "artifact",
+    "artefacts": "artifacts",
+    "authorise": "authorize",
+    "authorised": "authorized",
+    "authorises": "authorizes",
+    "authorising": "authorizing",
+    "axe": "ax",
+    "backpedalled": "backpedaled",
+    "backpedalling": "backpedaling",
+    "bannister": "banister",
+    "bannisters": "banisters",
+    "baptise": "baptize",
+    "baptised": "baptized",
+    "baptises": "baptizes",
+    "baptising": "baptizing",
+    "bastardise": "bastardize",
+    "bastardised": "bastardized",
+    "bastardises": "bastardizes",
+    "bastardising": "bastardizing",
+    "battleax": "battleaxe",
+    "baulk": "balk",
+    "baulked": "balked",
+    "baulking": "balking",
+    "baulks": "balks",
+    "bedevilled": "bedeviled",
+    "bedevilling": "bedeviling",
+    "behaviour": "behavior",
+    "behavioural": "behavioral",
+    "behaviourism": "behaviorism",
+    "behaviourist": "behaviorist",
+    "behaviourists": "behaviorists",
+    "behaviours": "behaviors",
+    "behove": "behoove",
+    "behoved": "behooved",
+    "behoves": "behooves",
+    "bejewelled": "bejeweled",
+    "belabour": "belabor",
+    "belaboured": "belabored",
+    "belabouring": "belaboring",
+    "belabours": "belabors",
+    "bevelled": "beveled",
+    "bevvies": "bevies",
+    "bevvy": "bevy",
+    "biassed": "biased",
+    "biassing": "biasing",
+    "bingeing": "binging",
+    "bougainvillaea": "bougainvillea",
+    "bougainvillaeas": "bougainvilleas",
+    "bowdlerise": "bowdlerize",
+    "bowdlerised": "bowdlerized",
+    "bowdlerises": "bowdlerizes",
+    "bowdlerising": "bowdlerizing",
+    "breathalyse": "breathalyze",
+    "breathalysed": "breathalyzed",
+    "breathalyser": "breathalyzer",
+    "breathalysers": "breathalyzers",
+    "breathalyses": "breathalyzes",
+    "breathalysing": "breathalyzing",
+    "brutalise": "brutalize",
+    "brutalised": "brutalized",
+    "brutalises": "brutalizes",
+    "brutalising": "brutalizing",
+    "busses": "buses",
+    "bussing": "busing",
+    "caesarean": "cesarean",
+    "caesareans": "cesareans",
+    "calibre": "caliber",
+    "calibres": "calibers",
+    "calliper": "caliper",
+    "callipers": "calipers",
+    "callisthenics": "calisthenics",
+    "canalise": "canalize",
+    "canalised": "canalized",
+    "canalises": "canalizes",
+    "canalising": "canalizing",
+    "cancelation": "cancellation",
+    "cancelations": "cancellations",
+    "cancelled": "canceled",
+    "cancelling": "canceling",
+    "candour": "candor",
+    "cannibalise": "cannibalize",
+    "cannibalised": "cannibalized",
+    "cannibalises": "cannibalizes",
+    "cannibalising": "cannibalizing",
+    "canonise": "canonize",
+    "canonised": "canonized",
+    "canonises": "canonizes",
+    "canonising": "canonizing",
+    "capitalise": "capitalize",
+    "capitalised": "capitalized",
+    "capitalises": "capitalizes",
+    "capitalising": "capitalizing",
+    "caramelise": "caramelize",
+    "caramelised": "caramelized",
+    "caramelises": "caramelizes",
+    "caramelising": "caramelizing",
+    "carbonise": "carbonize",
+    "carbonised": "carbonized",
+    "carbonises": "carbonizes",
+    "carbonising": "carbonizing",
+    "carolled": "caroled",
+    "carolling": "caroling",
+    "catalogue": "catalog",
+    "catalogued": "cataloged",
+    "catalogues": "catalogs",
+    "cataloguing": "cataloging",
+    "catalyse": "catalyze",
+    "catalysed": "catalyzed",
+    "catalyses": "catalyzes",
+    "catalysing": "catalyzing",
+    "categorise": "categorize",
+    "categorised": "categorized",
+    "categorises": "categorizes",
+    "categorising": "categorizing",
+    "cauterise": "cauterize",
+    "cauterised": "cauterized",
+    "cauterises": "cauterizes",
+    "cauterising": "cauterizing",
+    "cavilled": "caviled",
+    "cavilling": "caviling",
+    "centigramme": "centigram",
+    "centigrammes": "centigrams",
+    "centilitre": "centiliter",
+    "centilitres": "centiliters",
+    "centimetre": "centimeter",
+    "centimetres": "centimeters",
+    "centralise": "centralize",
+    "centralised": "centralized",
+    "centralises": "centralizes",
+    "centralising": "centralizing",
+    "centre": "center",
+    "centred": "centered",
+    "centrefold": "centerfold",
+    "centrefolds": "centerfolds",
+    "centrepiece": "centerpiece",
+    "centrepieces": "centerpieces",
+    "centres": "centers",
+    "channelled": "channeled",
+    "channelling": "channeling",
+    "characterise": "characterize",
+    "characterised": "characterized",
+    "characterises": "characterizes",
+    "characterising": "characterizing",
+    "cheque": "check",
+    "chequebook": "checkbook",
+    "chequebooks": "checkbooks",
+    "chequered": "checkered",
+    "cheques": "checks",
+    "chilli": "chili",
+    "chimaera": "chimera",
+    "chimaeras": "chimeras",
+    "chiselled": "chiseled",
+    "chiselling": "chiseling",
+    "circularise": "circularize",
+    "circularised": "circularized",
+    "circularises": "circularizes",
+    "circularising": "circularizing",
+    "civilise": "civilize",
+    "civilised": "civilized",
+    "civilises": "civilizes",
+    "civilising": "civilizing",
+    "clamour": "clamor",
+    "clamoured": "clamored",
+    "clamouring": "clamoring",
+    "clamours": "clamors",
+    "clangour": "clangor",
+    "clarinettist": "clarinetist",
+    "clarinettists": "clarinetists",
+    "collectivise": "collectivize",
+    "collectivised": "collectivized",
+    "collectivises": "collectivizes",
+    "collectivising": "collectivizing",
+    "colonisation": "colonization",
+    "colonise": "colonize",
+    "colonised": "colonized",
+    "coloniser": "colonizer",
+    "colonisers": "colonizers",
+    "colonises": "colonizes",
+    "colonising": "colonizing",
+    "colour": "color",
+    "colourant": "colorant",
+    "colourants": "colorants",
+    "coloured": "colored",
+    "coloureds": "coloreds",
+    "colourful": "colorful",
+    "colourfully": "colorfully",
+    "colouring": "coloring",
+    "colourize": "colorize",
+    "colourized": "colorized",
+    "colourizes": "colorizes",
+    "colourizing": "colorizing",
+    "colourless": "colorless",
+    "colours": "colors",
+    "commercialise": "commercialize",
+    "commercialised": "commercialized",
+    "commercialises": "commercializes",
+    "commercialising": "commercializing",
+    "compartmentalise": "compartmentalize",
+    "compartmentalised": "compartmentalized",
+    "compartmentalises": "compartmentalizes",
+    "compartmentalising": "compartmentalizing",
+    "computerise": "computerize",
+    "computerised": "computerized",
+    "computerises": "computerizes",
+    "computerising": "computerizing",
+    "conceptualise": "conceptualize",
+    "conceptualised": "conceptualized",
+    "conceptualises": "conceptualizes",
+    "conceptualising": "conceptualizing",
+    "connexion": "connection",
+    "connexions": "connections",
+    "contextualise": "contextualize",
+    "contextualised": "contextualized",
+    "contextualises": "contextualizes",
+    "contextualising": "contextualizing",
+    "cosier": "cozier",
+    "cosies": "cozies",
+    "cosiest": "coziest",
+    "cosily": "cozily",
+    "cosiness": "coziness",
+    "cosy": "cozy",
+    "councillor": "councilor",
+    "councillors": "councilors",
+    "counselled": "counseled",
+    "counselling": "counseling",
+    "counsellor": "counselor",
+    "counsellors": "counselors",
+    "crenelated": "crenellated",
+    "criminalise": "criminalize",
+    "criminalised": "criminalized",
+    "criminalises": "criminalizes",
+    "criminalising": "criminalizing",
+    "criticise": "criticize",
+    "criticised": "criticized",
+    "criticises": "criticizes",
+    "criticising": "criticizing",
+    "crueller": "crueler",
+    "cruellest": "cruelest",
+    "crystallisation": "crystallization",
+    "crystallise": "crystallize",
+    "crystallised": "crystallized",
+    "crystallises": "crystallizes",
+    "crystallising": "crystallizing",
+    "cudgelled": "cudgeled",
+    "cudgelling": "cudgeling",
+    "customise": "customize",
+    "customised": "customized",
+    "customises": "customizes",
+    "customising": "customizing",
+    "cypher": "cipher",
+    "cyphers": "ciphers",
+    "decentralisation": "decentralization",
+    "decentralise": "decentralize",
+    "decentralised": "decentralized",
+    "decentralises": "decentralizes",
+    "decentralising": "decentralizing",
+    "decriminalisation": "decriminalization",
+    "decriminalise": "decriminalize",
+    "decriminalised": "decriminalized",
+    "decriminalises": "decriminalizes",
+    "decriminalising": "decriminalizing",
+    "defence": "defense",
+    "defenceless": "defenseless",
+    "defences": "defenses",
+    "dehumanisation": "dehumanization",
+    "dehumanise": "dehumanize",
+    "dehumanised": "dehumanized",
+    "dehumanises": "dehumanizes",
+    "dehumanising": "dehumanizing",
+    "demeanour": "demeanor",
+    "demilitarisation": "demilitarization",
+    "demilitarise": "demilitarize",
+    "demilitarised": "demilitarized",
+    "demilitarises": "demilitarizes",
+    "demilitarising": "demilitarizing",
+    "demobilisation": "demobilization",
+    "demobilise": "demobilize",
+    "demobilised": "demobilized",
+    "demobilises": "demobilizes",
+    "demobilising": "demobilizing",
+    "democratisation": "democratization",
+    "democratise": "democratize",
+    "democratised": "democratized",
+    "democratises": "democratizes",
+    "democratising": "democratizing",
+    "demonise": "demonize",
+    "demonised": "demonized",
+    "demonises": "demonizes",
+    "demonising": "demonizing",
+    "demoralisation": "demoralization",
+    "demoralise": "demoralize",
+    "demoralised": "demoralized",
+    "demoralises": "demoralizes",
+    "demoralising": "demoralizing",
+    "denationalisation": "denationalization",
+    "denationalise": "denationalize",
+    "denationalised": "denationalized",
+    "denationalises": "denationalizes",
+    "denationalising": "denationalizing",
+    "deodorise": "deodorize",
+    "deodorised": "deodorized",
+    "deodorises": "deodorizes",
+    "deodorising": "deodorizing",
+    "depersonalise": "depersonalize",
+    "depersonalised": "depersonalized",
+    "depersonalises": "depersonalizes",
+    "depersonalising": "depersonalizing",
+    "deputise": "deputize",
+    "deputised": "deputized",
+    "deputises": "deputizes",
+    "deputising": "deputizing",
+    "desensitisation": "desensitization",
+    "desensitise": "desensitize",
+    "desensitised": "desensitized",
+    "desensitises": "desensitizes",
+    "desensitising": "desensitizing",
+    "destabilisation": "destabilization",
+    "destabilise": "destabilize",
+    "destabilised": "destabilized",
+    "destabilises": "destabilizes",
+    "destabilising": "destabilizing",
+    "dialled": "dialed",
+    "dialling": "dialing",
+    "dialogue": "dialog",
+    "dialogues": "dialogs",
+    "diarrhoea": "diarrhea",
+    "digitise": "digitize",
+    "digitised": "digitized",
+    "digitises": "digitizes",
+    "digitising": "digitizing",
+    "disc": "disk",
+    "discolour": "discolor",
+    "discoloured": "discolored",
+    "discolouring": "discoloring",
+    "discolours": "discolors",
+    "discs": "disks",
+    "disembowelled": "disemboweled",
+    "disembowelling": "disemboweling",
+    "disfavour": "disfavor",
+    "dishevelled": "disheveled",
+    "dishonour": "dishonor",
+    "dishonourable": "dishonorable",
+    "dishonourably": "dishonorably",
+    "dishonoured": "dishonored",
+    "dishonouring": "dishonoring",
+    "dishonours": "dishonors",
+    "disorganisation": "disorganization",
+    "disorganised": "disorganized",
+    "distil": "distill",
+    "distils": "distills",
+    "dramatisation": "dramatization",
+    "dramatisations": "dramatizations",
+    "dramatise": "dramatize",
+    "dramatised": "dramatized",
+    "dramatises": "dramatizes",
+    "dramatising": "dramatizing",
+    "draught": "draft",
+    "draughtboard": "draftboard",
+    "draughtboards": "draftboards",
+    "draughtier": "draftier",
+    "draughtiest": "draftiest",
+    "draughts": "drafts",
+    "draughtsman": "draftsman",
+    "draughtsmanship": "draftsmanship",
+    "draughtsmen": "draftsmen",
+    "draughtswoman": "draftswoman",
+    "draughtswomen": "draftswomen",
+    "draughty": "drafty",
+    "drivelled": "driveled",
+    "drivelling": "driveling",
+    "duelled": "dueled",
+    "duelling": "dueling",
+    "economise": "economize",
+    "economised": "economized",
+    "economises": "economizes",
+    "economising": "economizing",
+    "edoema": "edema",
+    "editorialise": "editorialize",
+    "editorialised": "editorialized",
+    "editorialises": "editorializes",
+    "editorialising": "editorializing",
+    "empathise": "empathize",
+    "empathised": "empathized",
+    "empathises": "empathizes",
+    "empathising": "empathizing",
+    "emphasise": "emphasize",
+    "emphasised": "emphasized",
+    "emphasises": "emphasizes",
+    "emphasising": "emphasizing",
+    "enamelled": "enameled",
+    "enamelling": "enameling",
+    "enamoured": "enamored",
+    "encyclopaedia": "encyclopedia",
+    "encyclopaedias": "encyclopedias",
+    "encyclopaedic": "encyclopedic",
+    "endeavour": "endeavor",
+    "endeavoured": "endeavored",
+    "endeavouring": "endeavoring",
+    "endeavours": "endeavors",
+    "energise": "energize",
+    "energised": "energized",
+    "energises": "energizes",
+    "energising": "energizing",
+    "enrol": "enroll",
+    "enrols": "enrolls",
+    "enthral": "enthrall",
+    "enthrals": "enthralls",
+    "epaulette": "epaulet",
+    "epaulettes": "epaulets",
+    "epicentre": "epicenter",
+    "epicentres": "epicenters",
+    "epilogue": "epilog",
+    "epilogues": "epilogs",
+    "epitomise": "epitomize",
+    "epitomised": "epitomized",
+    "epitomises": "epitomizes",
+    "epitomising": "epitomizing",
+    "equalisation": "equalization",
+    "equalise": "equalize",
+    "equalised": "equalized",
+    "equaliser": "equalizer",
+    "equalisers": "equalizers",
+    "equalises": "equalizes",
+    "equalising": "equalizing",
+    "eulogise": "eulogize",
+    "eulogised": "eulogized",
+    "eulogises": "eulogizes",
+    "eulogising": "eulogizing",
+    "evangelise": "evangelize",
+    "evangelised": "evangelized",
+    "evangelises": "evangelizes",
+    "evangelising": "evangelizing",
+    "exorcise": "exorcize",
+    "exorcised": "exorcized",
+    "exorcises": "exorcizes",
+    "exorcising": "exorcizing",
+    "extemporisation": "extemporization",
+    "extemporise": "extemporize",
+    "extemporised": "extemporized",
+    "extemporises": "extemporizes",
+    "extemporising": "extemporizing",
+    "externalisation": "externalization",
+    "externalisations": "externalizations",
+    "externalise": "externalize",
+    "externalised": "externalized",
+    "externalises": "externalizes",
+    "externalising": "externalizing",
+    "factorise": "factorize",
+    "factorised": "factorized",
+    "factorises": "factorizes",
+    "factorising": "factorizing",
+    "faecal": "fecal",
+    "faeces": "feces",
+    "familiarisation": "familiarization",
+    "familiarise": "familiarize",
+    "familiarised": "familiarized",
+    "familiarises": "familiarizes",
+    "familiarising": "familiarizing",
+    "fantasise": "fantasize",
+    "fantasised": "fantasized",
+    "fantasises": "fantasizes",
+    "fantasising": "fantasizing",
+    "favour": "favor",
+    "favourable": "favorable",
+    "favourably": "favorably",
+    "favoured": "favored",
+    "favouring": "favoring",
+    "favourite": "favorite",
+    "favourites": "favorites",
+    "favouritism": "favoritism",
+    "favours": "favors",
+    "feminise": "feminize",
+    "feminised": "feminized",
+    "feminises": "feminizes",
+    "feminising": "feminizing",
+    "fertilisation": "fertilization",
+    "fertilise": "fertilize",
+    "fertilised": "fertilized",
+    "fertiliser": "fertilizer",
+    "fertilisers": "fertilizers",
+    "fertilises": "fertilizes",
+    "fertilising": "fertilizing",
+    "fervour": "fervor",
+    "fibre": "fiber",
+    "fibreglass": "fiberglass",
+    "fibres": "fibers",
+    "fictionalisation": "fictionalization",
+    "fictionalisations": "fictionalizations",
+    "fictionalise": "fictionalize",
+    "fictionalised": "fictionalized",
+    "fictionalises": "fictionalizes",
+    "fictionalising": "fictionalizing",
+    "fillet": "filet",
+    "filleted": "fileted",
+    "filleting": "fileting",
+    "fillets": "filets",
+    "finalisation": "finalization",
+    "finalise": "finalize",
+    "finalised": "finalized",
+    "finalises": "finalizes",
+    "finalising": "finalizing",
+    "flautist": "flutist",
+    "flautists": "flutists",
+    "flavour": "flavor",
+    "flavoured": "flavored",
+    "flavouring": "flavoring",
+    "flavourings": "flavorings",
+    "flavourless": "flavorless",
+    "flavours": "flavors",
+    "flavoursome": "flavorsome",
+    "flyer / flier": "flier / flyer",
+    "foetal": "fetal",
+    "foetid": "fetid",
+    "foetus": "fetus",
+    "foetuses": "fetuses",
+    "formalisation": "formalization",
+    "formalise": "formalize",
+    "formalised": "formalized",
+    "formalises": "formalizes",
+    "formalising": "formalizing",
+    "fossilisation": "fossilization",
+    "fossilise": "fossilize",
+    "fossilised": "fossilized",
+    "fossilises": "fossilizes",
+    "fossilising": "fossilizing",
+    "fraternisation": "fraternization",
+    "fraternise": "fraternize",
+    "fraternised": "fraternized",
+    "fraternises": "fraternizes",
+    "fraternising": "fraternizing",
+    "fulfil": "fulfill",
+    "fulfilment": "fulfillment",
+    "fulfils": "fulfills",
+    "funnelled": "funneled",
+    "funnelling": "funneling",
+    "galvanise": "galvanize",
+    "galvanised": "galvanized",
+    "galvanises": "galvanizes",
+    "galvanising": "galvanizing",
+    "gambolled": "gamboled",
+    "gambolling": "gamboling",
+    "gaol": "jail",
+    "gaolbird": "jailbird",
+    "gaolbirds": "jailbirds",
+    "gaolbreak": "jailbreak",
+    "gaolbreaks": "jailbreaks",
+    "gaoled": "jailed",
+    "gaoler": "jailer",
+    "gaolers": "jailers",
+    "gaoling": "jailing",
+    "gaols": "jails",
+    "gasses": "gases",
+    "gage": "gauge",
+    "gaged": "gauged",
+    "gages": "gauges",
+    "gaging": "gauging",
+    "generalisation": "generalization",
+    "generalisations": "generalizations",
+    "generalise": "generalize",
+    "generalised": "generalized",
+    "generalises": "generalizes",
+    "generalising": "generalizing",
+    "ghettoise": "ghettoize",
+    "ghettoised": "ghettoized",
+    "ghettoises": "ghettoizes",
+    "ghettoising": "ghettoizing",
+    "gipsies": "gypsies",
+    "glamorise": "glamorize",
+    "glamorised": "glamorized",
+    "glamorises": "glamorizes",
+    "glamorising": "glamorizing",
+    "glamor": "glamour",
+    "globalisation": "globalization",
+    "globalise": "globalize",
+    "globalised": "globalized",
+    "globalises": "globalizes",
+    "globalising": "globalizing",
+    "glueing": "gluing",
+    "goitre": "goiter",
+    "goitres": "goiters",
+    "gonorrhoea": "gonorrhea",
+    "gramme": "gram",
+    "grammes": "grams",
+    "gravelled": "graveled",
+    "grey": "gray",
+    "greyed": "grayed",
+    "greying": "graying",
+    "greyish": "grayish",
+    "greyness": "grayness",
+    "greys": "grays",
+    "grovelled": "groveled",
+    "grovelling": "groveling",
+    "groyne": "groin",
+    "groynes": "groins",
+    "gruelling": "grueling",
+    "gruellingly": "gruelingly",
+    "gryphon": "griffin",
+    "gryphons": "griffins",
+    "gynaecological": "gynecological",
+    "gynaecologist": "gynecologist",
+    "gynaecologists": "gynecologists",
+    "gynaecology": "gynecology",
+    "haematological": "hematological",
+    "haematologist": "hematologist",
+    "haematologists": "hematologists",
+    "haematology": "hematology",
+    "haemoglobin": "hemoglobin",
+    "haemophilia": "hemophilia",
+    "haemophiliac": "hemophiliac",
+    "haemophiliacs": "hemophiliacs",
+    "haemorrhage": "hemorrhage",
+    "haemorrhaged": "hemorrhaged",
+    "haemorrhages": "hemorrhages",
+    "haemorrhaging": "hemorrhaging",
+    "haemorrhoids": "hemorrhoids",
+    "harbour": "harbor",
+    "harboured": "harbored",
+    "harbouring": "harboring",
+    "harbours": "harbors",
+    "harmonisation": "harmonization",
+    "harmonise": "harmonize",
+    "harmonised": "harmonized",
+    "harmonises": "harmonizes",
+    "harmonising": "harmonizing",
+    "homoeopath": "homeopath",
+    "homoeopathic": "homeopathic",
+    "homoeopaths": "homeopaths",
+    "homoeopathy": "homeopathy",
+    "homogenise": "homogenize",
+    "homogenised": "homogenized",
+    "homogenises": "homogenizes",
+    "homogenising": "homogenizing",
+    "honour": "honor",
+    "honourable": "honorable",
+    "honourably": "honorably",
+    "honoured": "honored",
+    "honouring": "honoring",
+    "honours": "honors",
+    "hospitalisation": "hospitalization",
+    "hospitalise": "hospitalize",
+    "hospitalised": "hospitalized",
+    "hospitalises": "hospitalizes",
+    "hospitalising": "hospitalizing",
+    "humanise": "humanize",
+    "humanised": "humanized",
+    "humanises": "humanizes",
+    "humanising": "humanizing",
+    "humour": "humor",
+    "humoured": "humored",
+    "humouring": "humoring",
+    "humourless": "humorless",
+    "humours": "humors",
+    "hybridise": "hybridize",
+    "hybridised": "hybridized",
+    "hybridises": "hybridizes",
+    "hybridising": "hybridizing",
+    "hypnotise": "hypnotize",
+    "hypnotised": "hypnotized",
+    "hypnotises": "hypnotizes",
+    "hypnotising": "hypnotizing",
+    "hypothesise": "hypothesize",
+    "hypothesised": "hypothesized",
+    "hypothesises": "hypothesizes",
+    "hypothesising": "hypothesizing",
+    "idealisation": "idealization",
+    "idealise": "idealize",
+    "idealised": "idealized",
+    "idealises": "idealizes",
+    "idealising": "idealizing",
+    "idolise": "idolize",
+    "idolised": "idolized",
+    "idolises": "idolizes",
+    "idolising": "idolizing",
+    "immobilisation": "immobilization",
+    "immobilise": "immobilize",
+    "immobilised": "immobilized",
+    "immobiliser": "immobilizer",
+    "immobilisers": "immobilizers",
+    "immobilises": "immobilizes",
+    "immobilising": "immobilizing",
+    "immortalise": "immortalize",
+    "immortalised": "immortalized",
+    "immortalises": "immortalizes",
+    "immortalising": "immortalizing",
+    "immunisation": "immunization",
+    "immunise": "immunize",
+    "immunised": "immunized",
+    "immunises": "immunizes",
+    "immunising": "immunizing",
+    "impanelled": "impaneled",
+    "impanelling": "impaneling",
+    "imperilled": "imperiled",
+    "imperilling": "imperiling",
+    "individualise": "individualize",
+    "individualised": "individualized",
+    "individualises": "individualizes",
+    "individualising": "individualizing",
+    "industrialise": "industrialize",
+    "industrialised": "industrialized",
+    "industrialises": "industrializes",
+    "industrialising": "industrializing",
+    "inflexion": "inflection",
+    "inflexions": "inflections",
+    "initialise": "initialize",
+    "initialised": "initialized",
+    "initialises": "initializes",
+    "initialising": "initializing",
+    "initialled": "initialed",
+    "initialling": "initialing",
+    "instal": "install",
+    "instalment": "installment",
+    "instalments": "installments",
+    "instals": "installs",
+    "instil": "instill",
+    "instils": "instills",
+    "institutionalisation": "institutionalization",
+    "institutionalise": "institutionalize",
+    "institutionalised": "institutionalized",
+    "institutionalises": "institutionalizes",
+    "institutionalising": "institutionalizing",
+    "intellectualise": "intellectualize",
+    "intellectualised": "intellectualized",
+    "intellectualises": "intellectualizes",
+    "intellectualising": "intellectualizing",
+    "internalisation": "internalization",
+    "internalise": "internalize",
+    "internalised": "internalized",
+    "internalises": "internalizes",
+    "internalising": "internalizing",
+    "internationalisation": "internationalization",
+    "internationalise": "internationalize",
+    "internationalised": "internationalized",
+    "internationalises": "internationalizes",
+    "internationalising": "internationalizing",
+    "ionisation": "ionization",
+    "ionise": "ionize",
+    "ionised": "ionized",
+    "ioniser": "ionizer",
+    "ionisers": "ionizers",
+    "ionises": "ionizes",
+    "ionising": "ionizing",
+    "italicise": "italicize",
+    "italicised": "italicized",
+    "italicises": "italicizes",
+    "italicising": "italicizing",
+    "itemise": "itemize",
+    "itemised": "itemized",
+    "itemises": "itemizes",
+    "itemising": "itemizing",
+    "jeopardise": "jeopardize",
+    "jeopardised": "jeopardized",
+    "jeopardises": "jeopardizes",
+    "jeopardising": "jeopardizing",
+    "jewelled": "jeweled",
+    "jeweller": "jeweler",
+    "jewellers": "jewelers",
+    "jewellery": "jewelry",
+    "judgement": "judgment",
+    "kilogramme": "kilogram",
+    "kilogrammes": "kilograms",
+    "kilometre": "kilometer",
+    "kilometres": "kilometers",
+    "labelled": "labeled",
+    "labelling": "labeling",
+    "labour": "labor",
+    "laboured": "labored",
+    "labourer": "laborer",
+    "labourers": "laborers",
+    "labouring": "laboring",
+    "labours": "labors",
+    "lacklustre": "lackluster",
+    "legalisation": "legalization",
+    "legalise": "legalize",
+    "legalised": "legalized",
+    "legalises": "legalizes",
+    "legalising": "legalizing",
+    "legitimise": "legitimize",
+    "legitimised": "legitimized",
+    "legitimises": "legitimizes",
+    "legitimising": "legitimizing",
+    "leukaemia": "leukemia",
+    "levelled": "leveled",
+    "leveller": "leveler",
+    "levellers": "levelers",
+    "levelling": "leveling",
+    "libelled": "libeled",
+    "libelling": "libeling",
+    "libellous": "libelous",
+    "liberalisation": "liberalization",
+    "liberalise": "liberalize",
+    "liberalised": "liberalized",
+    "liberalises": "liberalizes",
+    "liberalising": "liberalizing",
+    "licence": "license",
+    "licenced": "licensed",
+    "licences": "licenses",
+    "licencing": "licensing",
+    "likeable": "likable",
+    "lionisation": "lionization",
+    "lionise": "lionize",
+    "lionised": "lionized",
+    "lionises": "lionizes",
+    "lionising": "lionizing",
+    "liquidise": "liquidize",
+    "liquidised": "liquidized",
+    "liquidiser": "liquidizer",
+    "liquidisers": "liquidizers",
+    "liquidises": "liquidizes",
+    "liquidising": "liquidizing",
+    "litre": "liter",
+    "litres": "liters",
+    "localise": "localize",
+    "localised": "localized",
+    "localises": "localizes",
+    "localising": "localizing",
+    "louvre": "louver",
+    "louvred": "louvered",
+    "louvres": "louvers",
+    "lustre": "luster",
+    "magnetise": "magnetize",
+    "magnetised": "magnetized",
+    "magnetises": "magnetizes",
+    "magnetising": "magnetizing",
+    "manoeuvrability": "maneuverability",
+    "manoeuvrable": "maneuverable",
+    "manoeuvre": "maneuver",
+    "manoeuvred": "maneuvered",
+    "manoeuvres": "maneuvers",
+    "manoeuvring": "maneuvering",
+    "manoeuvrings": "maneuverings",
+    "marginalisation": "marginalization",
+    "marginalise": "marginalize",
+    "marginalised": "marginalized",
+    "marginalises": "marginalizes",
+    "marginalising": "marginalizing",
+    "marshalled": "marshaled",
+    "marshalling": "marshaling",
+    "marvelled": "marveled",
+    "marvelling": "marveling",
+    "marvellous": "marvelous",
+    "marvellously": "marvelously",
+    "materialisation": "materialization",
+    "materialise": "materialize",
+    "materialised": "materialized",
+    "materialises": "materializes",
+    "materialising": "materializing",
+    "maximisation": "maximization",
+    "maximise": "maximize",
+    "maximised": "maximized",
+    "maximises": "maximizes",
+    "maximising": "maximizing",
+    "meagre": "meager",
+    "mechanisation": "mechanization",
+    "mechanise": "mechanize",
+    "mechanised": "mechanized",
+    "mechanises": "mechanizes",
+    "mechanising": "mechanizing",
+    "mediaeval": "medieval",
+    "memorialise": "memorialize",
+    "memorialised": "memorialized",
+    "memorialises": "memorializes",
+    "memorialising": "memorializing",
+    "memorise": "memorize",
+    "memorised": "memorized",
+    "memorises": "memorizes",
+    "memorising": "memorizing",
+    "mesmerise": "mesmerize",
+    "mesmerised": "mesmerized",
+    "mesmerises": "mesmerizes",
+    "mesmerising": "mesmerizing",
+    "metabolise": "metabolize",
+    "metabolised": "metabolized",
+    "metabolises": "metabolizes",
+    "metabolising": "metabolizing",
+    "metre": "meter",
+    "metres": "meters",
+    "micrometre": "micrometer",
+    "micrometres": "micrometers",
+    "militarise": "militarize",
+    "militarised": "militarized",
+    "militarises": "militarizes",
+    "militarising": "militarizing",
+    "milligramme": "milligram",
+    "milligrammes": "milligrams",
+    "millilitre": "milliliter",
+    "millilitres": "milliliters",
+    "millimetre": "millimeter",
+    "millimetres": "millimeters",
+    "miniaturisation": "miniaturization",
+    "miniaturise": "miniaturize",
+    "miniaturised": "miniaturized",
+    "miniaturises": "miniaturizes",
+    "miniaturising": "miniaturizing",
+    "minibusses": "minibuses",
+    "minimise": "minimize",
+    "minimised": "minimized",
+    "minimises": "minimizes",
+    "minimising": "minimizing",
+    "misbehaviour": "misbehavior",
+    "misdemeanour": "misdemeanor",
+    "misdemeanours": "misdemeanors",
+    "misspelt": "misspelled",
+    "mitre": "miter",
+    "mitres": "miters",
+    "mobilisation": "mobilization",
+    "mobilise": "mobilize",
+    "mobilised": "mobilized",
+    "mobilises": "mobilizes",
+    "mobilising": "mobilizing",
+    "modelled": "modeled",
+    "modeller": "modeler",
+    "modellers": "modelers",
+    "modelling": "modeling",
+    "modernise": "modernize",
+    "modernised": "modernized",
+    "modernises": "modernizes",
+    "modernising": "modernizing",
+    "moisturise": "moisturize",
+    "moisturised": "moisturized",
+    "moisturiser": "moisturizer",
+    "moisturisers": "moisturizers",
+    "moisturises": "moisturizes",
+    "moisturising": "moisturizing",
+    "monologue": "monolog",
+    "monologues": "monologs",
+    "monopolisation": "monopolization",
+    "monopolise": "monopolize",
+    "monopolised": "monopolized",
+    "monopolises": "monopolizes",
+    "monopolising": "monopolizing",
+    "moralise": "moralize",
+    "moralised": "moralized",
+    "moralises": "moralizes",
+    "moralising": "moralizing",
+    "motorised": "motorized",
+    "mould": "mold",
+    "moulded": "molded",
+    "moulder": "molder",
+    "mouldered": "moldered",
+    "mouldering": "moldering",
+    "moulders": "molders",
+    "mouldier": "moldier",
+    "mouldiest": "moldiest",
+    "moulding": "molding",
+    "mouldings": "moldings",
+    "moulds": "molds",
+    "mouldy": "moldy",
+    "moult": "molt",
+    "moulted": "molted",
+    "moulting": "molting",
+    "moults": "molts",
+    "moustache": "mustache",
+    "moustached": "mustached",
+    "moustaches": "mustaches",
+    "moustachioed": "mustachioed",
+    "multicoloured": "multicolored",
+    "nationalisation": "nationalization",
+    "nationalisations": "nationalizations",
+    "nationalise": "nationalize",
+    "nationalised": "nationalized",
+    "nationalises": "nationalizes",
+    "nationalising": "nationalizing",
+    "naturalisation": "naturalization",
+    "naturalise": "naturalize",
+    "naturalised": "naturalized",
+    "naturalises": "naturalizes",
+    "naturalising": "naturalizing",
+    "neighbour": "neighbor",
+    "neighbourhood": "neighborhood",
+    "neighbourhoods": "neighborhoods",
+    "neighbouring": "neighboring",
+    "neighbourliness": "neighborliness",
+    "neighbourly": "neighborly",
+    "neighbours": "neighbors",
+    "neutralisation": "neutralization",
+    "neutralise": "neutralize",
+    "neutralised": "neutralized",
+    "neutralises": "neutralizes",
+    "neutralising": "neutralizing",
+    "normalisation": "normalization",
+    "normalise": "normalize",
+    "normalised": "normalized",
+    "normalises": "normalizes",
+    "normalising": "normalizing",
+    "odour": "odor",
+    "odourless": "odorless",
+    "odours": "odors",
+    "oesophagus": "esophagus",
+    "oesophaguses": "esophaguses",
+    "oestrogen": "estrogen",
+    "offence": "offense",
+    "offences": "offenses",
+    "omelette": "omelet",
+    "omelettes": "omelets",
+    "optimise": "optimize",
+    "optimised": "optimized",
+    "optimises": "optimizes",
+    "optimising": "optimizing",
+    "organisation": "organization",
+    "organisational": "organizational",
+    "organisations": "organizations",
+    "organise": "organize",
+    "organised": "organized",
+    "organiser": "organizer",
+    "organisers": "organizers",
+    "organises": "organizes",
+    "organising": "organizing",
+    "orthopaedic": "orthopedic",
+    "orthopaedics": "orthopedics",
+    "ostracise": "ostracize",
+    "ostracised": "ostracized",
+    "ostracises": "ostracizes",
+    "ostracising": "ostracizing",
+    "outmanoeuvre": "outmaneuver",
+    "outmanoeuvred": "outmaneuvered",
+    "outmanoeuvres": "outmaneuvers",
+    "outmanoeuvring": "outmaneuvering",
+    "overemphasise": "overemphasize",
+    "overemphasised": "overemphasized",
+    "overemphasises": "overemphasizes",
+    "overemphasising": "overemphasizing",
+    "oxidisation": "oxidization",
+    "oxidise": "oxidize",
+    "oxidised": "oxidized",
+    "oxidises": "oxidizes",
+    "oxidising": "oxidizing",
+    "paederast": "pederast",
+    "paederasts": "pederasts",
+    "paediatric": "pediatric",
+    "paediatrician": "pediatrician",
+    "paediatricians": "pediatricians",
+    "paediatrics": "pediatrics",
+    "paedophile": "pedophile",
+    "paedophiles": "pedophiles",
+    "paedophilia": "pedophilia",
+    "palaeolithic": "paleolithic",
+    "palaeontologist": "paleontologist",
+    "palaeontologists": "paleontologists",
+    "palaeontology": "paleontology",
+    "panelled": "paneled",
+    "panelling": "paneling",
+    "panellist": "panelist",
+    "panellists": "panelists",
+    "paralyse": "paralyze",
+    "paralysed": "paralyzed",
+    "paralyses": "paralyzes",
+    "paralysing": "paralyzing",
+    "parcelled": "parceled",
+    "parcelling": "parceling",
+    "parlour": "parlor",
+    "parlours": "parlors",
+    "particularise": "particularize",
+    "particularised": "particularized",
+    "particularises": "particularizes",
+    "particularising": "particularizing",
+    "passivisation": "passivization",
+    "passivise": "passivize",
+    "passivised": "passivized",
+    "passivises": "passivizes",
+    "passivising": "passivizing",
+    "pasteurisation": "pasteurization",
+    "pasteurise": "pasteurize",
+    "pasteurised": "pasteurized",
+    "pasteurises": "pasteurizes",
+    "pasteurising": "pasteurizing",
+    "patronise": "patronize",
+    "patronised": "patronized",
+    "patronises": "patronizes",
+    "patronising": "patronizing",
+    "patronisingly": "patronizingly",
+    "pedalled": "pedaled",
+    "pedalling": "pedaling",
+    "pedestrianisation": "pedestrianization",
+    "pedestrianise": "pedestrianize",
+    "pedestrianised": "pedestrianized",
+    "pedestrianises": "pedestrianizes",
+    "pedestrianising": "pedestrianizing",
+    "penalise": "penalize",
+    "penalised": "penalized",
+    "penalises": "penalizes",
+    "penalising": "penalizing",
+    "pencilled": "penciled",
+    "pencilling": "penciling",
+    "personalise": "personalize",
+    "personalised": "personalized",
+    "personalises": "personalizes",
+    "personalising": "personalizing",
+    "pharmacopoeia": "pharmacopeia",
+    "pharmacopoeias": "pharmacopeias",
+    "philosophise": "philosophize",
+    "philosophised": "philosophized",
+    "philosophises": "philosophizes",
+    "philosophising": "philosophizing",
+    "philtre": "filter",
+    "philtres": "filters",
+    "phoney": "phony",
+    "plagiarise": "plagiarize",
+    "plagiarised": "plagiarized",
+    "plagiarises": "plagiarizes",
+    "plagiarising": "plagiarizing",
+    "plough": "plow",
+    "ploughed": "plowed",
+    "ploughing": "plowing",
+    "ploughman": "plowman",
+    "ploughmen": "plowmen",
+    "ploughs": "plows",
+    "ploughshare": "plowshare",
+    "ploughshares": "plowshares",
+    "polarisation": "polarization",
+    "polarise": "polarize",
+    "polarised": "polarized",
+    "polarises": "polarizes",
+    "polarising": "polarizing",
+    "politicisation": "politicization",
+    "politicise": "politicize",
+    "politicised": "politicized",
+    "politicises": "politicizes",
+    "politicising": "politicizing",
+    "popularisation": "popularization",
+    "popularise": "popularize",
+    "popularised": "popularized",
+    "popularises": "popularizes",
+    "popularising": "popularizing",
+    "pouffe": "pouf",
+    "pouffes": "poufs",
+    "practise": "practice",
+    "practised": "practiced",
+    "practises": "practices",
+    "practising": "practicing",
+    "praesidium": "presidium",
+    "praesidiums": "presidiums",
+    "pressurisation": "pressurization",
+    "pressurise": "pressurize",
+    "pressurised": "pressurized",
+    "pressurises": "pressurizes",
+    "pressurising": "pressurizing",
+    "pretence": "pretense",
+    "pretences": "pretenses",
+    "primaeval": "primeval",
+    "prioritisation": "prioritization",
+    "prioritise": "prioritize",
+    "prioritised": "prioritized",
+    "prioritises": "prioritizes",
+    "prioritising": "prioritizing",
+    "privatisation": "privatization",
+    "privatisations": "privatizations",
+    "privatise": "privatize",
+    "privatised": "privatized",
+    "privatises": "privatizes",
+    "privatising": "privatizing",
+    "professionalisation": "professionalization",
+    "professionalise": "professionalize",
+    "professionalised": "professionalized",
+    "professionalises": "professionalizes",
+    "professionalising": "professionalizing",
+    "programme": "program",
+    "programmes": "programs",
+    "prologue": "prolog",
+    "prologues": "prologs",
+    "propagandise": "propagandize",
+    "propagandised": "propagandized",
+    "propagandises": "propagandizes",
+    "propagandising": "propagandizing",
+    "proselytise": "proselytize",
+    "proselytised": "proselytized",
+    "proselytiser": "proselytizer",
+    "proselytisers": "proselytizers",
+    "proselytises": "proselytizes",
+    "proselytising": "proselytizing",
+    "psychoanalyse": "psychoanalyze",
+    "psychoanalysed": "psychoanalyzed",
+    "psychoanalyses": "psychoanalyzes",
+    "psychoanalysing": "psychoanalyzing",
+    "publicise": "publicize",
+    "publicised": "publicized",
+    "publicises": "publicizes",
+    "publicising": "publicizing",
+    "pulverisation": "pulverization",
+    "pulverise": "pulverize",
+    "pulverised": "pulverized",
+    "pulverises": "pulverizes",
+    "pulverising": "pulverizing",
+    "pummelled": "pummel",
+    "pummelling": "pummeled",
+    "pyjama": "pajama",
+    "pyjamas": "pajamas",
+    "pzazz": "pizzazz",
+    "quarrelled": "quarreled",
+    "quarrelling": "quarreling",
+    "radicalise": "radicalize",
+    "radicalised": "radicalized",
+    "radicalises": "radicalizes",
+    "radicalising": "radicalizing",
+    "rancour": "rancor",
+    "randomise": "randomize",
+    "randomised": "randomized",
+    "randomises": "randomizes",
+    "randomising": "randomizing",
+    "rationalisation": "rationalization",
+    "rationalisations": "rationalizations",
+    "rationalise": "rationalize",
+    "rationalised": "rationalized",
+    "rationalises": "rationalizes",
+    "rationalising": "rationalizing",
+    "ravelled": "raveled",
+    "ravelling": "raveling",
+    "realisable": "realizable",
+    "realisation": "realization",
+    "realisations": "realizations",
+    "realise": "realize",
+    "realised": "realized",
+    "realises": "realizes",
+    "realising": "realizing",
+    "recognisable": "recognizable",
+    "recognisably": "recognizably",
+    "recognisance": "recognizance",
+    "recognise": "recognize",
+    "recognised": "recognized",
+    "recognises": "recognizes",
+    "recognising": "recognizing",
+    "reconnoitre": "reconnoiter",
+    "reconnoitred": "reconnoitered",
+    "reconnoitres": "reconnoiters",
+    "reconnoitring": "reconnoitering",
+    "refuelled": "refueled",
+    "refuelling": "refueling",
+    "regularisation": "regularization",
+    "regularise": "regularize",
+    "regularised": "regularized",
+    "regularises": "regularizes",
+    "regularising": "regularizing",
+    "remodelled": "remodeled",
+    "remodelling": "remodeling",
+    "remould": "remold",
+    "remoulded": "remolded",
+    "remoulding": "remolding",
+    "remoulds": "remolds",
+    "reorganisation": "reorganization",
+    "reorganisations": "reorganizations",
+    "reorganise": "reorganize",
+    "reorganised": "reorganized",
+    "reorganises": "reorganizes",
+    "reorganising": "reorganizing",
+    "revelled": "reveled",
+    "reveller": "reveler",
+    "revellers": "revelers",
+    "revelling": "reveling",
+    "revitalise": "revitalize",
+    "revitalised": "revitalized",
+    "revitalises": "revitalizes",
+    "revitalising": "revitalizing",
+    "revolutionise": "revolutionize",
+    "revolutionised": "revolutionized",
+    "revolutionises": "revolutionizes",
+    "revolutionising": "revolutionizing",
+    "rhapsodise": "rhapsodize",
+    "rhapsodised": "rhapsodized",
+    "rhapsodises": "rhapsodizes",
+    "rhapsodising": "rhapsodizing",
+    "rigour": "rigor",
+    "rigours": "rigors",
+    "ritualised": "ritualized",
+    "rivalled": "rivaled",
+    "rivalling": "rivaling",
+    "romanticise": "romanticize",
+    "romanticised": "romanticized",
+    "romanticises": "romanticizes",
+    "romanticising": "romanticizing",
+    "rumour": "rumor",
+    "rumoured": "rumored",
+    "rumours": "rumors",
+    "sabre": "saber",
+    "sabres": "sabers",
+    "saltpetre": "saltpeter",
+    "sanitise": "sanitize",
+    "sanitised": "sanitized",
+    "sanitises": "sanitizes",
+    "sanitising": "sanitizing",
+    "satirise": "satirize",
+    "satirised": "satirized",
+    "satirises": "satirizes",
+    "satirising": "satirizing",
+    "saviour": "savior",
+    "saviours": "saviors",
+    "savour": "savor",
+    "savoured": "savored",
+    "savouries": "savories",
+    "savouring": "savoring",
+    "savours": "savors",
+    "savoury": "savory",
+    "scandalise": "scandalize",
+    "scandalised": "scandalized",
+    "scandalises": "scandalizes",
+    "scandalising": "scandalizing",
+    "sceptic": "skeptic",
+    "sceptical": "skeptical",
+    "sceptically": "skeptically",
+    "scepticism": "skepticism",
+    "sceptics": "skeptics",
+    "sceptre": "scepter",
+    "sceptres": "scepters",
+    "scrutinise": "scrutinize",
+    "scrutinised": "scrutinized",
+    "scrutinises": "scrutinizes",
+    "scrutinising": "scrutinizing",
+    "secularisation": "secularization",
+    "secularise": "secularize",
+    "secularised": "secularized",
+    "secularises": "secularizes",
+    "secularising": "secularizing",
+    "sensationalise": "sensationalize",
+    "sensationalised": "sensationalized",
+    "sensationalises": "sensationalizes",
+    "sensationalising": "sensationalizing",
+    "sensitise": "sensitize",
+    "sensitised": "sensitized",
+    "sensitises": "sensitizes",
+    "sensitising": "sensitizing",
+    "sentimentalise": "sentimentalize",
+    "sentimentalised": "sentimentalized",
+    "sentimentalises": "sentimentalizes",
+    "sentimentalising": "sentimentalizing",
+    "sepulchre": "sepulcher",
+    "sepulchres": "sepulchers",
+    "serialisation": "serialization",
+    "serialisations": "serializations",
+    "serialise": "serialize",
+    "serialised": "serialized",
+    "serialises": "serializes",
+    "serialising": "serializing",
+    "sermonise": "sermonize",
+    "sermonised": "sermonized",
+    "sermonises": "sermonizes",
+    "sermonising": "sermonizing",
+    "sheikh": "sheik",
+    "shovelled": "shoveled",
+    "shovelling": "shoveling",
+    "shrivelled": "shriveled",
+    "shrivelling": "shriveling",
+    "signalise": "signalize",
+    "signalised": "signalized",
+    "signalises": "signalizes",
+    "signalising": "signalizing",
+    "signalled": "signaled",
+    "signalling": "signaling",
+    "smoulder": "smolder",
+    "smouldered": "smoldered",
+    "smouldering": "smoldering",
+    "smoulders": "smolders",
+    "snivelled": "sniveled",
+    "snivelling": "sniveling",
+    "snorkelled": "snorkeled",
+    "snorkelling": "snorkeling",
+    "snowplough": "snowplow",
+    "snowploughs": "snowplow",
+    "socialisation": "socialization",
+    "socialise": "socialize",
+    "socialised": "socialized",
+    "socialises": "socializes",
+    "socialising": "socializing",
+    "sodomise": "sodomize",
+    "sodomised": "sodomized",
+    "sodomises": "sodomizes",
+    "sodomising": "sodomizing",
+    "solemnise": "solemnize",
+    "solemnised": "solemnized",
+    "solemnises": "solemnizes",
+    "solemnising": "solemnizing",
+    "sombre": "somber",
+    "specialisation": "specialization",
+    "specialisations": "specializations",
+    "specialise": "specialize",
+    "specialised": "specialized",
+    "specialises": "specializes",
+    "specialising": "specializing",
+    "spectre": "specter",
+    "spectres": "specters",
+    "spiralled": "spiraled",
+    "spiralling": "spiraling",
+    "splendour": "splendor",
+    "splendours": "splendors",
+    "squirrelled": "squirreled",
+    "squirrelling": "squirreling",
+    "stabilisation": "stabilization",
+    "stabilise": "stabilize",
+    "stabilised": "stabilized",
+    "stabiliser": "stabilizer",
+    "stabilisers": "stabilizers",
+    "stabilises": "stabilizes",
+    "stabilising": "stabilizing",
+    "standardisation": "standardization",
+    "standardise": "standardize",
+    "standardised": "standardized",
+    "standardises": "standardizes",
+    "standardising": "standardizing",
+    "stencilled": "stenciled",
+    "stencilling": "stenciling",
+    "sterilisation": "sterilization",
+    "sterilisations": "sterilizations",
+    "sterilise": "sterilize",
+    "sterilised": "sterilized",
+    "steriliser": "sterilizer",
+    "sterilisers": "sterilizers",
+    "sterilises": "sterilizes",
+    "sterilising": "sterilizing",
+    "stigmatisation": "stigmatization",
+    "stigmatise": "stigmatize",
+    "stigmatised": "stigmatized",
+    "stigmatises": "stigmatizes",
+    "stigmatising": "stigmatizing",
+    "storey": "story",
+    "storeys": "stories",
+    "subsidisation": "subsidization",
+    "subsidise": "subsidize",
+    "subsidised": "subsidized",
+    "subsidiser": "subsidizer",
+    "subsidisers": "subsidizers",
+    "subsidises": "subsidizes",
+    "subsidising": "subsidizing",
+    "succour": "succor",
+    "succoured": "succored",
+    "succouring": "succoring",
+    "succours": "succors",
+    "sulphate": "sulfate",
+    "sulphates": "sulfates",
+    "sulphide": "sulfide",
+    "sulphides": "sulfides",
+    "sulphur": "sulfur",
+    "sulphurous": "sulfurous",
+    "summarise": "summarize",
+    "summarised": "summarized",
+    "summarises": "summarizes",
+    "summarising": "summarizing",
+    "swivelled": "swiveled",
+    "swivelling": "swiveling",
+    "symbolise": "symbolize",
+    "symbolised": "symbolized",
+    "symbolises": "symbolizes",
+    "symbolising": "symbolizing",
+    "sympathise": "sympathize",
+    "sympathised": "sympathized",
+    "sympathiser": "sympathizer",
+    "sympathisers": "sympathizers",
+    "sympathises": "sympathizes",
+    "sympathising": "sympathizing",
+    "synchronisation": "synchronization",
+    "synchronise": "synchronize",
+    "synchronised": "synchronized",
+    "synchronises": "synchronizes",
+    "synchronising": "synchronizing",
+    "synthesise": "synthesize",
+    "synthesised": "synthesized",
+    "synthesiser": "synthesizer",
+    "synthesisers": "synthesizers",
+    "synthesises": "synthesizes",
+    "synthesising": "synthesizing",
+    "syphon": "siphon",
+    "syphoned": "siphoned",
+    "syphoning": "siphoning",
+    "syphons": "siphons",
+    "systematisation": "systematization",
+    "systematise": "systematize",
+    "systematised": "systematized",
+    "systematises": "systematizes",
+    "systematising": "systematizing",
+    "tantalise": "tantalize",
+    "tantalised": "tantalized",
+    "tantalises": "tantalizes",
+    "tantalising": "tantalizing",
+    "tantalisingly": "tantalizingly",
+    "tasselled": "tasseled",
+    "technicolour": "technicolor",
+    "temporise": "temporize",
+    "temporised": "temporized",
+    "temporises": "temporizes",
+    "temporising": "temporizing",
+    "tenderise": "tenderize",
+    "tenderised": "tenderized",
+    "tenderises": "tenderizes",
+    "tenderising": "tenderizing",
+    "terrorise": "terrorize",
+    "terrorised": "terrorized",
+    "terrorises": "terrorizes",
+    "terrorising": "terrorizing",
+    "theatre": "theater",
+    "theatregoer": "theatergoer",
+    "theatregoers": "theatergoers",
+    "theatres": "theaters",
+    "theorise": "theorize",
+    "theorised": "theorized",
+    "theorises": "theorizes",
+    "theorising": "theorizing",
+    "tonne": "ton",
+    "tonnes": "tons",
+    "towelled": "toweled",
+    "towelling": "toweling",
+    "toxaemia": "toxemia",
+    "tranquillise": "tranquilize",
+    "tranquillised": "tranquilized",
+    "tranquilliser": "tranquilizer",
+    "tranquillisers": "tranquilizers",
+    "tranquillises": "tranquilizes",
+    "tranquillising": "tranquilizing",
+    "tranquillity": "tranquility",
+    "tranquillize": "tranquilize",
+    "tranquillized": "tranquilized",
+    "tranquillizer": "tranquilizer",
+    "tranquillizers": "tranquilizers",
+    "tranquillizes": "tranquilizes",
+    "tranquillizing": "tranquilizing",
+    "tranquilly": "tranquility",
+    "transistorised": "transistorized",
+    "traumatise": "traumatize",
+    "traumatised": "traumatized",
+    "traumatises": "traumatizes",
+    "traumatising": "traumatizing",
+    "travelled": "traveled",
+    "traveller": "traveler",
+    "travellers": "travelers",
+    "travelling": "traveling",
+    "travelog": "travelogue",
+    "travelogs": "travelogues",
+    "trialled": "trialed",
+    "trialling": "trialing",
+    "tricolour": "tricolor",
+    "tricolours": "tricolors",
+    "trivialise": "trivialize",
+    "trivialised": "trivialized",
+    "trivialises": "trivializes",
+    "trivialising": "trivializing",
+    "tumour": "tumor",
+    "tumours": "tumors",
+    "tunnelled": "tunneled",
+    "tunnelling": "tunneling",
+    "tyrannise": "tyrannize",
+    "tyrannised": "tyrannized",
+    "tyrannises": "tyrannizes",
+    "tyrannising": "tyrannizing",
+    "tyre": "tire",
+    "tyres": "tires",
+    "unauthorised": "unauthorized",
+    "uncivilised": "uncivilized",
+    "underutilised": "underutilized",
+    "unequalled": "unequaled",
+    "unfavourable": "unfavorable",
+    "unfavourably": "unfavorably",
+    "unionisation": "unionization",
+    "unionise": "unionize",
+    "unionised": "unionized",
+    "unionises": "unionizes",
+    "unionising": "unionizing",
+    "unorganised": "unorganized",
+    "unravelled": "unraveled",
+    "unravelling": "unraveling",
+    "unrecognisable": "unrecognizable",
+    "unrecognised": "unrecognized",
+    "unrivalled": "unrivaled",
+    "unsavoury": "unsavory",
+    "untrammelled": "untrammeled",
+    "urbanisation": "urbanization",
+    "urbanise": "urbanize",
+    "urbanised": "urbanized",
+    "urbanises": "urbanizes",
+    "urbanising": "urbanizing",
+    "utilisable": "utilizable",
+    "utilisation": "utilization",
+    "utilise": "utilize",
+    "utilised": "utilized",
+    "utilises": "utilizes",
+    "utilising": "utilizing",
+    "valour": "valor",
+    "vandalise": "vandalize",
+    "vandalised": "vandalized",
+    "vandalises": "vandalizes",
+    "vandalising": "vandalizing",
+    "vaporisation": "vaporization",
+    "vaporise": "vaporize",
+    "vaporised": "vaporized",
+    "vaporises": "vaporizes",
+    "vaporising": "vaporizing",
+    "vapour": "vapor",
+    "vapours": "vapors",
+    "verbalise": "verbalize",
+    "verbalised": "verbalized",
+    "verbalises": "verbalizes",
+    "verbalising": "verbalizing",
+    "victimisation": "victimization",
+    "victimise": "victimize",
+    "victimised": "victimized",
+    "victimises": "victimizes",
+    "victimising": "victimizing",
+    "videodisc": "videodisk",
+    "videodiscs": "videodisks",
+    "vigour": "vigor",
+    "visualisation": "visualization",
+    "visualisations": "visualizations",
+    "visualise": "visualize",
+    "visualised": "visualized",
+    "visualises": "visualizes",
+    "visualising": "visualizing",
+    "vocalisation": "vocalization",
+    "vocalisations": "vocalizations",
+    "vocalise": "vocalize",
+    "vocalised": "vocalized",
+    "vocalises": "vocalizes",
+    "vocalising": "vocalizing",
+    "vulcanised": "vulcanized",
+    "vulgarisation": "vulgarization",
+    "vulgarise": "vulgarize",
+    "vulgarised": "vulgarized",
+    "vulgarises": "vulgarizes",
+    "vulgarising": "vulgarizing",
+    "waggon": "wagon",
+    "waggons": "wagons",
+    "watercolour": "watercolor",
+    "watercolours": "watercolors",
+    "weaselled": "weaseled",
+    "weaselling": "weaseling",
+    "westernisation": "westernization",
+    "westernise": "westernize",
+    "westernised": "westernized",
+    "westernises": "westernizes",
+    "westernising": "westernizing",
+    "womanise": "womanize",
+    "womanised": "womanized",
+    "womaniser": "womanizer",
+    "womanisers": "womanizers",
+    "womanises": "womanizes",
+    "womanising": "womanizing",
+    "woollen": "woolen",
+    "woollens": "woolens",
+    "woollies": "woolies",
+    "woolly": "wooly",
+    "worshipped": "worshiped",
+    "worshipping": "worshiping",
+    "worshipper": "worshiper",
+    "yodelled": "yodeled",
+    "yodelling": "yodeling",
+    "yoghourt": "yogurt",
+    "yoghourts": "yogurts",
+    "yoghurt": "yogurt",
+    "yoghurts": "yogurts",
+    "mhm": "hmm",
+    "mmm": "hmm"
+}
\ No newline at end of file
diff --git a/whisper_service_deploy/whisper/normalizers/english.py b/whisper_service_deploy/whisper/normalizers/english.py
new file mode 100644
index 0000000..4932042
--- /dev/null
+++ b/whisper_service_deploy/whisper/normalizers/english.py
@@ -0,0 +1,550 @@
+import json
+import os
+import re
+from fractions import Fraction
+from typing import Iterator, List, Match, Optional, Union
+
+from more_itertools import windowed
+
+from .basic import remove_symbols_and_diacritics
+
+
+class EnglishNumberNormalizer:
+    """
+    Convert any spelled-out numbers into arabic numbers, while handling:
+
+    - remove any commas
+    - keep the suffixes such as: `1960s`, `274th`, `32nd`, etc.
+    - spell out currency symbols after the number. e.g. `$20 million` -> `20000000 dollars`
+    - spell out `one` and `ones`
+    - interpret successive single-digit numbers as nominal: `one oh one` -> `101`
+    """
+
+    def __init__(self):
+        super().__init__()
+
+        self.zeros = {"o", "oh", "zero"}
+        self.ones = {
+            name: i
+            for i, name in enumerate(
+                [
+                    "one",
+                    "two",
+                    "three",
+                    "four",
+                    "five",
+                    "six",
+                    "seven",
+                    "eight",
+                    "nine",
+                    "ten",
+                    "eleven",
+                    "twelve",
+                    "thirteen",
+                    "fourteen",
+                    "fifteen",
+                    "sixteen",
+                    "seventeen",
+                    "eighteen",
+                    "nineteen",
+                ],
+                start=1,
+            )
+        }
+        self.ones_plural = {
+            "sixes" if name == "six" else name + "s": (value, "s")
+            for name, value in self.ones.items()
+        }
+        self.ones_ordinal = {
+            "zeroth": (0, "th"),
+            "first": (1, "st"),
+            "second": (2, "nd"),
+            "third": (3, "rd"),
+            "fifth": (5, "th"),
+            "twelfth": (12, "th"),
+            **{
+                name + ("h" if name.endswith("t") else "th"): (value, "th")
+                for name, value in self.ones.items()
+                if value > 3 and value != 5 and value != 12
+            },
+        }
+        self.ones_suffixed = {**self.ones_plural, **self.ones_ordinal}
+
+        self.tens = {
+            "twenty": 20,
+            "thirty": 30,
+            "forty": 40,
+            "fifty": 50,
+            "sixty": 60,
+            "seventy": 70,
+            "eighty": 80,
+            "ninety": 90,
+        }
+        self.tens_plural = {
+            name.replace("y", "ies"): (value, "s") for name, value in self.tens.items()
+        }
+        self.tens_ordinal = {
+            name.replace("y", "ieth"): (value, "th")
+            for name, value in self.tens.items()
+        }
+        self.tens_suffixed = {**self.tens_plural, **self.tens_ordinal}
+
+        self.multipliers = {
+            "hundred": 100,
+            "thousand": 1_000,
+            "million": 1_000_000,
+            "billion": 1_000_000_000,
+            "trillion": 1_000_000_000_000,
+            "quadrillion": 1_000_000_000_000_000,
+            "quintillion": 1_000_000_000_000_000_000,
+            "sextillion": 1_000_000_000_000_000_000_000,
+            "septillion": 1_000_000_000_000_000_000_000_000,
+            "octillion": 1_000_000_000_000_000_000_000_000_000,
+            "nonillion": 1_000_000_000_000_000_000_000_000_000_000,
+            "decillion": 1_000_000_000_000_000_000_000_000_000_000_000,
+        }
+        self.multipliers_plural = {
+            name + "s": (value, "s") for name, value in self.multipliers.items()
+        }
+        self.multipliers_ordinal = {
+            name + "th": (value, "th") for name, value in self.multipliers.items()
+        }
+        self.multipliers_suffixed = {
+            **self.multipliers_plural,
+            **self.multipliers_ordinal,
+        }
+        self.decimals = {*self.ones, *self.tens, *self.zeros}
+
+        self.preceding_prefixers = {
+            "minus": "-",
+            "negative": "-",
+            "plus": "+",
+            "positive": "+",
+        }
+        self.following_prefixers = {
+            "pound": "£",
+            "pounds": "£",
+            "euro": "€",
+            "euros": "€",
+            "dollar": "$",
+            "dollars": "$",
+            "cent": "¢",
+            "cents": "¢",
+        }
+        self.prefixes = set(
+            list(self.preceding_prefixers.values())
+            + list(self.following_prefixers.values())
+        )
+        self.suffixers = {
+            "per": {"cent": "%"},
+            "percent": "%",
+        }
+        self.specials = {"and", "double", "triple", "point"}
+
+        self.words = set(
+            [
+                key
+                for mapping in [
+                    self.zeros,
+                    self.ones,
+                    self.ones_suffixed,
+                    self.tens,
+                    self.tens_suffixed,
+                    self.multipliers,
+                    self.multipliers_suffixed,
+                    self.preceding_prefixers,
+                    self.following_prefixers,
+                    self.suffixers,
+                    self.specials,
+                ]
+                for key in mapping
+            ]
+        )
+        self.literal_words = {"one", "ones"}
+
+    def process_words(self, words: List[str]) -> Iterator[str]:
+        prefix: Optional[str] = None
+        value: Optional[Union[str, int]] = None
+        skip = False
+
+        def to_fraction(s: str):
+            try:
+                return Fraction(s)
+            except ValueError:
+                return None
+
+        def output(result: Union[str, int]):
+            nonlocal prefix, value
+            result = str(result)
+            if prefix is not None:
+                result = prefix + result
+            value = None
+            prefix = None
+            return result
+
+        if len(words) == 0:
+            return
+
+        for prev, current, next in windowed([None] + words + [None], 3):
+            if skip:
+                skip = False
+                continue
+
+            next_is_numeric = next is not None and re.match(r"^\d+(\.\d+)?$", next)
+            has_prefix = current[0] in self.prefixes
+            current_without_prefix = current[1:] if has_prefix else current
+            if re.match(r"^\d+(\.\d+)?$", current_without_prefix):
+                # arabic numbers (potentially with signs and fractions)
+                f = to_fraction(current_without_prefix)
+                assert f is not None
+                if value is not None:
+                    if isinstance(value, str) and value.endswith("."):
+                        # concatenate decimals / ip address components
+                        value = str(value) + str(current)
+                        continue
+                    else:
+                        yield output(value)
+
+                prefix = current[0] if has_prefix else prefix
+                if f.denominator == 1:
+                    value = f.numerator  # store integers as int
+                else:
+                    value = current_without_prefix
+            elif current not in self.words:
+                # non-numeric words
+                if value is not None:
+                    yield output(value)
+                yield output(current)
+            elif current in self.zeros:
+                value = str(value or "") + "0"
+            elif current in self.ones:
+                ones = self.ones[current]
+
+                if value is None:
+                    value = ones
+                elif isinstance(value, str) or prev in self.ones:
+                    if (
+                        prev in self.tens and ones < 10
+                    ):  # replace the last zero with the digit
+                        assert value[-1] == "0"
+                        value = value[:-1] + str(ones)
+                    else:
+                        value = str(value) + str(ones)
+                elif ones < 10:
+                    if value % 10 == 0:
+                        value += ones
+                    else:
+                        value = str(value) + str(ones)
+                else:  # eleven to nineteen
+                    if value % 100 == 0:
+                        value += ones
+                    else:
+                        value = str(value) + str(ones)
+            elif current in self.ones_suffixed:
+                # ordinal or cardinal; yield the number right away
+                ones, suffix = self.ones_suffixed[current]
+                if value is None:
+                    yield output(str(ones) + suffix)
+                elif isinstance(value, str) or prev in self.ones:
+                    if prev in self.tens and ones < 10:
+                        assert value[-1] == "0"
+                        yield output(value[:-1] + str(ones) + suffix)
+                    else:
+                        yield output(str(value) + str(ones) + suffix)
+                elif ones < 10:
+                    if value % 10 == 0:
+                        yield output(str(value + ones) + suffix)
+                    else:
+                        yield output(str(value) + str(ones) + suffix)
+                else:  # eleven to nineteen
+                    if value % 100 == 0:
+                        yield output(str(value + ones) + suffix)
+                    else:
+                        yield output(str(value) + str(ones) + suffix)
+                value = None
+            elif current in self.tens:
+                tens = self.tens[current]
+                if value is None:
+                    value = tens
+                elif isinstance(value, str):
+                    value = str(value) + str(tens)
+                else:
+                    if value % 100 == 0:
+                        value += tens
+                    else:
+                        value = str(value) + str(tens)
+            elif current in self.tens_suffixed:
+                # ordinal or cardinal; yield the number right away
+                tens, suffix = self.tens_suffixed[current]
+                if value is None:
+                    yield output(str(tens) + suffix)
+                elif isinstance(value, str):
+                    yield output(str(value) + str(tens) + suffix)
+                else:
+                    if value % 100 == 0:
+                        yield output(str(value + tens) + suffix)
+                    else:
+                        yield output(str(value) + str(tens) + suffix)
+            elif current in self.multipliers:
+                multiplier = self.multipliers[current]
+                if value is None:
+                    value = multiplier
+                elif isinstance(value, str) or value == 0:
+                    f = to_fraction(value)
+                    p = f * multiplier if f is not None else None
+                    if f is not None and p.denominator == 1:
+                        value = p.numerator
+                    else:
+                        yield output(value)
+                        value = multiplier
+                else:
+                    before = value // 1000 * 1000
+                    residual = value % 1000
+                    value = before + residual * multiplier
+            elif current in self.multipliers_suffixed:
+                multiplier, suffix = self.multipliers_suffixed[current]
+                if value is None:
+                    yield output(str(multiplier) + suffix)
+                elif isinstance(value, str):
+                    f = to_fraction(value)
+                    p = f * multiplier if f is not None else None
+                    if f is not None and p.denominator == 1:
+                        yield output(str(p.numerator) + suffix)
+                    else:
+                        yield output(value)
+                        yield output(str(multiplier) + suffix)
+                else:  # int
+                    before = value // 1000 * 1000
+                    residual = value % 1000
+                    value = before + residual * multiplier
+                    yield output(str(value) + suffix)
+                value = None
+            elif current in self.preceding_prefixers:
+                # apply prefix (positive, minus, etc.) if it precedes a number
+                if value is not None:
+                    yield output(value)
+
+                if next in self.words or next_is_numeric:
+                    prefix = self.preceding_prefixers[current]
+                else:
+                    yield output(current)
+            elif current in self.following_prefixers:
+                # apply prefix (dollars, cents, etc.) only after a number
+                if value is not None:
+                    prefix = self.following_prefixers[current]
+                    yield output(value)
+                else:
+                    yield output(current)
+            elif current in self.suffixers:
+                # apply suffix symbols (percent -> '%')
+                if value is not None:
+                    suffix = self.suffixers[current]
+                    if isinstance(suffix, dict):
+                        if next in suffix:
+                            yield output(str(value) + suffix[next])
+                            skip = True
+                        else:
+                            yield output(value)
+                            yield output(current)
+                    else:
+                        yield output(str(value) + suffix)
+                else:
+                    yield output(current)
+            elif current in self.specials:
+                if next not in self.words and not next_is_numeric:
+                    # apply special handling only if the next word can be numeric
+                    if value is not None:
+                        yield output(value)
+                    yield output(current)
+                elif current == "and":
+                    # ignore "and" after hundreds, thousands, etc.
+                    if prev not in self.multipliers:
+                        if value is not None:
+                            yield output(value)
+                        yield output(current)
+                elif current == "double" or current == "triple":
+                    if next in self.ones or next in self.zeros:
+                        repeats = 2 if current == "double" else 3
+                        ones = self.ones.get(next, 0)
+                        value = str(value or "") + str(ones) * repeats
+                        skip = True
+                    else:
+                        if value is not None:
+                            yield output(value)
+                        yield output(current)
+                elif current == "point":
+                    if next in self.decimals or next_is_numeric:
+                        value = str(value or "") + "."
+                else:
+                    # should all have been covered at this point
+                    raise ValueError(f"Unexpected token: {current}")
+            else:
+                # all should have been covered at this point
+                raise ValueError(f"Unexpected token: {current}")
+
+        if value is not None:
+            yield output(value)
+
+    def preprocess(self, s: str):
+        # replace "<number> and a half" with "<number> point five"
+        results = []
+
+        segments = re.split(r"\band\s+a\s+half\b", s)
+        for i, segment in enumerate(segments):
+            if len(segment.strip()) == 0:
+                continue
+            if i == len(segments) - 1:
+                results.append(segment)
+            else:
+                results.append(segment)
+                last_word = segment.rsplit(maxsplit=2)[-1]
+                if last_word in self.decimals or last_word in self.multipliers:
+                    results.append("point five")
+                else:
+                    results.append("and a half")
+
+        s = " ".join(results)
+
+        # put a space at number/letter boundary
+        s = re.sub(r"([a-z])([0-9])", r"\1 \2", s)
+        s = re.sub(r"([0-9])([a-z])", r"\1 \2", s)
+
+        # but remove spaces which could be a suffix
+        s = re.sub(r"([0-9])\s+(st|nd|rd|th|s)\b", r"\1\2", s)
+
+        return s
+
+    def postprocess(self, s: str):
+        def combine_cents(m: Match):
+            try:
+                currency = m.group(1)
+                integer = m.group(2)
+                cents = int(m.group(3))
+                return f"{currency}{integer}.{cents:02d}"
+            except ValueError:
+                return m.string
+
+        def extract_cents(m: Match):
+            try:
+                return f"¢{int(m.group(1))}"
+            except ValueError:
+                return m.string
+
+        # apply currency postprocessing; "$2 and ¢7" -> "$2.07"
+        s = re.sub(r"([€£$])([0-9]+) (?:and )?¢([0-9]{1,2})\b", combine_cents, s)
+        s = re.sub(r"[€£$]0.([0-9]{1,2})\b", extract_cents, s)
+
+        # write "one(s)" instead of "1(s)", just for the readability
+        s = re.sub(r"\b1(s?)\b", r"one\1", s)
+
+        return s
+
+    def __call__(self, s: str):
+        s = self.preprocess(s)
+        s = " ".join(word for word in self.process_words(s.split()) if word is not None)
+        s = self.postprocess(s)
+
+        return s
+
+
+class EnglishSpellingNormalizer:
+    """
+    Applies British-American spelling mappings as listed in [1].
+
+    [1] https://www.tysto.com/uk-us-spelling-list.html
+    """
+
+    def __init__(self):
+        mapping_path = os.path.join(os.path.dirname(__file__), "english.json")
+        self.mapping = json.load(open(mapping_path))
+
+    def __call__(self, s: str):
+        return " ".join(self.mapping.get(word, word) for word in s.split())
+
+
+class EnglishTextNormalizer:
+    def __init__(self):
+        self.ignore_patterns = r"\b(hmm|mm|mhm|mmm|uh|um)\b"
+        self.replacers = {
+            # common contractions
+            r"\bwon't\b": "will not",
+            r"\bcan't\b": "can not",
+            r"\blet's\b": "let us",
+            r"\bain't\b": "aint",
+            r"\by'all\b": "you all",
+            r"\bwanna\b": "want to",
+            r"\bgotta\b": "got to",
+            r"\bgonna\b": "going to",
+            r"\bi'ma\b": "i am going to",
+            r"\bimma\b": "i am going to",
+            r"\bwoulda\b": "would have",
+            r"\bcoulda\b": "could have",
+            r"\bshoulda\b": "should have",
+            r"\bma'am\b": "madam",
+            # contractions in titles/prefixes
+            r"\bmr\b": "mister ",
+            r"\bmrs\b": "missus ",
+            r"\bst\b": "saint ",
+            r"\bdr\b": "doctor ",
+            r"\bprof\b": "professor ",
+            r"\bcapt\b": "captain ",
+            r"\bgov\b": "governor ",
+            r"\bald\b": "alderman ",
+            r"\bgen\b": "general ",
+            r"\bsen\b": "senator ",
+            r"\brep\b": "representative ",
+            r"\bpres\b": "president ",
+            r"\brev\b": "reverend ",
+            r"\bhon\b": "honorable ",
+            r"\basst\b": "assistant ",
+            r"\bassoc\b": "associate ",
+            r"\blt\b": "lieutenant ",
+            r"\bcol\b": "colonel ",
+            r"\bjr\b": "junior ",
+            r"\bsr\b": "senior ",
+            r"\besq\b": "esquire ",
+            # prefect tenses, ideally it should be any past participles, but it's harder..
+            r"'d been\b": " had been",
+            r"'s been\b": " has been",
+            r"'d gone\b": " had gone",
+            r"'s gone\b": " has gone",
+            r"'d done\b": " had done",  # "'s done" is ambiguous
+            r"'s got\b": " has got",
+            # general contractions
+            r"n't\b": " not",
+            r"'re\b": " are",
+            r"'s\b": " is",
+            r"'d\b": " would",
+            r"'ll\b": " will",
+            r"'t\b": " not",
+            r"'ve\b": " have",
+            r"'m\b": " am",
+        }
+        self.standardize_numbers = EnglishNumberNormalizer()
+        self.standardize_spellings = EnglishSpellingNormalizer()
+
+    def __call__(self, s: str):
+        s = s.lower()
+
+        s = re.sub(r"[<\[][^>\]]*[>\]]", "", s)  # remove words between brackets
+        s = re.sub(r"\(([^)]+?)\)", "", s)  # remove words between parenthesis
+        s = re.sub(self.ignore_patterns, "", s)
+        s = re.sub(r"\s+'", "'", s)  # when there's a space before an apostrophe
+
+        for pattern, replacement in self.replacers.items():
+            s = re.sub(pattern, replacement, s)
+
+        s = re.sub(r"(\d),(\d)", r"\1\2", s)  # remove commas between digits
+        s = re.sub(r"\.([^0-9]|$)", r" \1", s)  # remove periods not followed by numbers
+        s = remove_symbols_and_diacritics(s, keep=".%$¢€£")  # keep numeric symbols
+
+        s = self.standardize_numbers(s)
+        s = self.standardize_spellings(s)
+
+        # now remove prefix/suffix symbols that are not preceded/followed by numbers
+        s = re.sub(r"[.$¢€£]([^0-9])", r" \1", s)
+        s = re.sub(r"([^0-9])%", r"\1 ", s)
+
+        s = re.sub(r"\s+", " ", s)  # replace any successive whitespaces with a space
+
+        return s
diff --git a/whisper_service_deploy/whisper/version.py b/whisper_service_deploy/whisper/version.py
new file mode 100644
index 0000000..67426aa
--- /dev/null
+++ b/whisper_service_deploy/whisper/version.py
@@ -0,0 +1 @@
+__version__ = "20250625"