From dd704470729630d185b1f16005dd96ee805da2e1 Mon Sep 17 00:00:00 2001
From: panqian099 <11823801+panqian099@user.noreply.gitee.com>
Date: Sat, 22 Oct 2022 14:30:33 +0000
Subject: [PATCH 1/3] =?UTF-8?q?=E6=96=B0=E5=BB=BA=20=E6=BD=98=E5=80=A9-?=
 =?UTF-8?q?=E6=BD=98=E5=80=A9-=E5=9F=BA=E4=BA=8EWR-EL=E6=A8=A1=E5=9E=8B?=
 =?UTF-8?q?=E7=9A=84=E5=B0=8F=E9=BA=A6=E9=94=88=E7=97=85=E8=AF=86=E5=88=AB?=
 =?UTF-8?q?=E7=A0=94=E7=A9=B6?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 .../.keep"                                                        | 0
 1 file changed, 0 insertions(+), 0 deletions(-)
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.keep"

diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.keep" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.keep"
new file mode 100644
index 0000000..e69de29
-- 
Gitee


From 78d40953301b96b1a9e0b0f9a52a8eb7dea34c71 Mon Sep 17 00:00:00 2001
From: panqian099 <11823801+panqian099@user.noreply.gitee.com>
Date: Sat, 22 Oct 2022 14:33:11 +0000
Subject: [PATCH 2/3] =?UTF-8?q?=E6=BD=98=E5=80=A9-=E5=9F=BA=E4=BA=8EWR-EL?=
 =?UTF-8?q?=E6=A8=A1=E5=9E=8B=E7=9A=84=E5=B0=8F=E9=BA=A6=E9=94=88=E7=97=85?=
 =?UTF-8?q?=E8=AF=86=E5=88=AB=E7=A0=94=E7=A9=B6?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: panqian099 <11823801+panqian099@user.noreply.gitee.com>
---
 ...206\345\210\253\347\240\224\347\251\266.rar" | Bin 0 -> 55088 bytes
 1 file changed, 0 insertions(+), 0 deletions(-)
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266.rar"

diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266.rar" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266.rar"
new file mode 100644
index 0000000000000000000000000000000000000000..32d43afa2f08fadb2be8f172278a1396c8205715
GIT binary patch
literal 55088
zcmcG#V~{A(vNhQ5)3$Bfwr$(CZQHhO`?PJ_wl(Lzc`+07_kJ<;Ye!W@?97#!D>7H@
zQZ#TRzyt>b0DnF8h6V)v0{{*P_*EI?4`9e1ovjZ94Alq#(g=Vr<D;wT4*(#1(vyDd
zUqN%6*Ll<#tENaJB7IU`aGV=;ni+gtA9vnadY&A9nig_a7kV6Cb()`iTG2{JYhi3+
zKu2q4;cQ`MYv*VJ2@F&%0e1a$um(tkFKl9J;9~8J|9?L+#y7RFHgST1q;oPiu{Ndq
zw{Wv_v~sdHFfyTax3PwSw8wox*ZOBx4gg3F0KSNm=kq_a`2G*fvNf>zk42zE!hGNV
zEHX2(HE}d>Hqo=TFmyC<^w6_2)w8iPHnDc1wfB(uJ7?|tZ}TWXkXk@|NaW){e_#N^
z|9*1|TPJ%HBWDXcTX{#j|M8s8{(l#GPA1OI7Pe+i|8e`DebYYUS5;#xfI|r3^h!cR
zFY>~>?#fF3irWvFmMMf)Dk#AAg{VLv7zv2re#?m@@D3{*ei+#qZ6(c5pYfFk^Tqgd
z{7w9KnV{29c@Z$2sq)l7Urun<Y<Dn9e8L2>;Mqn&h2i8`=u#*tv>SduFMTY3eH893
zM6YJe**pSvEHYJZ+&S4KcD}tNT4W}pXGa(8KQ%qNb^oC;0|-(FgfE!85%CWV!T%l&
z8w)3+|MeDED)hzba1C%+d6ET`iSdPBc!${Dy)=K`hQ0uk76e~30pia<KV>T#pec~J
ze$Nx4V``Or!tg=NFVfNMZ-U&&?&iy^Lzw=PCg@&-@+i~L$PPhxX!!k<1uh=17Umo~
zUZ2Mj@na&%7E`+Q76B(%r2iBGKn#S$eQ#qs_T}~;%Xrw>gA@K=Bhs?CDpm3lNB_Wy
z9S40eYrbIGWUuzi4WN$iJf%u?MO9g6!pblhOe5oxsZx0#k*G3$(-D>Ye5a$SD&U{1
zVGB4L*W<brH(L#LZ`}ZWQiFehO#_1T0pYj5^OXDpOyvInFgs(Hf7SB8n2|Pg{7;4q
zDo&7sC=<TutK0_O`7nE_nbU@c-Onu{Ll82!3ervtRYq0o{)$>Rui6}F+M1r2o^l*@
zpMdjWyutO}zi8h?Lae=L@hCTIUzt+CNxMk!FgEMbU+YO|+tG?zHP)8J@QI&M#u8va
z5Z2T8Ka>*V4Sb<-6W-+C3PUGz4xo@mO}DKz;OCG%bWQ;0c3Lu|`odzpaKZ6-<*QH0
z5gf2lgh2#eDqbUKk^{Z<r^x95P~fTa8<h02);kY!{Y0`&BKc|n9HvqUv0@VG_gYgt
zyZ?i%2neJL7=Lp^%Ksl^+W$RdYUU;e&WbKh&O+7(PEHo47DfjDDL)z;11Dz_M_P;j
zsWtPQ;Qz_KnMgc9dSxMIIu#I^12J%m>N!kk|A0S&85#LW2#>r-vF0oysAB2-iR|J>
z5W7JXalA0>4&s@zb<1jQctox_R8^gQd2}9LF>P5FnGl}Tt<KBwJ?QTf!YTz!a;nuK
z(eLkFP+4?%NY25v1t6*@`E+mvpL@Py`Hp;SQeR$6D&fJ_yrm6W0t%F=DcK*qx~muz
zN(&(@V4EO={eD>0`ADVquq&Cvjmoj^Sn#=4_DRMAL-|$oo55Z(H&k?C>t<n~p5cwo
zhnHj|e%RH8uHd&3vvePfOVNA<7>k0dfD2SPsFlkHB!x@1_M2MIGhnM91GVKu#zL2;
zNY|f|R`TDx^ZU%Ww6C~AUGRJJy`QBkx-@LGIYu<vu%K%$W`PrnB^}=;|Kf8K2>>$R
z1As43^Lq9#LSp}Yg#JH#|Bv{92o#}$|53s8n<XfFhzgSs9jlA(N5bA!<P(snL=vxS
zg<~nl-IxYZu&pU&Wkng$(bxm~rc4LpU#b-#gn3m62q5?xL{8W!d6#f_Tvk(|>N_X;
zRlE_74D3-SK<dh5fjvg{_g*`#mDgUI+XP<sonMfj-&NagRX4i<A0zsN(q#kRrHaNU
z6LgX5`ogN4ml6%p*wDYVJF_(_*^|<Ch?@zhEhH!mp|VhJ>*)Oq$(?TFBb*od3Y3(a
zDPo33EoI9SjSULv1+-)VW*o3A{{A&8;1&uNR_7b@b1mf}%q+|<;{|IA^UHIKi*klG
zH9At7s?W##+qrLtcqA@`jFNl16mL^Fiv&~#X^;>72rxyjh<VZ*VGEGzN?L#lp^Vds
z>|o2BS}IV_$g#XF2pf*%#_*rFv*0yfh|FMRZT$4DF*TC@(t(UZ6QLo=)$;rwsr+{%
zWAH3P@}$Ak=kc@AJ3-4bRTB)oognsXXxp}xT{WcV%+>2vw$Ot%JyjXdrQ?%UJwi)M
zUdp94HM%xAm!F$qSMFG5UNV!pm(H~-Lt1u3k}M&u6>M`ZQ_&tbl?CL0u!iuSk<^ms
zo81%%9jI|^E-@dG8y2h$*oZ&~5A@uTG{!MtV{Dw*@h>yapQCatNM1Yl0+QX*;`R8h
z!5=B>3`-&sMnI=^)}i4k+LYmt8vBvm>T4#c>@bFPy{BFZ0rJ!^`Vr;W>2pw|n-M7c
zu$>XwDu%=pwmL3~ex{~;Y*}G@4~GxmdVQx)_TRg4Xrqrt@7~#ZXkdJcN2BjP+`4?-
z-mh=oy3*cwaF=iH?Y_I{6YsxvO_vbt1h%Z-WRr*B451~_`u<!*AYIn9P;B<hIr%|x
z5Ej%zv=CcpS*bJl`=Ea!L*^}9?s*^)K<QKrqyhLy5JDAy;k1T!Z?7IVa7mz1tr^j{
zEvQbkDvs)?o6f+va*fh6i1B47l!ZHhIS*VXQ4Xr_z?UVVLmZ?e;uS8dKWr9|whFiH
zb!7n8<?MP8`fWO#1LXN&+U*xkmcdwe=;nHv@7HjR`QKF{WE+{O)f)JF6nQ(Wg%TG%
z`dg*}_=Cf?+c%AW>3!+jlAzx9akj<Gku_?r*NcR)=JKW+p!!Ko>OItg^R~q4FXrNV
zvclGG{$$4Yq@f>)!rt)vkk*rO)gUZtfK=x3PD?_doW=Br;bHN5xDp6ASq_{QWoLQM
zo#B}dEOB2kf~%KlTen(EL@gRe^@$~ATyRjhm#)Jwadk<IE?3$Zz%P|)2B#X&Rhbs#
zClDqr1IAR4!v18W-7i$26R6EBbSht}HAbgXO-3k`DKgdrrPmh*k&`ZjXScRT7<WGx
z=rvSI@w5O*Ym2LQ2P`qIWnKM>S;8r#Mkuf$VTHCa6-mR1T%D4{o`|iq6`?vm1<vH1
zzt~@{*88<<3ec+Q&03dDM}1_OP`A5%A&&H@G*r<bBySMPU4j30{>VL4(Jf-pxj0Vl
zzV;o-m3zTGQNcA!8EsA^^P2uOObff93!X@k9`)omLE3zePvUKu@znb;f)(dWk;#E3
zW4O*?B(BI|VqxaE_56SihcT)Ra0})GM|3B~ebwck$t{1~51E_7CKxbx#G5xTz-a#U
z7`!ydm{#<Q!y<I>G+s9`r~PW64+5w%7C1w$Ff^g+ISIeWyxsz-5xO-EWSZi0cYI-v
z-h;{@N+2)YzR@FX&wY)S*?|u4&YTlA3KRqVyycT8V3<Q-O=yE7)qCkKm`P-2J%|3z
z=-lAkj_Azv%#U`V)&a8K&5#N^na3ReR`YJ<n=MeVNj~dlYhIcS?Vrdf<3yfQaWIi)
z2R12u<TY(BHt@^zGnSF=L<#`>N9@(r03hY3fcUp%-9!I~-RQp~c6$$7Lp>vN6C*2o
zI}2N9Cpr~dXA5U*6JwhH_a-?51+B$@Ui?S;YHxqo6aRI)W7XuS{9wY&Ow4j9Cc3Zt
zLI`24z^Z~>tO^o1CV#1I*kB1RM<Hsq!xy6eG`Uj8&(+fbpenHM3sc)%N*rp#IVp_}
zX3AlCZf?cMl))TYMbx1$x=LK^f|&)KS7NI<gGO;fDb3_u?&K$Q0aaR>TDn_W`lI;L
z@dbZa22XiYmYRaktITl|E#D{Zs_LZb>>8J(9TF$F6n{g`T4-ulhzYrjZX+$$74}%!
z(QisYO4)F$w7=aht?OL0T*Ms|K6_;T0HD>ty^mtV&2^Gz1D8XPtcIt>P8v=n)W8M+
zwKjLq@JgV#V%tv0u>wCX6Na-b(-TT-G?V8_P5JhRvk8QW?*<c`QZ`=)Oqy&^V9Q_*
z13;;c?j-VSa?t2T<x~$S55nk~tIX*BY&qs!7cD$~|FPcjsE!aAA(@SO=(v2k@%|Yp
z>$uvw@#`HIJ+2;O%vs0PD;N#DF9YAmU<iP@*YSd&Zi1n|dzV&gGK@fejC?fX)`Q(J
zu7uD}q32zlFG7T57hF-F*Kn;oX@t2d?ts<Oee<_Pum5iL4w9R<DVYnf-Wxua)*(>n
zEfo;o#ut3&2cyc#foGh4;blUkN`)x4>vb^YPj}&|b!{8I+MA+o{76B(hLEzI%&%tO
zAU+aW+eR#?FJFL;%2lR)SJRc()sSx>2ex=27VN;*xwM7);g1;1SQ>z#+?;8oh7+D8
zIuwT^nMWz)$!xCB&AeU@I_I>+twU0(SnZm9r7&Dvy#DmudEKqJ`^5CAHUIsU^vv1q
zz3p1Z;~gZ;Y)O-1prfFaypvV`+p+kg#<4wmg}HST0MXK8m-R@Aqc>X1zv1&!v)fcr
z_3nAL@=M3#pB6DVN2u;@k4MFzC;D?r4{t)r7J(v^xw9&)BRf3nRN~Y#{X6q2c>3{!
z#@pL6GrQ@w`$_w$z1gE@er-|l`jrp7olcE%7f{!@#{l~M%|Eih^Tk1Q_FVhiSr=vd
z<!`X|)fy%o-YG1-{N=kBdM}3NtFL(ZJD}L7YWT{xzH;G>nP=xr1b&?)^Ob59wR|iI
zh`w5VXxv5IMrj%@Sb}g-3n@dMfUQQIps$}-CD*agvG{uP<f|TiZZ=aZ>?wJ7O!ddz
zjZ$h1KjzEG+J#JpT0zu><HdCrLq`vMxa#Wz%@1$nv}3m8%M)`KPWD!<qw3N3Gk5h#
z^h97P+yk=)s@n{}6<2pkp#Uuhs=wDoI5+|vA9NuYIvq?N;$Gy!ic*Cn@D_!+T)x9%
zS14NZ3?(PDoILH7AcR+XB9+}6UdR``f`l>i56BaH8n154%&U{y<&~wIkNptAEA5kS
z{l!NwY{_g?^|!g}nEDg-pmxqo!Iq?8JsYWkw-fnw1aCX%H<1Jdp<LJ&%;&(b_wKIS
z$ERnXS*IGZ_p30wUeB>F?8-1=@2`LLznDH$_K))srl@8z=Uz*$vYBJKN(1>yjk6yC
z3YeG-=OA0Hwocs09-9W%GPk~QIP0ZXON<&*j7Z&3xE~UW16yB4Nw+>yn+A{ADxnKq
zAA#Za_oX>>y+>8C;i&8oA6si#HZb6*N{c<(BS8gS#uN_6@*l^Ft`P=8<OCS)gyaBg
z$(}|9W_*OyHU&b-BGBuf9U~Dqnji>Yqa)xqG<>G9DSkk!657j0sV-EHmru{MGY;Hb
ztc1jTSgpxK*>?;Rc@PZC++{S>|6Tt0WaIid-w3_!d4sktF@(v4kmL7zn5JqVdUoZf
z&Pw=9E(Zd4yi@~+?=t&4D%Z8J^6nEhy5COwtv7S3wYhlhO6&cNmHj(-J9>HP6h>xu
zhy9B)TejIT_ao;|K)qa}zJ{PEq&jKasgit-7qv$YM^E4|?K;^VrBjAAO-yTY`<aGA
zTbdnwnkFPz#mq!rqQIXQ2{8|@4w-V1K0(Qu#F@^iOm%=)c^5ll#nj014iCRE(~I9X
zLUJV#bEg!D;V)_Lf4)9hhJ;cu_;>bB6oO#rvFiMN$LNxCS)Kp^2?jm`^Qt3RhMO{y
zhu<tkLYFW-S@F1b#MQk$pSMMasFGk56bw#&KsbPCAV80%ezEFa6K`1qJ7XsdhdqYb
zrsD{?Ng+Fp=8G<dXb>Csp84|^bmn#5>ASb8N4K}mY_;p=`;`k%G|%p6>D;OO{EfG(
zM|LOAYZv<4Y44-?Qzvj>kn-rDf|^v$3RbG-?is}VpkV2<t7Iwk12LB4yP*{HvfvPL
zF@)I$j(SpT`NX(ngr&H{9r#y3F+xQ_+W`e_>})<fIWaR*|Brh!NX65+7-$L{7>*)+
zUT4PvQHmT)e{C3($^uMmzfuTpl!$hBescWCj4K0BKluI@ssh+ReEo}gaZ3hF$nP?{
zq9yvz)2hb@{}n@Ak^B@SDU<mQ!YY-&A?wUYSt}t&Q+5CNqz(FNAXeJkiZGf+8@6uK
zC|o)WuN+4_fRD%i*OV(;K5lzX$-mk*lIwcJm7<7}Xc}NvO#g$XGW};M^~tT0v3shY
zxa_ZliNH7}UON;wD0NNQdL94)+F22M()z*o5ZvFUFcMb~?(Kff{g8^H{)Xy`eG>oJ
zcl{DQ5NOQ7HVWz&9CSDs8ZLn9MUt2KO)@~6<9gbmH!4&(Fm}DXa2&SU%2GgutgusT
zi2muBjAh59D8VIBI)L&kcE|Oq!fc><6<J3D>J0E;1YB{0<0u&+K~A1Ea+;aiW<;n|
zjq1do-QKyG0C7AR$n3vuds-*4rB^(}ub5Gw#EeOW5XA`T`21^10}8ycKWH?;QtT~#
zpzWr9F^+k(Gq$pX8U;n;wTQcedvwh?;hO4&0}TGAhjCy_m?u>bnb}95i7&C;Jr&QX
z6m+bJeyHigL;~U5Os?0QD4{xF+65T_sQUW6N&|mj8MFXK9qIL;-R)YWgu)HHjs(Yr
zoya64_DZqw-0pa*TCYk!ABQj)`x#vjlz4;D{Uo^D5gjA@0(1W;dWW-(>N9P&68xww
zg0Q6H4N$^WO-aoWd7I-JSH<Z&I|zv4x|>vG46mBRDi`5?Uk%_@$|lq!p=vOh<glWT
zL{B{^KSTf)T43<JGng3lb8~w)I0PeQ;i6=y%Ikgv3Ta`{J+#<pYU?cB8M7Z4V??=Q
zC^Yryss&tDtwZMkfjz2Nq{_`@V>me_(Z@HI0eD?AMqPU6w(Yn+3~D^AfWOVj>-L=Z
zd?}qicf$o#zQc=ZHo4jW0gvS3SyZT>Xarrai#Gr5qq<{<Fjg$`5Iamkc22<W7zw<@
zm`o%iS~IZwHVA?uc%qHMnIRAmb5Pf4GMo~QG!qKW3DBEXzMkEGuzdY|_3Yl}{oV2b
z2R;+l)rvFZ4XC*ZD3F;3s7`1$HRL}MGkL_25;c_xQ$Ti>GmKp!%U66N`Bp&>A}^r~
zKg*~kX)X=Ncz~*lh~v@5l!-LnMakha9{+~6j&v?Qi$zh!Bw{;TJ91YjY7(D%SDaH#
z2%;#;@=Am;Le%RR6Ynx3QU&t#Y2*Hydib$X$hdeGD9S~oC@z-6L0&|1#OqDTH)97o
zAc2GaJ}<8^Skfb1{*jR=p+*O*?_CaZW!4)XkcYOnuD4!vOdX8h0VtN4)29XOa-3(_
zNUSO6(qdWRttv38&Z$}o*2y@64uUl%raKuZvXr7nPrb_GK(y(sB5+kR7$n$5AZS)M
zH<e_`Ww%|_1cH9{5giF0F51JRDTcMmPR;|`+Wq4tG9BU_?$r!yBxM+UXaaFMR#IRL
z=gJgPFUlnl^6utyYp2%tD|W}z@tu`J+rC`C{2|rgMo#N7O%d9xZIxz%NdPfW>sUt=
zb?WKiT~N80Uyx8n-XVdJvZ>r<chV5*Pc0x??7_K8zPw=bh@=AV5zOJtXB3i0zoE<w
zrkY(`sZYE{1WZ0cFc0MX=3E|MIhR~S-i>;mLu3(HE83a9MdRy);_i9jKxQKHWJPCj
zo-hW_dDUc{d_AsWlGHR=JS)O7RD2>@)v^9!a9f+^=DtzO=~H`BHoVbs0|#JxI{W0$
zn>`u*M_g!|pIiop#(<a&Z|2(>S>+Ls`HlI$JnbP_&)Jr)XzcN=*Z=p!u~uJRAdn>d
zv!+2n5Z`M~y{oYP)Hy3O!l<Ove$KFyQbS#JH>qAebwE&36QRjIf*+3#EPwABc9RE=
z>5Fgns|S|{Sch$*?;Dx4w17GmQ7K$ZvQ2g)aDl;?o~Apbp~OwdFbRxOJT&E?mdP(`
z;0K&B{|Hm|tQw3QI!vl8D8&D62xlu78dXG+^RrjQwJ9cm^7%z$uVY%mXKy;I777qf
z#ywi$ppzCjl*l0)hz*Y>v%8DI3?MS%URg}>L1;j;#$-~#kWfO=(Pp7Uj*T|&b*an@
z&6VBXMskW|b5Q1ZKAeFfmjgbp0%Cf^zM$j^XPffxmh>_*Bq8yk#)EoG@rygLVHP1R
zt%$UyHWs#I3N>9k@A<nzlpyxq6BMu5sY&B9#zPV0{ncNLaZ7qtrqdcug%AbD7=^&@
zdlylZrN)|P@}nx<_A#^^+i@LdSLV5eAl0ypr17J^eL+~oVaq@+7i2P6&#Mhlj1+Py
zYF-(TLc%PCeErz5X_{bUSJ24TbuUOtAp%=?5i-DB_%M82B4tUUlAHx1LrTxzl#NcP
z#IYS$QU*{W_!uJQ*pCkCp|`pqX`X=s;R~5HKS-SmV-r8|hWmnW!h&NGIZTmAh81+4
z&Cd`|SDO_(cFP5Ec2t`LU&L55rKX6gg)gy8qQ&R@yX?{Uri~xWT?`K^Jv+@?w)1Ub
z3(QzD@A4RgnVnNy%FW3_>oEA%doI{LzT>GQcWV!WryepaOzTv!B7*Uwd(BFcHZb%a
z>{Rt>JV4D1JvcI2HCV0VUNZ{5bXH4)2Sq_nz)1yC4fw@<qr2k62m-{C6wX6;gbCoR
zusT%+0Z|sy9eYew30%4Sd0KYCpnwq`I*PX&6sJyM5c}!z?Qn;1`@4p>EZqG=u&5Hm
zo!}oL#EA_}!mZM*<CYNmrFLE1y5t@9WTkz3Bbfv^gk>&NdWtHywLp$8Ol1yi5h;_c
z^N&X@3o0AvSea91rVuqr2m^3CzfsrFyGop!xSB^jo3-z|_9SXUg<lT-4Y;`?r`OXb
zb@%2N$&w+Jv`0+|ES5?(vMuEB*<4(}&XMf6;DYO+vO13TTz)Rlz@2NMZ@cSsfBXBo
zHANSA>5tlFk5jU8Iyg-zzqq2&Ll3{EnJ4exr$dh@2lpShQeF=1SiZ1^y6DJ5*(#qA
z7*BkYlqWHl#_neV7%Mt7rMSp9pc<}z41+A6(6V;0L&dn{8j8yt_P*=Z{x?jlkPFqW
zeXwq#vY-Zmx`om~1eT<Hb<G%BR;}YyUGg=(XetAq7<|H@gAXoY;jy(k<I#L$--Tdf
zL}NtFV+8CmT4d|mO~c>2JNC`>stKIHNt5XIWs~3y@f|ryjYGDzF>PI?wI>NN{Y*7U
ztzk0|#uNTsz)qd2oH}1VL`Lu4u%eFZP18=W3!T8E78L$u2FU4=#?8d!LxJt(xx>2!
zOre(q!E)i8l<hXH(E|dLp8Ma`8Bt(*3?KUp(x}l)&2_5r8=y(965c^?w*=SFr}Nj{
zJ~9v&d<id8U7<zKF2vN{e&=KlhJ&cqZAF?>IR<<037?(2AXX?K`XAHl)=?1WDe@^2
zC~@B_ShK4r_Qk&;p{7gsqiHeO`P0$n!O@IJfp|jo@+;u1Q=uC7cR4iQ7s+}K`XqiJ
z=x@;K1{vL;jx1McNnb<C4u=Nok{&jkF<Z%LcmyDhG9%ERLVvJ1X@YtRx->Q>PH#qF
zqHrR2op7eB`GW)P*jFKjcnwzN`&$Kj?qV_4zG$3=GTB1pU|0JY3sZ0OtskX`e&JnK
zUki!LF%8L+WP}3?VSw3&>u3*3cJ|rl4npLnh2M?yCrCu#=fTXBwRy>#Y%YyRV-p7*
zL<JyXKxZ#?%LBns*g0RZv3Ce9QE@^q_v3c?EU&jbyOwGa+Hq^kpmeNxf2MfM&c)>8
zd{1R_Q(10Gye~VSyD`q?u~7J$(!RoBk-^>FKhY%aX4vd>VMaDR%+#2Le<cLmiIy<!
z*(oNDXQvmn9H!Uiz7j@*Rrl5>Q)#+%^sA}NX72=CcWjZ;0ZfkQeopTLz11cQ+6tRw
zoz}JktYnpUaPS?6oPqP+zO}lJG>w#!Qw%8;S6X)1K!)Zmq8qQQxNmO6GBY{+kMIMx
zRPNd#&~FD(cQir7Y%15p!^Tz<9g2hq8C~J6M5of<<UIAo25QDej0;F;UCl!8Hp<un
zj(V}y^djJJftzZGLbJSntt5mAs-^-T*zbxVsp|IItL#Lvdc~hxLkB+b<|<`0^$O}Y
ze($YWQ!`hL3zI&E;;o?Y2GeMIWrr}C*!`9&hO#N6WV$@isPU}aQOhRH4FOmRSb0Mo
zz{5T1LZcNz`Uja;0mhOB*~^wH{7!&}{<>mS>dqeNcVHr%%jXuo|3GsIbjj&bXN-FB
zQ9$m|m+Da1Tp5IXlr-pt|M|)wGOA!*(1Oz2=TgFDvk!x{rgrw#+y3e7ki_f?-I&X+
zvYdMW$H0TS1<$by{w(*vN4eV?`gVeOWb1qXla@*|iT@R+EoAo<y2~5Wm9_Q51MCZc
zj2Ez2wtX%RfOR&?ceL`x>(Cqd^^MS*&S!r8gXs$wdup`XvHTIW0l0MeJJZ(pEzSk8
zdusQ3Cr&Su)MuBrXVTYaKfa+@KE6&UZ=G_A5mQczP32zp;P*|mRHm(E#XjJEv91gS
z8${gDT%H00N*6*o@^ZVFbUo46gOfykEOgT>`hte{szqMxXY~d=(g`ZbQ)OM?v^69J
z2#8OWLfjz>V%9bx?-fs<(t6u&fW4wBKSOa8IOxC3SC^}7<hhgMr6`VW0^an9Vz89a
zO49Z|{m#vnrE3taV&oED-;-Nh!8&OGuhy11)^}*C6y8Mi(57Bj_y?x?I0u4@JsU5%
z+~(!ZCaxh?g~4LSdYlyQnPoiYo0CF!9pOj&ZtWApl%4;phP%4BXwEYuyvB<?+Hb-J
zhC-BhFg04>Tl3DgPtnZBvyb2-JgJ~<_>N1`2B-XImeZbmXqbp>@kZoj!{}wxM<+wu
zc&^aELIYW}XX;>qdi8RQ^bu%mgCu&Aw3LZrUJ__wKF(o_*TaB=^7)KH6sN+FXN_)j
zUdbU5C@PC6#pfOKeQhsT*fL0g|8nGnZbqMV367N0t_zw+jQs>j5K;NzL)ctOcg<=^
z$tE`P4j(O%duVt7aU3~KJAm|Py)UyB$aVHeAiIz+M$V%t=e9$teU4rS^kPmvlJW|+
zRN$-WlLM|;iw;u-pA}R|W98!1>SXm4RG8p)0pQGzTNR;#%Ah@bq{mgALM_T|D>Ror
zCjxkWIUhFR9hgK(ktWaE{jU^SMFXhmo8%B+K6r_jo)jHC9<Ugy@m**V^QR0TjHFD#
zq%R03)I+0=Mo#><Gr01fDE+o&lzmEPltb(Xg7?R66jSclme5rEBT8*MbWnhjx=-Z)
z9L{(d2LRc50m5&y#5(_{SnL1)r&#|#a`n33<7ME~3gC?31F<}~u#gZFABwA9=UsJ}
zyQ%-!96u&o{*?hobz-(6dBHv@+c<Ns8CXD&8^pjo7+~?Gf-L%idE&fq>V^#IxDHJ7
zg3NufWCU{p339igIab$$FLeLnN9AL+uKqv|Nb1p1E)cu;)MRI8Wmo5au5PS@<u9^}
zDl`@Cu5W%FsM%1HM{cB<2jpEvjZlnTGm3&E@&b?YrO4J6Yih-!!5)ScMUs`%8p-XF
z%N1owts3(((-sC5D`ks$yoZFx?w4YJ&??}sfIgZERN<6(4(oVb_X)eTTJi}Tf_VHi
zBefKbLjJ4@(;gKXHNX$Yv3!_w(?5mJ)(}*Z4w(6*FiSd#Y}tQWOk+$fVK=HL6oAZ<
zN&Bs|a9osu=GP7XbYw~;Px>!?0mUY+F&`oCEwJzCr5|HWauUXTjdZMi)OzAr<=z(6
ztv=Pd`Oz-nx-RHrNMFO^WsKT3jM-IZpZbSA%UDKIRz9aS^BzIGP)ZW!CaBs_i9Ah*
zLh9E6lmxuA@hY_<`-fQ8&8k=~Ije&`gL1Ii)cEuBMJ-uo@(z@ptX4P5)08kiHPObG
ze+UOAbn=J1#6!z5Vd9(PDLhZgSE`dH^v(}Y+OD~L_~7J@mW0taLlu*!^%GEZsY$R0
zG_@WHt>(fP)M{IRaqU8)1-Q%+<0~i@kONmQAd>V-QE9drt380kA2*E;!H{n*RFb3S
zh$TE%g2%sz=*L;Ekfd0Y=~np=%Opm0+xV$%b}d{^jb+trJ-zt&rRhC+a_)-p`}H|*
z`@T-Ptz5_Ti7?8j&t*(1jVl^=7o#%Ny7aNk#JN<0&5D%^+T;y<{3eKBdH<{fTe*F*
zdFj@UE}xP0{J3TwtVRJ^P^ZkM3OrOky}z+=I2a_6@HcCavvm6ASLVUY^r@aM&i(Mo
zDW2}X*Lv$N&+Lx6tzGM!KheH;noG&}r9A_{Y2w$b3^e!7S^q_O{RoJujC>0e>v^O7
z)JY9#^XaF5^Q{AT{u5OxJRO&Jgvu7rtMjd2a{V)yVrf0OO($ix{PppR=dKZ?JXZXp
zBs#I=QSUEmwa}zp*YGcOSc0WU;uIQ5SacfMEHyOr`SXIbx|TauUT8n^Di%4hFqX#R
z7<H|Q_u&u%OLzxB>~XP-CWMz6K{9EbTDAm_U3q&)M$a#F-haGud{bt}m)5^=ta@^9
zMlXM}a`?7kf(UWBv!ucTv}@=rENg3nRh`=d%r9L3Mh+zcU9^LV1OEY;D}-rDu1?o)
zN0Wg>rqqrGb%@mEP*|;rBayq5tn@z1O6B*m6tWPjq(HI(0gAx3%B$MwczSi^r3=&R
zIRn8!-|D*X7J~f5)2ni7xMty`Upv5!nxZ<o`6%yCx~%7&(eUwjXdS|1D$zZIiC~-Z
zYVjwn{QZ97JGs4kdg0P-P!@RIf!wofLva(-Oq}TY)sU}s_Bv(>G4yz<?p%(Evz%*M
zz5=$b@Goiv#+LKXDr_)O70z6`u5U@-q^_CnHr<4s36enpW3P<L((mk?MVb`0X06#b
zh<r)c6V}3b$q44$5I5LGr@metFmFw9Yf}SSELm#POdvo(g^oKmz(fUKMa%7x6~Fcr
z?E_K--1kvd4bcGBf<BK7!v9K7hR!08ViiwkbZR2M(oPVTV4~<EMrQf>7(Hq&d<xH|
zMbOl2PUm2{RmaeOHPHeJU1~;=fwZsqJLI5#)btVWv3h9jt<#nBs{tY8w)!6GN?!#!
z4^kzQ_jPEpRQ;nn2by$cSt13%f1{zPe}H0xk#6PI*~M2iZ|;i}csuZ6>!q99Hh;JE
z>LZmymwDTCX8j~)=5&JdGd=6K*s=6A4M0LYUn2fMke+H2dDW<tWJ+`bfMk{qzux4*
zj89WJocTdwc<TG>V{kAP0ar6Cgr^vV{?#FAJdXM$DdWtU->Id+R6L2Rh~V_6MX;6#
zNB+;1c>@-aegYtDYbqdUUHY6Vq@VB_Mm_{9EPYOKbC`S$ttxn2Aid-$!Y*CHD|?Kt
zOC-J^Dk&nN0OplOz63XEgb%)%pm8s9w!!JY@Y66T@cU45mUj|7Qjp5gQV0j&Rs`z7
zwo5EGUBZ2%f}5sgjK-gr=CJV&Ffl%Ybi$Z;T62|*Ut($Xz2EroKF^Nc>0Et!;q(dK
z);w&z`CQwccGNAOO{{$UQ1#Y1X`R0S@o9hkg8f4&Rpw?^+|{@-U4&Xc&Fsmuz}a@}
z@+;@yK}rYmFy2pZ2rBZF*GRolM%3u;-{5!4$019xAk`MgxArh@qAdIXqL@ik)&xI}
zz0PPK%pGk%;k+IvU?4OQCLB##S(~nYj7(*(Gj~%X!=HW`l}+-bPH|iSQeuRq+FP>&
zz=z<kE(K#Jph-T{V1bBv!gricDV)UD0i|gvZxFKo<j4RN9-|5W=MsgVk#$9k&YIfE
z3Oyht%O!f%Mk^h1MH)@ACyPq4+5gRd$^0P#0DLu-;l#|uM<DWGNV+GkZhz(0x0%5N
zL4wDgB_au+@Pd<YTJkmbp#0`BEx5TfSW@1u;h@hx*>ysh8*PIZUXW^@)<H_su&2Fl
zm^sXZWw3jFP;pD8qAozGaO|%X(9C;(i5xUE;$Rw!49hwO9Jr<mpk@>3N+S{(gv==|
z<&XvysvMYo^X?x`{e4wg#IxB$nOG=$=F|A7+9dR1Dst01$motV8cptZm?+_;wXxu4
zqXKzh4-&Oh;dRpT(-!^YHq6TDdRk3Wr>Nh43xW)wcwSga4oiE0wh3I73r%0Iu*jpo
z%<20e^hEIdHuOb-g(Go*Q>g7E(TW?2n=yRd6*6c>EMyom@Ciq0koMU9P!ePce^W{7
zg^e-VR|octDjK(#j78<z-kMDpR*06=)Vu$d*O)Hc6<Zpjht}B+5neGftTn|=4o1;e
z;*stJhM|=N3$?A&1#|N@j*tGCc)ene5OxyJAZ4d;BHICWUpC#=fHBG$9#0iWD2dEv
z#+rP8-yHBOR;f<D8%#DA=;s2=(_K@0SCgh+!S~%H=XJ|1OBa2!_sF635!la#q;+eb
z)|R#H5)UjcYq_-u|4zbe(kG>1G%Iy}7-5+j#gHS3$DGF1+Ff^(!bO6=bsp}97d3ZJ
zEm0TnQUghKImtVBwpLeHSx0xonrD;|>BfX(>Eq7>UQw+?;OGK-C|NNGb{DskwM`B}
zL`~j<^S{ai%Ke?LmhI0^?QRkH!6v7X1?%|=wtDu*-B5YZ(r%%Awgv(%vc>C9qV#yy
zx|kx=G}2Fb<3?dTkqjW~qabUIf9n{wt-^&$rbgHmkoz<Mh7x#87x4-TTS1{%vi270
z=NK;&1}Om4Tvnc*9lo)4`+W7{fwBD^3xoaRfB=*oiHMCkU~ewt#7afQ>K-jYFC47L
zM4=Ww?)#ubCxT2?VHg2HFR1C6e`Z`lU>&Pa@|UTi5X?MqQqY)9b+dGwV-3_)q0w+L
z+Pb3&uLGN+bSrvqz46bDMiGOmX$*x0nfO4kQx>=|W~qc_842Na4vf0!`OBu$J<a61
zvS?X38(2=em}a8)mQ#(iNWJfJ6>F?b%-{qKjL!q{WHEQt)y}VY5{lyhfCc2WIGKsz
z@0hJV`rf+sv&^6pM3C!X5}81Ny_~Je5?2yrW{{O(X1V8?#-(@z*j$vU3n6|-1q3bk
zD$7R72{7We(uxRS7vmb*x{MdGP(LAuc!j#9Bq!J0E|5A942So`{@>wXV?5|H><h=Z
zJ({?4+IbrXxACxEfFMjPW-69(rD+6-;dq|a1?2wrYmt>xo#$_!*xy@W@H(kBr>g6)
zRxI3iYn?%;8ja;}DUqOu3$XeKm_ClQ@!8Mp7+7=9#R}r9#Y~+NQxnzW)Xn?3L4H*I
zf+G(Dt4yca^LrQ-It5T2XupR>%}1DJSoVcQ0qJZaED(KpCInOgLR;%`Vaj?8@_c!k
z2HKcH$PTwk*{PoSlbwpem=pjQl4{Ke>;PcuA<}WytrAppLGDKJCmC%<bw-Y^oIZ)^
z=fUl5w%UJ<nVVk2llEmeU{&sIFTlq<J33^%lk-;!GmaG)LHn=M_2GExd;p=S^7=y5
zFR7Bc=<K5hudZJ@-&VLbI%2$*hM~{OsB-=pH@upg2=>Z3Y7OG7l=E_a5A#BOWz7JY
zK?+T%ax!C);WA=8FOG~5z;oy(FCWsE@9XxX``fAV2IcRUWM-stql<(n3JzormxYwX
za4dN%P)$slb2Hh6VS2iS%=)3+W572!bG;+D?0GquJSJ>vLYRtw5R{o?fB%SK9=y@}
z*eBwtBD5_EKfesD4y<z}nP3Iwu0v#%Y^2l_n%b|G6+y>e6$Q48B#zF^WjWjOGHM#r
z>D%}C0HSFx%<ehjYUaHLVGQTy@x9t|(6Pniv=yNmeM%`#rKHi>hp%P#jIZox_f4VC
z9Gy)yF1Wj4n<MA7s;IG}>v?fr7gTeJvDwdV<>aSlY1IM~m#f%7VJ`Q&5_Nsk%w+3x
ztx7E_K|R94Z0;L1s;_k{;h|C=&Vs*h&Ge$kf)5S)BDt_LJf0*BCEiXn0qH(g&A>f~
z0T0K_N@J7ht;Qe8<wvG76-tEO7akeNV#!|-&8Md3CnmCaX>6i;44Kb`w+W`VX$h6M
z_sdWdF6eP~3`j`@cF!nH7L)RYe?KfP08Updlu0_3wJK1~6<cv77n4isZ-o4;gp*&m
zR9X0zT2YA4a<w+?j<quARd^u!){|JPhFUr_QuO4^^jzqX;(reI*@w3JKc-C-MIR2a
zNh|uFi|h+ek1nvuvh2G!w{NQU06~xzURnM4I`IQ~cI5@&Tf}d?H+J=!oc3R6U44O<
z_Uo>Hw)2eM;GG72zx4iU6oduW4Fsb?5DKBH`hg$Fchv~I_b+u&0y9}?4De62#I*CJ
z<u3R*BrsxgHnJK_1Ctvix)j!=mNF)cCizpOiKdgc1+AmMK~iF!ds;E*LuTE<;Lo=X
z7vh-NYlgV*we?;|SnwOvGKx4h91sf@7#g!7cb~l-*-bgwgpv@zZ)F}^C&q?Kky;>+
ztX+H^Fyz&|VFgUUzu<`-tiMQ74jsPKKJh%YrH1xoZ>;(GinlwP@~(P}DhfE@n^L!<
zS&$-UDccPN(yx_iXIFGn_r~R8`R51dXtB89v(?(tW(WF{k<-Cy5IGkD6N}<a^?<%1
znNWa*=lKsxtJ4q71>n&L2F4k4WOq<lnBDLHLKU}Jq?KRSV1R$rF}yz#;-e|Tzem)}
z?TPnai+6D)?t~%(a^p&P*wyep4j2^SDnv;}knnsaT9<#s1o;YXPNDpy6p5tKM^K1P
zN_O)M{DmX2q?^^uWw(?W$<SR6j80B!9|t-(7(3<*6zw0v;Ohzqc}TQ_8r{-nWVW32
zZC9-yd{;$4ol8!^u8qxvUv3$HXLMaVs{#`6&^V!`MV0(g)U^xF%IK!fQ0tl!!m5uZ
zzTtw&8tiH5$wb+ngK2#NsPB!*WOb4AE|FuJ{Pgzz;rpN(d-z?AYW-WzNY@sViR7-y
zY^pR_qJ;0al9#|oX=^HKLtUNl%$uqkJzRR%L^`wa^nJ8qtX446q-_(YBC=~1Q15ss
z8!LYECzUq?c8`%7MV^cecyjX4GIE91EEy1B%O`q2$YQC6uAB^5@(Pj2IIUlRnB;TG
zI@KC@C`mg<UUP<vb(D7>h3;o%eu*Ay2%HE5TWrOc^-l7iu+$r;I7}|E7?cwdkU9*}
z=sJ>Kpl*e`1XQz++9MPoZfpVaJI^i_6N)q_NRjq2%#AU<y-&?n7D^X<5ZO4!j~}Ku
zKZuU^T%qZu;Q>;GazWd|clS_LAODF8*$k{f2kC{j?<9cUVlxw?ePStmNhUG)B*Z$`
z3Jv<AR(#EEWRDH%n=dB(@3^;lr-A5DO_+=hLw$w0lSC!x1zfLqe#~P#Gd-h0m63DQ
zkE&qtO4(nyuUs(FYnpUb2?<9?+ygjJE{Vtw!UdR2K%hNV2=FAI5(6R6F{L<f;$}Kz
zK%3d2JMFg*G|ruoS+VM}$RGX+@p<!Y(*TxfnC4>vSmOdJeG==)iL4V-bC|>v=j)~N
zUCuzout{5X225ZIMSD;aOyJlN+vm=YT{Q#P$TxE5aPJna-r1{;g{NBQ-}A4XS)&DB
zm_hLiY`ep!r`jJn!&9v-a{w(3tDCd0G%ml`+lGy7ZZ5z@4jBHvdEkIa$&S_6TGy^L
zFCBg6@6kz_t{d-HFTNh^lDxFl)$?cAsNJucWyh`9t4}u{zw|J;%DJsBfIOVp$qQy*
z^rEVFS~u~&MSq6@Zb2404P%0e{rg%Q8f21%xOdX9aVmdePR815Wv$g50DvDB)W7#;
z3vc@0^>RPWh~Rjm`_!;~J3R%Ne*+>1GfWq-+d$anMjgc+<eIOmUyQqDm4t1sEH+RI
zQgx@_@BxY^hSI&ATCYArOaV7&Q&LjBX>7$1P#k7WSMs1KEQQY(51;&Qu#$JZt71^O
zOVN<4KnGeT$UtaURV<Nb_^8_@hB@dhVn+UT`IEh`d#e{u?yFvYG&85;b<eL^0T!{T
zzGoXwy2%21iubrFwO&GHM#p_}NqUL(w>S$#s2nG0YphY$c>JhlC&!RSI*g6Cnx^Xf
zf~h%NAs(*5pSz5S*Ak`Oh>=a35uQ<J2?Ud*P&L@IK`F#J!D&etwcZSoDb)vmf(}8;
z(PnvQb!*bp4AdCcyx@KgV_x3i5Ze7uLSs%ufD<`GVGl&_y3^THs)^XX(Ba$&k`wzQ
z8RUQ?O`^kKnKeuD@X8%V<mjZ)DruI1pW4ZE6wZyRP|a#=$4I?BKts(s;8~~IW187x
z-GaMexdR0_ZXH?=rydBM2*9K72QeSA#vYU*+knYj+dt#z^FQZk`?Oan7b1R835-fB
zg1cLZOC$UPs~O$cGA?LDzN7e`(I2L!AK?)QzDs-RFuC{@V8otCkK4ygKp2sUk~sP;
zvDujfov|fI-hy6{8DWFqjgXa40pn1N(wC~MSBdC_!$LXtMwvS>5)>|hjPmINQrQJZ
zjj9CucZLKmGUmWswGO$`f;*E-qJGOoYw9`VpjV#gQ|L7H(napf#_V+^_5SW9C}D_P
z_>9+Eu}2g?M1*I_(Rj?9?GjD*(j^Jcr-;+u3$X;L{E739JksjfPf7*ncpq&@{KkRk
zPE^+8A)^bM?G}&hPUUjY5bZEq#~b`WWklq8qMk0+KWH2AfeI9IL_5qT$b?0}mB3&i
zBdmE-1v${AkLpm9f*psl!nsZ*$;VaKhJ=)nV2=K~xe4cEJuU+G*fQr}U6f$@aYWc`
z7fWBhvXd_k&Z-OCBs4K8&$+swE~BiJX#{qu4G4l_j~*{X*f-`wJFg_57_|~GWXuE@
z8C1%Wc?~ULdyP4pde5sYxLtZYKG=8<9;6vD`h`WVt=rOpredHC6*@P3Jjzo*loTHf
z8M<G=Y0Mf@FY`urX^sNGGOq4O$$ErixVARTObR8|I$SDCnYtbSN$f%(UrU{d!;d6%
z9m%uiE?6vYPC4J#R~FYNdevMGv5C1CGCFapTptpyr7lJ{Xh%r|ZI{E|X3ubW#(Y#_
zn}cDc(qN9O9N9Gx4o&@wBVLwdVF>+i3K9Yn&xAa_Ed+PKock!>;Rewo5yO@b8Y7Qz
zk47EZZc>v5A~H{LB6y%7o*KZP&uw8mz=p|uIhO;>P47^`VgO^koeKi%FBr>3Rop)=
zS;HzDWS}3(<wSTz4oa2`3|CExa<>|@Zb_6ilg5!475ZDNO{`mhdX`70yyYHm5@mGG
z{$9ubtGPBxz#Z;|>VH~~K9JdufV#myghXnjS}jS5;(!&Ke`|5tiW*YL^*5i<I%x%I
z%#jx10TtmGb6bjlIWm}&$TC0zGGQPrkTw-;qy1Ld@+8gtT@v$89EeA_0NCi{&75In
zl{D%>kYK)kq*_@x#2=bYcRP4{oy9$#so*{Kayv?K#@R-&AlAp$#v0&w;HyeKF?ZIV
z9J(e3l&~CrN-}Al?)*446bhlTIeYiqGPVkG*t2yuvLH|CWk@fsT2FOhPb76%h7H*K
zS2I}Ae;-)2qvIU_LmU;1bUmTI11`!)`$HD9D819&JZrz=pS!~jS0ApSAH+<xvp&{f
z-uB`Y{SS&ATgHH0kDM`gm`i28R5h2*lm4|dveuWwlmGO&20gmn?Dn&AGv!bWC6w%Q
zlnROt5u#2I(cw7Dp%8RN45mA1=(PoCewW;j^?K7M_|v7;0W93+pXaybb#85~GWWKf
zZ`T(!y{KGCwop)ZB_avQllef<L&CyE${%$jke@~)m&%Pf5g73ULAES~cu=_$27=fU
zv@D3a>K*6_>bu>2SVEv6144Sf<Fc`-*TI-@r7O9_zj*4j4qBFMDV1CI$yw~2jQYKK
z_#fT#V@%?RS@?ips%+QEU?cIH0st)%PO^5_8?{?#ipjdoCx2Kd|7dc_v^qGIe7bz=
z1xyIC1%=zy-Fv?J%Bg=ZwI@H*x_pG4dGCB~Dn1D(LjoL#NG`$or79eC2;ezJZ{-WF
zTW>C40@6rU2^&f4MA4e`P@ql<)W%uqGQ>oAsI_bn0mL%0VzP*}d|mod<%ONBz4t+k
z3Ke+-e_PqyFX66<HQejGa*|z#{L`<|xO}W~<fa6Xy&&9v*6Lq7cLIA$Q=j98KUlhQ
zxX#_`*0B~ef7`m6F271>cN1Zx1A5ZWbLSxBS^eak?X;R<W8SB8!@P_xzRaWv{km?i
zT*DWxn?*$eQ0&&;ex<H&irM|CY@`DmuFb@pVHmZqRHII~h{yYS*jQDrHVk%B5<n8<
zEC6LUMk)6H)YfreY((HSF5wKuF%`*UOG=++(J?hXl-)CyI<O}(c_*06qG;rJkV#KW
zPG=+~aWlycPEL`~a<+0`PVzB@aFGK6(i-1EaYHrz0d2VifL?sCA`&==j-IwM#|50a
zm&(ihgYS|G0fquh&t@*IVV<cm)nkGzX~IU}KRTqO882m*>PTV+bXk`Uqu9judsBMl
z4b+j1T@mDxYij@Nlh+Il*KDyrAXeg}_%QdW-E$oQf1)<;NVdqH3TG3NJgTBuQsSWP
z7ztjZjo*Rwfo~hp%72ma<Yjcfmetdn#?any^$MFywN{5lIM%>9+IFy*E8=8qJ34YO
zg(T9}UpEs<f%KTT|B>%aX(t|*Vz3Z@q90SjrKp5xrt%xBMit(~y{U2qJKIyr+<$-V
z-;1RspByJeE23ETXP&*hv`khYrz#g9C{69E@S-~{MB`Zp`dHN+g#*RcR(mwCESWoN
zKC)D%+oHuyC1g~rL3D<I{G}jg<jB~AZNcdRF*XD-IcaSf8pB{inx%~Xh?s#=5M-`f
zXu5~pxN4YG+UPK+;x@Ssm75*yBAZg@fO#fcBl6AB+IDox-G8a54O2A^F3aMX-8*MN
zk}T17zmF`%i{js3Kw}p?XM>xe7rAz@75C3nL)3M^!G^1VHnklb@$PUn74fdA#D%Jg
zPPeL(%={BV^#OB9BZ{&36(27Gsm?_P3G!MnK|r6#6Z-SnmaODuLB(N?lr5AMBjIr(
zCm>&+s4765Ez8AZ;CsBeN9%Da4#AB`Fl##S0tB3_b`jseg7jPQURIW!<f?O$Ozu2O
za4zZUvo(y=;0XeSdvf#05e#TNd@KjwdEUL}({DH64v@w*n_>nkJ0?Wlw>7#vN{7f2
zT5ZYY<*#=>_Z!SucX7&V+qkL^RM09}_fAL2ZAk~m_~tR0SWmML<gE({`9?G?r0w}v
zBI0(5ZAkLHTAMC?e<SNs{DM&MXn!dwXH357;j(_G>Kjt%by>HU@=_R%uk%JXl9#Hn
zZDB-;^1qp9Tr*b`CY32;9hpV^z#pM4%flk6l<q={=G^b<8F0^B5-bt63G^Qqi)_*l
z-8YvDk|=HW><KC2G5owtp_V+@!fBv5Urxi9z<wUL^*zbJjfxrl03fYMnzkfhN>!M^
z;f1Hi?I<XJp+LQL!K-UHpHt&H93V5do;P%f_jY8j%ddc?WpZ^ld8#*es9>BZgJ0<q
z-0SIZ?wqTJ=M8&kEsN%U84$z^Iw)ySU91!r8^K$y-)yG+;Up@#+!Qq^!D{5w#qNG2
z>Ztx+F6IKj_nG;-dm$gYzx07S0KCfUWl@Mz3?RirpXt!eS-%O_y@|Lg6*y;g+<!82
zU;tKIzQ8@w!>5X%q)C6*>Tj~+I<MkI1C7)LW8~x@{F`Z8SBUsEmfNrx?gO$k*1H1Q
zJcdhTQQ!Y*usR1U2xLNP6SNWXho`@7#~I2kKw|5{WZDXwCHuYDZ!&4f)#zvyx-}M<
zgAO$LyKvTCd)}Tgsnr*^f_J#Kc6#Q6{c&NlctTsE?-r}@;0$|PGFHUKYnsVE@CHgm
z9xvAqhvwvAtPpavRJM*f34UUQ6PkFz>G&uXC$i_gT%5xWU7M!d{&4z+_8ouZaPIJr
z-L4+m#uv%_+*gW<$Fx+F!Y4+Kx#b#a5R%`Xg{aQal-w5pMsYP%JJIR^A|^iBf|K|Y
z(;j%PhON{}8}n1EIe~6u1PRKCc?jc>UY-6g*6u0DmbGgXblG;TvTd!hZQHhO+qUgh
zSY_L`ZJS+te|<XoM0E7U_n-S2kt5$689C;QXBhCZ1w-a&Wufi@v_2op+}66J+AB1=
z0<*HH>8l=z2C1}W_PG~CqPI9jruTC-Xs3KaZ#WF)SE+(iyyzo(s8U0kS+y)zhk-8D
z<M1^cWLz2h*+}Jmh{HABp|L=tS}ra07XqcHg)BNq7>JhOGQVLSpH-5mkhz!c)(Tp~
z6f`jzXb-nfD#qw>eIgJ$^Tpa6d#DKAH5W||E|LY2$s}q)K^EVgP2Sqj#Lj#T7z9%~
zcMXQze?vba+au5SHFvdwp5D)Ic##ZCZ4S`nTwd?u-0S%Q%63lOgOF!2lP6(cTB!cX
zWMJqcj5VS5O_)g8Iz}j`0o)@ZBQztf8hqyq?`QYyPoLlzBI{$p3rMQHM21L4R5?kW
z(RMqVM-2Ccr1|Cba#cfa#2{;ECwTpWa^644y$k7qVUn7~p(qTwfh$z&uPp8ge2uiN
znw<`@7sJwoUyE<4r*FgXX7kJ~SWb;>-3hnlRFU;qvW;Mk)&A!`-5-q3yU^wQi{L=w
z!{k$tEXL*9uhD(A-3R8*#rm`-@Ld5y&<im1T%&t>WsdGhul8v9Dr{iilF|Bll|j#!
zSp^2F3c5LT2XGblkJMjMFMtdxrN310582>U7oZU_rN5O@P*81_c6!F##U!FLJvF)G
z8>4qn&(VL?vP%dQ<X$=Fy%a2JX=Liu_<F9U--&(p8`jCU&R#Ayl;P6blpV<g+00vW
z3iIk}*i+VDe8?{04lA!b1ny#I+>=RfX5xv~Ms9KbI(C~G(N-8Y7(7OTeIBTaz5>~|
za-YnEFCV=r#KwYF2D^L6l=o93C#lkj04c;)DUn2-$Nf{1l5wGkRqOWN9FA-<(3+%V
zsadG884FJ?ytzP8{>WwBcwEP8qv*E1P&wqodD|4<+Mr-zxSpH`rqbAH{5R2X2z0h-
z6%H%F_^JdH$->E}$Hu4F`84Xos#+dzoExtWABqY_lYG0N0vlu(8V04zuGYRmaljIq
zV)lf-xobKx85+h)pxPVPNF~pGCDZco!zP=u5_S=%yGJsJ(bOQm^LGv#u38L14EN>&
zBqzxuG|B=%tZ_-D7zxgRoorbf8c}?L{JlnzVeb}VyY|eYB?s@QrMQ%I+c!tK=VD`O
z!+`LG-7QuTPOcR<c|^V~lG!%{m70^-wCaOI67EtObdS{c1rCc2&@Kay`$NTK1lR8K
zlupCQxoSVt$2f6HIJ}vsXAliLr}XS$);9yih?Col39a7^CU#W$dYtO$cer8+koD!T
zUHzaS&njuZ3UktMI-w}4FTky|yDf<UGeGhB`>jNKj*VN%jxIO5N}aC%WtcfQG2DQ<
zB=2e-_lp}y^+<70RSLX1&Tg?lIj`-Fv{G(n@Vt=D#Ro?zZlz7%6Pn<mG~%^LdXts|
zy=(u#cN#q#RrGG{a5OV%7$T*3n4l1aQf&&w;*#-&?poSMvp2n&=G?p}B8A9j0#pK=
zL!19>On&z%7W`tUzGWUDjKvj*uq*|H51&E!mV|t-w59>?*nz1EcYb;Pf(_uo70I-2
zCZ3=F`qgX$9Wu}`UKup3VehZuI=`5{9^w{il8UAa7NT_Q0uCBfbnh^^W955&c3sVh
zD>lZb-)gm5EVIctT$s)lPxTs??O9M(Bd+FsA~{5Xg8B;=)Q!y9A^LW{AyQ6HP0i<1
zL>DdgyyouVzQ9e$XrnEiy|8-j_07I@ocy9A{FlHRbhfcC;Dy!FBjJ>mv83c@70XRz
zJWyxpG?_pa&B<*Xs?~<CZ1g*x3V!TkQ<oqtt_tl%hSkohROy`f4%$1}uKlxzWvpZ9
zT>+*f+YX@eE&v+^MVeXLs0~d<<H^<*OOhhC;JNWgZtN5Vu?H#;78>2dHV_CmbgQ_=
zDy)HuliAX%s4upEjE^LRgWVyx!hC$E8Bq?Nw=AM3@>diZI{Z-N*KwK6el}he1Ar6*
z;6r?;Vg8UjQvcPqD?4{5V@D@lL;as)!T-A$#2sg(6gR!}BL)dN{16_-M)(Hq`2YDi
zzP_`Qt@V%NVW@9q<xUF=MHzG}<Y@5|Q6V5mA0R#*R>11dMlhNGc|?Nzf?~o#;tI4%
z{Gvanl#fA-|BY#gD~|+(iRs7W*|de?cbRS1H?oldbc5{x{S@o-iZ4@ez1wfEMe#ru
zFFl}@4aXN#GcgU^a9yaK4DFg}NtuRHvTTwqx4o)JYByXPtQibzT3nva_v}+7AICuD
z)Mt<Bk$MHPKm7}0)!`liMvN<sYT1)u62Ugo2K+nhfM#5(f4$`Ri~W(J$4sDA&$N0}
zws(Lv#)Fs8=L9k*KMb#<{-{-Pyw~(+ME3r*dZ@W`<bFm(@_#-e3c~zC(!#XXMnA(5
zg7u$A)F-;lqQoSis|+HwClYFo!M(%p$euQ%A0;59aF~jei&n#Q?1KlD<riPFXa9Fu
zCa1WRC>gozjK{UjlR<xDor#ntwXxcTMt>alFq2lQl}+dLYu>29(n%p<K;ps}h<WrS
z{*$-718+KxOl)|$Crb)=cF9jk%h|$vL)^YOx8`0F*iMm~Ft=2t^C4QILap1e9gn`(
z81wTNt5_isq6j(vP%x37ZDgT+20>+>C7%4HZOI#9vs5np`@~I76J^5c?OUBASJF!P
z`Ex~kB;X3~?^U!2;C^a*@izd7YIMI_4&qU}pV7<#0LcNshthL#`^k8v|M_UT(ETU#
ze{32j6a60v>z`r0GBEi!%RFJB@bL+qoy`BGs5m(Oc~b~E@SPl-jbWj@q0nJuehzLy
zrvZSZ0pQcD?`Z#oB>#UdB%2=t%iPHEC!XI||H*pI|F3vpp|mAJ**k20A_)Kh2>`&S
zXqSThiA3T5TqHkGkDaTz<3DIuo5KGciuQ%a42#CkML&Q+W}xtOy&w&K@|)+sxu}u8
zlm3sr^`m|LT%bvuc={A=8F0k%1&|eliJ49XMCZ9Le}D#-a^5d4C@{=^p$S5O{cKTr
z^Uu|Q?pTW4`t?yq>OFNMmxNuya@y^I%{D>oZ44V$EYJRP^icXCW2wv+-9nML<gGxN
zRJd6-C0p5j(V2a~w^=drq}!PklOm?uxJ;lrufq;$S!>lGs}71LyUYUwdL%kC7rgs8
znl;09rS{>Yn)2iw5|G&9Y3<&T;a`X6$dF%Qw;2$Kj3}CI@<p+HGIlUrD6A#n08ll7
zVS{SCG%(y%Odw^4fJdQDp=|M8it<-FwPG==J_ZJ!=yooEI;m3LPW6mkOdv3F7~235
zG`wf9kq6CddYdHjn$7Kh-!xW!PjXlpeQgYrL32aGcrfd{D?#fY&1$70mZ4!AqM%?_
z4`6g4IV&Z2`)zg&aV;#c*d>LiV2WF-6B?hN0kgBfhk8&PwugG=m=YA4WS88troP|q
z#7>dJ6R}%xJV=k^32FVQ!DnGUvA%m!pOtrEU3G9v>jX*HF{bd?NbG-%LLI-BIzRvT
zIeqp`>gD4MB|&1vNMF?4k03((yc>f-I^keb$7_f(_vG5LSqyu0>xN|?m6ui_5KD{K
zO8FSoZgCD<+^9$58bp0WZ=E)}9zZpjY%$g-msxZ09xm|+T$WkB?{xGTILtVG!QeVW
z^Ezc5M+AVJDRul0anytW(#wKh7c7hUlUm~cO==n2I2v0USQ-D=ZdU(c-~N+jT%H-g
z6Q0X}Q`HO*WMGO+LPB?@#_v-RCVfxrniONnp6>Y~Qn0L%&}`7uO!V{wg4ovVsO!-k
z=4}<}+wPRQS$onY8xA!!H~k!>2=UCVC-*r;iKkP(kgv4Rp^=A!ciJe~^OM4~iYy41
zgO0tY#txvrd2LN5xrrto&;wDNq}#5djT7YFLNo{~gyao?G(x`~i;j1<J(SXmn<Gp_
zlLtZZCs8VyGE2gd&_9{^DGPsjCL7?|$0L$LU)&lD>zM4=b0sPgI(MKQ2YcKpEbe>@
zryW^PBJ^MH+sW6Fnn4eNE#ATC-6kLU2Xe)bn>Ukvgi&+ti&^u2H=7d<nt26Rs06-v
zc4I+*_VCu3Z2Q=XDpgAI6GV3A`B4!8^0^fAgbRcP)?80uSaqgl(IvL#ihZ~%+qjCT
zmI*ZXa^RfKrn%K>8O_UKg3wEi4iXUH7~~nO>LB1G3j*g8q%x&CxZZHtjc!x{xe;)(
zMvLXX+t~)*$P4;K$cd=ogL|~)JbCoKeCi&nI?onW37tI<O=>=zhPDF+H1G@&<?G=H
zQuLBd3G4|ZN$B(tJsy&5ksJ9@U=uakEEDol?n9WOp^<^T?~F3XE(`fYNrj=zq*>a^
z?jCbZ%$cwHx243@O#t>F#BeEbz~9%b;V|S_=DHud3zxFL&MGkIjz9Gh#s^e0B#31A
zdZOV~-rv|Z*L(x-+kbf~lv1V1`{4?4m9P0Bg^w2M;Fu}Xx+d_+tj13Zt?ewOgNJ-D
zN8$elVVrvlA58;|2`W1A*Qkv7(j>7*a%P*z=Wr7VPU~=}GkMn9*d<4qDy7+&a#-Xd
zo!z17%#g#MT8GJ>w2tcn=&{s4Ts9QQyt!$pA!RpUMEMRdUwgsN5hLI;gwa^Zew)8n
z^r1?k@HS80wT$>nXcf1iW(8_;nFICL3tf%GA&_&?AYp!*QFVgAuI%GAc$Cl@_IAkO
zWYrx0SS?U8rJi08`%z&skP+Nlp{Q7z`(`a2!HH`bW)SbKrN|2Mv@=a%w>`)`Va=Bz
zx|QG+C7~CEZ{84Sz^}x(4hE;ushK-4jvUBNj_zG(^F&>()mvzwg%8rCG+8<D_f_J_
z2{~j7xCaN$Y&N^YUDl`9?i=Q?W$&g>lqBcP>8xK0JJ-ovn-3Cj2wy83qhfE60?rvd
zaz;-xeDZ54@4mxTX{m@7d!7wvFnT=ZisbaV$(<~BW9bZ2x1LJ3)HuIMW7zSEs1_D0
zq-TGf**m5-|FvE(MN(OZ$6K{1JMyrih%#}QLg4a3=Cf-n-=u$0EKt4?W3ykQlQ+8f
zb<ecdTp{4n_Dol+-rU3Nevn9y+X0w3qJ3ti+JVS^k!n(TzSFSNq`Pcw2tlK_7S(<~
z(<@(f$~x(LeY*^?L{B~uWehuzUaZ4TVdS%_u`SqLH!f_Cs*&PBlm}G{1KuV(#c#wl
z7w*JbFPs>OyfDB`qh!_np9v|cpyCV1MKl9h^bdD#nWEdZVyKHL`C`i7g%h>6BcT(u
z-qbiAP}eo1ISy=vkT?$6x7C-+Zpf?D#Xx}WeW@|dNtle^0%L4YzEyvc=cOg(Kd;P0
zWj8mDlD~IIoY~+B6fKw~JJPglq^@R)6Xd~thDPzGF2$K>CN^m$9TMTZwIcIbymfQF
z8?DMpmUlYQ4M1*!?JhQ%+Z<eC`Ug6N_W4c%Ni}BRVGncz8sacdPjO897Vj>yUI<EU
zWN11RgCJjUACGG^cheE-4*X^Lk)RuU*iu2BV#istclC>)%Z@9(p|aGH`>j?tOr1cU
z;v7B_E`C==E_Wi$%7F-ld1<k{3?tYWd<F$O(u^vkPRF@_dQ`cp7Dl?SJytLpIET3!
z??~$~VyqGpwyP_0o2rCY*K%ho*9?@tnmXsgwCu&XlX*vKlU;c2i5G}JEVGS1_={qz
zCO%{tlJWe#(yRUb`_Hv3fVhI@xJtY3ZLN~_k+s!^-&Fh8r?&?@i<4Jrb)S}6ee`63
zRgYCE(rp(#g{SQ({j!Uml|gD36D*lZ@E=D2RkE*(={E#S^$(Bxr+i&XlU%OvW?z8$
zXjk$w7^X}W<McBq-L*VQi|9AfA`W;2GB?WS12pvEC+Kg1Buz>GHGk%?S(>wxm2in+
z0aID$GD;Y~vWyn&^^O>_V&nY)6UHRZ-&{B<FG-p0*OTj+%Sle|5WZxc16frLa^g8p
zDf^}sO{2Z6^hX4(rx?dAZF{jL{u6knaNHKA%SYekHZ8V^NnQtgCpvK#zEiN?9)od*
z$to+R9*Evm=&FmofqOK=JBz&2tG!^pX;|;5Da3;|Wp$WW-u_~T&Ww*t=u8DoHS_|J
z+}3bxYpEcazriURuKc`u+1dSAu|Y7`5s{geb|j&o*Gkx3LoD|Mwq$iA|8l4D!D-<`
zG&RgjQ^`7q{jGhpXfc?(m{K>F)YfEnR$YR=dctx5SBhiE>`TqVup(=8N+L~=TZvZJ
z=kLM4Gy0{u!B&L#9D8%(`<{2}vJ*%4Q!?KWKY!qD9~u#|>nE7w_(%b3qs9u7=U3uT
z0cr-E)a5EDQ^S*im<s8mKxXlpBS^f~#=>vNer8^PfGtsYJ=a*l;>32=(IvlYAy^>5
zc$QZli5$U|#GnV3F8Ae}SpnF5zeFNhOd}kZ0=Pn5_f$#Yat1Wen5sm%#YWm0j4mc_
zuE#Cy<=u+ch#rE5ZU-e6%gd97kX}1tfrskxJwbF-q4+y0b|6;^G<3w5%TABGYRXKE
zURk@vGq^db=h!08nQ9gdDH)}@3gi_&yFd(}u((zJQ~1{-g9LKm%NtQ|{gfyFRpI~t
zOW*z!!;n0%{}Q>QY}JtTuFTBLf_JXMm*~L+Da;zJXw}WP&7}0`u$YM;t5sILi6Hy>
zL8xoRE`=%ykvAS2fo?Jh5FqDVKn_`^?)!86Z2h)aZ_>&n;<`JRNB&(X4LC*7$PbB1
zX0sJKtiD07H(&Vo_%B`=Cpj7GH6BhD7SDzAQX4F+ymqKL5yP!3WW8mlx!UijIg(PS
zKxPr6^3r+8YMjqSG9mKsMit^mtMP(r^uo>@TdjLST{KSpEb=cIXBIONfEZyI-48-b
zBuPt1W+@(riJZL>J`{gb+B0`I(O##OyXfbOz~+^gIDjL!N*ro>Hr*1rQn1+rUw=^<
zmwg6}$GRQBqc30PsAcGP3n)_nKYMGd<bLw!&PB2HJh({3s~NNYU7L$fwo4`Pc<cNc
zL{jebhtjBSBE=QY)06-ruaxr%UMr}$Y1w5>XV!|s4G*}fJs=TK!5QOP+!g@$pw+g>
z3-Bj+#jPvC>jtp<m67BGK%zXVMfP?ILqZ<8g3{}f6WXZn0|&?kdM8*@l}*E{ZJy$F
z@RBJCiUqS{*yu?&u(4Bs;10lFyh?brN;pAR*@#H&iRInpi&^HR<3`WjbG|6_M^*9D
zPo`Q$6vGl?;Wrv`8i5G@P9W3}8m|HnV#U3c#{yeerd_Q-T0_s$Zi7rC7!*q9DKq1Y
zHgXXAhl1AIkQ+s8aaf1vGS2`N_N0CJqgS8&_9NH`qf-mLQ<Wf}K!rDxVPgtU-3~Z6
z%t~r`(zH~DvkI1B%emEu6U-mZE^W`L1D>?WM?2mOa{)Rap#m;G_)b*u4@C4T0pzTX
zjxG7;Q(Rw!Y%}AhaeV%~6m^|r8?hCayr?RH-Fo~Y)Dm5C$)|#hgV`R0%hml|-d^Ad
zclI8}cjyPo(K|<asH2aMtGHM;UI9<z6AV(}R(EerqMoBwt+Nop)65x|8OZxMgUl=Y
z@$B9tb(T}Vu_TCV=z|^huX=2;(((T5AzI5T^WvxxY)K;n1_KLR;J~rGZ{l*!U2~N(
zGd>K0Z7~@e%7F$YN9=}v-O^Md0?WjB3{v~0kQAF1{xR-G;oS>H!*$Xia(mwJ<WroZ
zbJ;ehalITO3XEh1Ph;^I;_A#`QDGA}c?d;ThXL{M1Wh)2#w)3b`GZwr$N;XyX3b|6
z-;$=u-E<o|5&lGz(Q)KfeIed81>N2RqSWlsIkdDJ8v4VtRu`sXp4(t&M+{fs!~k>k
z6pI19B%SlbHV|nzcM1AU%R9FAlfs5*gNeh)p^dsD15}GJlIdACb^kXh$ecgdn3+TE
zXX5>-SA(gJyncxAjIno4n#EM7b5RQF;Ol!XNW~wFMK^02biccqae<!KSa#NR8SH*H
zD)^WGDac|6RuXR{JU2=SV0CK23CN<x&Lb**6%u!|nHHnkp7-8j+3*3>h0UK9&EE~V
zO*rd;SH3@z8s|>rUXb8xbeJ)iyc~hJf;C3Hh5AJ5R~8m<<`UQp>>}_PoL)beh_hjv
zl?t#YQS=ZBItYk-fC2zj+!Q|8CzBXV94HLexk{I3#dbll&T*X*t_%}4Lax1PVX`jc
z(9cA^6sb>5BO%#&@}Ez>Jk7tN3~`tdwQ)RZ=Wv*lhZ$1R?m0xS->7-x`Jk^Y6!i%g
zKz@p@6I(Dniy)h=$Xg2LsG`fZ)-XKp4I$(F8^3|gUSOY@p|^3I{hDi=Xc&3^y*>Z_
zJ^lQ}d;Rp~>-+lE{r%}n6|I;zJ~PW7wAE<>RC~&wu`2f%dEDvqY_=a}>_MQ}ZU7@G
z_poSbEXtKLa~=kbCjWQEeB3ExJ;%=9K@ugV+s6!g(Vr7_k?9~C?psf-(3`2KY%Zi4
zBlljRLJ`YD*=y>iSg|9o?qj1JNxNgT26Hi)_`+hi8gQHjY^cy#c~16sKFVlI7XL`Y
zdQ#*dHB}QMbl(S9O0t>?We5sGW&AmiTMF^LQPYBZcg&dO0v&z=u(AB!o1%CQ^E0r<
zBOs^ZcALG?&So2T$Sp!mX%{ec)(eRxP?~H8wJ%ek=}eg_2xlh}6$(Fl{3VK-ZfVzx
z>*z7+(2hIQ2fB)iX_953&C@Q)7TP^<3_7zq3LTL)Y6v?2Z-QdEQFDDKz;Xa{3S%by
z!Cn(DpkUPT9emGITr{*DItMGE3;s<;ICzR9UrZ5%UKB`tTpYuZI)J4C%~X_xBi`I&
z@^#J(gaXWPGf2}D>CQJDsDs5m9na$s7Bt&fqDLgTQaD5)ewMul%HnPChwG)tUcxcZ
zaR;0!SxBgogULV;lzu|OnIu6@O9{Qz<YOM?njtA3?rpCr4yo{=lq0TW>b<Qs%WjH@
zp){gS`(VWK0*qCsCsw0JP8l!jB#X|h2NZ0l!w0l|FalD3(9a?+Wp0pxD|T=t^%MMH
zEKAtVK+@u#q@5&;poJR-ZgbSk<4{q($gwCL^GIsIbZja>J~s8r09QB1EcXC^NpU-m
ztj?Yo0x2%>zWvN|Rpp879AldG*WBTY(>HXM>AX5vHHw&?UZ0ybe?JV0oED<Q7PeY|
zoDV~)r6v0p{yu;W*tNr2Ntf!`{(!12Dq8Uk3f1D}F~|hrd?evJ0Hct`U0J6>Y$yWj
z)G=468e6ExPC8{I3o{g@&RXfJhPt_zKPZzG`0(}Sxzyj5`z&jr2$mn|iu2VOOi$q{
zz?7YJ$gpIsSBMFWlaP1NIHaIx6o+q+<{R{ds;m1Sj!eNar&NJ(fXCF;RNKtCNp`Kq
z2VzCRq>QEykvgL8f2{2!H|L|~v%M5vok2ZBLq%j`IhCHRBW$GHdiUC+0LO{a85=O~
zQC=q^UL8Vbn{Sd1P9AMV+_b%{<RtX8K5BLeC2~_yF6T=F12Ndfw|!VI?#Wz6$G(He
zYdAP}Q<3-mkw+Z;kX5F~4_>;UdW8q~eX2SzkpHF};`iwfs-xcZ6NbK??y^4-L3xDT
zQb;)h?f`b{4+QBPh;}--St&Y#57$nQ5OO(?Tn~oyiB$3iw4MR^fHM3k!)O@>t5<}N
z(Ks55*=1^LJ~LNE2gI_xh>&n41=3tn3V1PTM$r&u8X@ITZqUBDL-7rnGbja#QqOM$
zAI6|&Oz14d9G3{4-XlG5<ne_=m(jZ}eVmu$kw~p*5G(2X>)gi(N1IS-vT>Hnwz!hM
z)N;0vP^%I?{6|Z64)>YPOpm!_upPlv0^3KA9xQiwUak0#2rD<${_T}KTD|tYF0TAv
z5M)>w(evv!E?DE*2i%yH#L@S*4%XEUk*8YJ&cS`QfZMs>THnw{8!W=biFd*yOK^S#
zVEsu3KP6%@2*@`yK0Zac`A>-`^54v0**O^7={p$f{%@0{8OL*kXFrpqlZ^yGY|czf
zf)9E^TQ3R)C)SZfhO{Ffv%*t=PMdP{eJ$(p66=!4veem+?Ou`LgBYy=nprIMAmT{A
zF9nhZWEUObyfF#fwf!61XcScX)m?**3|GHUfm;$?JY+i0@M=n+QZivZ5!KusN#;Dr
z45lJY^DcNw?Gzmuw?Qz4sn<P15>J>xK3N+s6EVu*xKe_FM*mvz&;&Eqlhd=?^<(*q
zK?j06X~~^0-Bo=z{1!laQPl8bvjRQieyp>7yGXrSIWIoEPaPoZrj{sH&Fap34U>K!
zn5Pw3hnUjWHQVaeSLCU^URa}QQ)u~0I^om$eIjQAapx#sgC;+Y>5H^D3nu5KX*Z#y
z=Gif5e^>pWhW+Sd$n%6igMLQ{O&ZBIQyh-O2lPE?c8BfM;S|vm!47bm?c!m$UJ_7L
zMUT4%kIkWo>a>nLhwbq}v<jHH%Vpn(cz1(zlnUobhbT9Rgxj`l+lmazUE&)UHQk3E
zD**yZj-5w5oIh66C>QEQf42p8f_e7yhBTtIe}eVJAI>R^{Bd&gV9!&S+4$YH;J!Uq
zpRBzKR{*n&*xX|SiB-SOi9mkJ9L@%VVhZ`Y6iro;D`LSEH~UnHHk{QXek_9^R$fIp
z#Os~RdQzSh{*{Z8uGzWz{`W;29{Fj?<lIFQjkNuT!Fejl$di_fcWXA4o2xs|qv~U6
z*7`8sE3(!1R{(UEYw!A2J3pJJ9W#9&vz7W!E{?#03=`oq47&gGcZK`E$;JQo(i6bx
zm&qLM65yD^02sZqATtv$i;HgiO##u%NWaFju5#_szQgZ<>JY4K6cv1D3sD7@nv4kP
zLGCLW+w`E~a+~ut8Z<e&z?MF<-KiK;{s~IABVWk(tm8d?fR~y!9=qZ-#6^h{QEZuN
zi;+@)P)R~YS!%N<vBu#T;=*d8{s}r-KF8UaPHS7M#%Uq)G$JmxWoo5E<6MgI$-xD#
zpjUt9Xk8LI))5V<7-8>za)J&KNQh{SY2ZIp7pb!`d6svZd!l|M8!cilfgc;bd8(JY
zcoz^afB{6#LV>X_^syd3JMdXGFbEQI!Zl%P9AOYUksxUpzB|J!9SM9;K}>vv^03NQ
zv3wC?RKR)PQ*wGCmm~-bqp5oES6yu!r%_bL743t}y{pbA{B}192lV}EL91|6#0eJl
z)B5)ZuvvT<hu#cy<NyfqT<L&(VVFBJlz>RC6u-i&iN1vCMKsAs(fybVA{O#oP%7eb
z`8dO9{Say8F#P<&?pB`RJunJKghdWntx@Q`s0Kpv>t%;CV3z3lc;xH}tmLp0sP8e8
zs;KpZSfs!jfJFrmWm-+JW-BiS8iLwRj#)-9x17Ox8~;ciNbXoj)<)MP&MF}K(MY3e
zg<7N71*7$G1eE4wZ%yb*I(bXVY`PxhS=u}uy2cebve+!q(~Zd9K<jwieS12}prghK
z=6|HMfbmZX!E(ff$Roi8{o0XO+<k|*BTwJ$gG^gT!WW23(}4Eeq~2e_20jU}lF*Mw
zeFeiDSs(b++k-HRBcXlySa2%kOU0!2t5d~w_i^?%&v+Gbv1P@Iu^iGErNTFw=VWu-
z9bf*7rlsu!-V;qGYY;1^EJ4vCWG%&rt>l&!80_sNadW`HeR2pMTh^jD!iLLxU0R!(
zxM(4#fO+AcN9`$MBk*LK3~bAO2_(9*6_@Ysh?m4l&XA|)M-IrdoLXW(GJO<pnI~4g
zQr>Stpg&ezj6|H}&*b-8{h{9%+sTidF7mE7)ETt4qypGRy6<r>ugZc5jG9089IulX
z#o695h^H}3=qsQHkbksyuGxv?47`ph>6riAg*|*!`*;6Dv$nKwmDW(??88Gwg@=xj
z?*-bS?EE~0`;ji((o8YiqW?x>1m{|-qwysWqRe5cyCXg1gy0q^2WyV?!^U+jyOv$W
z=<r>;2-PE;mluK3+h6yGkTRgy<)q}y08j^%IONJ_yo`BD>xtnhnZS6#UL}hg`N#Z_
z=%l+8z$uWI$$%D&qJ^hb>(z=~;$vOyhM7IJj7DsAAVSUYEk4y#R3U~poxw4z&}U-6
z%(yqk%5;h-Ixr%@7GcmjqKv>^?pHKGIfdSOPI9_oOb;;dcD_7ZfYVoU?yPAN<h?)D
zLIf_^aLZU+AA_tocz+Yr{b&8srx-L*Zfxx7>%7>*P^CM-txc`X2R&U=eX+LphU!A@
z#~=hN>H=xEKjq8O_)H(-hTjnYD<pO6-3D&bub7}s+`~;ESggwiLjoEkN#8fIYMV|b
zuu-$#GAGO(+{Zg3t*1bmoy*w;#@p#6mx&qzPv}qq`g$}(P5C)VBi}?QI)Mv_F4*8B
zuC%R=C4m)Qa0i$1sZ7a?HqKNM&gRs?Ocj{;NCNA%nq&7~BEOQs8E-n9rsf(-Ji28)
zct25+7FS0-WAU<$j<8N~H-og>9}{`{tgh*fG?*uDe))Nku}+kpRRAQCFsgBPB>x%k
z+LX%n8GW%Tk*tE?dT?ffvFaTnL!6sH)o|P3W@7V@B0B0XS&=-ng<{%ciT`z1sqCZ~
zsXdIa9k}jRXOmps#*lIkFFpOPYS^;7g`nb7ccl^Q;39wRET6?g=l@biG&>Uaf@ey0
ziY(^Tl$EbP)DTwO{7@%WLi-7t>Q{r32UA_NWvlnW(=BbBoOE%b7qP?v6bliRy10Mr
ziuRJT-+9KXjNn-_+Me1W$A^H=DQ$uT1q@#gLFCE;-DSfZUgEHghi5v^bYpM@b#-WZ
zangpv;wmA?XA5MpYQQGok1+Ne98}xBfm*baI*{GFIkVmnI^J>G=PWMhY`3@F-x&xG
zMzi_gApz$RQ4c=2a~*Zc@1ttmPCM`wQRpbJxff)7zO~)T`;sT}=zs>s-P4Flxnsll
zBDpJC%>uWDU#3~s&gqwS*L46D@Rzg8336vJ6bx_L6>ZZqpf3IW%?&Ysd&3Fl1vNrG
z>=&O%{hy`YfNVidxJy05kl5Hl5-&ry4B&cSg`ALmk&UX_EALqKsU{`}nq!*`p4lCS
zdc}0;-5}kGLgea{<%8Vc-y!gbtaoOGyHJ1p{(KWq{?mNx>Sqld5T7Mb-~Q(-`Cm2P
zaxk`cHg_<#HvUofX`S4he&+H1*T=H=#*#!>k<J-}`z|dI&);bbCvOfaa5w5XoUqWR
z%tMU%*jax&qZ?)t!;W3Xde`EXyxN(xaV5a8(Fm=Hcv38^o4W)poe<OC{rSlQ0-(*g
zO}0IUyAnPvP=h;Nmzu)5kDX10C$Z{{Sr8%%z?5*n>YwIEYXBgAfcUIpnvOrW^j|eU
za&q|D8sSLiXlU#3)BMQh--UWlX<+A~AAzZNLX}lT2rp4uV^>oL$r>6For*9SDv=-)
z1wTEk7@J~cLBH2E<XiTB>zVWYiuW*8)%unx#R(J1KwnTGbqdam*KlW83q6ImD}NSx
z&4e>`n0P8#vbpD^M>iY*<)25gI0gU`^a_Me5-|?+a|i!bzpMY%{owxrq#%0scvJrY
zq;}*<egG*ZK_=NBK&t&Fzw-BA$go|Ztcw+`jc_ueP5Z%tl+noz$<R<paC%~RA^fBU
zqU8?xhoVer5^Niupxhh8JLbfteQV-*9|roYiC0;$puJCJ!>Sb!>{AKptfLo5YXs0u
zZ71!;#l_9Vz#JOgKA(yPXBC@nIX02u9>Ct69MPbXyMW4?p?e0YYe-?}Nf9GcwfU-Y
ziCCycQFVw^RdEe8D;kxeT?&)pfb5isK^807LN4D3(UCj#1OPND_)lOTK}Kpw${mMU
z{EnNXty(4dBo09Ue$wV#+(Kf1Hi=FR3$+U12cvjCtT`z};qoOU6_lL@J}Inm^#Ha^
zh<Y7w5em39E3qFsN{Q6ZLLJ9x;fIb=e`$??Mj6*z@a*#m_}h4ZvOV|Phc0RpYw(AT
z^6G0^e6M!*=F|mGuUxpVcKLQ0f@?RUg(+bL2OipEjvunD@*3s~eH7M^kjZKiu7Ypr
z^Y&1HD9saR4GwrS_;X0F;-}E_t|}pzjN%u*ur!xdE&h5Q=wXp%@ub{=k2hG(B#oyx
ze|xr|Opex<?jfEkrsPfVZ@YtcF)*^Ua0I5<vbO_rv47~O(ZLam7H1FaOP|_oU;2i~
zB8b$EpPex1|Iks%KXjB10zvhPIT-sk5W2hZbP=AsLOD5b*;oQev&7j-gBIBVBxc)j
ztj{mmhI}bG>fg0Q=Th+aXA$+di&WyoCD9tzPSLa+q&H38yf)js<y1&IrIs_xk3QMD
zh7$|#?1kU2#%IqwnLedP-ml>)7R^pI0y)1DiuUvBgDp$nFCE*hW>{V~k&rE4b-o`d
zaBOSW@vUX9HMR}sm)<)XpSmYz)l_b^Xh;<P-nH<U(uhht!l@}4mBUw}alqPy79h%I
zEGrgUULJkR{w=hee5-c;tkllU*{)vsJ9D&t?u5{s(Jk8nz%3@wOA1T#<S2qK>pOz6
zW+sau_R>-5v#cI(-E9ZMW82=rjYN1F&!l<&1}eP=pYnO5@brb0Y+^OO@y)3_|M=+R
zvr`FX9yk6`;_YmDmj@8FoUc=^X}EY8Wet%ajhm|_Vcx1_Ltw1zFW@6Bvn{bKyY%<u
zeF}3%b*ezfUBMxb48koknc@Y4$m?VsSH8GBjCj^CJ7x}xs`~lSxpOPJyY{?k;dt$#
zjpJ)rW?S=b_4U`dHhc=ykr0<NYsxhsV_j3QdCM9?VsQsFnPM`rGIkugbRDA{1^}rr
z{CQQeVJ~RcA2Wo!;GqsvjN;v79E&vjg>UAn(uv@DBk;xnRyB`;xTYCC^7y`{YfqJ~
zt*yVhM~heP`Xa*5G;iGY&fd1&^G%CS>DlO3*0RDSsDE5OnYF|Z)9_63yE@Xb2&U5$
zXdS=~GKaZWTad}W?~lI5eaB~(E<Az);JbEAfhp`sPc#+GRb8(sQEObCN@{ClxsnqV
z@>?$zz&xt;tmo-bTzDMAcaW8o&!_anD@Df6mc!e|7S>CxYp+fiZ{-{#((0SWD`Rds
zszd_KH&9b6^_?3A?Q_l31gnEWaArNOHZ_+c&6Hh@Z7VsXsaVL7;$^i)zDRC&^}OvG
zQ;LUhYDSjrNbzx&j|s@LdVVb#3gI)vA5b_^g-|aO>h(@_^%PKQB@1g@h*^|_!rCsC
zzfz5B_#+1g$mEyuJ0|ix@~6PNkg1@s8t48oQw+eCVMCP>=PetK_Cfor+hEy#Mqp<B
z)-5+Q1e({nldNC(oo3UPtUI~_@W(qUUx5#SGdWubn;%Pr^jG#&N_XcGBQJEX?ZLCP
zr%p)e>b=GbH<s3S#dgEt`F*IV<=GD#RU13gH2ZnsM?yWF&v!$Rlw$2sMX#7}NOZV^
zESiR}Tx-vi%Tm$g(R!$GZ1Y_$zc*kuO*PF4pUUmUBNWVF81Y3)+JYmdP7|dbe-KMP
zQXiy}PfGYH8+LwNi%g{R!$y?^?}Fy0$*%pdQ7@QfkUVg-*?Co_DAlxR;J!a>R4L*%
zS>)?)>NRKc9f4n@XoL*E&X+SJf7qx_d{b7Xtd<wS>9mTqUj(3l4`inWKdbakX<W^N
zNFXjlkX%strpl9Lyz|RA8MtgnOd%Q08y_GS<CO`=Ou?ruSE=AtG&0w<mX|NpKHCY~
zn(Ip!uP&dmCCihy7A~EWI;k8YrQ6GYPoJ82w$-Q}YjqD6-gyjqJ4a8Zq*9O}iX#)w
z-*MBdu@X4CMeRBgN`m!Acxw*5I~{%LD^<w~>0iJtctjJ^N+C{3*1Pe(uoKLcZn^wX
zLc~tb1~^1%xjc8_kzqKpX%*|#5N`CRi6E0yg_s>YFk7fkT$Kv)w48%ukUgkS?&^RP
z%A+H-RjL8-!dj?7MVjKOQB1KLgMVP1J2})}?717tR!royeJ5}182mz{*5Cuq;sA`S
z18_PcW=BJIfs{`cm^JfkbeL6Xw80vWi~0J1H~wX#hfo0UwJ5q1Q)3<V9((Ol*|yzB
z3s;^e4r53O?mHI9C4j;6_k$YAwVa*utH*?|>QZ4z8Ow$xKHKD%C?(Ieb)GvxWJ0Th
z)Mi1K%{~-)oa9xQM}v1M>on3G$T5*vfYea!M?lF~H}slN*u~5$dU{;Ujpaa1!sFFK
zw9<E(6KX^-J5_32n0@0P_@}<kYR}_Dyir}hsd|=k7^&O11mZdP26s>~Y~eH<9}1ai
zkQMd_kw#-e_z@41bJP;{Gl{YbK?YkAsOJL?Q%vc**N;>%{V0HU^W{ery+FJ0?<ytw
z=Qk{<660p~f)IA12mp)wBA_DPIFGJ0R#FHgjU%l)zD}Wx+7SiplBS_?CGARH+x`Ch
zamrLTumm9m%#`lHIbvp&G_f)Ze|ma~Qd4Qvh}cg}b*trJij*vfuMgAtRh=;bO+VGI
z(}hR)M%`8Cl^Xzqtr2`Hv!&OCa`QGy1BV!Vy<@Eqb`(v?C8w|j-+om4gRI%7cXL`N
zQUeo8qjH|Ggx=5p^?4?*{7tbPOg<9m<^s%99XEX^P1mo+|1OjFiYt-BfVIze&nUw}
zu$uuz^W1!?MN8Wb5l|FrzSWQTj!UA_Bc*OMO?|o>Wt|?ukt2=In!!=q-FKwIdAPfI
z80mx`HgiKgRtNNC1xbBg)jemrR8=Nws5@daFv^5>ebhGf@#Y1urjn*~RDm^+DDOPG
zi&@p$h8QKLuIR^ol+Jy~eVxgc?afd9>gfN)jDlM5&htgc^wtG$Zt*h7su8T=><`r2
z4ZkI>)RAHBYyfCYUn}wc`Ie4U#FMy{l-M@nrnSo)i8G!=6~D<Z`=KugO>plj?CKbr
zj7l(Q<tEs{K6?DaM)9jpDh<~1ZY*6rTR*n8WpC&If`ga|^Og~Y`~_C~01J6#5^St8
zl^jko5?A>{)Hqe8+Tt)d$qTqNlgGt;kYp=^0~1pKD>=a0;Q!MEApsgDJnmCdLpuJR
z4r+qPXfPYbD%>IGBnD+(gNUuk>c??68(>i{q|^|fqt|10)IA8yGD?v<DAG=huXy?7
z8Tau)ef33hn|S!@neT?!pOeLlu|ExSg<q{UNuJDaF&FmF?P7Lk29Nj7seLfJS)xV(
z^OG(Ey(rT?nUn&?#-hSCZ-LVPZ8M`_d_qt2dO6loYQLi@XKf{)6|E{TtifXX7^KNm
zjtbf_BL35Epn0MCrwf94jsy9kT0%?n7?mm^&Msn`Fl8X4{Oo4Pb3z@0+33OIFJ#ns
zKQFX6_WUt!zh>T|cHt)BTvFWU02niBxynUSSw>+RXnfD=JPQB%ji%%y%|~xoo9o*j
zIEuHyviLGu6V^7WnRW+siR!eErl3%(&AgG2N#mH&ZM8cwGJEV-HCM~AN>;%oWdpH|
zxOF!U3^kwu6bsSVxmID0r)-c%S>S|VcIfMQidX+!em|y$NqMA~qh1tbi2-35k>nW6
zjjBYQQ-L{ObKfbJ0ICD6QhK6e;#jL9EvXU!GitQ~iUSA?14ujZyF=o#Hq>1&)hxGN
zuhz)XrHj|m^D?xn)1#I%&(v0}>P6E60<cOKd=%irt_?jV{?_R$qnWuSI6+Hp;_dEm
z!l?f*R<70r=ucchHj&BOmn~Y`DBdq>*a^rm8s>VQbHgY(JDeA)%!24+r%;oh9th1=
zGa3=*YG^83iM10c-HZ~D@pTvBVt8Of`F0+G*17P3uidgIw=ZcX;eV9BU(#7%MvX2K
z!WcO~JzNyg5+kwnEsr#@YEI2$)dnf(=OPn_x{rb1;LP!VS|I!*M`6MxNBbK8$WewC
zJv{^e%25#SNJl{!8;pK_>DVpUyGk-43W{xek1E(GscE#;Z&ekc#|RPO9a-wp1XcN$
zgcRdY%e9VQdgNmF8qK-nkd!bjRMTMEut*E2^8hUwTmf$b_7{)?^Toe6q_)wAvwJ7^
zO{R7);K%lDR_j-IL#Ql)OM<m%>99;9`0vM+<H^w39_$hn=L_^HQB_r1PKZS^2eqLB
z_L-G(bp<9FS2d7cl9F~z>|efbde%sBn9SyXA@9}oV#q^}HR+?dasEy=O&SWG9sL7l
zsIjaE%l#uq?f+367)h#;VWC9+IghPS93IUWBSU<97#&U<HIem^<({^RbboNCR;hry
zifXay)@G<0lV+nBG2rPQmX!`{TPDeh-2g?YF)y<K&RR&FPcv7rDOM>E9A8c)B9J1I
z74kIX&T!#UVdYzF#3<g&mFm1b+(0QObYJ{kLfvK)vs7%X=+2$)ywIT|_#PUx|Eo1n
zV^V#T$lWdmS!u^(wng4@?j(yWQ?8vOdn&Q#hmM+lX8H8>;|KP_4GSPJiC=qb;`ugN
zo;=q$_X<t#S6%=5_4W4-|5)hr<ZqZG{|AY;SrC4mIub@*<q9uS@TLj;<dg9ze0HY9
z6%?3pmUaC{){^RZNO*kSS<h-H3ruE+<Wi7T14B{_wr8(Y`UAgkWzJB4hrZml<gBts
zkJ7TAJD7L!1;RO#&ji)bXZ5p!rg*`nJrb2UE|3r<8knGU<P?02$i*=0jDZ3o?C6zB
zIQfauAD4J38k^K5BFr}*MLBi>-2RLE$Ipa_y4TN2ukC1>;{0oqr)J#s1Zo_N*oH0p
zh1dsTn8Ad%Nur)a^XL~l5iFE*w1~`t@sT&c7AOc*Q`OROudmh<Z8lH<4RYETwSA}j
z`yvs%(bj<VQwJ1ZA=w^%GHP_~6~5S%OcqfG>>MGd#)g-ZZ~rmPx^V@E=BU6gRg8~U
z`Z(xv2(P5Y6FXu(mm+TMVM`(?fNXf8w%R4<j%zKm{y2lV0#R9;gLX=;LY?)vntM>s
zWV1!%_5v~af`wzzmf!w|90_X@v;FYm&G%G|dW9rJmyCRMEe`CmTk*BDLmJPvHC+nQ
z+HUWWAtF($wwh3rD>{sQNNq?71>uDuK}Z@KA8zl|PU1gyl=2ZAEvm%Bj~xY)ss2H&
zj?OVblvxKye8~}=n(Aoj$Vk;1hh=;KsBQIje}0<tDv`q$`SOz9A^Vy=c<Xx}{O8Ax
zqW$Zij_h!MJXRPlkb?hB6&)TOOGk!G8Nx&UHk2QqOQG)W&Uj?yk+WmUNF}$sT$!&0
zSTG%60^gG|n;XFyP#jwVmIX#)AOtP4<L=hMO`jZzx$novhG&SO_vQ=@?J%(8uI->?
zrSJklT-06S?J(%)W?zw#3>wv6X~lGfsPJ7e#t{na25_K@9B@|^vIVJVX#X5<p;ipR
zVdq#=#<-kG%QG^Lg7ho236Q<AV6n)K0tO~8cSi<spFg-vOqlWeT)s1FeG92y-A(^R
zhFnAq3Iyd$;<q61M>RCuoDK!!poJur=;wz-Wf9MyK*h_FWM5A2i`MiI^|E82%4<ek
z>*$<aC&FbTtRSGba~ZoV`UW_lW*Oo%MKsxiHbTGtQbR~h*osmed(lFmGx#=!oXG3?
z8wMurq<xF<$74V7Vqx#EFPT7Omy`)Akb#y#c)Ra_?7U8`TXnEY0Dn6ykigNJxw8CD
zBZ}cJBz5!%fac}QJB`=Z<~E<^ZqAZYpD!MP_=MS_^oM0~p81#$_JE*jucrD68Ez93
z(&&UEW~;fPWsU%Wf9)vDLP&)fpqeQEVMW$AZ)w@~ifF>lH|BWwluFyx7i-6lb8Yxf
zPqx~{``AFc1ddSXH}8&jAKz-r(kn&)s;Y)($F4~2+;!Wf3n^S2`JyR7Il6m*9f!GU
zXC5l;n8mJPDOeuG!;<x9o_25By%>e*`LnYJ4l?1ZucP=MJL=ri<x{sT6dp0Z)di5p
zQ@4HA=z~F2a;|Be@SFd3=;Ib@p-EV-q||lDy|H2DuLSo&22q0QsS`ij9GEWavgq?S
zkL|_stAO~*XAsyy;T}Z?)V_3Z^eem{=h%c6R!QyX4v0s+73?uF9B-_EDvs}}B`|tg
z-)k}BV365xu-a6=jhMaagH^@VKd2O!^~cFdStSf-=1LEsXuL1|OQ`;9V??7!CTp;>
zvKx(sC?Kjmys1Vk0K*mYXxgj??;%#&cU2<h`AcLKg>qzIb&Ljt3~56-$`l@ci}WEo
zxp7V3!j%{67af&q@zn16#V2dKu1M|nh6NZUrTQnZg}enqu#a{hCzZxis7&p=OYW~)
zO!+O&LLo}WVY)hNv{ey1>V0FwppiDi1Fgm?zr0`?MicG96^vyb!{W6@X)_Yk@&>d=
zwAo)w&Qe%e?1f~MlAM5axP<99pm)i&ew6Tj&iMz$1L_55<5RZqsrDeSCw?qF=|cS8
zcle$0g=+a-2e|};Bk~>aqzS0C<U;{ZAo?w)%yp!dvmJZoyF>yKp3DsgpfT*Tdz1x%
zz)ZbC9t_8B=1hc%DM$uriG~lopo(S>ZB5V^BCgK!mi*lgO<ORRoE=0yQ^c*aw9m0M
zCAIC#P0LTej`h$eI=3!s*>-IqRo!@ssBquLHE75J+ua!3+)V(%;+BvYaTgRXm@p40
zI6A4O`AZ%<KRltxUR2j2GD?un1{5+#fuQPC1uZd&)qvMhyw7Bf(lht*37OU=U9wCs
z0EJj3q*LS8>lEcK0Al)%0Lx~M46KHqk0o!xuON(Yzz~MWt0Vhyf0fdgsVe74AcP0Q
z)h*;z*Z8U|J2pqAYMj4o<htf1Bt*cR$!vJw0G7ed_T_NYDkF%M9c|}3&_jq6^7jI@
zYuWYPBuWIu6)uE5G-xZ`?7;Oy<~tL=4442rFd5vL<bZ=Vrj5}lW5`Y}_ehf|Nr9|6
zYg<1Hy)wYs=dY4RiSoM!;o@UXh60FSC8S-4x~X!`O>ie4SCXouhi7t#v&U>n_X!Y<
z8g7`1aRKG%n*yKsK!{z>d^R=^ML{Ay@Rh@k;6>k5mNW}F6B`X=C4uLwCWgB$Oc@9h
zn{TA7?YC7)@^MN3omWapFvsX``YU$Vnh=58qL{U=4^BdG7$#`!gl&<Z2a%@+_7o_R
zBfbYX{>wT8D!f9%Axdvf<%RDv_lOu_UDl3AvdL#eDd4dIK@FS7$?i02UzH8Ws)4j&
zXnbER$f~L%?0MM}?<Ow(_j6Dzu;-Gi7)1p#(C$~Qz1_4$*1<l3KwgX26(%f&<-Pao
z+gvaYtc?mxM0B63OaO9$Nl|avlWJY>V77V<%W|j!D#&1<VriclcU!6_L77nmEaXJy
z``T~M1zXe+ug-c}SP-omv*eq4cTHFka&zJ?!)x>my-76~jWt+za`i$UiN<p^@MIwA
z?E$&OKpSes^2U0M>v8Sdz*cFntm|1B)@rk5UT(RW%7c;ph+=0jPBOqiYuCR?yM97e
zcTyu>5u;4>VvMGhLM`cq=7<%Fd8Q_`TqRNb-tCQey>Dd>WX?%{-Hn^UO9(|sUgrA+
zx_07%DU{@4=B#UBUSacq)F(y-%gt&yG!6p>??(P|y+;2S7W~Oou|@kw-Vnu-gdf2b
z99MM8-TIoIEN+A}owF{aFhk{XX-bF#Ho?H1*=p^V&|I$Jd`8POx`sUtxF{lEDD0ST
zl7P%nKPZ)yzc_SzcVIw&0@#|H&5Fg=&p8wcD5$*$cRzl4;qfau{dg5EnqIJAzD|^C
zQ2-<aO{Jf0%)4M@EuL<WU9;GhP@;F!^kBZM=gmb&{3C!zsB}cT8Y^6q7q<a|6B(Gu
z1uRjdg-h@wa;Nv2xMYGE6La8>=Q+`S+vKbQIrE7=X~wvr<g3(Y86!^1o|1?Sg~)FQ
zMuA~4l!?%o_fwiT@G6Eh4@^R>7_&Q%NLsI$>#y*lwZ^NMdodGX&(5`Yx1G4ffW2ae
znz135V#iFKi9)$AW!2#c#Ut_`S*@wi@u@zP{a)=p2K_}lSLHAahcaoqsOc1LheXXF
zBEd1UA!Bq04ED7!Fl7l*Jabp>TFVXYBXcOMH$uZz7Z{ztTFC+<oovv~H*1!rB4-`T
zMQVCVw8|3Z{x*sEyY-a)8S6|z@8zvh%TP8}-3u|YBfzSz1Me$sz(5$dPiv|VxU1ft
zD5!<kTd`+R!a|P+{3hT5GwtTF5y&<0Lc@C->yNaPucA?&!w|3cOeFdGYYskt2qz-I
zm$U)^!37PhkAd|IU7!aooz=Vy92!+4&{x6pZ&oi@G69Uee(4mA<9A4H=^dMx$Rj4~
zqsLcuci{rBiy8?xmbOeJM;ct`oxBHCdNE+;uVnFJfJgirt&G@KFdJzCs}$-J7{3%^
z#T>d4>cNnkt+<H)tG&01j$~1m#l+0a%*@Qp%*;}YnHfeSrV%qUGxJCqF*A)=8ZooA
z_w}{U@%!@jwLiR0uU@PFGPAn6qoT4hBMvgAB>@J=F;PC1^=p62Qem^pdJdKzt=;-p
zJ5)|`T@N|<5Y~Wn@Q|*nx_LjmPdU9>XZqRO)lADdFvN!u4Aoz@^z%PK8r4!4_rxqe
zu=DD}mOGR$;qJfm4z|{wzLPU7e~D5FH|Xr<j6y51zi+K<%bDTrz_~qlPdaoyWHUv4
zLXK3gkjgfUVWWdd`?pSCbZrMngaT_?o5My~^6(Ow&mEDh=5VhP@NYh=Fo>t(A8R7%
zHn}6A{i$%y%*XiLC&g7JebNq<SMHKgg9J%zidptb$9mepb3mmXN>AR6taANTH6BGv
z%Ai5lm1P@Y`T^{@*A=*KgB%#4r`P4{#6`V|Qhaa#Q~(6%2=>=93(>+xL^Cq6(-x<(
zFrcdfh-5}ZB_#4fE_)H0fTu-c{xbGH#{mVwv@eK|b1!1Vd$O=edF!|?CDFtpMX+JQ
zB{z2nIG1mI0#|9>y+H14rDY2E48H~iy~cTRm^}$bYikU^By}+Xc(J>zM@CFggskeT
zGaV_EXlqg$|4=)S$(NAYo(6T5uY*)$DD23M9Hz3pZ;u#WJ#2Bk${rr%#B>NM@)2N+
zS<v00*zmPxL}xQ1-GA;ICiZvA6jZ|*D6J6Vt3dbx$*yHby%L+2#FG2r^z??Vb)Z$L
z<?f%2oB|nP27ZOT7Mka!?L?i-C9xnX9?<3tXQz7wh8GzShM~UOL9?go$Qq|Sqc%jk
zn_wTNaH?phN7x;sF<jb2PK3mrO1^=K`G}9mPBb)hbx}h53ndmD2QQr($i=}t98gj|
z!8HaaqXf)qk8}plWfm1ivmX`fPT7gA_G^k{Fv#X)plE}X1|EZ1S7jythe|uzPb_8Q
z8}wgPyy5V1Egp`gJ|0>2{G$l-UM)1W+*j<IEv1=-l?3}nC-L%KUAvWy{Xv#B46+~a
zxbiIc#a>Vvn|QA+>k!&f%5bF|1Eb2#4uZyL%!cvy-c#UpcY%zxVkPjx-)x<ua(JW`
zNU|oX_F_2+;?1D%DmAmfTu-O+zND#Ub54xoEIAH$p}$?VcoM~D<4PaYnT$r8k>ZRl
z(G-N4gB5aqYouDT)K?W~lB#2u7+fLB&7nr#rSrk1QSsHk=y36;b?O31C$(Q(A-SkG
z?C?mg#Ss0CCXMS6nkIax|LS%9i}lqpinyMKFJ{OBOTx*)_8X6`^v}yPB-Wp^$!nI+
z+_===r)yK@QFV>nL54V^bc5;)i8sABuZc0;LfuEapajIXX^^urJB$cv-)w*XJm#v8
zqorUp9ixt|w*BCn=qMI0SY!2DQ1dTLWfj*sKIBJ1aAqCwM?6@bSH{VbncajNF}fb_
z>hcL%9JUg74h@->@Jk_y+q2h@kn?hG3CkPzJ$vn_(@84t0xr`|<>Wo=<7~;<3z=!<
z#V1OA0#a6|H`tM}qe)NW0x7YRd6tSh*lLaHx~EjGRMzT=@-$momM?l76lk>N1)57T
zO>LOAe(-=P)E&s@kNT}gdnW0LPqOQ>Tq&B&<;^0#OO5JY0Lvf4^MlPM=v<e#wa^H3
z&LVbwF-0c7VvZ8|F#v2?dT2Z;yANsGQpn@w$KDgiHY6j^C6i$svE*4lY#`hyUhVYh
zlh~7elx>O70jI;>vXv7v=*prE;o&W@Z6v*O#H#~wYa-uql^|ZWAODQIyBUSb=C@VN
z7+{N-mgm@u&1_N}3v)Zj%|mm&$)o8@3{8*8Q0axmJ=lCLuERw>`ot6?sm`*n;^D2O
z9X`B^%o=JvCPsB=l>ye!9bNX!()vp@dnh*7N8&bwR!)W{;XKJ9&#3YXP}?T!?{H;x
z_7IDSEH<<v3}sBT32m?L&&cd(5;4K5<DpBjvWbjVw=1%uo0L!>Y-f<J+j!XqWt(TU
z7)p?iqM93E%;7%AB6eqZe!tzuwl>(mm$N`L7w)jsb1S_5u@NyA;Maax17Enu5yIL>
zowCgLL9x<fY-pG>Ls9vI=)>YB)1;mS?UJ{EX>EawC`_srNvF6z`Teb`4bFLQs}%}-
zPd{6r&|s6}cd}IVPa8l@*HMYYakC3810&E}Vp>8Y=eaC<n&m9|ur+{Rrewe$O3WaA
zi#{A@kqM|Hxu62WH_NVS2EII$(2w|KQ)D>5pz7t%6T6WrCD2-pw48K>@s45vqxY`{
za}z&#{<S-)D>+(K8cYTLDF-r@-J|1TlPWL|OWwpeN@;63Jv|do71uNNo{S7blB}MO
zt(A8)SMp<(THWy#d%>%Hq!#4p9Q5Xc@AQ(={Na6C&Q>u#9VZBq^#0GKlU{EK_D6t2
zC}l1xWy+_uxk`Nw64NkCj3KR08fJRG31Sr;&|(P%u_>t}%=2VQxP5yBN+vt3pMl;2
z_W*+>@@s42%2|eZeXUAWWG}xtYYq-JO-?kFw9@j{-A5l7_uZYGd%SL>PWmZ)s)Epi
zX=2@e9w&|+za8DvTNWcWs^7XQsBMM3Ag5Q;^L?DB!V!5fAa)n8T?#6|)ULM<z8o<d
zt6pPWA($PrJoj&dD&+RK28GN0Jda)0d=u5V5MHb_`#ZuP<zU1gKzsKwg6C%^n78Vf
zCexR=1KTzf`qQi2Mggpg2rw0}Es#4umI<2TTPb|{a%i)3lfz$JL{qUq1`tY_h-u4^
zJ8?z!s1JycY%4QXtKf_5uPT+=j?yEQpr}8tA`v706h_WkE@9dGGzwr7ww^t}aI(;M
zv{*p~;#P385)n47>R6TjL5wBWppT1qnK-WG#gF%pgy>37d^jwf{R7E~P|}_Q#)BzW
zw^Hpi6rz7sP3qyqm6_vw(0}OcNi_&D(A3}643TDp0|RZHT8Kh5f^L;0?kZW$lV8wq
zCvmn#uvsD2?%2VS^v$jPo^EOL5URZ+j9$PpCZ+2>{6gQJpj++&cXm6dCT2V(iIreN
z2FP3%R!!5volkWY|BmBsd|ft@Y~uL7Z(~bey)4?-rc#=4e1@>D9*!14muk6?_iuSB
ztiR-`SXF!M{*tFsd*I4pq`|?O`R1ngA;)3lm4<tE@a8cIV;!@IpW7!J$YNpiB~OL_
zh_4n4?9bml0nL2GZ*nGf3q-owpkOXm&>9PSk0{dT$RZ<P$ri1}>uS4f%c?UGb)z$<
zifZ3GQDX2|PhNGUY=Z{}F$QV5hf|ESKtG^Q_=my<Ap-yO<HRJSsrc3r)#g2<Br{L<
z;Aom4eL^F5g3x!k!ek<F`GRqBMOrJ~ExfT!WFEC^69X{)fQlC`sR&_oits!;8in6h
zY@7R*!ehX&H8CF+$%1`2#{NtyE$0?Yrkr{ihv`-uw0|kv`na9q!LxBK-`C^rT&dLy
zK?w8<K?XmtBPF0Y+UM#rLRUEoTAz-fDZo7)ri5d9C!;(j*>7CT;8F;n8mFnMZT+iI
zTb0jet%Xh<FB*U6TJTUkPTZ-o`H49tbP)tSf+^DK8%=V<AVA-2WAYpjHz6?#5ta>)
z=7jRYW)(n<bwpFf7VTwOJI6s-fiN&F1cEug1{0a10`%s4fSUEAFgT;3%P@KLXO=Fh
zDrs33=-7&^f4?^)Flgi5Y=b?@O({Vg%)4s8&PY9S<d;5G9qoa2;~#C4w7>MJpcrm}
z@+V67`kZ@OaTPcz4n0<Cb?+1Y)~AwbVNzRn;6htv^8V7NiUoBMby_&3YhcTkFtiWw
z*K@sZZkW%4Tx~b<U*rV&fCIsxPXhQv(FEQanthrK1jmJjtUk6~ZV_IwK_s)2kd;;O
zrRJ7$=h`^>qAxP*aDXk!!?NBn=(OHj8`nLU1GxZre<0m&($pZw(3aF|jV*#YUxc3V
z5yUZQtyAfoLT736Ph^v)f;vYgSi5+tPfe=QN!zeDXtR2}XvLR?IHw<fNI!im&U|y<
z3PJ>VFOr~Sti#T+I<+pV+3gxmN+L%`-zEAo2K*F6Q%#%ZJA(gPpNfwNZK-OsLF;RB
zD-RK@kd^3O(9-#9Zj0}qjGX_s_|Cr!(M=&{L4Hk~v$863xEVa|N<p!)R5?k~ue-0X
z-^pJ65Te+eHN)+O6B>v}SzF1FBFjoX$K2BONDuuH3E-sb0qy}Ca{8g*%bn)Jv}?ef
zJK*w&Kma6tvI9A2JWZFRbbulcVUyfN5X(uJ-T%2ZnWzJqeoPBKktChd69r3>iMB`2
zbtjfyE+m=ff1;wGulG2}$0D&t?S*s5c?<z4@R_nq!ZBf7qwwdxf#<v3*fFN?lhG<F
zpz<+*VI8n`dKH(60kG{QdOP+(c-wNsna$#%p}arEKu*>5-p|fU3Fgt<6afnmN;$hl
z6Jt7pvtY?!T~i6+2D;Hb2SUT#fHTR%L#HK@qNy6F5`ZDbP?vVga>4C>Ci+XmZ-8YX
zRHES+eZQH;5!r{N9HwG5aOrBrV+>q?NjwB+ttEU$*w6^K;>r))h5>~Rl9yda63qes
zMgS)4nWZ(*2cz#+Xt@&@&DiZ%mB)}Zla;!7ok6Dv@?5fD{71vAQF_f7)GH{FfHfa|
zjgQl3(8r^WWh>n62Oyap(+oi)PBMFDXf%u~EX?FQ>S|6y1(=XPmueQ`0(a0EYXd1E
znX3Gu`Yd-a`nMu3SCxaCuiu#m%eTq5U0+Ucu%<{`@Qytxb4ej_qn^k5E#RY_H0MJQ
zB16y@wL8dZ*=EX<gmY78oyRsdAeJ1fk(po7*xvqDW#|sTImYieIAMC63Z(WAJabL!
z>bPfTPB3?GJ1Iu`A`GuF%kTahHNJ3tb=PrcwP^j|4Zvg8TRJ|D#4Zt268*_ZdYXE-
zP+EGr^S6OCgycLwxkCp9%8e2DOyY1ue2FQB%4Df+DB5?MttXdO_JE9)WSeFPN1d5k
zAt14|t)%5V`3lf@OazK%y87YIA*U(Eo{Bp>#?;~+=Q8>D-a!B@=0CvJNZB(ff|A&6
z#le!~(x{Z#Qz7F{9CCoWLAd1uC_N(+ze!dlqF%yayP{ha1hgtC?zY8q_O7-$EIG?A
z?yP5=5n|=_KW>_wIbZc9!(|mR8I@;hvP5I;qH6QEC$nt0+6|e1-z=5xx`Oug$yNr5
z_u$UJ9wuY^!8V2`{)&^=vN1n^M61vED~qCg2wb=-f(W)1?P9%~K2{!${w#|;U(s!I
z)t`2t)mWULop0)M7ET17L4LQN<-6K^yunfcSJ3Wcrs-T-CTVwV*nk8w%0}^Hm?OJ@
z<$xyq$J7Yd*9NhdbYkRhJ(Dk?_FbJP`7QEQnDskoX-jx!L+QgmV=kQXfz_}}m0PfF
zc;`sIN?qu8LlwKQ<V=xjE{#fxze*9eu*Oa`t+t?O$82|VQ-|p)WA$BYgDHMMa)F4M
z&C=~Y?RKIq`*1?>4~-<}h?o)wZ+qIPb2%WlIkl_~TXZd427%1_ZN7kHv^X}Xmd+qW
zrN98()k&HP`hS7a7VD9Z82r9pG;|JBZ+ieC22ccvHJEWyjZ-bTsNTPF&j~4L3Ut-_
zw?l?ADbB}cf>d(}tx&UiwHTTaqqf8X;h-;4x}`7eoP_iqZj=?MncHpo-|Krh;7@M0
z%=TPWj&J2RU9C-6-;U&_sEx!B@QB+NU7GCmEsL;Qf}x#EaayE^W*`5K(uf})bk3Xl
zb%vyHx*yx$9Kb%Z771JMK!k$Nkrdl2Ny(KwN3!RryYx#Qf|D;FUw9phsO3Yk#od3}
z0KSc`-1BMU-mtql^#wy^>wu;<1VR#Z3QhV9xV6bHK~GsurxhK2&1O+ZmXHVnJ=J<z
zW3lG#K%m*IvUK3ZCppiSm?OV>>!7yB)HwWWu&@(e-##Uh`iu9+pN>+<<50vD@J`vM
zjgNraoC|43i7GUT>vBv@Dcsa^C_&;?olc+b_Zn7916Zrs<vS9aIU3hFB0IgS22?fw
z*tI%ykV?{t%+ZeO={nf<{$3~C(?y(m>gBfowY4l$m0fLfuc5r|B-P3=d2hm{q!P5i
zj?fKs!}$E7Jab+-h070QmWR}i?I+3xvH-lx`$3zYiW*I9kAWx|Gw`QW<EW!d3`cl?
zc6LDzzd*f>Y3v}UdZOe@n!_HoFkj$D{L>3HgeNOAe8dRc(03Dx#J9kJtYfoP%8ehm
zhd%G|Zh^^a!|UfpCMb}I^6<VlNJxT(rUVtv7S+lgDNXCu6LV;b)fD<>Le6-36Uw?h
zXiu6n1$G=p(xeOWfk*jp_gLLu16`+IYjQ+>7v)tKbQFv6$$x}p(2JBfvnrnW;lR;0
zN2>KcCgb+yEV1lASuy1sFL+(`i*N(PdkKeX`fSLXsP^w8&Y-X*soHoBX90`v%noS&
z)#}fmc7_Js>aU!?i1R_r3uX~uOLOf}WM(Z1Bgq{$4tm4PY!6T=O95~shC?a6#h~NV
za#9yoTzSuLey#dm05H<|v7;_S{jRpS!)LcU7uUFp>M}2CHFFSIf~A7u*lN49_CD5c
zu6p)%<3SF@95Z_kgDxzt1+6U4d$K<_+CRuzS5oRB>f>yZw*#)fHme!4=TYg|Kbu)J
z8u4{Zx}pbZa6V}aY=};h^~)y&^wy|rnK;*lO)J@cgP$BkqJVb+lQtKc9VTPUljD|m
z6Bb3c*&BrW|Fv&cj|MhJOLPHIAOCgn|3~{~|9$Z{{kr-uI$Kaxk1PR?Hq*m!La2Z3
zJl<GJkd-9Kvn44mW!<-j&})%Lo41OAz+p45+z11u-7>W&>?F5USp7(sCrKh}#cigc
zLS@%2WYaA<X_a3VUv%iRk7UU72Nhp*ofYaRUe2)Mr}mn1{IL;i7XT|&Q}dX8<aKh%
zPr29if%=%^(tSdiM;P(zbAi$;_an#qO!_+b{aWPng${=iQ)g}5tG#ik39M_DY>`^-
zyNS5gxBX@kMvA-sLQ1TK*&OQzmFkK7GB0PxpX3v53CveV<tX$sDrD)$U!R9Ku-5UQ
z8C9dIT*xlF1bQ=ALtu45?_!x&<;*Mq>cpyf73+rdM2-p*#Mpp)`HT=XDasU#t%`TQ
zhqd1=>x<-A+7&Sx!HwsAZ$O0+&S&!L*rfz}i`o9e%Qb>Z>H9!80-iM*4Qo#R0+MuI
z5P^!3uVF7g_@4~lhqOar78DgIfV+X4EB7H(YKyG<e*g%eS8CvztKysPS%_1i?3LM}
zz3O-vbimW}*w4W;-kJ`T9w4e1qYbP(3R)@8On|I=eD7O014OY{nMy@dzMI`E%{nAx
zxoh77g;80FN+2gbK|`1c8CxMWLj~oZkTtTQy)RnGG?$b(f{O>Ek}-^9nf$Vhxjw}1
zWA83&&!iUVY>dCn$bb$!rUEvam2(Xun8E$wTP-AhTiYga&V0ZGWICk|;dA%=66Kzl
zKB`0kKFv2Q(0c`v(DxmJv&=S06w1to4j6>O_=I{soNEyTT5g`uDTMf!j@%u@R-&gA
z3pr1Q+t0%)94#Fy*__aW<qO$!6i)xRCc<;?m(I{;9;u3>Nd{<k|5EO>MM&m~MLsU;
ziy|)<^?`%)?|9E2nft}@)6#0=n_RLL!l5m$%Jw89a;vEt926WsM-Yd4R&QD0Tda15
zh}ZSaNcSL@bkl?16QHE%vS1~%a^}|sY7tlb?XEuSU*4`8mmdn+3f=^ZS_d`LT3vJ5
z*m4HQGb)srfo8@c5!^nyBCjn+U^EY_6(Nmw796>0=p?dZW@{nM;SS==;ZpDAfqI%O
zap!)JTmy}>k_XzYl?jGVX0pfyvG_Yw>KcH=PvhIz%OvvagrwX$q6N`+IXyd{|J)3T
zJjlI#k2`w5B&tgYiET5pL{N~N28<`eh!)jifQkYVyVp0$DO=XN!(kK<`*k2rW8pnX
zDfPX3Ud-R@x?SzfhUs)Wz&<{=#ONII>44pT!9-|k+pz=(;V8W~@2-d}o({Uk8O6TR
znzXlB`l1c!ZFLr2)0-ymnR9Ur)J^;WgcyfFoyP!Injx(P7Ee&<WY7e4tPcjxIWuVN
zi|)PNhtN>~?}&}Pwh-|*>b6#D?%aV<7cm%vpR;fkkM^~}9m@tpGJ$g!W;JShexFGX
za>dLBKu@i=sxeOY;yAfaU`KpNjj+l{ynf4iNRY3&#tT2o2V_*H5*Ixa$M=h71SiX*
z__IOxFe%us+2j*c8m4qNtj;7UsnjfQ`<F@N3>O(I@aW-CbRjLO3DLZO<5=v0>vX${
zsfti*d(Oo(aT<88upiZ2O4MKKiVRuMmGhP*awaM8*ymM{08$>F%fpl?+p=9f%q%+e
zuFx++OpW0rQug$|x5q?nC@+|7+8wBz<L2Q&Fzi)g`#NX%bH@15c!!Tptm>`B*i&?@
z=OeW7its;Q_bn<JReN2XtZJ=bGof;O<2p~hBx9(_5nj=Fymsaa5qiB(y4OH?fu&YL
zv2Co_eE@&$WjgrPSLYDn_Q32+9E8&@eLL&!43WA#@5|>xhc`y_MiJle6L~EPaoFUP
z_wUct%=sC4i+Tz0jykHX=ruB>3z#+tK~0lSVOyRUO6Db%3Z6GC>$R<)0n);;#N2p`
zE!(Y^L@{T<$CjKyED>cZfoEfuTO3B095g7EHc@=m)<NIJsQV>%z$bnsw*Ns+-yO5P
z)dI{*_x-T1T~0$ME*Z1H5ufXjipH3!dbY=R4}@n$s}^_}_v{iT2Nb-jkhD0g+6+DN
z8{j+5qLs!@D9i71<KYVAX@1utL&Uz8)xrfiuktXHgj@3}^;xV(Xy)Csh?ify=%`5r
zDRZ{%UE~=qwq~8>SzRtdluFg3dq?c7EvQ+XN;y}D2<L|~+7m|v_*^c##fTbGZ&w;E
zueJ|x1U0yq1OpOH=%Fa-GHrg^0PNjid5BuKnM1n{!}JsEc^d4M3rf!1bi8a_uiQ?u
z_Z+(iAM1AJW9DzNO{<ixmU}FQ!UEc}Y0|8ozwRb4PQvxG(WVi3rFe6@dLr*MYve<V
z5tB$;HuX-SgItHh+W9JDXGiqVs<(b#dY-5MP&gmEtR=YKquTWSBYannwMgO!wOP7?
zOzTiAajn-|m(D7i{8)6?D8GD@>`uyqx{dUm=2X%w{8l((wwn#o<{Cx+xjpQVR`7Rw
zV)ZDavVpvYMNRbqZN80co{^>gOt2q`+Cv$!NVI#C&IE0Myiqa9p`I`zMx$w{k%lpw
zOnCS{MvS|$z6qjn1YNK|WHRN^qEA?$Ic|4j)D10DMIy`&cxu5ca(G5UVpm=VV>$Yj
zpu%QS@K5tGs2uqKsFh?gEE}IA-WsrxxV#+vo8Y+Mar@RWugmZoY{zL<l<HjCDOkm%
zc&C)3n?zD6@5X8_udD*8+t51(uc9yuU%{Z^v_xSWW(suHf>H@%Ok)USy(lsF3|(?G
zw7m+Y^=-M7nlQJt3INO)2^L`O#ZsnVq3&B)FtpRENHT=?iB?05@&(1}^L`RgEX+ok
zJqp_%dQh}iLp?BZwO=k}l;KalHE#X|R_r@4_n?i2S6;0CwrLM+gTZ3Ep|YWPt677P
zghLYwlKK5IfDXiUTF>klUitW{;ijp2g!relx3m8#1N%Pn!%(`bicC>I`S6swyJw`q
zc`fcvb(8`4P|}Ewj^vu6J%VczSR!{okW_r&6)5Odf%7#MIVm&ZGqc&EFET2NqH7oD
zLtg&~s49$Sg!BiYCR2}0<4a|d`dd!ZR#`g_p2<g7OtdCF|M$pJI}4IYp?5CxtSGBw
zv7z{AOAek&Y0`d*3gqJ-73#-R+7%Q~E$pafZL6o1x6+Q`wul_DMVS=Wyo40nspdJ6
zNvgb&;S?4giimvda)gp&QTMd-f)T12GJlr%dgf&4E?uNiq&$}%%o;S|8)kPlxREXg
zO>mjo;mbUUTj59T2fYcf>72PS8!m7~q;~!9TTz#PwsmoLgj*4dc9JhZ6N~jIKX^tX
zV(%Uw+aQ(g#Z!CBpalu?2|5`dcF4IEB)tPb-4?TriB4hf!lKdH$yzX=8+<IzcuuB_
zdK;dDa!XKoBzP<+{3EahxfEkY#+RBs%R8Mq$=>`p>=*@e%$Mqi&RsNZba`#s*V6w;
z&3`yx#_;lOT^>Y!V0jV0HL&M|*t$6BrphkW?y{?YG!)(33BgnNs+RG?$6YwudFg7a
zHIxUQNlC#wf6>ttD6m2VBI!~sn=d-5@lSeDhK7z_f2&0s8Z!R39Ea(bAi9-<Jv}SO
zm$2+#Jk?jF%ozytOIbD;1&qmDRg_hb158E1KXte5Jd$g_<)mN+XI80T(^Y04xQ@z|
zX%MA@9Et{A-GZFCYC<||SVWGzoQZ7RU3bT0lp&EJ5`5te!$vE4m#dM2+kBRfAI<yg
zrRcRquzQ2>ff$g*|9aU5DJ8&ocW$Tp(VN?Pb7&O$2`zbZ(wih_7_78YZ%78-pi=H4
zV1?1p1iNaWcP22Lq8fSG4C=0%o=wYEq~)K5PPY*M$`-ECF`(nx0Wqv}^oY{w7x_(-
z2IqG>{%Ce=dx8S%#y8p(MKCE-JO@Ecnyh@MV+KPvTIQd{3=g3gNmWuELR83=6wZ0*
zP@LLsKy1JuY+xKff9m$^RTN0+*@CPKN(m7uQ;abAmr4%QBj*hmTnm}DQ9&&??^2oR
zljg*V>MmbkJ%wO#tLtlr<WWvupi*q4pE|~*8<~q}ri%H%h8V|P8^vPk9xui;5_-}o
z_cXaxVHpE+%R4l28F#|%$n)cE2zg+xNK?IGuVt1O@2}o9dag+cXtf2R^c04J`M9dR
zx-KGI#$&0T3{mkE4ZM#h9<cnaAhZTIOF0xdn9c81uLZ6wJ>am~I2I78D8_hNk5QNC
z*S#?dq3Z~lV#%M{=#;j*=VX5X*aGi<baPv=+guMfs#GL2lWyc8E*ssWt%|UzMxmYn
z#W7GS%Xo83tdP_9J;>S*y4?D$ofA9q`QO8jEshZfk+)rlUQma3i{a8{OZQ3Hdo7CG
z$vBF%Xoyd0HAgdJoK~O9NRNKLQuz$lsBOim&gg@qff*>TMgh~Y41F_wy$$v=Bf-J5
z)*IlRN80;wdhv=Z?RHTuQmdImiI*nYj6gw~Z97o%26`$7n~Nd;>?()#lO`(PsWxx*
z>rUK4kLQpBJZ|@pQZalYKa;t6W-U9pG+)mYfiXL4cw*AH22+7|2<CAtSEl$Uw+}jh
zwp4E&>-*E?$7e&YyXV=QaQ@IYNnP%5z$}%R`O)tu(=f+juHU4~i#>OL@7^8M%tgA=
z@#cg75{$)T#>LojA;DEi6`e7o_<UnwFzp2`fbC|NLuytv#i(~5F}+B3tz73QS>0qu
zl+ak-NF<6?cW6wfNs()^IM!tj;w!N^yP~?H#qbTpA1=4DI@H4dxop5hH~H%uxM?@M
z`!Q1Vm@jHV_t<IY9$Dlqa<HG^KuIrVQxQ+f`8En<p*v9NX#UnlW2B3&Bhse+$HKMv
zllS|SfEXd@pfjWyvYY`NgQ=hFV*JU7(|ZP`bcNsQpyjh)pl0p4+QO}OF>xJ8#oFXz
za<yNPy=3U4?*sbfLFls}cX0x_-3vd`&E3hdXl;!No}<YxzZ1>KvpG8H;(&a6djV-V
zirzaM@?4BDND%7+a+V5cA`EEx3FAK^+HPN9R15{yfI!q*I_&rbM%w@5V8ltqO#Q#(
zQA5axTf^6V^He=C*jLy53y(O#OcXw4pX_-}V!7$Ryt#LlVZzSE5M*^hmO$Zz&I*+C
zb%HD1sH^Qy<t{34j2YyED0Szyr0;r>dZGEsFq59>4x71uyV(yNB!xbC)9nt1Lr+`A
z0s;sprtv1YT=-nJdp!~1rla21))q!{rzAi8yX25^ugOopFo`c9L+QBQh>YB^SmiKq
zrt0@h&F@=F9SA&gu3Au~mOYooq62CPqEdKC8d-t2-9dd~dOm}oMm3%fvYKt3e79fU
zTC2zGIg4t{RO26h$6kOA;$RJvUP%s=N@e@wdgRDVIF!p~=c5-dkrSYlivb3Cm0&^%
zi#mwlfuv%9@rg9C+*$#Y$SD~@GzBFemyl>X3cN9jqZ+h|o=<bpA>Oj*xt4Q!p-qa?
zD2c<Z$MmxxboMDfd#-sX=I@Z{mQsJ8kyDKRbgbk7yHZ8zg&&wk<4D@;q}5TmG1@NH
zE0*Ik1QrZ&uHndCMOc$HC0Y__LcfxzdE+chudTe=e5Lykk>qptUpDrq+=I|dL$vfd
zzKnKUz4h9{eyWF9uY+#jyy$hb2mmuwBc%&*Fkk(S!L;`UOhVb2X~cu=HF?${(Mh=u
zO0`<e044y4Eo2vHY(K$4jQs5`xz{(#)Y78jlSxrpgG7FwWVIN8!eO4Ee#nBVIkd6@
z9}^jboTq^}4K)aRKg9y$t&TTsKog=S05)g?(~Wb!@`oshBF`i91KQy4MF_bwl|T~Q
z0>^x3c5Y%_T9U&mou61yE=!N(>|g;)WnP^t75373`;5Zk;HjbR1AQXY1tH8IJiyz=
zaNokQQk7!w_Jt1GUwamG?0llEHDuYu%~HPG8*A`HCEEs<e<2g+vH*cydy^d2C{1mG
zBl3c6Ww8be1D+ZCecbL%QtIHL)G{jzdwMf(rDD0BE<$Z`R@mr&L6g`QG^O$sw{Uqc
z3dR;0%n^L{@NO^P)NcCo+{Ihf<tHp%?1ajpM_o(#-$N$5%5=z*&8T!Z-r7E2F9XCy
zF!1Ms_Xt<?;3OtXHxHefsU7(dtmHr6%+`m!F+(91?pZ_@<)d|mm+=!C)aPn#*S?;v
zt>wfUHag0Hu&a0dGON(Gp2KBXg@ayboanXK2dBR`nT`kbNGe!pB<VZntnM(<$qa*o
z8TF?*KSi4O_4Wff`xi~?5m~`)f{Cx-uaaKQwj!OBY6BF^eQ%`Fx%^iHz$Ev19hATS
zr?uCjAHhi5$mrnC_v(SWdbHu)z6&a}Zf#2jT1UVvk38gRo`B%&IY@V|mApc0z6vgo
zOm)?!cE<9OMBmNiG=Eq?FdgOC^CVjgtHb8gNNTveheEH(kXHU+^s6c({>TAV%EYJ0
zmi`BHv_$tYlk7G|I>K-3680~662km~C+JCEl)t#rzroZ0JXhlU4>tV=oBo4M|G}pJ
zJJ{4lXxI9e1JD90*as0&3?a4E7j^na4nY5UANOC682*>LIAzPZ+`n9CrsQX!nJmP_
zSXpn3B&Kr?-=TRT<}`<k7J{*GNyLU6p%zrgqO7RU=vh?7`9ixhJXl>X@HZycC&&RV
zY!Dzy5Lhtc{a0B;S*CE@Tne&HyFf{yF#eES3arsY45|f3=^cKRlXH>sQx4bR@>}Z5
za~G!-pZo^PBkPqFowa<A>lIa?g0z1WbD6_?dk5cKW<)=TM?R$LpG|y@=+X%jT0*cs
z3fAPviKtr&VI+#NoixxTvE9;=BWnh(%YW*n40Q%2u<S2U>~mxd9shZt{N03=kB-!7
zpM<TIOn<O3VajsAL7m7_RlNhhB}(kt?l)<n%Stk3VZ&NbhSD5Ao7?9xRT9ZG)nZ+%
zxemx)x5E|Z68+YAp96Tz*V(ITb_h*-uH?E8mv=fx%h#-TZcYiBZuMV`T>+!i9I5-L
zve(l$8Qfm6tvu!k_k)A&jR=Iu(!X<rWHK?OHp(<Z&U%Td1PSJz{t|c*FE;w{6l0_T
zmpTo=st$FmydImmCii7D@2f+UM?~Qjd<Dlm<YY@nEhLhTFQgsFfl-{c4@JxS@s5RR
z?iuuw)@OGBhWVLJSuB~`3!x4DoSsdQn+jou9r)gs&=p!;j1`H%Aao!J{i$kUfFNwj
zme3w#hQCWDDr>~hSf>`kOFEeiE*BW&nmNwnnHL>Z$NBInT*UzV6m=B;;%M2(#A{YE
z(?1r%2*tJ_gbhpa0T#p;I32>)3jt#P?WZ$gHMtU@;iLoC$)UxQ${Q6$xD*4ZIAWQr
zsP9XEU%(6ra!^A3g$AcEW55Jn$tnJCMSI^|=fa1!KQ3FM>IfI4p%81xD3wWAx*0Nc
zjPXi9g=17BbXTje`vx~>S#SBamt-~vAr{Y};TX5fgAmt!P?H-%pN{f(-s(O-(HQ-}
z%CfdOXGjDg#m^fE1;E>l)%gmyl`^|_tB{csLtTsxvkwW){y<@tf4_&+^;&-3)hzio
z^$o$-YJ~5!jJw|-M*%T{cDpSVSj;8^-5rgZdZ97Vs|a-O3gAYp=-u1{HZGNPQ`Ot+
zRw;NkYPhAyxR1h<+$hpLV$Zhs+Dfb@y;%wuiy4GxRo-a5!vZ)c8m9CXsG<=7XJf57
z-@In5W_K5|;Cr`Rt3bTpCdZUFwDo<;V67FU8EzG3G%dgmbd&~eH$z61qcB&>BG)pt
zal0F^TkA8@hEe4+DTZ?H=%?e3al^wx0s_yOs)e@YGjqvTnH6Ok=S8WG4@(2UfR4f|
z+2~qfgQ&LaJQ?}e*F}rOwc&+l*F6t2HtX~pd9nhBfu0ifXP4C=xd+#g@^KK06g1JY
zT%1tF<+D-aK}NOPZz_H8U2`LR2S*7DVT{3rPb@zHj&Pfh1qC4i6g_HXxcMIa{+5;-
zo7}809jP0X+0R@;!G&YJyj_Exx0F7JkZN<@t;P`(?Ib*|fPC|9h6$C-lIAfNGHR?!
zBze$6NZ|$|I;`&&OZcqB!z$MpJdA_gUHfq-vjpl*T^5|oIxf|HY}3T4RnO6(hrO7A
zo_!M*9;(iky~H8>4P1Yd9#*%yN!|aKD<EEkT?>ZjX%t)<)k3QCXJu+MYfs~vBWo+f
z?O2B;rrP&oT85yidPY11PYg}3N)7jc=^LR<<qp2BW!>v9m)ezSTf^f&fbU06*HA0V
zNY0Hfowb5XFhW($CW}sGj-0iQ06c;fgvq?CIs8Jo_rsC4S}wWCR>J}fsQQ7t$P*=T
zX+am{1JaZjHfy*=kYx!Ip{!7qN#MA#nqm@Msi^lMp$sfNk*BO?W^|M+sOnHvsCc-r
z-<EXwGl`&pB$^J-!x}x?x0Kj-C`n}tZHpx8CzZbJ&qQEdzuY*Gcq%nbF@;Z%*4H{%
z7~6+1<E4HY@Jw)VRv?E8dOjtMWf|sY`C5=L%U%5lnxF9+da$ibG`N}D{Q11c*QYi<
zOTbCizA(xADB$8Ti|FJE54O_1LS>)(c~sl!<^4?*Wg`KuV)QJW8|-l2ux{|N-zzkF
ztqkb4*6<{M7)y_r>=POQZVrdD(Ye}ZT}fV*y|0VO6_@a{6}mT9)KO%(c{w_35Ufy-
zxx<;x_FE{VcgWsZni8+<BNigkOJg2KG04T74<BNCU8hs`qACe7!C_n@)Yga$%KYP~
z5d><(@HQ0OG4luFN){Lt3JK7q6bAG1H4wmLp$%^ZLQ}A4q3~QjO%y>=KUaUIK&A@2
zE%Kg7X}%e$wNpm(830%gCSPlRnPaG20o!o2?MpjlIMas<n@j(;)rlJwx4`6ygB2}&
z<j)En?<dTKPbIXHOd<`aa<}!V2?+zuee464?I1$c&61#O<YLv$n*{MH%f3R$!#Tm!
zaY91mXQFU)@FDq4BlaEBn0Q`>CrejPk%A!BZwpQAmG8PCUtfmg^5XYTsZWXcONHDE
z)HqM?OmZ2Xkn+U1q3)hiLoIJ8S=mbwB*gbi*@UjS7Y!Y!%18P6cS%c_UuvSikJ9sX
z+o`}Raf|0#&b^nmPENe0uO=Vmt@QNnBn0iA_{>vnJp}4;G@Nq4>srdtsQdJ;`}$v|
zE0<5NFk3`C{X^Oz$0cd7J>}P9APUbRt7=F#OLHw|H5_5qv0IKi`xo;l#WRVO9;FP*
zT&9PuhiH<^#kJDs8UY<gJZc``c2_kK5L+O*Bryal4KEqlJtL|Og7<Nhljyr*)sxQS
zKTq&sh#Q~C<m+}i{n19FY48(H9v|Y0n;~`$UM=HKzn4*8=cQOn!wrSZjYT(U5GI0H
zAW?LVB<wo*+vArK&JFmJ!AS09%m@bXwW0Y@%oCHJYvM)iX!^er4P++9-1Rh)6p9F|
z+pEcrj0;iUhiQ-^t3Kw)m^rqMjmNwvtR_5!dU0@`#7@OIevsrt-TK+vaEsk35IBKO
zuuc1ixJqS{RM>Mv=S*S)D==wHsS;%nDR`s6YAD*aHjyPfanz+<EEyES0zUlw4@+uS
zo%J|7w5>|)X9pkaCJieUMQx1z?b&|Z7G8U5fA0UlZB6VbdE8I9?hK!;O_VIT$kpug
zjg<5R{9%RR5j)?CHHczbXZB^LN4KgNelelxXLs43kQcY0Z>u(|m)`fb!75RHow{Fl
z4=-EW8zj|XiC!JOAS4~tx{klDAymy`-ma<>N0c=FP+8s9jT9dF(^S5z`7#Dw!bfO<
z0UpsBhjNtjwca0K>aD;HBOcLqmpc(+g<q%p<S;o5zR$MRbgcs8X_7#TPI}NdjKQ2f
z)G=4zao}8h<wKtS6K)fl)PC8tk>Vz|P{-4vNYiBNcH~mlsCKUH#GAo&JZ>6&9aAuR
zY~=b(&hx;4i*yE|;V!kzl^S-_<%>vep@Ibw6Y1SadVUc}&Htx8xc>{0{Fe^g8FpY<
z#a9PzT4L>sOR9-}O|B@3_cU6Tqa^^)ayYKLoS}>+L|mccp|G6k>^boes1r)AxkQwW
zCI0-7p%_zlPof@gCxF<Qz#_%Evxq2wqK^Otk;g8-$RI+o3ZV?#iYNhmhpe#mzW9^-
ztRZ7VV78otWxJ`b$xG8dYrsFH$s>5A?sD1t(XN+AJM33}dq`wh(fa-=c|9YH_!MeT
zWbgr{`wR;K`Tc8+<6YuZT}6_>^`cx|szRA&QIp86;@6;NMVE{druky3ySa9%#bRcB
zqMP=t_PynZm3lsck|OxN`5L?PKDElB`Nmp%D?q(SXVKUJ$ixfI1e39SLe|8L`;WYd
zD_j3V1v=$b_TWNUG-?g5dB>}5eYaDRngY<zCFe%lwhdVQu0c*_R6KTz%zDpd-LfAr
z_ccomqm*x4^jd$S^G<wy2=5Cl2j>`dKL#!FgzxyQ_&UG$Bo*>FKFxYEr$t1Q-K`VJ
z9ig>URzThgW!;BH$@F(!6}^?Jlcc{Hw;;zD#xjIQ&;j$)FLoZ^&0Le#f9CvhHg7K~
z$woSr1A{g|+bY<9JmwW&RWRqtgdKndKp-*P^yeTROEwftZ#@_TKdq!v<c9rOb`iG!
zE%7XkFF}lkqt|dS4BcHcG|~=!;e_KnAVeD-KaB5~KJA^^{){8{2-yDxJR87tl>ak;
z#1{OiN4SbS$iWwJhUhMP#>C2E47%Wvd}Fk~Bm*c=it~z;?Z!Ok!pSSwHnra_C;X4J
z8ssu-^-<P+#1>l5DRlk}v{7UJZKM=qp%w-FSy_mXXb<{U35Q(L_QAIDi=`E+em><!
zTYD98>Mik#Xlv*Yh&-0`;eM!`IcY(kfH0Ra5C`@80scCXZ_GAF{kR%^&VGz<p42B@
zwDBb)Zj}RV-j|g+;{8c42aJbrK4j-U8QalwE|b(1i%+K+flsA#FokD8?!C0%Bu-!a
z6JGa(N1A)QUn0?_K@RF7(&l9U6jV<JE)Jthfn?WSf&O9W@tkiDS!;k3c`(p28_6Fk
z6~%H(vWKyx?wz^$ApVzSE>(JV<RA^aou0|u`<vXy%*zM@ncX~t^Rud^!}}F<6c0e&
zSC5y1nubXYaf514uw=NhncoV{(bvq>;Qj-caiY;*t`XO{G5SyNy$3YNb>(?n_L96;
z8@X1yC(LVr6)ljFJ!`WWG^!MZzLKL<T|(vJNyyoKpd688^lD_XqxoIXiYfNIl{X%9
z;<RBM3<E@H^onSMLKn(Z!D<5>GdQx4Hp^FRyF4Lag>zeAiLFhe=e{^4rpc|k-TuG1
zr6VmcPEaqzm8msuan6hcsKn1fDmAqr3q}VNpclko=sc}XO$OU<*DYh;y72f4+(2rh
zU*O}*Pav>$4$0E)_~6;j)#eM2=qW9rpHB1}+<a9Qz_uZSWSR5o8x4cI2Oy=M2TAVz
zS{_99R|ZiN00mU}O;G*&kuf<tr42VW2+5AENpvx28zFrzSlq3XK317R2cZ;fYK@om
zjDcgm>ax)Z{)CL@N&H}Y4s0{qdH@<e9h)s^4$&13lzA*rZ(zeMdRV<LZt0!u35W;b
zB@2YX!d8oaF^lN3YWKX&?o~9;tLWfN#Y!KB0lh&^?b85;v!~v;tJ&kIIuDII3~H7*
zd49=e?5<q{`{rOMAGib8=ZjMp&hMW|e}UL~e;-oD{=gY%A1@vnr9rQW6At+vRVcL^
z<2btpZxJ7LPdXAnaM&_SXSd}-RYh)8CyYLq7n$faF1aAdj3YIVRRLTRJrvjuR-DBe
zhpI0_h7p1IG8)K6>KL`_f;ONk<A~B^k`Mv6j|j_cj$jl$xkfJUdOg|ImflWJzb`^<
zhk;C|V$2AmNEx!@s_|;tbrV~rW+0|0`)lXCgO0QQJaUBYu2jPp9&LuTt@4ZsI^n4)
z%*9++e-UOTno1+n698)7(Rs^^ILT<~IW0~QnRD^-GGeF*hG@-XQC5FbP+Wd-B$D|9
z6a%^I2RMP|LRoTkKF1+%xDwCs9Hi0$^YO1Su^=&VTBD{-xiS>%=XOZMEMM;Uqd^%0
zTfQg%B7q@kY1QOVWlDpOJD3PEyfBXL!*KeDTfD*)H?%ei=^5aiT>$yUw&O9`#JI&P
z<v0!5-sB>W$8|136HmQ^*_|1*#>N_5^*kTsaV(1goAW#OY}cFAi-T)BG*-Q@ZxlW#
zOg$orL2<~`Zxk)VpAm+nrXR6%@&`--jbQ}qa!zH9snJgT@Slh@dR{*f7}>aq584>p
zsYA77`+%!n_F@i(Kz5Qa(|-IZ=TU%F^98bNk2c#2%Xn~xWe6ku#mhiH8$}`Zsa3Vo
zSrsDoJ9m1DrFb0yZ}|vi=z$V~{sm4Ep1iBD#i)3;<~xdyW+GkrFsEOF(><3m&>u3Q
zoNuKV3zSJqzjF;y2=rqeSFgm$p(R`Xi)2pXeO{o#b|itrmknhoKXtEuDXGNvodEp4
z1E+oVo;XwH7>bP3vj#tR2o{|NHw=9D@cL<{aU=asGX`*FFsRkp^TTHp_J{8Mu@p8O
zCz|xL74b*ipYPLjV(J;DVJ6GgK|FfoQyc7xaNV)BO=MJ^En{JF^@_yk(UHmxIL>W$
z{W6F2=oUcq^)a3OVa&6l?7J4<(fDVUSK&nC*Xluv0Z*00wdfg>RnEy`1{yn1`GoIR
z&;wXw=K-mu<*R3a`;m9mv10KZvFdL3af>%erti9UW&(9HoodkWX=pr|^Kn(RvQ{w>
z4yQT7kvVs>v$EV|3A4!Wl9cV=tN5gV)3oI;o%U5U6iwR`d~njG!e^;5hMsejhjQgy
zI^rr5MGZ7VK!IZ9$0TmJ8ljoe;+hSUVXm#Gf@mZv;jcJW_uRs+Dh8wV+H%*?(=m>R
zlT3UQv}fG~Bdp*MT52av0s+-i`1T=EY}3vm&teP_YM-%!rS-`aYQMn&H!C2sK(0vf
zI1L_pV<f=jp+gQ0_e8p-&+Ofk3M2E*64t|EViRg3?4Ems0@)N6oO^pxj~on`;*1|s
zZ^O>4qhEH)gLb*%Rf$wKmWyq^4ES*-cMmW@_%!JyHgMfAEBX1o2!H4Y_IYP62ZITq
zeA?;QywEsZ;KupvTm296*D%_St0tu-sGnI&`)*)(^f??B{0wW+F8#p$t-KW<2Xxub
zt|gCeBGn2tU!XpUcu*!7#Y26L{O%C#s4xYc=qWOSNFMC@DbEn>P^+VSacWUh<D>lN
z%_W=mB1Nj$6hrhB&Gojv1gB#TE0bv<olv<}C#Mtc9$q%RIJ-R2&n5X+{Eq!TMPE;B
z!Cx>PFpak&oHEVbiRVrgVM`8(=UL$pCy`Uty=mu5^`EK#YES+ftNfpCPyY9VlmAf4
z|4_>RP|E*M%KuQx|4_>RP|E*M%KuQx|36Yn&Esxh(=SR11P+7*^z{k^0+RkGZOX4{
zn=cz~SHr(&ml@p||Ni_MRc`CPnZP|<0gj{<rMLL^JaYdi55Uaam=R#_;%IK_3UIJj
za(1u)*qXb1rLto@82>AkiG%VV{eMp-`j1lm`yqX0Fi8FN_rcV{?r5q0dj{@*l;N*~
z`HJhbx$t-8By1+^f0@zz%gOjhevbcTKx1fXWo~NY=m7Zg2>MFWqfGvH@p#%+9=iXY
zMED;S|F2)qSNs#5_P^uPR*I;p{d;_#e-i({{D8g!Gn8~q{&ijn9z)Cj+WGvKZ%6Q-
z1pb%NjIZdynwft`UxHz)lK=PUtp6bTS425z^{Kz=T|1@;K>&qF00sJbfBmmOloiWY
zDU4!(fzA^_!-3`HI#7dvzcN-;0fChr0TW%8b`5?#B{2O5SsDM=GtT(mZbbC|=g;7;
ziCuuBm%YhXm6}gbT#2Jgz(0XnK_h^HIu1ZWfQxHi(fq$EdKd!&+j;;YYPH5a`Krk1
q|GkR-k8)N6ZqI_htdbIR$U&`v)@Ojt|37hN!uQZn1_cKL`riQYziV{>

literal 0
HcmV?d00001

-- 
Gitee


From ad2be063466fa76917055f1a996a36bcbc5db44a Mon Sep 17 00:00:00 2001
From: panqian099 <11823801+panqian099@user.noreply.gitee.com>
Date: Sat, 22 Oct 2022 14:33:46 +0000
Subject: [PATCH 3/3] =?UTF-8?q?=E6=BD=98=E5=80=A9-=E5=9F=BA=E4=BA=8EWR-EL?=
 =?UTF-8?q?=E6=A8=A1=E5=9E=8B=E7=9A=84=E5=B0=8F=E9=BA=A6=E9=94=88=E7=97=85?=
 =?UTF-8?q?=E8=AF=86=E5=88=AB=E7=A0=94=E7=A9=B6?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: panqian099 <11823801+panqian099@user.noreply.gitee.com>
---
 .../.idea/WheatRustClassification-master.iml" |  11 +
 .../.idea/misc.xml"                           |   7 +
 .../.idea/modules.xml"                        |   8 +
 .../.idea/workspace.xml"                      | 156 +++++
 .../Untitled-checkpoint(1).ipynb"             | 331 ++++++++++
 .../Untitled-checkpoint.ipynb"                | 620 ++++++++++++++++++
 .../.pytest_cache/CACHEDIR.TAG"               |   4 +
 .../.pytest_cache/README.md"                  |   8 +
 .../README.md"                                | 127 ++++
 .../Untitled.ipynb"                           | 620 ++++++++++++++++++
 .../__pycache__/dataset.cpython-36.pyc"       | Bin 0 -> 1827 bytes
 .../__pycache__/dataset.cpython-37(1).pyc"    | Bin 0 -> 1842 bytes
 .../__pycache__/dataset.cpython-37.pyc"       | Bin 0 -> 1842 bytes
 .../dataset.py"                               |  43 ++
 .../ensemble_selection.py"                    | 169 +++++
 .../generate_library_of_models.py"            | 191 ++++++
 .../prepare_dataset.py"                       |  38 ++
 .../requirements.txt"                         |   8 +
 .../utils.py"                                 | 163 +++++
 .../utils.pyc"                                | Bin 0 -> 5586 bytes
 20 files changed, 2504 insertions(+)
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/WheatRustClassification-master.iml"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/misc.xml"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/modules.xml"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/workspace.xml"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.ipynb_checkpoints/Untitled-checkpoint(1).ipynb"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.ipynb_checkpoints/Untitled-checkpoint.ipynb"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.pytest_cache/CACHEDIR.TAG"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.pytest_cache/README.md"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/README.md"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/Untitled.ipynb"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/__pycache__/dataset.cpython-36.pyc"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/__pycache__/dataset.cpython-37(1).pyc"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/__pycache__/dataset.cpython-37.pyc"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/dataset.py"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/ensemble_selection.py"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/generate_library_of_models.py"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/prepare_dataset.py"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/requirements.txt"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/utils.py"
 create mode 100644 "code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/utils.pyc"

diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/WheatRustClassification-master.iml" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/WheatRustClassification-master.iml"
new file mode 100644
index 0000000..2490720
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/WheatRustClassification-master.iml"
@@ -0,0 +1,11 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<module type="PYTHON_MODULE" version="4">
+  <component name="NewModuleRootManager">
+    <content url="file://$MODULE_DIR$" />
+    <orderEntry type="jdk" jdkName="Python 3.6 (pq)" jdkType="Python SDK" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+  <component name="TestRunnerService">
+    <option name="PROJECT_TEST_RUNNER" value="pytest" />
+  </component>
+</module>
\ No newline at end of file
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/misc.xml" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/misc.xml"
new file mode 100644
index 0000000..404edc0
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/misc.xml"
@@ -0,0 +1,7 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.6 (pq)" project-jdk-type="Python SDK" />
+  <component name="PyCharmProfessionalAdvertiser">
+    <option name="shown" value="true" />
+  </component>
+</project>
\ No newline at end of file
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/modules.xml" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/modules.xml"
new file mode 100644
index 0000000..d2ee608
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/modules.xml"
@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectModuleManager">
+    <modules>
+      <module fileurl="file://$PROJECT_DIR$/.idea/WheatRustClassification-master.iml" filepath="$PROJECT_DIR$/.idea/WheatRustClassification-master.iml" />
+    </modules>
+  </component>
+</project>
\ No newline at end of file
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/workspace.xml" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/workspace.xml"
new file mode 100644
index 0000000..cf80099
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.idea/workspace.xml"
@@ -0,0 +1,156 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ChangeListManager">
+    <list default="true" id="def273b0-b4ed-407e-9205-33a81b54b2c7" name="Default Changelist" comment="" />
+    <option name="SHOW_DIALOG" value="false" />
+    <option name="HIGHLIGHT_CONFLICTS" value="true" />
+    <option name="HIGHLIGHT_NON_ACTIVE_CHANGELIST" value="false" />
+    <option name="LAST_RESOLUTION" value="IGNORE" />
+  </component>
+  <component name="MarkdownSettingsMigration">
+    <option name="stateVersion" value="1" />
+  </component>
+  <component name="ProjectId" id="1l6P1UjmjWO8xspj3tfQsGhNvNV" />
+  <component name="ProjectViewState">
+    <option name="hideEmptyMiddlePackages" value="true" />
+    <option name="showLibraryContents" value="true" />
+  </component>
+  <component name="PropertiesComponent"><![CDATA[{
+  "keyToString": {
+    "RunOnceActivity.OpenProjectViewOnStart": "true",
+    "RunOnceActivity.ShowReadmeOnStart": "true",
+    "last_opened_file_path": "D:/第八章：对抗生成网络架构原理与实战解析",
+    "settings.editor.selected.configurable": "com.jetbrains.python.configuration.PyActiveSdkModuleConfigurable"
+  }
+}]]></component>
+  <component name="RunManager" selected="Python.gan">
+    <configuration name="gan" type="PythonConfigurationType" factoryName="Python" temporary="true" nameIsGenerated="true">
+      <module name="WheatRustClassification-master" />
+      <option name="INTERPRETER_OPTIONS" value="" />
+      <option name="PARENT_ENVS" value="true" />
+      <envs>
+        <env name="PYTHONUNBUFFERED" value="1" />
+      </envs>
+      <option name="SDK_HOME" value="" />
+      <option name="WORKING_DIRECTORY" value="D:/第八章：对抗生成网络架构原理与实战解析/gan" />
+      <option name="IS_MODULE_SDK" value="false" />
+      <option name="ADD_CONTENT_ROOTS" value="true" />
+      <option name="ADD_SOURCE_ROOTS" value="true" />
+      <option name="SCRIPT_NAME" value="D:/第八章：对抗生成网络架构原理与实战解析/gan/gan.py" />
+      <option name="PARAMETERS" value="" />
+      <option name="SHOW_COMMAND_LINE" value="false" />
+      <option name="EMULATE_TERMINAL" value="false" />
+      <option name="MODULE_MODE" value="false" />
+      <option name="REDIRECT_INPUT" value="false" />
+      <option name="INPUT_FILE" value="" />
+      <method v="2" />
+    </configuration>
+    <configuration name="generate_library_of_models (1)" type="PythonConfigurationType" factoryName="Python" temporary="true">
+      <module name="WheatRustClassification-master" />
+      <option name="INTERPRETER_OPTIONS" value="" />
+      <option name="PARENT_ENVS" value="true" />
+      <envs>
+        <env name="PYTHONUNBUFFERED" value="1" />
+      </envs>
+      <option name="SDK_HOME" value="" />
+      <option name="WORKING_DIRECTORY" value="D:/datasets/WheatRustClassification-master" />
+      <option name="IS_MODULE_SDK" value="false" />
+      <option name="ADD_CONTENT_ROOTS" value="true" />
+      <option name="ADD_SOURCE_ROOTS" value="true" />
+      <option name="SCRIPT_NAME" value="D:/datasets/WheatRustClassification-master/generate_library_of_models.py" />
+      <option name="PARAMETERS" value="" />
+      <option name="SHOW_COMMAND_LINE" value="false" />
+      <option name="EMULATE_TERMINAL" value="false" />
+      <option name="MODULE_MODE" value="false" />
+      <option name="REDIRECT_INPUT" value="false" />
+      <option name="INPUT_FILE" value="" />
+      <method v="2" />
+    </configuration>
+    <configuration name="generate_library_of_models" type="PythonConfigurationType" factoryName="Python" temporary="true">
+      <module name="WheatRustClassification-master" />
+      <option name="INTERPRETER_OPTIONS" value="" />
+      <option name="PARENT_ENVS" value="true" />
+      <envs>
+        <env name="PYTHONUNBUFFERED" value="1" />
+      </envs>
+      <option name="SDK_HOME" value="" />
+      <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" />
+      <option name="IS_MODULE_SDK" value="true" />
+      <option name="ADD_CONTENT_ROOTS" value="true" />
+      <option name="ADD_SOURCE_ROOTS" value="true" />
+      <option name="SCRIPT_NAME" value="$PROJECT_DIR$/generate_library_of_models.py" />
+      <option name="PARAMETERS" value="" />
+      <option name="SHOW_COMMAND_LINE" value="false" />
+      <option name="EMULATE_TERMINAL" value="false" />
+      <option name="MODULE_MODE" value="false" />
+      <option name="REDIRECT_INPUT" value="false" />
+      <option name="INPUT_FILE" value="" />
+      <method v="2" />
+    </configuration>
+    <configuration name="prepare_dataset (1)" type="PythonConfigurationType" factoryName="Python" temporary="true">
+      <module name="WheatRustClassification-master" />
+      <option name="INTERPRETER_OPTIONS" value="" />
+      <option name="PARENT_ENVS" value="true" />
+      <envs>
+        <env name="PYTHONUNBUFFERED" value="1" />
+      </envs>
+      <option name="SDK_HOME" value="" />
+      <option name="WORKING_DIRECTORY" value="D:/datasets/WheatRustClassification-master" />
+      <option name="IS_MODULE_SDK" value="false" />
+      <option name="ADD_CONTENT_ROOTS" value="true" />
+      <option name="ADD_SOURCE_ROOTS" value="true" />
+      <option name="SCRIPT_NAME" value="D:/datasets/WheatRustClassification-master/prepare_dataset.py" />
+      <option name="PARAMETERS" value="" />
+      <option name="SHOW_COMMAND_LINE" value="false" />
+      <option name="EMULATE_TERMINAL" value="false" />
+      <option name="MODULE_MODE" value="false" />
+      <option name="REDIRECT_INPUT" value="false" />
+      <option name="INPUT_FILE" value="" />
+      <method v="2" />
+    </configuration>
+    <configuration name="prepare_dataset" type="PythonConfigurationType" factoryName="Python" temporary="true">
+      <module name="WheatRustClassification-master" />
+      <option name="INTERPRETER_OPTIONS" value="" />
+      <option name="PARENT_ENVS" value="true" />
+      <envs>
+        <env name="PYTHONUNBUFFERED" value="1" />
+      </envs>
+      <option name="SDK_HOME" value="" />
+      <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" />
+      <option name="IS_MODULE_SDK" value="true" />
+      <option name="ADD_CONTENT_ROOTS" value="true" />
+      <option name="ADD_SOURCE_ROOTS" value="true" />
+      <option name="SCRIPT_NAME" value="$PROJECT_DIR$/prepare_dataset.py" />
+      <option name="PARAMETERS" value="" />
+      <option name="SHOW_COMMAND_LINE" value="false" />
+      <option name="EMULATE_TERMINAL" value="false" />
+      <option name="MODULE_MODE" value="false" />
+      <option name="REDIRECT_INPUT" value="false" />
+      <option name="INPUT_FILE" value="" />
+      <method v="2" />
+    </configuration>
+    <recent_temporary>
+      <list>
+        <item itemvalue="Python.gan" />
+        <item itemvalue="Python.generate_library_of_models (1)" />
+        <item itemvalue="Python.generate_library_of_models" />
+        <item itemvalue="Python.prepare_dataset" />
+        <item itemvalue="Python.prepare_dataset (1)" />
+      </list>
+    </recent_temporary>
+  </component>
+  <component name="SpellCheckerSettings" RuntimeDictionaries="0" Folders="0" CustomDictionaries="0" DefaultDictionary="application-level" UseSingleDictionary="true" transferred="true" />
+  <component name="SvnConfiguration">
+    <configuration />
+  </component>
+  <component name="TaskManager">
+    <task active="true" id="Default" summary="Default task">
+      <changelist id="def273b0-b4ed-407e-9205-33a81b54b2c7" name="Default Changelist" comment="" />
+      <created>1606914352714</created>
+      <option name="number" value="Default" />
+      <option name="presentableId" value="Default" />
+      <updated>1606914352714</updated>
+    </task>
+    <servers />
+  </component>
+</project>
\ No newline at end of file
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.ipynb_checkpoints/Untitled-checkpoint(1).ipynb" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.ipynb_checkpoints/Untitled-checkpoint(1).ipynb"
new file mode 100644
index 0000000..7b2cb97
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.ipynb_checkpoints/Untitled-checkpoint(1).ipynb"
@@ -0,0 +1,331 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 44,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# %load dataset.py\n",
+    "from torch.utils.data import Dataset\n",
+    "from torchvision import transforms, utils\n",
+    "import numpy as np\n",
+    "from scipy import ndimage\n",
+    "import torch\n",
+    "from PIL import Image        #图像处理库\n",
+    "\n",
+    "class ICLRDataset(Dataset):\n",
+    "    def __init__(self, imgs, gts, split_type, index, transform, img_mix_enable = True):\n",
+    "        if index is None:\n",
+    "            self.imgs = imgs\n",
+    "            self.gts = gts\n",
+    "        else:\n",
+    "            self.imgs = [imgs[i] for i in index]    #图片集\n",
+    "            self.gts = [gts[i] for i in index]      \n",
+    "                   \n",
+    "        self.split_type = split_type\n",
+    "        self.transform = transform\n",
+    "        self.img_mix_enable = img_mix_enable\n",
+    "    \n",
+    "    def __len__(self):\n",
+    "        return len(self.imgs)\n",
+    "    \n",
+    "    def augment(self, img, y):        \n",
+    "        p = np.random.random(1)     #生成0-1之间的一个1维数组\n",
+    "        if p[0] > 0.5:    #取出数组里的那个数跟0.5作比较\n",
+    "            while True:\n",
+    "                rnd_idx = np.random.randint(0, len(self.imgs))   #前闭后开，其实就是所有图片索引\n",
+    "                if self.gts[rnd_idx] != y:       #如果图片标签不是y就跳出---检查是不是有分错类的图片\n",
+    "                    break\n",
+    "            rnd_crop = self.transform(Image.fromarray(self.imgs[rnd_idx]))  #用于变换的图片集\n",
+    "            d = 0.8\n",
+    "            img = img * d + rnd_crop * (1 - d)   #对图像进行混合和随机裁剪\n",
+    "        return img\n",
+    "\n",
+    "    def __getitem__(self, idx):\n",
+    "        img = self.imgs[idx]\n",
+    "        y = self.gts[idx]\n",
+    "        img = Image.fromarray(img)\n",
+    "        img = self.transform(img)\n",
+    "        if (self.split_type == 'train') & self.img_mix_enable:\n",
+    "            img = self.augment(img, y)  \n",
+    "        return img, y     #增强训练集数据，返回增强后的图片和对应标签\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 54,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# %load utils.py\n",
+    "#用于训练、测试和读取数据集图像的使用函数\n",
+    "import torch\n",
+    "import torch.nn as nn\n",
+    "import torch.optim as optim\n",
+    "from torch.optim import lr_scheduler\n",
+    "import numpy as np\n",
+    "import torchvision\n",
+    "from torchvision import datasets, models, transforms\n",
+    "#import matplotlib.pyplot as plt\n",
+    "import time\n",
+    "import os\n",
+    "import copy\n",
+    "import torch.nn.functional as F\n",
+    "from PIL import Image, ExifTags\n",
+    "\n",
+    "def train_model_snapshot(model, criterion, lr, dataloaders, dataset_sizes, device, num_cycles, num_epochs_per_cycle):\n",
+    "    since = time.time()    #记录训练时间\n",
+    "\n",
+    "    best_model_wts = copy.deepcopy(model.state_dict())   #从预训练的模型中复制权重并初始化模型\n",
+    "    best_acc = 0.0\n",
+    "    best_loss = 1000000.0\n",
+    "    model_w_arr = []\n",
+    "    prob = torch.zeros((dataset_sizes['val'], 3), dtype = torch.float32).to(device)    #预测\n",
+    "    lbl = torch.zeros((dataset_sizes['val'],), dtype = torch.long).to(device)          #标签\n",
+    "    for cycle in range(num_cycles):\n",
+    "        optimizer = optim.SGD(model.parameters(), lr=lr, momentum=0.9)#, weight_decay = 0.0005)\n",
+    "        scheduler = lr_scheduler.CosineAnnealingLR(optimizer, num_epochs_per_cycle*len(dataloaders['train']))    \n",
+    "        #余弦周期调整学习率,修改优化器中的学习率,(优化器，调整间隔数，调整系数)\n",
+    "        for epoch in range(num_epochs_per_cycle):\n",
+    "            #print('Cycle {}: Epoch {}/{}'.format(cycle, epoch, num_epochs_per_cycle - 1))\n",
+    "            #print('-' * 10)\n",
+    "\n",
+    "            # Each epoch has a training and validation phase\n",
+    "            for phase in ['train', 'val']:\n",
+    "                if phase == 'train':\n",
+    "                    model.train()  # Set model to training mode\n",
+    "                else:\n",
+    "                    model.eval()   # Set model to evaluate mode\n",
+    "\n",
+    "                running_loss = 0.0\n",
+    "                running_corrects = 0\n",
+    "                idx = 0\n",
+    "                # Iterate over data.迭代数据\n",
+    "                for inputs, labels in dataloaders[phase]:\n",
+    "                    inputs = inputs.to(device)\n",
+    "                    labels = labels.to(device)\n",
+    "\n",
+    "                    # zero the parameter gradients\n",
+    "                    optimizer.zero_grad()\n",
+    "\n",
+    "                    # forward\n",
+    "                    # track history if only in train\n",
+    "                    with torch.set_grad_enabled(phase == 'train'):\n",
+    "                        outputs = model(inputs)\n",
+    "                        _, preds = torch.max(outputs, 1)\n",
+    "                        if (epoch == num_epochs_per_cycle-1) and (phase == 'val'):\n",
+    "                            prob[idx:idx+inputs.shape[0]] += F.softmax(outputs, dim = 1)\n",
+    "                            lbl[idx:idx+inputs.shape[0]] = labels\n",
+    "                            idx += inputs.shape[0]\n",
+    "                        loss = criterion(outputs, labels)\n",
+    "                        # backward + optimize only if in training phase\n",
+    "                        if phase == 'train':\n",
+    "                            loss.backward()\n",
+    "                            optimizer.step()\n",
+    "                            scheduler.step()\n",
+    "                            #print(optimizer.param_groups[0]['lr'])\n",
+    "                    \n",
+    "                    # statistics\n",
+    "                    running_loss += loss.item() * inputs.size(0)\n",
+    "                    running_corrects += torch.sum(preds == labels.data)\n",
+    "\n",
+    "                epoch_loss = running_loss / dataset_sizes[phase]\n",
+    "                epoch_acc = running_corrects.double() / dataset_sizes[phase]\n",
+    "\n",
+    "                #print('{} Loss: {:.4f} Acc: {:.4f}'.format(\n",
+    "                #    phase, epoch_loss, epoch_acc))\n",
+    "\n",
+    "                # deep copy the model\n",
+    "                if phase == 'val' and epoch_loss < best_loss:\n",
+    "                    best_loss = epoch_loss\n",
+    "                    best_model_wts = copy.deepcopy(model.state_dict())\n",
+    "            #print()\n",
+    "        model_w_arr.append(copy.deepcopy(model.state_dict()))\n",
+    "\n",
+    "    prob /= num_cycles\n",
+    "    ensemble_loss = F.nll_loss(torch.log(prob), lbl)  \n",
+    "    ensemble_loss = ensemble_loss.item()\n",
+    "    time_elapsed = time.time() - since\n",
+    "    #print('Training complete in {:.0f}m {:.0f}s'.format(\n",
+    "    #    time_elapsed // 60, time_elapsed % 60))\n",
+    "    #print('Ensemble Loss : {:4f}, Best val Loss: {:4f}'.format(ensemble_loss, best_loss))\n",
+    "\n",
+    "    # load best model weights加载最佳模型权重\n",
+    "    model_arr =[]\n",
+    "    for weights in model_w_arr:\n",
+    "        model.load_state_dict(weights)   \n",
+    "        model_arr.append(model) \n",
+    "    return model_arr, ensemble_loss, best_loss, prob\n",
+    "\n",
+    "def test(models_arr, loader, device):\n",
+    "    res = np.zeros((610, 3), dtype = np.float32)\n",
+    "    for model in models_arr:\n",
+    "        model.eval()\n",
+    "        res_arr = []\n",
+    "        for inputs, _ in loader:\n",
+    "            inputs = inputs.to(device)\n",
+    "            # forward\n",
+    "            # track history if only in train\n",
+    "            with torch.set_grad_enabled(False):\n",
+    "                outputs = F.softmax(model(inputs), dim = 1)    \n",
+    "                res_arr.append(outputs.detach().cpu().numpy())\n",
+    "        res_arr = np.concatenate(res_arr, axis = 0)\n",
+    "        res += res_arr\n",
+    "    return res / len(models_arr)\n",
+    "\n",
+    "def read_train_data(p):\n",
+    "    imgs = []\n",
+    "    labels = []\n",
+    "    for i, lbl in enumerate(os.listdir(p)):\n",
+    "        for fname in os.listdir(os.path.join(p, lbl)):\n",
+    "            #read image\n",
+    "            img = Image.open(os.path.join(p, lbl, fname))\n",
+    "            #rotate image to original view旋转图像到原始视图\n",
+    "            try:\n",
+    "                exif=dict((ExifTags.TAGS[k], v) for k, v in img._getexif().items() if k in ExifTags.TAGS)\n",
+    "                if exif['Orientation'] == 3:\n",
+    "                    img=img.rotate(180, expand=True)\n",
+    "                elif exif['Orientation'] == 6:\n",
+    "                    img=img.rotate(270, expand=True)\n",
+    "                elif exif['Orientation'] == 8:\n",
+    "                    img=img.rotate(90, expand=True)\n",
+    "            except:\n",
+    "                pass\n",
+    "            #resize all images to the same size将所有图像调整为相同的大小\n",
+    "            img = np.array(img.convert('RGB').resize((512,512), Image.ANTIALIAS))\n",
+    "            imgs.append(img)\n",
+    "            labels.append(i)\n",
+    "    return imgs, labels\n",
+    "\n",
+    "def read_test_data(p):\n",
+    "    imgs = []\n",
+    "    labels = []\n",
+    "    ids = []\n",
+    "    for fname in os.listdir(p):\n",
+    "        #read image\n",
+    "        img = Image.open(os.path.join(p, fname))\n",
+    "        #rotate image to original view\n",
+    "        try:\n",
+    "            if not('DMWVNR' in fname):\n",
+    "                exif=dict((ExifTags.TAGS[k], v) for k, v in img._getexif().items() if k in ExifTags.TAGS)\n",
+    "                if exif['Orientation'] == 3:\n",
+    "                    img=img.rotate(180, expand=True)\n",
+    "                elif exif['Orientation'] == 6:\n",
+    "                    img=img.rotate(270, expand=True)\n",
+    "                elif exif['Orientation'] == 8:\n",
+    "                    img=img.rotate(90, expand=True)\n",
+    "        except:\n",
+    "            pass\n",
+    "        #resize all images to the same size\n",
+    "        img = img.convert('RGB').resize((512,512), Image.ANTIALIAS)\n",
+    "        imgs.append(np.array(img.copy()))\n",
+    "        labels.append(0)\n",
+    "        ids.append(fname.split('.')[0])\n",
+    "        img.close()\n",
+    "    return imgs, labels, ids\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 63,
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "ename": "AssertionError",
+     "evalue": "",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[1;31mAssertionError\u001b[0m                            Traceback (most recent call last)",
+      "\u001b[1;32m<ipython-input-63-05f933fe65f1>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m     10\u001b[0m \u001b[0mparser\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0madd_argument\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'D:\\datasets\\test'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mhelp\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m'path to test data folder'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdefault\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m'test_data'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtype\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mstr\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     11\u001b[0m \u001b[0mparser\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0madd_argument\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'D:\\datasets\\savepath'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mhelp\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m'save path for training and test numpy matrices of images'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdefault\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m'.'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtype\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mstr\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 12\u001b[1;33m \u001b[0margs\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mparser\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mparse_args\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m  \u001b[1;31m#获取参数，调用上面的属性\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     13\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     14\u001b[0m \u001b[1;31m#read training data\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\argparse.py\u001b[0m in \u001b[0;36mparse_args\u001b[1;34m(self, args, namespace)\u001b[0m\n\u001b[0;32m   1747\u001b[0m     \u001b[1;31m# =====================================\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1748\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mparse_args\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0margs\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnamespace\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1749\u001b[1;33m         \u001b[0margs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0margv\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mparse_known_args\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnamespace\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1750\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0margv\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1751\u001b[0m             \u001b[0mmsg\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0m_\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'unrecognized arguments: %s'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\argparse.py\u001b[0m in \u001b[0;36mparse_known_args\u001b[1;34m(self, args, namespace)\u001b[0m\n\u001b[0;32m   1779\u001b[0m         \u001b[1;31m# parse the arguments and exit if there are any errors\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1780\u001b[0m         \u001b[1;32mtry\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1781\u001b[1;33m             \u001b[0mnamespace\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0margs\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_parse_known_args\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnamespace\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1782\u001b[0m             \u001b[1;32mif\u001b[0m \u001b[0mhasattr\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mnamespace\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0m_UNRECOGNIZED_ARGS_ATTR\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1783\u001b[0m                 \u001b[0margs\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mextend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mgetattr\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mnamespace\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0m_UNRECOGNIZED_ARGS_ATTR\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\argparse.py\u001b[0m in \u001b[0;36m_parse_known_args\u001b[1;34m(self, arg_strings, namespace)\u001b[0m\n\u001b[0;32m   2014\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mrequired_actions\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   2015\u001b[0m             self.error(_('the following arguments are required: %s') %\n\u001b[1;32m-> 2016\u001b[1;33m                        ', '.join(required_actions))\n\u001b[0m\u001b[0;32m   2017\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   2018\u001b[0m         \u001b[1;31m# make sure all required groups had one option present\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\argparse.py\u001b[0m in \u001b[0;36merror\u001b[1;34m(self, message)\u001b[0m\n\u001b[0;32m   2497\u001b[0m         \u001b[0mshould\u001b[0m \u001b[0meither\u001b[0m \u001b[0mexit\u001b[0m \u001b[1;32mor\u001b[0m \u001b[1;32mraise\u001b[0m \u001b[0man\u001b[0m \u001b[0mexception\u001b[0m\u001b[1;33m.\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   2498\u001b[0m         \"\"\"\n\u001b[1;32m-> 2499\u001b[1;33m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mprint_usage\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0m_sys\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mstderr\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   2500\u001b[0m         \u001b[0margs\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m{\u001b[0m\u001b[1;34m'prog'\u001b[0m\u001b[1;33m:\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mprog\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'message'\u001b[0m\u001b[1;33m:\u001b[0m \u001b[0mmessage\u001b[0m\u001b[1;33m}\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   2501\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mexit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m2\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0m_\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'%(prog)s: error: %(message)s\\n'\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m%\u001b[0m \u001b[0margs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\argparse.py\u001b[0m in \u001b[0;36mprint_usage\u001b[1;34m(self, file)\u001b[0m\n\u001b[0;32m   2467\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mfile\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   2468\u001b[0m             \u001b[0mfile\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0m_sys\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mstdout\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 2469\u001b[1;33m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_print_message\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mformat_usage\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfile\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   2470\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   2471\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mprint_help\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfile\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\argparse.py\u001b[0m in \u001b[0;36mformat_usage\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m   2433\u001b[0m         formatter.add_usage(self.usage, self._actions,\n\u001b[0;32m   2434\u001b[0m                             self._mutually_exclusive_groups)\n\u001b[1;32m-> 2435\u001b[1;33m         \u001b[1;32mreturn\u001b[0m \u001b[0mformatter\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mformat_help\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   2436\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   2437\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mformat_help\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\argparse.py\u001b[0m in \u001b[0;36mformat_help\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    282\u001b[0m     \u001b[1;31m# =======================\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    283\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mformat_help\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 284\u001b[1;33m         \u001b[0mhelp\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_root_section\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mformat_help\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    285\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mhelp\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    286\u001b[0m             \u001b[0mhelp\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_long_break_matcher\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msub\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'\\n\\n'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mhelp\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\argparse.py\u001b[0m in \u001b[0;36mformat_help\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    213\u001b[0m                 \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mformatter\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_indent\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    214\u001b[0m             \u001b[0mjoin\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mformatter\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_join_parts\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 215\u001b[1;33m             \u001b[0mitem_help\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mjoin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mfunc\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m*\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mfunc\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0margs\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    216\u001b[0m             \u001b[1;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mparent\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    217\u001b[0m                 \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mformatter\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_dedent\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\argparse.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[1;34m(.0)\u001b[0m\n\u001b[0;32m    213\u001b[0m                 \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mformatter\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_indent\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    214\u001b[0m             \u001b[0mjoin\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mformatter\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_join_parts\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 215\u001b[1;33m             \u001b[0mitem_help\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mjoin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mfunc\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m*\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mfunc\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0margs\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    216\u001b[0m             \u001b[1;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mparent\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    217\u001b[0m                 \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mformatter\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_dedent\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\argparse.py\u001b[0m in \u001b[0;36m_format_usage\u001b[1;34m(self, usage, actions, groups, prefix)\u001b[0m\n\u001b[0;32m    338\u001b[0m                 \u001b[0mpos_parts\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0m_re\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfindall\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mpart_regexp\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mpos_usage\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    339\u001b[0m                 \u001b[1;32massert\u001b[0m \u001b[1;34m' '\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mjoin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mopt_parts\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m==\u001b[0m \u001b[0mopt_usage\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 340\u001b[1;33m                 \u001b[1;32massert\u001b[0m \u001b[1;34m' '\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mjoin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mpos_parts\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m==\u001b[0m \u001b[0mpos_usage\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    341\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    342\u001b[0m                 \u001b[1;31m# helper for wrapping lines\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;31mAssertionError\u001b[0m: "
+     ]
+    }
+   ],
+   "source": [
+    "# %load prepare_dataset.py\n",
+    "#读取训练数据和测试数据，从训练数据中删除重复的数据并保存在numpy矩阵中\n",
+    "import numpy as np\n",
+    "import os\n",
+    "import argparse\n",
+    "from utils import read_train_data, read_test_data\n",
+    "\n",
+    "parser = argparse.ArgumentParser(description='Data preperation')\n",
+    "parser.add_argument('D:\\datasets\\train', help='path to training data folder', default='train_data', type=str)\n",
+    "parser.add_argument('D:\\datasets\\test', help='path to test data folder', default='test_data', type=str)\n",
+    "parser.add_argument('D:\\datasets\\savepath', help='save path for training and test numpy matrices of images', default='.', type=str)\n",
+    "args = parser.parse_args()  #获取参数，调用上面的属性\n",
+    "\n",
+    "#read training data\n",
+    "train_imgs, train_gts = read_train_data(args.train_data_path)\n",
+    "\n",
+    "#remove dublicate training imgs\n",
+    "idx_to_rmv = []\n",
+    "for i in range(len(train_imgs)-1):\n",
+    "    for j in range(i+1, len(train_imgs)):\n",
+    "        if np.all(train_imgs[i] == train_imgs[j]):\n",
+    "            idx_to_rmv.append(i)\n",
+    "            if train_gts[i] != train_gts[j]:\n",
+    "                idx_to_rmv.append(j)\n",
+    "\n",
+    "idx = [i for i in range(len(train_imgs)) if not(i in idx_to_rmv)]\n",
+    "print('unique train imgs:',len(idx))\n",
+    "\n",
+    "#save unique training imgs\n",
+    "np.save(os.path.join(args.save_path, 'unique_train_imgs_rot_fixed'), np.array(train_imgs)[idx])\n",
+    "np.save(os.path.join(args.save_path, 'unique_train_gts_rot_fixed'), np.array(train_gts)[idx])\n",
+    "\n",
+    "#read test data\n",
+    "test_imgs, test_gts, ids = read_test_data(args.test_data_path)\n",
+    "\n",
+    "#save test data\n",
+    "np.save(os.path.join(args.save_path, 'test_imgs_rot_fixed'), np.array(test_imgs))\n",
+    "np.save(os.path.join(args.save_path, 'test_gts'), np.array(test_gts))\n",
+    "np.save(os.path.join(args.save_path, 'ids'), np.array(ids))\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.7.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.ipynb_checkpoints/Untitled-checkpoint.ipynb" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.ipynb_checkpoints/Untitled-checkpoint.ipynb"
new file mode 100644
index 0000000..acb1078
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.ipynb_checkpoints/Untitled-checkpoint.ipynb"
@@ -0,0 +1,620 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# %load dataset.py\n",
+    "from torch.utils.data import Dataset\n",
+    "from torchvision import transforms, utils\n",
+    "import numpy as np\n",
+    "from scipy import ndimage\n",
+    "import torch\n",
+    "from PIL import Image        #图像处理库\n",
+    "\n",
+    "class ICLRDataset(Dataset):\n",
+    "    def __init__(self, imgs, gts, split_type, index, transform, img_mix_enable = True):\n",
+    "        if index is None:\n",
+    "            self.imgs = imgs\n",
+    "            self.gts = gts\n",
+    "        else:\n",
+    "            self.imgs = [imgs[i] for i in index]    #图片集\n",
+    "            self.gts = [gts[i] for i in index]      \n",
+    "                   \n",
+    "        self.split_type = split_type\n",
+    "        self.transform = transform\n",
+    "        self.img_mix_enable = img_mix_enable\n",
+    "    \n",
+    "    def __len__(self):\n",
+    "        return len(self.imgs)\n",
+    "    \n",
+    "    def augment(self, img, y):        \n",
+    "        p = np.random.random(1)     #生成0-1之间的一个1维数组\n",
+    "        if p[0] > 0.5:    #取出数组里的那个数跟0.5作比较\n",
+    "            while True:\n",
+    "                rnd_idx = np.random.randint(0, len(self.imgs))   #前闭后开，其实就是所有图片索引\n",
+    "                if self.gts[rnd_idx] != y:       #如果图片标签不是y就跳出---检查是不是有分错类的图片\n",
+    "                    break\n",
+    "            rnd_crop = self.transform(Image.fromarray(self.imgs[rnd_idx]))  #用于变换的图片集\n",
+    "            d = 0.8\n",
+    "            img = img * d + rnd_crop * (1 - d)   #对图像进行混合和随机裁剪\n",
+    "        return img\n",
+    "\n",
+    "    def __getitem__(self, idx):\n",
+    "        img = self.imgs[idx]\n",
+    "        y = self.gts[idx]\n",
+    "        img = Image.fromarray(img)\n",
+    "        img = self.transform(img)\n",
+    "        if (self.split_type == 'train') & self.img_mix_enable:\n",
+    "            img = self.augment(img, y)  \n",
+    "        return img, y     #增强训练集数据，返回增强后的图片和对应标签\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# %load utils.py\n",
+    "#用于训练、测试和读取数据集图像的使用函数\n",
+    "import torch\n",
+    "import torch.nn as nn\n",
+    "import torch.optim as optim\n",
+    "from torch.optim import lr_scheduler\n",
+    "import numpy as np\n",
+    "import torchvision\n",
+    "from torchvision import datasets, models, transforms\n",
+    "#import matplotlib.pyplot as plt\n",
+    "import time\n",
+    "import os\n",
+    "import copy\n",
+    "import torch.nn.functional as F\n",
+    "from PIL import Image, ExifTags\n",
+    "\n",
+    "def train_model_snapshot(model, criterion, lr, dataloaders, dataset_sizes, device, num_cycles, num_epochs_per_cycle):\n",
+    "    since = time.time()    #记录训练时间\n",
+    "\n",
+    "    best_model_wts = copy.deepcopy(model.state_dict())   #从预训练的模型中复制权重并初始化模型\n",
+    "    best_acc = 0.0\n",
+    "    best_loss = 1000000.0\n",
+    "    model_w_arr = []\n",
+    "    prob = torch.zeros((dataset_sizes['val'], 3), dtype = torch.float32).to(device)    #预测\n",
+    "    lbl = torch.zeros((dataset_sizes['val'],), dtype = torch.long).to(device)          #标签\n",
+    "    for cycle in range(num_cycles):\n",
+    "        optimizer = optim.SGD(model.parameters(), lr=lr, momentum=0.9)#, weight_decay = 0.0005)\n",
+    "        scheduler = lr_scheduler.CosineAnnealingLR(optimizer, num_epochs_per_cycle*len(dataloaders['train']))    \n",
+    "        #余弦周期调整学习率,修改优化器中的学习率,(优化器，调整间隔数，调整系数)\n",
+    "        for epoch in range(num_epochs_per_cycle):\n",
+    "            #print('Cycle {}: Epoch {}/{}'.format(cycle, epoch, num_epochs_per_cycle - 1))\n",
+    "            #print('-' * 10)\n",
+    "\n",
+    "            # Each epoch has a training and validation phase\n",
+    "            for phase in ['train', 'val']:\n",
+    "                if phase == 'train':\n",
+    "                    model.train()  # Set model to training mode\n",
+    "                else:\n",
+    "                    model.eval()   # Set model to evaluate mode\n",
+    "\n",
+    "                running_loss = 0.0\n",
+    "                running_corrects = 0\n",
+    "                idx = 0\n",
+    "                # Iterate over data.迭代数据\n",
+    "                for inputs, labels in dataloaders[phase]:\n",
+    "                    inputs = inputs.to(device)\n",
+    "                    labels = labels.to(device)\n",
+    "\n",
+    "                    # zero the parameter gradients\n",
+    "                    optimizer.zero_grad()\n",
+    "\n",
+    "                    # forward\n",
+    "                    # track history if only in train\n",
+    "                    with torch.set_grad_enabled(phase == 'train'):\n",
+    "                        outputs = model(inputs)\n",
+    "                        _, preds = torch.max(outputs, 1)\n",
+    "                        if (epoch == num_epochs_per_cycle-1) and (phase == 'val'):\n",
+    "                            prob[idx:idx+inputs.shape[0]] += F.softmax(outputs, dim = 1)\n",
+    "                            lbl[idx:idx+inputs.shape[0]] = labels\n",
+    "                            idx += inputs.shape[0]\n",
+    "                        loss = criterion(outputs, labels)\n",
+    "                        # backward + optimize only if in training phase\n",
+    "                        if phase == 'train':\n",
+    "                            loss.backward()\n",
+    "                            optimizer.step()\n",
+    "                            scheduler.step()\n",
+    "                            #print(optimizer.param_groups[0]['lr'])\n",
+    "                    \n",
+    "                    # statistics\n",
+    "                    running_loss += loss.item() * inputs.size(0)\n",
+    "                    running_corrects += torch.sum(preds == labels.data)\n",
+    "\n",
+    "                epoch_loss = running_loss / dataset_sizes[phase]\n",
+    "                epoch_acc = running_corrects.double() / dataset_sizes[phase]\n",
+    "\n",
+    "                #print('{} Loss: {:.4f} Acc: {:.4f}'.format(\n",
+    "                #    phase, epoch_loss, epoch_acc))\n",
+    "\n",
+    "                # deep copy the model\n",
+    "                if phase == 'val' and epoch_loss < best_loss:\n",
+    "                    best_loss = epoch_loss\n",
+    "                    best_model_wts = copy.deepcopy(model.state_dict())\n",
+    "            #print()\n",
+    "        model_w_arr.append(copy.deepcopy(model.state_dict()))\n",
+    "\n",
+    "    prob /= num_cycles\n",
+    "    ensemble_loss = F.nll_loss(torch.log(prob), lbl)  \n",
+    "    ensemble_loss = ensemble_loss.item()\n",
+    "    time_elapsed = time.time() - since\n",
+    "    #print('Training complete in {:.0f}m {:.0f}s'.format(\n",
+    "    #    time_elapsed // 60, time_elapsed % 60))\n",
+    "    #print('Ensemble Loss : {:4f}, Best val Loss: {:4f}'.format(ensemble_loss, best_loss))\n",
+    "\n",
+    "    # load best model weights加载最佳模型权重\n",
+    "    model_arr =[]\n",
+    "    for weights in model_w_arr:\n",
+    "        model.load_state_dict(weights)   \n",
+    "        model_arr.append(model) \n",
+    "    return model_arr, ensemble_loss, best_loss, prob\n",
+    "\n",
+    "def test(models_arr, loader, device):\n",
+    "    res = np.zeros((610, 3), dtype = np.float32)\n",
+    "    for model in models_arr:\n",
+    "        model.eval()\n",
+    "        res_arr = []\n",
+    "        for inputs, _ in loader:\n",
+    "            inputs = inputs.to(device)\n",
+    "            # forward\n",
+    "            # track history if only in train\n",
+    "            with torch.set_grad_enabled(False):\n",
+    "                outputs = F.softmax(model(inputs), dim = 1)    \n",
+    "                res_arr.append(outputs.detach().cpu().numpy())\n",
+    "        res_arr = np.concatenate(res_arr, axis = 0)\n",
+    "        res += res_arr\n",
+    "    return res / len(models_arr)\n",
+    "\n",
+    "def read_train_data(p):\n",
+    "    imgs = []\n",
+    "    labels = []\n",
+    "    for i, lbl in enumerate(os.listdir(p)):\n",
+    "        for fname in os.listdir(os.path.join(p, lbl)):\n",
+    "            #read image\n",
+    "            img = Image.open(os.path.join(p, lbl, fname))\n",
+    "            #rotate image to original view旋转图像到原始视图\n",
+    "            try:\n",
+    "                exif=dict((ExifTags.TAGS[k], v) for k, v in img._getexif().items() if k in ExifTags.TAGS)\n",
+    "                if exif['Orientation'] == 3:\n",
+    "                    img=img.rotate(180, expand=True)\n",
+    "                elif exif['Orientation'] == 6:\n",
+    "                    img=img.rotate(270, expand=True)\n",
+    "                elif exif['Orientation'] == 8:\n",
+    "                    img=img.rotate(90, expand=True)\n",
+    "            except:\n",
+    "                pass\n",
+    "            #resize all images to the same size将所有图像调整为相同的大小\n",
+    "            img = np.array(img.convert('RGB').resize((512,512), Image.ANTIALIAS))\n",
+    "            imgs.append(img)\n",
+    "            labels.append(i)\n",
+    "    return imgs, labels\n",
+    "\n",
+    "def read_test_data(p):\n",
+    "    imgs = []\n",
+    "    labels = []\n",
+    "    ids = []\n",
+    "    for fname in os.listdir(p):\n",
+    "        #read image\n",
+    "        img = Image.open(os.path.join(p, fname))\n",
+    "        #rotate image to original view\n",
+    "        try:\n",
+    "            if not('DMWVNR' in fname):\n",
+    "                exif=dict((ExifTags.TAGS[k], v) for k, v in img._getexif().items() if k in ExifTags.TAGS)\n",
+    "                if exif['Orientation'] == 3:\n",
+    "                    img=img.rotate(180, expand=True)\n",
+    "                elif exif['Orientation'] == 6:\n",
+    "                    img=img.rotate(270, expand=True)\n",
+    "                elif exif['Orientation'] == 8:\n",
+    "                    img=img.rotate(90, expand=True)\n",
+    "        except:\n",
+    "            pass\n",
+    "        #resize all images to the same size\n",
+    "        img = img.convert('RGB').resize((512,512), Image.ANTIALIAS)\n",
+    "        imgs.append(np.array(img.copy()))\n",
+    "        labels.append(0)\n",
+    "        ids.append(fname.split('.')[0])\n",
+    "        img.close()\n",
+    "    return imgs, labels, ids\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "unique train imgs: 732\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "usage: ipykernel_launcher.py [-h] [--train_data_path TRAIN_DATA_PATH]\n",
+      "                             [--test_data_path TEST_DATA_PATH]\n",
+      "                             [--save_path SAVE_PATH]\n",
+      "ipykernel_launcher.py: error: unrecognized arguments: -f C:\\Users\\Administrator\\AppData\\Roaming\\jupyter\\runtime\\kernel-60e74fca-82ff-42d7-afc4-1d27b752461b.json\n"
+     ]
+    },
+    {
+     "ename": "SystemExit",
+     "evalue": "2",
+     "output_type": "error",
+     "traceback": [
+      "An exception has occurred, use %tb to see the full traceback.\n",
+      "\u001b[1;31mSystemExit\u001b[0m\u001b[1;31m:\u001b[0m 2\n"
+     ]
+    }
+   ],
+   "source": [
+    "# %load prepare_dataset.py\n",
+    "#读取训练数据和测试数据，从训练数据中删除重复的数据并保存在numpy矩阵中\n",
+    "%run prepare_dataset.py\n",
+    "import numpy as np\n",
+    "import os\n",
+    "import argparse\n",
+    "from utils import read_train_data, read_test_data\n",
+    "\n",
+    "parser = argparse.ArgumentParser(description='Data preperation')\n",
+    "parser.add_argument('--train_data_path', help='path', default='D:/datasets/train', type=str)\n",
+    "parser.add_argument('--test_data_path', help='path', default='D:/datasets/test', type=str)\n",
+    "parser.add_argument('--save_path', help='save', default='D:/datasets/savepath', type=str)\n",
+    "args = parser.parse_args()  #获取参数，调用上面的属性\n",
+    "\n",
+    "#read training data\n",
+    "train_imgs, train_gts = read_train_data(args.train_data_path)\n",
+    "\n",
+    "#remove dublicate training imgs\n",
+    "idx_to_rmv = []\n",
+    "for i in range(len(train_imgs)-1):\n",
+    "    for j in range(i+1, len(train_imgs)):\n",
+    "        if np.all(train_imgs[i] == train_imgs[j]):\n",
+    "            idx_to_rmv.append(i)\n",
+    "            if train_gts[i] != train_gts[j]:\n",
+    "                idx_to_rmv.append(j)\n",
+    "\n",
+    "idx = [i for i in range(len(train_imgs)) if not(i in idx_to_rmv)]\n",
+    "print('unique train imgs:',len(idx))\n",
+    "\n",
+    "#save unique training imgs\n",
+    "np.save(os.path.join(args.save_path, 'unique_train_imgs_rot_fixed'), np.array(train_imgs)[idx])\n",
+    "np.save(os.path.join(args.save_path, 'unique_train_gts_rot_fixed'), np.array(train_gts)[idx])\n",
+    "\n",
+    "#read test data\n",
+    "test_imgs, test_gts, ids = read_test_data(args.test_data_path)\n",
+    "\n",
+    "#save test data\n",
+    "np.save(os.path.join(args.save_path, 'test_imgs_rot_fixed'), np.array(test_imgs))\n",
+    "np.save(os.path.join(args.save_path, 'test_gts'), np.array(test_gts))\n",
+    "np.save(os.path.join(args.save_path, 'ids'), np.array(ids))\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "unique train imgs: 732\n"
+     ]
+    }
+   ],
+   "source": [
+    "%run prepare_dataset.py"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "usage: ipykernel_launcher.py [-h] [--data_path DATA_PATH]\n",
+      "                             [--library_size LIBRARY_SIZE]\n",
+      "                             [--library_path LIBRARY_PATH]\n",
+      "ipykernel_launcher.py: error: unrecognized arguments: -f C:\\Users\\Administrator\\AppData\\Roaming\\jupyter\\runtime\\kernel-9ba0d6cf-6ce4-4517-8c53-84c6b0e19712.json\n"
+     ]
+    },
+    {
+     "ename": "SystemExit",
+     "evalue": "2",
+     "output_type": "error",
+     "traceback": [
+      "An exception has occurred, use %tb to see the full traceback.\n",
+      "\u001b[1;31mSystemExit\u001b[0m\u001b[1;31m:\u001b[0m 2\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\IPython\\core\\interactiveshell.py:3333: UserWarning: To exit: use 'exit', 'quit', or Ctrl-D.\n",
+      "  warn(\"To exit: use 'exit', 'quit', or Ctrl-D.\", stacklevel=1)\n"
+     ]
+    }
+   ],
+   "source": [
+    "# %load generate_library_of_models.py\n",
+    "import torch\n",
+    "import torch.nn as nn\n",
+    "import torch.optim as optim\n",
+    "from torch.optim import lr_scheduler\n",
+    "import numpy as np\n",
+    "import torchvision\n",
+    "from torchvision import datasets, models, transforms\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import time\n",
+    "import os\n",
+    "import argparse\n",
+    "import copy\n",
+    "from sklearn.model_selection import StratifiedKFold\n",
+    "import datetime\n",
+    "from PIL import Image\n",
+    "import torch.nn.functional as F\n",
+    "\n",
+    "from dataset import ICLRDataset\n",
+    "from utils import train_model_snapshot, test\n",
+    "from sklearn.metrics import confusion_matrix\n",
+    "from hyperopt import hp, tpe, fmin, Trials\n",
+    "from collections import OrderedDict\n",
+    "\n",
+    "\n",
+    "def score(params):\n",
+    "    global test_prob, val_prob, trails_sc_arr,idx    # 全局变量                                                                                                                              \n",
+    "    print(params)\n",
+    "    k = 5 \n",
+    "    sss = StratifiedKFold(n_splits=k, shuffle = True, random_state=seed_arr[idx])    #提供训练/测试索引来分割训练/测试集中的数据\n",
+    "    #（折叠的数量，至少是2；在分组前是否对每个类的样本进行洗牌；当shuffle为真时，random_state将影响索引的排序）\n",
+    "    \n",
+    "    #define trail data augmentations  训练集数据增强\n",
+    "    data_transforms = {\n",
+    "        'train': transforms.Compose([\n",
+    "            transforms.ColorJitter(contrast = params['contrast'], hue = params['hue'], brightness = params['brightness']),\n",
+    "            transforms.RandomAffine(degrees = params['degrees']),\n",
+    "            transforms.RandomResizedCrop(224),\n",
+    "            transforms.RandomHorizontalFlip(p = 0.5 if params['h_flip'] else 0.0), #以概率P水平翻转图像\n",
+    "            transforms.RandomVerticalFlip(p = 0.5 if params['v_flip'] else 0.0),#以概率P垂直翻转图像\n",
+    "            transforms.ToTensor(),\n",
+    "            transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n",
+    "        ]),\n",
+    "        'val': transforms.Compose([\n",
+    "            transforms.Resize((params['val_img_size'], params['val_img_size'])),\n",
+    "            transforms.CenterCrop(224),\n",
+    "            transforms.ToTensor(),\n",
+    "            transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n",
+    "        ]),\n",
+    "    }\n",
+    "\n",
+    "    trail_test_prob = np.zeros((test_imgs.shape[0], 3), dtype = np.float32)\n",
+    "    trail_val_prob = torch.zeros((train_imgs.shape[0], 3), dtype = torch.float32).to(device)\n",
+    "    \n",
+    "    sc_arr = []\n",
+    "    models_arr = []\n",
+    "    fold = 0\n",
+    "    #train a model for each split\n",
+    "    for train_index, val_index in sss.split(train_imgs, train_gts):\n",
+    "        #define dataset and loader for training and validation\n",
+    "        image_datasets = {'train': ICLRDataset(train_imgs, train_gts, 'train', train_index, data_transforms['train'], params['img_mix_enable']),\n",
+    "\t                      'val': ICLRDataset(train_imgs, train_gts, 'val', val_index, data_transforms['val'])}\n",
+    "\n",
+    "        dataloaders = {'train': torch.utils.data.DataLoader(image_datasets['train'], batch_size=16, shuffle=True, num_workers=2),\n",
+    "                       'val': torch.utils.data.DataLoader(image_datasets['val'], batch_size=16, shuffle=False, num_workers=2)}\n",
+    "\n",
+    "        #create model instance\n",
+    "        model_ft = params['arch'](pretrained=True)\n",
+    "        try:\n",
+    "            num_ftrs = model_ft.fc.in_features\n",
+    "            model_ft.fc = nn.Linear(num_ftrs, 3)\n",
+    "        except:\n",
+    "            num_ftrs = model_ft.classifier.in_features\n",
+    "            model_ft.classifier = nn.Linear(num_ftrs, 3)\n",
+    "        model_ft = model_ft.to(device)\n",
+    "\n",
+    "        criterion = nn.CrossEntropyLoss()\n",
+    "\n",
+    "        dataset_sizes = {x:len(image_datasets[x]) for x in ['train', 'val']}\n",
+    "        \n",
+    "        model_ft_arr, ensemble_loss, _, fold_val_prob = train_model_snapshot(model_ft, criterion, params['lr'], dataloaders, dataset_sizes, device,\n",
+    "                               num_cycles=params['num_cycles'], num_epochs_per_cycle=params['num_epochs_per_cycle'])\n",
+    "        models_arr.extend(model_ft_arr)\n",
+    "        fold += 1\n",
+    "        sc_arr.append(ensemble_loss)\n",
+    "        trail_val_prob[val_index] = fold_val_prob\n",
+    "    \n",
+    "    #predict on test data using average of kfold models\n",
+    "    image_datasets['test'] = ICLRDataset(test_imgs, test_gts, 'test', None, data_transforms['val'])\n",
+    "    test_loader = torch.utils.data.DataLoader(image_datasets['test'], batch_size=4,shuffle=False, num_workers=16)\n",
+    "    trail_test_prob = test(models_arr, test_loader, device)\n",
+    "\n",
+    "    print('mean val loss:', np.mean(sc_arr))\n",
+    "\n",
+    "    test_prob.append(trail_test_prob)\n",
+    "    val_prob.append(trail_val_prob)\n",
+    "\n",
+    "    #save validation and test results for further processing \n",
+    "    np.save(os.path.join(args.library_path, 'val_prob_trail_%d'%(idx)), trail_val_prob.detach().cpu().numpy())\n",
+    "    np.save(os.path.join(args.library_path, 'test_prob_trail_%d'%(idx)), trail_test_prob)\n",
+    "    idx += 1\n",
+    "    \n",
+    "    trails_sc_arr.append(np.mean(sc_arr))\n",
+    "\n",
+    "    torch.cuda.empty_cache()\n",
+    "    del models_arr\n",
+    "\n",
+    "    return np.mean(sc_arr)\n",
+    "\n",
+    "parser = argparse.ArgumentParser(description='Data preperation')\n",
+    "parser.add_argument('--data_path', help='path to training and test numpy matrices of images', default='.', type=str)\n",
+    "parser.add_argument('--library_size', help='number of models to be trained in the library of models', default=50, type=int)\n",
+    "parser.add_argument('--library_path', help='save path for validation and test predictions of the library of models', default='trails', type=str)\n",
+    "args = parser.parse_args()\n",
+    "\n",
+    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
+    "\n",
+    "torch.manual_seed(0)\n",
+    "np.random.seed(0)\n",
+    "\n",
+    "torch.backends.cudnn.deterministic = True\n",
+    "torch.backends.cudnn.benchmark = False\n",
+    "\n",
+    "#read train data\n",
+    "train_imgs = np.load(os.path.join(args.data_path, 'D:/datasets/savepath/unique_train_imgs_rot_fixed.npy'))\n",
+    "train_gts = np.load(os.path.join(args.data_path, 'unique_train_gts_rot_fixed.npy'))\n",
+    "\n",
+    "#read test data\n",
+    "test_imgs = np.load(os.path.join(args.data_path, 'test_imgs_rot_fixed.npy'))\n",
+    "test_gts = np.load(os.path.join(args.data_path, 'test_gts.npy'))\n",
+    "ids = np.load(os.path.join(args.data_path, 'ids.npy')).tolist()\n",
+    "\n",
+    "test_prob = []\n",
+    "val_prob = []\n",
+    "trails_sc_arr = []\n",
+    "\n",
+    "n_trails = args.library_size\n",
+    "seed_arr = np.random.randint(low=0, high=1000000, size=n_trails)\n",
+    "\n",
+    "#create search space for hyperparameter optimization\n",
+    "space = OrderedDict([('lr', hp.choice('lr', [i*0.001 for i in range(1,4)])),\n",
+    "                     ('num_cycles', hp.choice('num_cycles', range(3, 6))),\n",
+    "                     ('num_epochs_per_cycle', hp.choice('num_epochs_per_cycle', range(3, 6))),\n",
+    "                     ('arch', hp.choice('arch', [models.densenet201, models.densenet121, models.densenet169,\n",
+    "                                                 models.wide_resnet50_2, models.resnet152, \n",
+    "                                                 models.resnet101, models.resnet50, models.resnet34, models.resnet18])),\n",
+    "                     ('img_mix_enable', hp.choice('img_mix_enable', [True, False])),\n",
+    "                     ('v_flip', hp.choice('v_flip', [True, False])),\n",
+    "                     ('h_flip', hp.choice('h_flip', [True, False])),\n",
+    "                     ('degrees', hp.choice('degrees', range(1, 90))),\n",
+    "                     ('contrast', hp.uniform('contrast', 0.0, 0.3)),\n",
+    "                     ('hue', hp.uniform('hue', 0.0, 0.3)),\n",
+    "                     ('brightness', hp.uniform('brightness', 0.0, 0.3)),\n",
+    "                     ('val_img_size', hp.choice('val_img_size', range(224, 512, 24))),\n",
+    "                     ])\n",
+    "\n",
+    "trials = Trials()\n",
+    "\n",
+    "idx = 0\n",
+    "if not os.path.exists(args.library_path):\n",
+    "    os.mkdir(args.library_path)\n",
+    "\n",
+    "#use tpe algorithm in hyperopt to generate a library of differnet models 利用hyperopt中的tpe算法生成不同模型库\n",
+    "best = fmin(fn=score,space=space,algo=tpe.suggest,max_evals=n_trails,trials=trials)\n",
+    "#fmin是对不用的算法集及其超参数进行迭代，使目标函数最小化的优化函数\n",
+    "#（最小化的目标函数；定义的搜索空间；搜索算法-为超参数空间的顺序搜索提供逻辑；最大评估数；trials对象）\n",
+    "print(best)\n",
+    "\n",
+    "np.save(os.path.join(args.library_path, 'scores.npy'), np.array(trails_sc_arr))\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{'arch': <function densenet121 at 0x000002654D50DD08>, 'brightness': 0.20568033410096712, 'contrast': 0.047925566574344027, 'degrees': 87, 'h_flip': True, 'hue': 0.2804254868966057, 'img_mix_enable': True, 'lr': 0.001, 'num_cycles': 5, 'num_epochs_per_cycle': 5, 'v_flip': True, 'val_img_size': 248}\n",
+      "  0%|                                                                           | 0/50 [00:00<?, ?trial/s, best loss=?]"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "job exception: module '__main__' has no attribute '__spec__'\n",
+      "\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    },
+    {
+     "ename": "AttributeError",
+     "evalue": "module '__main__' has no attribute '__spec__'",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
+      "\u001b[1;32m~\\Desktop\\WheatRustClassification-master\\generate_library_of_models.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m    160\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    161\u001b[0m \u001b[1;31m#use tpe algorithm in hyperopt to generate a library of differnet models\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 162\u001b[1;33m \u001b[0mbest\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mfmin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfn\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mscore\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mspace\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mspace\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0malgo\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mtpe\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msuggest\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mmax_evals\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mn_trails\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    163\u001b[0m \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mbest\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    164\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\fmin.py\u001b[0m in \u001b[0;36mfmin\u001b[1;34m(fn, space, algo, max_evals, timeout, loss_threshold, trials, rstate, allow_trials_fmin, pass_expr_memo_ctrl, catch_eval_exceptions, verbose, return_argmin, points_to_evaluate, max_queue_len, show_progressbar)\u001b[0m\n\u001b[0;32m    480\u001b[0m             \u001b[0mcatch_eval_exceptions\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcatch_eval_exceptions\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    481\u001b[0m             \u001b[0mreturn_argmin\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mreturn_argmin\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 482\u001b[1;33m             \u001b[0mshow_progressbar\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mshow_progressbar\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    483\u001b[0m         )\n\u001b[0;32m    484\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\base.py\u001b[0m in \u001b[0;36mfmin\u001b[1;34m(self, fn, space, algo, max_evals, timeout, loss_threshold, max_queue_len, rstate, verbose, pass_expr_memo_ctrl, catch_eval_exceptions, return_argmin, show_progressbar)\u001b[0m\n\u001b[0;32m    684\u001b[0m             \u001b[0mcatch_eval_exceptions\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcatch_eval_exceptions\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    685\u001b[0m             \u001b[0mreturn_argmin\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mreturn_argmin\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 686\u001b[1;33m             \u001b[0mshow_progressbar\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mshow_progressbar\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    687\u001b[0m         )\n\u001b[0;32m    688\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\fmin.py\u001b[0m in \u001b[0;36mfmin\u001b[1;34m(fn, space, algo, max_evals, timeout, loss_threshold, trials, rstate, allow_trials_fmin, pass_expr_memo_ctrl, catch_eval_exceptions, verbose, return_argmin, points_to_evaluate, max_queue_len, show_progressbar)\u001b[0m\n\u001b[0;32m    507\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    508\u001b[0m     \u001b[1;31m# next line is where the fmin is actually executed\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 509\u001b[1;33m     \u001b[0mrval\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mexhaust\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    510\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    511\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mreturn_argmin\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\fmin.py\u001b[0m in \u001b[0;36mexhaust\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    328\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mexhaust\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    329\u001b[0m         \u001b[0mn_done\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 330\u001b[1;33m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmax_evals\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mn_done\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mblock_until_done\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0masynchronous\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    331\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mrefresh\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    332\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\fmin.py\u001b[0m in \u001b[0;36mrun\u001b[1;34m(self, N, block_until_done)\u001b[0m\n\u001b[0;32m    284\u001b[0m                 \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    285\u001b[0m                     \u001b[1;31m# -- loop over trials and do the jobs directly\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 286\u001b[1;33m                     \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mserial_evaluate\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    287\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    288\u001b[0m                 \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mrefresh\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\fmin.py\u001b[0m in \u001b[0;36mserial_evaluate\u001b[1;34m(self, N)\u001b[0m\n\u001b[0;32m    163\u001b[0m                 \u001b[0mctrl\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mbase\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mCtrl\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcurrent_trial\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mtrial\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    164\u001b[0m                 \u001b[1;32mtry\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 165\u001b[1;33m                     \u001b[0mresult\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdomain\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mevaluate\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mspec\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mctrl\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    166\u001b[0m                 \u001b[1;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[1;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    167\u001b[0m                     \u001b[0mlogger\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0merror\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"job exception: %s\"\u001b[0m \u001b[1;33m%\u001b[0m \u001b[0mstr\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0me\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\base.py\u001b[0m in \u001b[0;36mevaluate\u001b[1;34m(self, config, ctrl, attach_attachments)\u001b[0m\n\u001b[0;32m    892\u001b[0m                 \u001b[0mprint_node_on_error\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mrec_eval_print_node_on_error\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    893\u001b[0m             )\n\u001b[1;32m--> 894\u001b[1;33m             \u001b[0mrval\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfn\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mpyll_rval\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    895\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    896\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mrval\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mfloat\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mint\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnumber\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\Desktop\\WheatRustClassification-master\\generate_library_of_models.py\u001b[0m in \u001b[0;36mscore\u001b[1;34m(params)\u001b[0m\n\u001b[0;32m     78\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     79\u001b[0m         model_ft_arr, ensemble_loss, _, fold_val_prob = train_model_snapshot(model_ft, criterion, params['lr'], dataloaders, dataset_sizes, device,\n\u001b[1;32m---> 80\u001b[1;33m                                num_cycles=params['num_cycles'], num_epochs_per_cycle=params['num_epochs_per_cycle'])\n\u001b[0m\u001b[0;32m     81\u001b[0m         \u001b[0mmodels_arr\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mextend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmodel_ft_arr\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     82\u001b[0m         \u001b[0mfold\u001b[0m \u001b[1;33m+=\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\Desktop\\WheatRustClassification-master\\utils.py\u001b[0m in \u001b[0;36mtrain_model_snapshot\u001b[1;34m(model, criterion, lr, dataloaders, dataset_sizes, device, num_cycles, num_epochs_per_cycle)\u001b[0m\n\u001b[0;32m     40\u001b[0m                 \u001b[0midx\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m0\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     41\u001b[0m                 \u001b[1;31m# Iterate over data.\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 42\u001b[1;33m                 \u001b[1;32mfor\u001b[0m \u001b[0minputs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mlabels\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mdataloaders\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mphase\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     43\u001b[0m                     \u001b[0minputs\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0minputs\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mto\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdevice\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     44\u001b[0m                     \u001b[0mlabels\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlabels\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mto\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdevice\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\site-packages\\torch\\utils\\data\\dataloader.py\u001b[0m in \u001b[0;36m__iter__\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    277\u001b[0m             \u001b[1;32mreturn\u001b[0m \u001b[0m_SingleProcessDataLoaderIter\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    278\u001b[0m         \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 279\u001b[1;33m             \u001b[1;32mreturn\u001b[0m \u001b[0m_MultiProcessingDataLoaderIter\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    280\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    281\u001b[0m     \u001b[1;33m@\u001b[0m\u001b[0mproperty\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\site-packages\\torch\\utils\\data\\dataloader.py\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self, loader)\u001b[0m\n\u001b[0;32m    717\u001b[0m             \u001b[1;31m#     before it starts, and __del__ tries to join but will get:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    718\u001b[0m             \u001b[1;31m#     AssertionError: can only join a started process.\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 719\u001b[1;33m             \u001b[0mw\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mstart\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    720\u001b[0m             \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_index_queues\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mindex_queue\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    721\u001b[0m             \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_workers\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mw\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\multiprocessing\\process.py\u001b[0m in \u001b[0;36mstart\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    110\u001b[0m                \u001b[1;34m'daemonic processes are not allowed to have children'\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    111\u001b[0m         \u001b[0m_cleanup\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 112\u001b[1;33m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_popen\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_Popen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    113\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_sentinel\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_popen\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msentinel\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    114\u001b[0m         \u001b[1;31m# Avoid a refcycle if the target function holds an indirect\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\multiprocessing\\context.py\u001b[0m in \u001b[0;36m_Popen\u001b[1;34m(process_obj)\u001b[0m\n\u001b[0;32m    221\u001b[0m     \u001b[1;33m@\u001b[0m\u001b[0mstaticmethod\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    222\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m_Popen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprocess_obj\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 223\u001b[1;33m         \u001b[1;32mreturn\u001b[0m \u001b[0m_default_context\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mget_context\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mProcess\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_Popen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprocess_obj\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    224\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    225\u001b[0m \u001b[1;32mclass\u001b[0m \u001b[0mDefaultContext\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mBaseContext\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\multiprocessing\\context.py\u001b[0m in \u001b[0;36m_Popen\u001b[1;34m(process_obj)\u001b[0m\n\u001b[0;32m    320\u001b[0m         \u001b[1;32mdef\u001b[0m \u001b[0m_Popen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprocess_obj\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    321\u001b[0m             \u001b[1;32mfrom\u001b[0m \u001b[1;33m.\u001b[0m\u001b[0mpopen_spawn_win32\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mPopen\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 322\u001b[1;33m             \u001b[1;32mreturn\u001b[0m \u001b[0mPopen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprocess_obj\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    323\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    324\u001b[0m     \u001b[1;32mclass\u001b[0m \u001b[0mSpawnContext\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mBaseContext\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\multiprocessing\\popen_spawn_win32.py\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self, process_obj)\u001b[0m\n\u001b[0;32m     44\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     45\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mprocess_obj\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 46\u001b[1;33m         \u001b[0mprep_data\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mspawn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mget_preparation_data\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprocess_obj\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_name\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     47\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     48\u001b[0m         \u001b[1;31m# read end of pipe will be \"stolen\" by the child process\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\multiprocessing\\spawn.py\u001b[0m in \u001b[0;36mget_preparation_data\u001b[1;34m(name)\u001b[0m\n\u001b[0;32m    170\u001b[0m     \u001b[1;31m# or through direct execution (or to leave it alone entirely)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    171\u001b[0m     \u001b[0mmain_module\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0msys\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmodules\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'__main__'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 172\u001b[1;33m     \u001b[0mmain_mod_name\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mgetattr\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmain_module\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m__spec__\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m\"name\"\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    173\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mmain_mod_name\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    174\u001b[0m         \u001b[0md\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'init_main_from_name'\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mmain_mod_name\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;31mAttributeError\u001b[0m: module '__main__' has no attribute '__spec__'"
+     ]
+    }
+   ],
+   "source": [
+    "%run generate_library_of_models.py"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.7.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.pytest_cache/CACHEDIR.TAG" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.pytest_cache/CACHEDIR.TAG"
new file mode 100644
index 0000000..381f03a
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.pytest_cache/CACHEDIR.TAG"
@@ -0,0 +1,4 @@
+Signature: 8a477f597d28d172789f06886806bc55
+# This file is a cache directory tag created by pytest.
+# For information about cache directory tags, see:
+#	http://www.bford.info/cachedir/spec.html
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.pytest_cache/README.md" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.pytest_cache/README.md"
new file mode 100644
index 0000000..b10f023
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/.pytest_cache/README.md"
@@ -0,0 +1,8 @@
+# pytest cache directory #
+
+This directory contains data from the pytest's cache plugin,
+which provides the `--lf` and `--ff` options, as well as the `cache` fixture.
+
+**Do not** commit this to version control.
+
+See [the docs](https://docs.pytest.org/en/latest/cache.html) for more information.
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/README.md" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/README.md"
new file mode 100644
index 0000000..def227c
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/README.md"
@@ -0,0 +1,127 @@
+# Wheat Rust Classification from Ensemble Selection of CNNs
+
+Fourth place solution for CGIAR Computer Vision for Crop Disease competition organized by CV4A workshop at ICLR 2020. The main objective of the competition is to classify a given image of wheat whether it is healthy, has a stem rust or has a leaf rust.
+
+## Summary of Approach
+
+Create an ensemble from a library of diverse models with different architectures and augmentations. All models are initially pre-trained on imagenet and fine-tuned on the dataset. The models and augmentations are chosen automatically using hyperparameter optimization.
+
+### Model Architectures
+
+The following architecturs are included in the library of models:
+
+* ResNet [1]
+* ResNext [2]
+* WideResNet [3]
+* DenseNet [4]
+
+### Data Augmentations
+
+The following augmentations are included in the search space of hyperparameter optimization to choose from:
+
+* Rotation
+* Random cropping and resizing
+* Horizontal flipping
+* Vertical flipping
+* Brightness augmentation
+* Hue augmentation
+* Contrast augmentation
+* Mixup augmentation [5]
+
+### Common Configuration
+
+The following configurations is applied on all trails in hyperparameter optimization process:
+
+* Stochastic Gradient Descent (SGD) optimizer
+* Snapshot ensemble [6]
+* 5-Fold training
+
+## Getting Started
+
+### Prerequisites
+
+Firstly, you need to have 
+
+* Ubuntu 18.04 
+* Python3
+* At least 11 GB GPU RAM
+
+Secondly, you need to install the challenge data and sample submission file by the following the instructions [here](https://zindi.africa/competitions/iclr-workshop-challenge-1-cgiar-computer-vision-for-crop-disease/data).
+
+Thirdly, you need to install the dependencies by running:
+
+```
+pip3 install -r requirements.txt
+```
+
+### Project files
+
+* prepare_dataset.py: reads training and test data, removes duplicates from training data and saves them in numpy matrices. It has the following arguments:
+
+    --train_data_path: path to training data folder
+
+    --test_data_path: path to test data folder
+
+    --save_path: save path for training and test numpy matrices of images
+
+* generate_library_of_models.py: generates a library of models with different architectures and augmentations through hyperparameter optimization search. It has the following arguments: 
+
+    --data_path: path to training and test numpy matrices of images
+
+    --library_size: number of models to be trained in the library of models
+
+    --library_path: save path for validation and test predictions of the library of models
+
+* ensemble_selection.py: applies Ensemble Selection [7] algorithm on the generated library of models to find the best ensemble with the lowest validation error and use it to create the final submission. It has the following arguments:
+
+    --train_data_path: path to training data folder
+
+    --data_path: path to training and test numpy matrices of images
+
+    --sample_sub_file_path: path to sample submission file
+
+    --library_size: number of models to be trained in the library of models
+
+    --library_path: save path for validation and test predictions of the library of models
+
+    --final_sub_file_save_path: save path for final submission file
+
+* dataset.py: has the dataset class for training and test data.
+
+* utils.py: utility functions for training, testing and reading dataset images. 
+
+## Running
+
+### 1- Prepare dataset
+
+```
+python3 prepare_dataset.py
+```
+
+### 2- Generate the library of models
+
+```
+python3 generate_library_of_models.py
+```
+
+### 3- Create ensemble and generate submission file
+
+```
+python3 ensemble_selection.py
+```
+
+## References
+[1] He, Kaiming, et al. "Deep residual learning for image recognition." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.
+
+[2] Xie, Saining, et al. "Aggregated residual transformations for deep neural networks." Proceedings of the IEEE conference on computer vision and pattern recognition. 2017.
+
+[3] Zagoruyko, Sergey, and Nikos Komodakis. "Wide residual networks." arXiv preprint arXiv:1605.07146 (2016). 
+
+[4] Huang, Gao, et al. "Densely connected convolutional networks." Proceedings of the IEEE conference on computer vision and pattern recognition. 2017.
+
+[5] Zhang, Hongyi, et al. "mixup: Beyond empirical risk minimization." arXiv preprint arXiv:1710.09412 (2017).
+
+[6] Huang, Gao, et al. "Snapshot ensembles: Train 1, get m for free." arXiv preprint arXiv:1704.00109 (2017).
+
+[7] Caruana, Rich, et al. "Ensemble selection from libraries of models." Proceedings of the twenty-first international conference on Machine learning. 2004.
+
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/Untitled.ipynb" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/Untitled.ipynb"
new file mode 100644
index 0000000..c8c0a43
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/Untitled.ipynb"
@@ -0,0 +1,620 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# %load dataset.py\n",
+    "from torch.utils.data import Dataset\n",
+    "from torchvision import transforms, utils\n",
+    "import numpy as np\n",
+    "from scipy import ndimage\n",
+    "import torch\n",
+    "from PIL import Image        #图像处理库\n",
+    "\n",
+    "class ICLRDataset(Dataset):\n",
+    "    def __init__(self, imgs, gts, split_type, index, transform, img_mix_enable = True):\n",
+    "        if index is None:\n",
+    "            self.imgs = imgs\n",
+    "            self.gts = gts\n",
+    "        else:\n",
+    "            self.imgs = [imgs[i] for i in index]    #图片集\n",
+    "            self.gts = [gts[i] for i in index]      \n",
+    "                   \n",
+    "        self.split_type = split_type\n",
+    "        self.transform = transform\n",
+    "        self.img_mix_enable = img_mix_enable\n",
+    "    \n",
+    "    def __len__(self):\n",
+    "        return len(self.imgs)\n",
+    "    \n",
+    "    def augment(self, img, y):        \n",
+    "        p = np.random.random(1)     #生成0-1之间的一个1维数组\n",
+    "        if p[0] > 0.5:    #取出数组里的那个数跟0.5作比较\n",
+    "            while True:\n",
+    "                rnd_idx = np.random.randint(0, len(self.imgs))   #前闭后开，其实就是所有图片索引\n",
+    "                if self.gts[rnd_idx] != y:       #如果图片标签不是y就跳出---检查是不是有分错类的图片\n",
+    "                    break\n",
+    "            rnd_crop = self.transform(Image.fromarray(self.imgs[rnd_idx]))  #用于变换的图片集\n",
+    "            d = 0.8\n",
+    "            img = img * d + rnd_crop * (1 - d)   #对图像进行混合和随机裁剪\n",
+    "        return img\n",
+    "\n",
+    "    def __getitem__(self, idx):\n",
+    "        img = self.imgs[idx]\n",
+    "        y = self.gts[idx]\n",
+    "        img = Image.fromarray(img)\n",
+    "        img = self.transform(img)\n",
+    "        if (self.split_type == 'train') & self.img_mix_enable:\n",
+    "            img = self.augment(img, y)  \n",
+    "        return img, y     #增强训练集数据，返回增强后的图片和对应标签\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# %load utils.py\n",
+    "#用于训练、测试和读取数据集图像的使用函数\n",
+    "import torch\n",
+    "import torch.nn as nn\n",
+    "import torch.optim as optim\n",
+    "from torch.optim import lr_scheduler\n",
+    "import numpy as np\n",
+    "import torchvision\n",
+    "from torchvision import datasets, models, transforms\n",
+    "#import matplotlib.pyplot as plt\n",
+    "import time\n",
+    "import os\n",
+    "import copy\n",
+    "import torch.nn.functional as F\n",
+    "from PIL import Image, ExifTags\n",
+    "\n",
+    "def train_model_snapshot(model, criterion, lr, dataloaders, dataset_sizes, device, num_cycles, num_epochs_per_cycle):\n",
+    "    since = time.time()    #记录训练时间\n",
+    "\n",
+    "    best_model_wts = copy.deepcopy(model.state_dict())   #从预训练的模型中复制权重并初始化模型\n",
+    "    best_acc = 0.0\n",
+    "    best_loss = 1000000.0\n",
+    "    model_w_arr = []\n",
+    "    prob = torch.zeros((dataset_sizes['val'], 3), dtype = torch.float32).to(device)    #预测\n",
+    "    lbl = torch.zeros((dataset_sizes['val'],), dtype = torch.long).to(device)          #标签\n",
+    "    for cycle in range(num_cycles):\n",
+    "        optimizer = optim.SGD(model.parameters(), lr=lr, momentum=0.9)#, weight_decay = 0.0005)\n",
+    "        scheduler = lr_scheduler.CosineAnnealingLR(optimizer, num_epochs_per_cycle*len(dataloaders['train']))    \n",
+    "        #余弦周期调整学习率,修改优化器中的学习率,(优化器，调整间隔数，调整系数)\n",
+    "        for epoch in range(num_epochs_per_cycle):\n",
+    "            #print('Cycle {}: Epoch {}/{}'.format(cycle, epoch, num_epochs_per_cycle - 1))\n",
+    "            #print('-' * 10)\n",
+    "\n",
+    "            # Each epoch has a training and validation phase\n",
+    "            for phase in ['train', 'val']:\n",
+    "                if phase == 'train':\n",
+    "                    model.train()  # Set model to training mode\n",
+    "                else:\n",
+    "                    model.eval()   # Set model to evaluate mode\n",
+    "\n",
+    "                running_loss = 0.0\n",
+    "                running_corrects = 0\n",
+    "                idx = 0\n",
+    "                # Iterate over data.迭代数据\n",
+    "                for inputs, labels in dataloaders[phase]:\n",
+    "                    inputs = inputs.to(device)\n",
+    "                    labels = labels.to(device)\n",
+    "\n",
+    "                    # zero the parameter gradients\n",
+    "                    optimizer.zero_grad()\n",
+    "\n",
+    "                    # forward\n",
+    "                    # track history if only in train\n",
+    "                    with torch.set_grad_enabled(phase == 'train'):\n",
+    "                        outputs = model(inputs)\n",
+    "                        _, preds = torch.max(outputs, 1)\n",
+    "                        if (epoch == num_epochs_per_cycle-1) and (phase == 'val'):\n",
+    "                            prob[idx:idx+inputs.shape[0]] += F.softmax(outputs, dim = 1)\n",
+    "                            lbl[idx:idx+inputs.shape[0]] = labels\n",
+    "                            idx += inputs.shape[0]\n",
+    "                        loss = criterion(outputs, labels)\n",
+    "                        # backward + optimize only if in training phase\n",
+    "                        if phase == 'train':\n",
+    "                            loss.backward()\n",
+    "                            optimizer.step()\n",
+    "                            scheduler.step()\n",
+    "                            #print(optimizer.param_groups[0]['lr'])\n",
+    "                    \n",
+    "                    # statistics\n",
+    "                    running_loss += loss.item() * inputs.size(0)\n",
+    "                    running_corrects += torch.sum(preds == labels.data)\n",
+    "\n",
+    "                epoch_loss = running_loss / dataset_sizes[phase]\n",
+    "                epoch_acc = running_corrects.double() / dataset_sizes[phase]\n",
+    "\n",
+    "                #print('{} Loss: {:.4f} Acc: {:.4f}'.format(\n",
+    "                #    phase, epoch_loss, epoch_acc))\n",
+    "\n",
+    "                # deep copy the model\n",
+    "                if phase == 'val' and epoch_loss < best_loss:\n",
+    "                    best_loss = epoch_loss\n",
+    "                    best_model_wts = copy.deepcopy(model.state_dict())\n",
+    "            #print()\n",
+    "        model_w_arr.append(copy.deepcopy(model.state_dict()))\n",
+    "\n",
+    "    prob /= num_cycles\n",
+    "    ensemble_loss = F.nll_loss(torch.log(prob), lbl)  \n",
+    "    ensemble_loss = ensemble_loss.item()\n",
+    "    time_elapsed = time.time() - since\n",
+    "    #print('Training complete in {:.0f}m {:.0f}s'.format(\n",
+    "    #    time_elapsed // 60, time_elapsed % 60))\n",
+    "    #print('Ensemble Loss : {:4f}, Best val Loss: {:4f}'.format(ensemble_loss, best_loss))\n",
+    "\n",
+    "    # load best model weights加载最佳模型权重\n",
+    "    model_arr =[]\n",
+    "    for weights in model_w_arr:\n",
+    "        model.load_state_dict(weights)   \n",
+    "        model_arr.append(model) \n",
+    "    return model_arr, ensemble_loss, best_loss, prob\n",
+    "\n",
+    "def test(models_arr, loader, device):\n",
+    "    res = np.zeros((610, 3), dtype = np.float32)\n",
+    "    for model in models_arr:\n",
+    "        model.eval()\n",
+    "        res_arr = []\n",
+    "        for inputs, _ in loader:\n",
+    "            inputs = inputs.to(device)\n",
+    "            # forward\n",
+    "            # track history if only in train\n",
+    "            with torch.set_grad_enabled(False):\n",
+    "                outputs = F.softmax(model(inputs), dim = 1)    \n",
+    "                res_arr.append(outputs.detach().cpu().numpy())\n",
+    "        res_arr = np.concatenate(res_arr, axis = 0)\n",
+    "        res += res_arr\n",
+    "    return res / len(models_arr)\n",
+    "\n",
+    "def read_train_data(p):\n",
+    "    imgs = []\n",
+    "    labels = []\n",
+    "    for i, lbl in enumerate(os.listdir(p)):\n",
+    "        for fname in os.listdir(os.path.join(p, lbl)):\n",
+    "            #read image\n",
+    "            img = Image.open(os.path.join(p, lbl, fname))\n",
+    "            #rotate image to original view旋转图像到原始视图\n",
+    "            try:\n",
+    "                exif=dict((ExifTags.TAGS[k], v) for k, v in img._getexif().items() if k in ExifTags.TAGS)\n",
+    "                if exif['Orientation'] == 3:\n",
+    "                    img=img.rotate(180, expand=True)\n",
+    "                elif exif['Orientation'] == 6:\n",
+    "                    img=img.rotate(270, expand=True)\n",
+    "                elif exif['Orientation'] == 8:\n",
+    "                    img=img.rotate(90, expand=True)\n",
+    "            except:\n",
+    "                pass\n",
+    "            #resize all images to the same size将所有图像调整为相同的大小\n",
+    "            img = np.array(img.convert('RGB').resize((512,512), Image.ANTIALIAS))\n",
+    "            imgs.append(img)\n",
+    "            labels.append(i)\n",
+    "    return imgs, labels\n",
+    "\n",
+    "def read_test_data(p):\n",
+    "    imgs = []\n",
+    "    labels = []\n",
+    "    ids = []\n",
+    "    for fname in os.listdir(p):\n",
+    "        #read image\n",
+    "        img = Image.open(os.path.join(p, fname))\n",
+    "        #rotate image to original view\n",
+    "        try:\n",
+    "            if not('DMWVNR' in fname):\n",
+    "                exif=dict((ExifTags.TAGS[k], v) for k, v in img._getexif().items() if k in ExifTags.TAGS)\n",
+    "                if exif['Orientation'] == 3:\n",
+    "                    img=img.rotate(180, expand=True)\n",
+    "                elif exif['Orientation'] == 6:\n",
+    "                    img=img.rotate(270, expand=True)\n",
+    "                elif exif['Orientation'] == 8:\n",
+    "                    img=img.rotate(90, expand=True)\n",
+    "        except:\n",
+    "            pass\n",
+    "        #resize all images to the same size\n",
+    "        img = img.convert('RGB').resize((512,512), Image.ANTIALIAS)\n",
+    "        imgs.append(np.array(img.copy()))\n",
+    "        labels.append(0)\n",
+    "        ids.append(fname.split('.')[0])\n",
+    "        img.close()\n",
+    "    return imgs, labels, ids\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "unique train imgs: 732\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "usage: ipykernel_launcher.py [-h] [--train_data_path TRAIN_DATA_PATH]\n",
+      "                             [--test_data_path TEST_DATA_PATH]\n",
+      "                             [--save_path SAVE_PATH]\n",
+      "ipykernel_launcher.py: error: unrecognized arguments: -f C:\\Users\\Administrator\\AppData\\Roaming\\jupyter\\runtime\\kernel-60e74fca-82ff-42d7-afc4-1d27b752461b.json\n"
+     ]
+    },
+    {
+     "ename": "SystemExit",
+     "evalue": "2",
+     "output_type": "error",
+     "traceback": [
+      "An exception has occurred, use %tb to see the full traceback.\n",
+      "\u001b[1;31mSystemExit\u001b[0m\u001b[1;31m:\u001b[0m 2\n"
+     ]
+    }
+   ],
+   "source": [
+    "# %load prepare_dataset.py\n",
+    "#读取训练数据和测试数据，从训练数据中删除重复的数据并保存在numpy矩阵中\n",
+    "%run prepare_dataset.py\n",
+    "import numpy as np\n",
+    "import os\n",
+    "import argparse\n",
+    "from utils import read_train_data, read_test_data\n",
+    "\n",
+    "parser = argparse.ArgumentParser(description='Data preperation')\n",
+    "parser.add_argument('--train_data_path', help='path', default='D:/datasets/train', type=str)\n",
+    "parser.add_argument('--test_data_path', help='path', default='D:/datasets/test', type=str)\n",
+    "parser.add_argument('--save_path', help='save', default='D:/datasets/savepath', type=str)\n",
+    "args = parser.parse_args()  #获取参数，调用上面的属性\n",
+    "\n",
+    "#read training data\n",
+    "train_imgs, train_gts = read_train_data(args.train_data_path)\n",
+    "\n",
+    "#remove dublicate training imgs\n",
+    "idx_to_rmv = []\n",
+    "for i in range(len(train_imgs)-1):\n",
+    "    for j in range(i+1, len(train_imgs)):\n",
+    "        if np.all(train_imgs[i] == train_imgs[j]):\n",
+    "            idx_to_rmv.append(i)\n",
+    "            if train_gts[i] != train_gts[j]:\n",
+    "                idx_to_rmv.append(j)\n",
+    "\n",
+    "idx = [i for i in range(len(train_imgs)) if not(i in idx_to_rmv)]\n",
+    "print('unique train imgs:',len(idx))\n",
+    "\n",
+    "#save unique training imgs\n",
+    "np.save(os.path.join(args.save_path, 'unique_train_imgs_rot_fixed'), np.array(train_imgs)[idx])\n",
+    "np.save(os.path.join(args.save_path, 'unique_train_gts_rot_fixed'), np.array(train_gts)[idx])\n",
+    "\n",
+    "#read test data\n",
+    "test_imgs, test_gts, ids = read_test_data(args.test_data_path)\n",
+    "\n",
+    "#save test data\n",
+    "np.save(os.path.join(args.save_path, 'test_imgs_rot_fixed'), np.array(test_imgs))\n",
+    "np.save(os.path.join(args.save_path, 'test_gts'), np.array(test_gts))\n",
+    "np.save(os.path.join(args.save_path, 'ids'), np.array(ids))\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "unique train imgs: 732\n"
+     ]
+    }
+   ],
+   "source": [
+    "%run prepare_dataset.py"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "usage: ipykernel_launcher.py [-h] [--data_path DATA_PATH]\n",
+      "                             [--library_size LIBRARY_SIZE]\n",
+      "                             [--library_path LIBRARY_PATH]\n",
+      "ipykernel_launcher.py: error: unrecognized arguments: -f C:\\Users\\Administrator\\AppData\\Roaming\\jupyter\\runtime\\kernel-9ba0d6cf-6ce4-4517-8c53-84c6b0e19712.json\n"
+     ]
+    },
+    {
+     "ename": "SystemExit",
+     "evalue": "2",
+     "output_type": "error",
+     "traceback": [
+      "An exception has occurred, use %tb to see the full traceback.\n",
+      "\u001b[1;31mSystemExit\u001b[0m\u001b[1;31m:\u001b[0m 2\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\IPython\\core\\interactiveshell.py:3333: UserWarning: To exit: use 'exit', 'quit', or Ctrl-D.\n",
+      "  warn(\"To exit: use 'exit', 'quit', or Ctrl-D.\", stacklevel=1)\n"
+     ]
+    }
+   ],
+   "source": [
+    "# %load generate_library_of_models.py\n",
+    "import torch\n",
+    "import torch.nn as nn\n",
+    "import torch.optim as optim\n",
+    "from torch.optim import lr_scheduler\n",
+    "import numpy as np\n",
+    "import torchvision\n",
+    "from torchvision import datasets, models, transforms\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import time\n",
+    "import os\n",
+    "import argparse\n",
+    "import copy\n",
+    "from sklearn.model_selection import StratifiedKFold\n",
+    "import datetime\n",
+    "from PIL import Image\n",
+    "import torch.nn.functional as F\n",
+    "\n",
+    "from dataset import ICLRDataset\n",
+    "from utils import train_model_snapshot, test\n",
+    "from sklearn.metrics import confusion_matrix\n",
+    "from hyperopt import hp, tpe, fmin, Trials\n",
+    "from collections import OrderedDict\n",
+    "\n",
+    "\n",
+    "def score(params):\n",
+    "    global test_prob, val_prob, trails_sc_arr,idx    # 全局变量                                                                                                                              \n",
+    "    print(params)\n",
+    "    k = 5 \n",
+    "    sss = StratifiedKFold(n_splits=k, shuffle = True, random_state=seed_arr[idx])    #提供训练/测试索引来分割训练/测试集中的数据\n",
+    "    #（折叠的数量，至少是2；在分组前是否对每个类的样本进行洗牌；当shuffle为真时，random_state将影响索引的排序）\n",
+    "    \n",
+    "    #define trail data augmentations  训练集数据增强和归一化、验证集归一化\n",
+    "    data_transforms = {\n",
+    "        'train': transforms.Compose([\n",
+    "            transforms.ColorJitter(contrast = params['contrast'], hue = params['hue'], brightness = params['brightness']),\n",
+    "            transforms.RandomAffine(degrees = params['degrees']),\n",
+    "            transforms.RandomResizedCrop(224),\n",
+    "            transforms.RandomHorizontalFlip(p = 0.5 if params['h_flip'] else 0.0), #以概率P水平翻转图像\n",
+    "            transforms.RandomVerticalFlip(p = 0.5 if params['v_flip'] else 0.0),#以概率P垂直翻转图像\n",
+    "            transforms.ToTensor(),\n",
+    "            transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n",
+    "        ]),\n",
+    "        'val': transforms.Compose([\n",
+    "            transforms.Resize((params['val_img_size'], params['val_img_size'])),\n",
+    "            transforms.CenterCrop(224),\n",
+    "            transforms.ToTensor(),\n",
+    "            transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n",
+    "        ]),\n",
+    "    }\n",
+    "\n",
+    "    trail_test_prob = np.zeros((test_imgs.shape[0], 3), dtype = np.float32)\n",
+    "    trail_val_prob = torch.zeros((train_imgs.shape[0], 3), dtype = torch.float32).to(device)\n",
+    "    \n",
+    "    sc_arr = []\n",
+    "    models_arr = []\n",
+    "    fold = 0\n",
+    "    #train a model for each split  为每一个分割训练一个模型\n",
+    "    for train_index, val_index in sss.split(train_imgs, train_gts):\n",
+    "        #define dataset and loader for training and validation  确定数据集，载入训练集、验证集\n",
+    "        image_datasets = {'train': ICLRDataset(train_imgs, train_gts, 'train', train_index, data_transforms['train'], params['img_mix_enable']),\n",
+    "\t                      'val': ICLRDataset(train_imgs, train_gts, 'val', val_index, data_transforms['val'])}\n",
+    "\n",
+    "        dataloaders = {'train': torch.utils.data.DataLoader(image_datasets['train'], batch_size=16, shuffle=True, num_workers=2),\n",
+    "                       'val': torch.utils.data.DataLoader(image_datasets['val'], batch_size=16, shuffle=False, num_workers=2)}\n",
+    "\n",
+    "        #create model instance创建模型实例\n",
+    "        model_ft = params['arch'](pretrained=True)   #预训练\n",
+    "        try:\n",
+    "            num_ftrs = model_ft.fc.in_features\n",
+    "            model_ft.fc = nn.Linear(num_ftrs, 3)\n",
+    "        except:\n",
+    "            num_ftrs = model_ft.classifier.in_features\n",
+    "            model_ft.classifier = nn.Linear(num_ftrs, 3)\n",
+    "        model_ft = model_ft.to(device)\n",
+    "\n",
+    "        criterion = nn.CrossEntropyLoss()\n",
+    "\n",
+    "        dataset_sizes = {x:len(image_datasets[x]) for x in ['train', 'val']}\n",
+    "        \n",
+    "        model_ft_arr, ensemble_loss, _, fold_val_prob = train_model_snapshot(model_ft, criterion, params['lr'], dataloaders, dataset_sizes, device,\n",
+    "                               num_cycles=params['num_cycles'], num_epochs_per_cycle=params['num_epochs_per_cycle'])\n",
+    "        models_arr.extend(model_ft_arr)\n",
+    "        fold += 1\n",
+    "        sc_arr.append(ensemble_loss)\n",
+    "        trail_val_prob[val_index] = fold_val_prob\n",
+    "    \n",
+    "    #predict on test data using average of kfold models\n",
+    "    image_datasets['test'] = ICLRDataset(test_imgs, test_gts, 'test', None, data_transforms['val'])\n",
+    "    test_loader = torch.utils.data.DataLoader(image_datasets['test'], batch_size=4,shuffle=False, num_workers=16)\n",
+    "    trail_test_prob = test(models_arr, test_loader, device)\n",
+    "\n",
+    "    print('mean val loss:', np.mean(sc_arr))\n",
+    "\n",
+    "    test_prob.append(trail_test_prob)\n",
+    "    val_prob.append(trail_val_prob)\n",
+    "\n",
+    "    #save validation and test results for further processing \n",
+    "    np.save(os.path.join(args.library_path, 'val_prob_trail_%d'%(idx)), trail_val_prob.detach().cpu().numpy())\n",
+    "    np.save(os.path.join(args.library_path, 'test_prob_trail_%d'%(idx)), trail_test_prob)\n",
+    "    idx += 1\n",
+    "    \n",
+    "    trails_sc_arr.append(np.mean(sc_arr))\n",
+    "\n",
+    "    torch.cuda.empty_cache()\n",
+    "    del models_arr\n",
+    "\n",
+    "    return np.mean(sc_arr)\n",
+    "\n",
+    "parser = argparse.ArgumentParser(description='Data preperation')\n",
+    "parser.add_argument('--data_path', help='path to training and test numpy matrices of images', default='.', type=str)\n",
+    "parser.add_argument('--library_size', help='number of models to be trained in the library of models', default=50, type=int)\n",
+    "parser.add_argument('--library_path', help='save path for validation and test predictions of the library of models', default='trails', type=str)\n",
+    "args = parser.parse_args()\n",
+    "\n",
+    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
+    "\n",
+    "torch.manual_seed(0)\n",
+    "np.random.seed(0)\n",
+    "\n",
+    "torch.backends.cudnn.deterministic = True\n",
+    "torch.backends.cudnn.benchmark = False\n",
+    "\n",
+    "#read train data\n",
+    "train_imgs = np.load(os.path.join(args.data_path, 'D:/datasets/savepath/unique_train_imgs_rot_fixed.npy'))\n",
+    "train_gts = np.load(os.path.join(args.data_path, 'unique_train_gts_rot_fixed.npy'))\n",
+    "\n",
+    "#read test data\n",
+    "test_imgs = np.load(os.path.join(args.data_path, 'test_imgs_rot_fixed.npy'))\n",
+    "test_gts = np.load(os.path.join(args.data_path, 'test_gts.npy'))\n",
+    "ids = np.load(os.path.join(args.data_path, 'ids.npy')).tolist()\n",
+    "\n",
+    "test_prob = []\n",
+    "val_prob = []\n",
+    "trails_sc_arr = []\n",
+    "\n",
+    "n_trails = args.library_size\n",
+    "seed_arr = np.random.randint(low=0, high=1000000, size=n_trails)\n",
+    "\n",
+    "#create search space for hyperparameter optimization\n",
+    "space = OrderedDict([('lr', hp.choice('lr', [i*0.001 for i in range(1,4)])),\n",
+    "                     ('num_cycles', hp.choice('num_cycles', range(3, 6))),\n",
+    "                     ('num_epochs_per_cycle', hp.choice('num_epochs_per_cycle', range(3, 6))),\n",
+    "                     ('arch', hp.choice('arch', [models.densenet201, models.densenet121, models.densenet169,\n",
+    "                                                 models.wide_resnet50_2, models.resnet152, \n",
+    "                                                 models.resnet101, models.resnet50, models.resnet34, models.resnet18])),\n",
+    "                     ('img_mix_enable', hp.choice('img_mix_enable', [True, False])),\n",
+    "                     ('v_flip', hp.choice('v_flip', [True, False])),\n",
+    "                     ('h_flip', hp.choice('h_flip', [True, False])),\n",
+    "                     ('degrees', hp.choice('degrees', range(1, 90))),\n",
+    "                     ('contrast', hp.uniform('contrast', 0.0, 0.3)),\n",
+    "                     ('hue', hp.uniform('hue', 0.0, 0.3)),\n",
+    "                     ('brightness', hp.uniform('brightness', 0.0, 0.3)),\n",
+    "                     ('val_img_size', hp.choice('val_img_size', range(224, 512, 24))),\n",
+    "                     ])\n",
+    "\n",
+    "trials = Trials()\n",
+    "\n",
+    "idx = 0\n",
+    "if not os.path.exists(args.library_path):\n",
+    "    os.mkdir(args.library_path)\n",
+    "\n",
+    "#use tpe algorithm in hyperopt to generate a library of differnet models 利用hyperopt中的tpe算法生成不同模型库\n",
+    "best = fmin(fn=score,space=space,algo=tpe.suggest,max_evals=n_trails,trials=trials)\n",
+    "#fmin是对不用的算法集及其超参数进行迭代，使目标函数最小化的优化函数\n",
+    "#（最小化的目标函数；定义的搜索空间；搜索算法-为超参数空间的顺序搜索提供逻辑；最大评估数；trials对象）\n",
+    "print(best)\n",
+    "\n",
+    "np.save(os.path.join(args.library_path, 'scores.npy'), np.array(trails_sc_arr))\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{'arch': <function resnet152 at 0x000002654D4F8F28>, 'brightness': 0.006903174616102503, 'contrast': 0.10395561286019817, 'degrees': 14, 'h_flip': True, 'hue': 0.12930405533670436, 'img_mix_enable': False, 'lr': 0.001, 'num_cycles': 5, 'num_epochs_per_cycle': 5, 'v_flip': True, 'val_img_size': 416}\n",
+      "  0%|                                                                          | 0/100 [00:00<?, ?trial/s, best loss=?]"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "job exception: module '__main__' has no attribute '__spec__'\n",
+      "\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    },
+    {
+     "ename": "AttributeError",
+     "evalue": "module '__main__' has no attribute '__spec__'",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
+      "\u001b[1;32m~\\Desktop\\WheatRustClassification-master\\generate_library_of_models.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m    161\u001b[0m \u001b[1;31m#use tpe algorithm in hyperopt to generate a library of differnet models\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    162\u001b[0m \u001b[1;31m#best = fmin(fn=score,space=space,algo=tpe.suggest,max_evals=n_trails,trials=trials)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 163\u001b[1;33m \u001b[0mbest\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mfmin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfn\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mscore\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mspace\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mspace\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0malgo\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mtpe\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msuggest\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mmax_evals\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m100\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    164\u001b[0m \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mbest\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    165\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\fmin.py\u001b[0m in \u001b[0;36mfmin\u001b[1;34m(fn, space, algo, max_evals, timeout, loss_threshold, trials, rstate, allow_trials_fmin, pass_expr_memo_ctrl, catch_eval_exceptions, verbose, return_argmin, points_to_evaluate, max_queue_len, show_progressbar)\u001b[0m\n\u001b[0;32m    480\u001b[0m             \u001b[0mcatch_eval_exceptions\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcatch_eval_exceptions\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    481\u001b[0m             \u001b[0mreturn_argmin\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mreturn_argmin\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 482\u001b[1;33m             \u001b[0mshow_progressbar\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mshow_progressbar\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    483\u001b[0m         )\n\u001b[0;32m    484\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\base.py\u001b[0m in \u001b[0;36mfmin\u001b[1;34m(self, fn, space, algo, max_evals, timeout, loss_threshold, max_queue_len, rstate, verbose, pass_expr_memo_ctrl, catch_eval_exceptions, return_argmin, show_progressbar)\u001b[0m\n\u001b[0;32m    684\u001b[0m             \u001b[0mcatch_eval_exceptions\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcatch_eval_exceptions\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    685\u001b[0m             \u001b[0mreturn_argmin\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mreturn_argmin\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 686\u001b[1;33m             \u001b[0mshow_progressbar\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mshow_progressbar\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    687\u001b[0m         )\n\u001b[0;32m    688\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\fmin.py\u001b[0m in \u001b[0;36mfmin\u001b[1;34m(fn, space, algo, max_evals, timeout, loss_threshold, trials, rstate, allow_trials_fmin, pass_expr_memo_ctrl, catch_eval_exceptions, verbose, return_argmin, points_to_evaluate, max_queue_len, show_progressbar)\u001b[0m\n\u001b[0;32m    507\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    508\u001b[0m     \u001b[1;31m# next line is where the fmin is actually executed\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 509\u001b[1;33m     \u001b[0mrval\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mexhaust\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    510\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    511\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mreturn_argmin\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\fmin.py\u001b[0m in \u001b[0;36mexhaust\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    328\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mexhaust\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    329\u001b[0m         \u001b[0mn_done\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 330\u001b[1;33m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmax_evals\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mn_done\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mblock_until_done\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0masynchronous\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    331\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mrefresh\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    332\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\fmin.py\u001b[0m in \u001b[0;36mrun\u001b[1;34m(self, N, block_until_done)\u001b[0m\n\u001b[0;32m    284\u001b[0m                 \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    285\u001b[0m                     \u001b[1;31m# -- loop over trials and do the jobs directly\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 286\u001b[1;33m                     \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mserial_evaluate\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    287\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    288\u001b[0m                 \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mrefresh\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\fmin.py\u001b[0m in \u001b[0;36mserial_evaluate\u001b[1;34m(self, N)\u001b[0m\n\u001b[0;32m    163\u001b[0m                 \u001b[0mctrl\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mbase\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mCtrl\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtrials\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcurrent_trial\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mtrial\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    164\u001b[0m                 \u001b[1;32mtry\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 165\u001b[1;33m                     \u001b[0mresult\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdomain\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mevaluate\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mspec\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mctrl\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    166\u001b[0m                 \u001b[1;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[1;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    167\u001b[0m                     \u001b[0mlogger\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0merror\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"job exception: %s\"\u001b[0m \u001b[1;33m%\u001b[0m \u001b[0mstr\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0me\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\hyperopt\\base.py\u001b[0m in \u001b[0;36mevaluate\u001b[1;34m(self, config, ctrl, attach_attachments)\u001b[0m\n\u001b[0;32m    892\u001b[0m                 \u001b[0mprint_node_on_error\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mrec_eval_print_node_on_error\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    893\u001b[0m             )\n\u001b[1;32m--> 894\u001b[1;33m             \u001b[0mrval\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfn\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mpyll_rval\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    895\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    896\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mrval\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mfloat\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mint\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnumber\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\Desktop\\WheatRustClassification-master\\generate_library_of_models.py\u001b[0m in \u001b[0;36mscore\u001b[1;34m(params)\u001b[0m\n\u001b[0;32m     78\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     79\u001b[0m         model_ft_arr, ensemble_loss, _, fold_val_prob = train_model_snapshot(model_ft, criterion, params['lr'], dataloaders, dataset_sizes, device,\n\u001b[1;32m---> 80\u001b[1;33m                                num_cycles=params['num_cycles'], num_epochs_per_cycle=params['num_epochs_per_cycle'])\n\u001b[0m\u001b[0;32m     81\u001b[0m         \u001b[0mmodels_arr\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mextend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmodel_ft_arr\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     82\u001b[0m         \u001b[0mfold\u001b[0m \u001b[1;33m+=\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32m~\\Desktop\\WheatRustClassification-master\\utils.py\u001b[0m in \u001b[0;36mtrain_model_snapshot\u001b[1;34m(model, criterion, lr, dataloaders, dataset_sizes, device, num_cycles, num_epochs_per_cycle)\u001b[0m\n\u001b[0;32m     40\u001b[0m                 \u001b[0midx\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m0\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     41\u001b[0m                 \u001b[1;31m# Iterate over data.\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 42\u001b[1;33m                 \u001b[1;32mfor\u001b[0m \u001b[0minputs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mlabels\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mdataloaders\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mphase\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     43\u001b[0m                     \u001b[0minputs\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0minputs\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mto\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdevice\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     44\u001b[0m                     \u001b[0mlabels\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlabels\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mto\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdevice\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\site-packages\\torch\\utils\\data\\dataloader.py\u001b[0m in \u001b[0;36m__iter__\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    277\u001b[0m             \u001b[1;32mreturn\u001b[0m \u001b[0m_SingleProcessDataLoaderIter\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    278\u001b[0m         \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 279\u001b[1;33m             \u001b[1;32mreturn\u001b[0m \u001b[0m_MultiProcessingDataLoaderIter\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    280\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    281\u001b[0m     \u001b[1;33m@\u001b[0m\u001b[0mproperty\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\site-packages\\torch\\utils\\data\\dataloader.py\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self, loader)\u001b[0m\n\u001b[0;32m    717\u001b[0m             \u001b[1;31m#     before it starts, and __del__ tries to join but will get:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    718\u001b[0m             \u001b[1;31m#     AssertionError: can only join a started process.\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 719\u001b[1;33m             \u001b[0mw\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mstart\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    720\u001b[0m             \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_index_queues\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mindex_queue\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    721\u001b[0m             \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_workers\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mw\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\multiprocessing\\process.py\u001b[0m in \u001b[0;36mstart\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    110\u001b[0m                \u001b[1;34m'daemonic processes are not allowed to have children'\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    111\u001b[0m         \u001b[0m_cleanup\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 112\u001b[1;33m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_popen\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_Popen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    113\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_sentinel\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_popen\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msentinel\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    114\u001b[0m         \u001b[1;31m# Avoid a refcycle if the target function holds an indirect\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\multiprocessing\\context.py\u001b[0m in \u001b[0;36m_Popen\u001b[1;34m(process_obj)\u001b[0m\n\u001b[0;32m    221\u001b[0m     \u001b[1;33m@\u001b[0m\u001b[0mstaticmethod\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    222\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m_Popen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprocess_obj\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 223\u001b[1;33m         \u001b[1;32mreturn\u001b[0m \u001b[0m_default_context\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mget_context\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mProcess\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_Popen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprocess_obj\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    224\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    225\u001b[0m \u001b[1;32mclass\u001b[0m \u001b[0mDefaultContext\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mBaseContext\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\multiprocessing\\context.py\u001b[0m in \u001b[0;36m_Popen\u001b[1;34m(process_obj)\u001b[0m\n\u001b[0;32m    320\u001b[0m         \u001b[1;32mdef\u001b[0m \u001b[0m_Popen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprocess_obj\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    321\u001b[0m             \u001b[1;32mfrom\u001b[0m \u001b[1;33m.\u001b[0m\u001b[0mpopen_spawn_win32\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mPopen\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 322\u001b[1;33m             \u001b[1;32mreturn\u001b[0m \u001b[0mPopen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprocess_obj\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    323\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    324\u001b[0m     \u001b[1;32mclass\u001b[0m \u001b[0mSpawnContext\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mBaseContext\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\multiprocessing\\popen_spawn_win32.py\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self, process_obj)\u001b[0m\n\u001b[0;32m     44\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     45\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mprocess_obj\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 46\u001b[1;33m         \u001b[0mprep_data\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mspawn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mget_preparation_data\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprocess_obj\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_name\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     47\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     48\u001b[0m         \u001b[1;31m# read end of pipe will be \"stolen\" by the child process\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\multiprocessing\\spawn.py\u001b[0m in \u001b[0;36mget_preparation_data\u001b[1;34m(name)\u001b[0m\n\u001b[0;32m    170\u001b[0m     \u001b[1;31m# or through direct execution (or to leave it alone entirely)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    171\u001b[0m     \u001b[0mmain_module\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0msys\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmodules\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'__main__'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 172\u001b[1;33m     \u001b[0mmain_mod_name\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mgetattr\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmain_module\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m__spec__\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m\"name\"\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    173\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mmain_mod_name\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    174\u001b[0m         \u001b[0md\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'init_main_from_name'\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mmain_mod_name\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
+      "\u001b[1;31mAttributeError\u001b[0m: module '__main__' has no attribute '__spec__'"
+     ]
+    }
+   ],
+   "source": [
+    "%run generate_library_of_models.py"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.7.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/__pycache__/dataset.cpython-36.pyc" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/__pycache__/dataset.cpython-36.pyc"
new file mode 100644
index 0000000000000000000000000000000000000000..fb3841946c3e952ba38f0a23c969491afe094e13
GIT binary patch
literal 1827
zcmb7FOK&7K5cZ>A^T-m|A-h@Nf<Q_i7$LZ!kd2f`C@6}e-JnISAaB#|neC*zXJUKM
zB<eX$q}?k=_y_zTE*1wQPWcOaqRKOQiM`;F%U_r6a@AL5&(|v}?Z5u~{^g-h$Uo#x
z^8jx{*FS=gMADK}G@%qRD_O-8UU`Y<zIo{<0qDIltfC}Z>RL$)bbi^cI!T9;?}!X!
zcu8cK(>YJNl0PSLbOqK)Ol`QEnpEc|X0~lqTI<0`RoeO!Q<VAu7u2$-(qRtJ-$Rh$
z&g;ejY(v+55IISxBngup$t1s|374MqFG=DpJV<;Q$rgMEvJF0TZD((1|Jl-kqbvjy
z*l<L<*!2;J#*^eUQfqH$Lj-w4K+<~SH62h%CA*9oZ<oy7+-m|leosdm4(nOtgXB#B
z60E`L-a$seHZ}$wX@3Y^_dpEch@ZMM3P;g(@+&*@o|8X_ieL(OV*M{IEr1WNsD|2U
z-+sIE&G8R9SNeGR)$wkwPt9n2{Nh!fnr9PjcFI)iVo+qMDMs}dRjN&{j^%>+{`mZD
zd$TOG$wt+9>j64}9tGrL;|@Q4A&R;%Li9JwQI?i^3&8cJ{}+mLc^Qpr!BMJ>kx|Qs
zW)Y9-qILzud(fBe#J=jF8a=daJubl;b3V>(>!xt*1E54y#Z=^V`g57bq4jiL4sJIp
zG-&;zmibhzqPh1`EU0y+7i-Z08(qhYPHJ3DXhZ`Vu^#QQ0MJG6T@>BWff+_Fcc5zw
z(m-ZU=$t|U&;!UNFZ0?R7*q8D%->SuD6pXrutJCr(EwT^^y1<DBi($4#flA!7i-C#
zXaE>E?*s~UPG`J<(p)E!pL%LM-EcEZG6MyA;?3!t%{)MU<E@j1opSZ$keoevfgT4Y
zJc)1>&cRAL^QAWn8rlRs()bd_$D~Idkr{0;HvixXLsuw<(4qbPbol#k*sh+&Ez9b$
z4dA-VQDs9Mi`pm*@=m5)b^_4Z)<BJ_RH^hl4joOFL!K=?w{#5KYbgpjwGo0$jmDPB
z|L~YjhE-mh$6(G~F+#iWW3)#fKpNMm2k1Vt%N?0x9kwB>$g*R09{}*-?&pjFg1Fbo
zj2m{s8wUKK&bvXt&?#zyZK8Hz3>(&P;G5w%V89qVus)P=QOD683NU*05ei2J*6yNS
zKLr$f*IdY{`WUA^L4l>b5Uz$hLJV_Ln7k6=GnmqNeF$wYf2+X4#iRQbb+aDc-y^s}
za1H9T%7w6PA*zv_l!&`RoK4blF{AL=QfnyCHT5})TTdKvcN*&xzch-7w!0CJx(c=N
zid{7$mA&e_x4)0ilS1-s$EklQG(27kQ=l*x)~_ek_*`L1t*^5JF#MSTcFQe4+}l@<
bK(ziYXKb{&h-M4xRHJ`gx(YYEi=F)o99M^J

literal 0
HcmV?d00001

diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/__pycache__/dataset.cpython-37(1).pyc" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/__pycache__/dataset.cpython-37(1).pyc"
new file mode 100644
index 0000000000000000000000000000000000000000..0fbc904d20541e8000ff00da46a22a48026ce44a
GIT binary patch
literal 1842
zcmb7FOK%)S5bnpk_F-el;=}=@NO15TSaC*{W3VN#WLc486f`olhUwYOcxPtU-Mz@R
zW>2=1TsgwYKJr8O6MaJB<iC&;RlU1@h+ODV)znmVS66*i-Sh3zQj@@l^<O_9_=Nm}
zAB*9Eumz-l1SW~31u1DvDax#1C69UO#h&}-g&zl?_lmHL;%KgG#0}8-MYC+hElR#4
zGLYdVkzqzJc-)r!CFw>t;GJ}-4R;cg=*)DPZ5owSx<68-w*JKAh2F;rmCVazkb&s$
zqLAXs#o!=p0qGvFjKoxun8}u8l3&u8OHcZjB=%+z#J-GV1HJ><gc#bkwY$A{Fc07`
z4Z#Hte4}3^eF&`fBpF6(^)0Q5Aa4n9TCKmKeJZJBmr?EQkPByfLqNwL=xEJhKC6A;
zybgeaH`v|VPbv7u!GI9<M?ks*YyeyQ)Sgn<ims90*opU&{7F;<CcugHzOpn2JiNRd
zXs3Pm{r1;KKj=*9qi<xHSGfi(%}5>XWO`^u<D*xvv&0-swAn5at@D1KCMF+MUzUkB
znL3iQBJ{>*@0uG0tWHPecyk%!!QcXNzJ3p?o)EBw5u&$IjMAjgn;<L}{lBbS^z+12
z1GZAwnUEVFm|1G7joKw(9{|n6>H4aLYK+h}^|*j&%-J}zjoTiw%Ycq3^HY&k$uC9L
z4XvlMqJOtgp+oEEmCR1n3cC9c*$idpdNvmWu+buC3{vB4LL(Z`h;?X(1t6Vw-Urbw
z99V3`avMlvk_L)=OfM+(0^NtQ@*=C;hFz*Y0{a~_4gwnr0W*a77#*M|LeC#P+|t=o
z>{}ez#{rPr(STrJzbok3YdYmMbm<zAeCVn1>AJHp$qe-Bv3EhQ+0+Bcue~)=vmsYc
zUyzfhuOQ;9z=T&3&O+NY`tqeW4Qg5k9a8%e-cLw}JSJ0GV{ZQb4W_P;4S-<${e1Ap
z?>KIrcN>;fV;jK9m!r~#_|7Y%Fv)wBa@E0HY@@G6WujDa)(ss@mP4H_J+pKihpLhy
zm#1-rLaIh%OXYuXOeTXetIQK{=S~};ZFm^%&}As&D)m5mi0pht>QIMecvz7fvU?x^
z9`2}48Ay=#8kus#u6WG=AJAQS3+WgT8se;Ai>Mv2gf*)<;LY&eXMh;nw?4FSUUj2;
z5McJ|6J!ny?A=+vJ_D)iEwGTM>Qgj*h74PI##|4#gcxKdH(4pf7huwOd<bpMpA|si
z{PDw%x}6VK@gW=`I0jWxW<uDe5amct3Y6PIoJ^8pW>I*9sa0gpkbREqE)plXGmZU;
zUm96No9&24ZH3zIl3jt@D}CK_zkwcJPYS`eEvNo9*Km0$Y;=Wnuzocu$7c#lYJHvN
nAj5wRWH;UN7rT4P0f^q;7mSTIX4!0FpK6S+O;_NAw~^StRd0<R

literal 0
HcmV?d00001

diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/__pycache__/dataset.cpython-37.pyc" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/__pycache__/dataset.cpython-37.pyc"
new file mode 100644
index 0000000000000000000000000000000000000000..0fbc904d20541e8000ff00da46a22a48026ce44a
GIT binary patch
literal 1842
zcmb7FOK%)S5bnpk_F-el;=}=@NO15TSaC*{W3VN#WLc486f`olhUwYOcxPtU-Mz@R
zW>2=1TsgwYKJr8O6MaJB<iC&;RlU1@h+ODV)znmVS66*i-Sh3zQj@@l^<O_9_=Nm}
zAB*9Eumz-l1SW~31u1DvDax#1C69UO#h&}-g&zl?_lmHL;%KgG#0}8-MYC+hElR#4
zGLYdVkzqzJc-)r!CFw>t;GJ}-4R;cg=*)DPZ5owSx<68-w*JKAh2F;rmCVazkb&s$
zqLAXs#o!=p0qGvFjKoxun8}u8l3&u8OHcZjB=%+z#J-GV1HJ><gc#bkwY$A{Fc07`
z4Z#Hte4}3^eF&`fBpF6(^)0Q5Aa4n9TCKmKeJZJBmr?EQkPByfLqNwL=xEJhKC6A;
zybgeaH`v|VPbv7u!GI9<M?ks*YyeyQ)Sgn<ims90*opU&{7F;<CcugHzOpn2JiNRd
zXs3Pm{r1;KKj=*9qi<xHSGfi(%}5>XWO`^u<D*xvv&0-swAn5at@D1KCMF+MUzUkB
znL3iQBJ{>*@0uG0tWHPecyk%!!QcXNzJ3p?o)EBw5u&$IjMAjgn;<L}{lBbS^z+12
z1GZAwnUEVFm|1G7joKw(9{|n6>H4aLYK+h}^|*j&%-J}zjoTiw%Ycq3^HY&k$uC9L
z4XvlMqJOtgp+oEEmCR1n3cC9c*$idpdNvmWu+buC3{vB4LL(Z`h;?X(1t6Vw-Urbw
z99V3`avMlvk_L)=OfM+(0^NtQ@*=C;hFz*Y0{a~_4gwnr0W*a77#*M|LeC#P+|t=o
z>{}ez#{rPr(STrJzbok3YdYmMbm<zAeCVn1>AJHp$qe-Bv3EhQ+0+Bcue~)=vmsYc
zUyzfhuOQ;9z=T&3&O+NY`tqeW4Qg5k9a8%e-cLw}JSJ0GV{ZQb4W_P;4S-<${e1Ap
z?>KIrcN>;fV;jK9m!r~#_|7Y%Fv)wBa@E0HY@@G6WujDa)(ss@mP4H_J+pKihpLhy
zm#1-rLaIh%OXYuXOeTXetIQK{=S~};ZFm^%&}As&D)m5mi0pht>QIMecvz7fvU?x^
z9`2}48Ay=#8kus#u6WG=AJAQS3+WgT8se;Ai>Mv2gf*)<;LY&eXMh;nw?4FSUUj2;
z5McJ|6J!ny?A=+vJ_D)iEwGTM>Qgj*h74PI##|4#gcxKdH(4pf7huwOd<bpMpA|si
z{PDw%x}6VK@gW=`I0jWxW<uDe5amct3Y6PIoJ^8pW>I*9sa0gpkbREqE)plXGmZU;
zUm96No9&24ZH3zIl3jt@D}CK_zkwcJPYS`eEvNo9*Km0$Y;=Wnuzocu$7c#lYJHvN
nAj5wRWH;UN7rT4P0f^q;7mSTIX4!0FpK6S+O;_NAw~^StRd0<R

literal 0
HcmV?d00001

diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/dataset.py" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/dataset.py"
new file mode 100644
index 0000000..b68a127
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/dataset.py"
@@ -0,0 +1,43 @@
+from torch.utils.data import Dataset
+from torchvision import transforms, utils
+import numpy as np
+from scipy import ndimage
+import torch
+from PIL import Image
+
+class ICLRDataset(Dataset):
+    def __init__(self, imgs, gts, split_type, index, transform, img_mix_enable = True):
+        if index is None:
+            self.imgs = imgs
+            self.gts = gts
+        else:
+            self.imgs = [imgs[i] for i in index]
+            self.gts = [gts[i] for i in index] 
+                   
+        self.split_type = split_type
+        self.transform = transform
+        self.img_mix_enable = img_mix_enable
+    
+    def __len__(self):
+        return len(self.imgs)
+    
+    def augment(self, img, y):        
+        p = np.random.random(1)
+        if p[0] > 0.5:
+            while True:
+                rnd_idx = np.random.randint(0, len(self.imgs))
+                if self.gts[rnd_idx] != y:
+                    break
+            rnd_crop = self.transform(Image.fromarray(self.imgs[rnd_idx]))
+            d = 0.8
+            img = img * d + rnd_crop * (1 - d)
+        return img
+
+    def __getitem__(self, idx):
+        img = self.imgs[idx]
+        y = self.gts[idx]
+        img = Image.fromarray(img)
+        img = self.transform(img)
+        if (self.split_type == 'train') & self.img_mix_enable:
+            img = self.augment(img, y)
+        return img, y
\ No newline at end of file
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/ensemble_selection.py" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/ensemble_selection.py"
new file mode 100644
index 0000000..9da4b64
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/ensemble_selection.py"
@@ -0,0 +1,169 @@
+import torch
+import torch.nn as nn
+import torch.optim as optim
+from torch.optim import lr_scheduler
+import numpy as np
+import torchvision
+from torchvision import datasets, models, transforms
+import pandas as pd
+#import matplotlib.pyplot as plt
+import time
+import os
+import copy
+import argparse
+from sklearn.model_selection import StratifiedKFold
+import datetime
+from PIL import Image
+
+import torch.nn.functional as F
+
+def cross_entropy(y, p):
+    p /= p.sum(1).reshape(-1,1)
+    return F.nll_loss(torch.log(torch.tensor(p)), torch.tensor(y)).numpy()
+
+def weighted_cross_entropy(y, p):
+    p /= p.sum(1).reshape(-1,1)
+    w_arr = np.array([0.53, 0.3, 0.0])
+    return np.sum([F.nll_loss(torch.log(torch.tensor(p[y==c])), torch.tensor(y[y==c])).numpy()*w_arr[c] for c in range(3)])
+
+class ensembleSelection:
+
+    def __init__(self, metric):
+        self.metric = metric
+        
+    def _compare(self, sc1, sc2):
+        if sc1 < sc2:
+            return True
+        return False
+        
+    def _initialize(self, X_p, y):
+        """
+        This function finds the id of the best validation probabiltiy
+        """
+        current_sc = self.metric(y, X_p[0])
+        ind = 0
+        for i in range(1, X_p.shape[0]):
+            print(i)
+            sc = self.metric(y, X_p[i])
+            print(sc)
+            if self._compare(sc, current_sc):
+                current_sc = sc
+                ind = i
+        return ind, current_sc
+        
+    def es_with_replacement(self, X_p, Xtest_p, y):
+        best_ind, best_sc = self._initialize(X_p, y)
+        current_sc = best_sc
+        sumP = np.copy(X_p[best_ind])
+        sumP_test = np.copy(Xtest_p[best_ind])
+        i = 1
+        # find the best combintation of input models' reuslts
+        while True:
+            i += 1
+            ind = -1
+            for m in range(X_p.shape[0]):
+                #check if adding model m to the combination of best models will improve the results or not
+                sc = self.metric(y, (sumP*X_p[m])**(1/i))
+                if self._compare(sc, current_sc):
+                    current_sc = sc
+                    ind = m
+            if ind>-1:
+                sumP *= X_p[ind]
+                sumP_test *= Xtest_p[ind]
+            else:
+                break
+        sumP = sumP**(1/(i-1))
+        sumP_test = sumP_test**(1/(i-1))
+
+        sumP /= sumP.sum(1).reshape(-1,1)
+        sumP_test /= sumP_test.sum(1).reshape(-1,1)
+        
+        return current_sc, sumP, sumP_test
+        
+    def es_with_bagging(self, X_p, Xtest_p, y, f = 0.5, n_bags = 20):
+        list_of_indecies = [i for i in range(X_p.shape[0])]
+        bag_size = int(f*X_p.shape[0])
+        sumP = None
+        sumP_test = None
+        for i in range(n_bags):
+            #create a random subset (bag) of models
+            model_weight = [0 for j in range(X_p.shape[0])]
+            rng = np.copy(list_of_indecies)
+            np.random.shuffle(rng)
+            rng = rng[:bag_size]
+            #find the best combination from the input bag
+            sc, p, ptest = self.es_with_replacement(X_p[rng], Xtest_p[rng], y)
+            print('bag: %d, sc: %f'%(i, sc))
+            if sumP is None:
+                sumP = p
+                sumP_test = ptest
+            else:
+                sumP *= p
+                sumP_test *= ptest
+                
+        #combine the reuslts of all bags
+        sumP = sumP**(1/n_bags)
+        sumP_test = sumP_test**(1/n_bags)
+
+        sumP /= sumP.sum(1).reshape(-1,1)
+        sumP_test /= sumP_test.sum(1).reshape(-1,1)
+
+        sumP[sumP < 1e-6] = 1e-6
+        sumP_test[sumP_test < 1e-6] = 1e-6
+
+        final_sc = self.metric(y, sumP)
+        print('avg sc: %f'%(final_sc))
+        return (final_sc, sumP, sumP_test)
+
+parser = argparse.ArgumentParser(description='Data preperation')
+parser.add_argument('--train_data_path', help='path to training data folder', default='train_data', type=str)
+parser.add_argument('--data_path', help='path to training and test numpy matrices of images', default='.', type=str)
+parser.add_argument('--sample_sub_file_path', help='path to sample submission file', default='.', type=str)
+parser.add_argument('--library_size', help='number of models to be trained in the library of models', default=50, type=int)
+parser.add_argument('--library_path', help='save path for validation and test predictions of the library of models', default='trails', type=str)
+parser.add_argument('--final_sub_file_save_path', help='save path for final submission file', default='.', type=str)
+args = parser.parse_args()
+
+np.random.seed(4321)
+
+n = args.library_size
+
+#read training gt
+train_gts = np.load(os.path.join(args.data_path, 'C:/Users/x2/Desktop/pq/wheat rust/savepath/unique_train_gts_rot_fixed.npy'))
+#train_gts = np.load(os.path.join(args.data_path, 'unique_train_gts_rot_fixed.npy'))
+#read validation probability on training data generated from automatuic hypropt trails
+#and create a matrix of (N,D,3) where N i the number of models and D is the data size
+train_prob = np.array([np.load(os.path.join(args.library_path, 'C:/Users/x2/Desktop/pq/wheat rust/trails/val_prob_trail_%d.npy'%(i))) for i in range(n)])
+#train_prob = np.array([np.load(os.path.join(args.library_path, 'val_prob_trail_%d.npy'%(i))) for i in range(n)])
+
+#read test probability generated from hypropt trails
+#and create a matrix of (N,D,3) where N is the number of models and D is the data size
+test_prob = np.array([np.load(os.path.join(args.library_path, 'C:/Users/x2/Desktop/pq/wheat rust/trails/test_prob_trail_%d.npy'%(i))) for i in range(n)]) 
+#test_prob = np.array([np.load(os.path.join(args.library_path, 'test_prob_trail_%d.npy'%(i))) for i in range(n)]) 
+
+ids = np.load('C:/Users/x2/Desktop/pq/wheat rust/savepath/ids.npy').tolist()
+
+#use ensemble selection algorithm to find best combination of models using geometric average
+es_obj = ensembleSelection(cross_entropy)
+sc, es_train_prob, es_test_prob = es_obj.es_with_bagging(train_prob, test_prob, train_gts, n_bags = 10, f = 0.65)
+
+#detect samples with high confidence for healthy wheat
+idx = (np.max(es_test_prob, 1) > 0.7) & (np.argmax(es_test_prob, 1) == 2)
+
+#create another ensemble with more weights for leaf and stem classes
+es_obj = ensembleSelection(weighted_cross_entropy)
+sc, es_train_prob, es_test_prob = es_obj.es_with_bagging(train_prob, test_prob, train_gts, n_bags = 10, f = 0.65)
+
+#increase the probability of confident samples for healthy wheat
+es_test_prob[idx, 0] = 1e-6 
+es_test_prob[idx, 1] = 1e-6
+es_test_prob[idx, 2] = 1.0
+
+#create submission
+sub = pd.read_csv(os.path.join(args.sample_sub_file_path, 'sample_submission.csv'))
+sub['ID'] = ids
+lbl_names = os.listdir(args.train_data_path)
+for i, name in enumerate(lbl_names):
+    sub[name] = es_test_prob[:,i].tolist()
+sub.to_csv(os.path.join(args.final_sub_file_save_path, 'final_sub.csv'), index = False)
+
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/generate_library_of_models.py" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/generate_library_of_models.py"
new file mode 100644
index 0000000..d397095
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/generate_library_of_models.py"
@@ -0,0 +1,191 @@
+import torch
+import torch.nn as nn
+import torch.optim as optim
+from torch.optim import lr_scheduler
+import numpy as np
+from torchvision import datasets, models, transforms
+
+import pandas as pd
+#import matplotlib.pyplot as plt
+import time
+import os
+import argparse
+import copy
+from sklearn.model_selection import StratifiedKFold
+import datetime
+from PIL import Image
+import torch.nn.functional as F
+
+from dataset import ICLRDataset
+from utils import train_model_snapshot, test
+from sklearn.metrics import confusion_matrix
+from hyperopt import hp, tpe, fmin, Trials
+from collections import OrderedDict
+
+
+
+
+
+def score(params):
+    global test_prob, val_prob, trails_sc_arr,idx
+    print(params)
+    k = 5
+    sss = StratifiedKFold(n_splits=k, shuffle = True, random_state=seed_arr[idx])
+    #define trail data augmentations
+    data_transforms = {
+        'train': transforms.Compose([
+            transforms.ColorJitter(contrast = params['contrast'], hue = params['hue'], brightness = params['brightness']),
+            transforms.RandomAffine(degrees = params['degrees']),
+            transforms.RandomResizedCrop(224),
+            transforms.RandomHorizontalFlip(p = 0.5 if params['h_flip'] else 0.0),
+            transforms.RandomVerticalFlip(p = 0.5 if params['v_flip'] else 0.0),
+            transforms.ToTensor(),
+            transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
+        ]),
+        'val': transforms.Compose([
+            transforms.Resize((params['val_img_size'], params['val_img_size'])),
+            transforms.CenterCrop(224),
+            transforms.ToTensor(),
+            transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
+        ]),
+    }
+
+    trail_test_prob = np.zeros((test_imgs.shape[0], 3), dtype = np.float32)
+    trail_val_prob = torch.zeros((train_imgs.shape[0], 3), dtype = torch.float32).to(device)
+    
+    sc_arr = []
+    models_arr = []
+#    fold = 0
+    #train a model for each split
+    for train_index, val_index in sss.split(train_imgs, train_gts):
+        #define dataset and loader for training and validation
+        image_datasets = {'train': ICLRDataset(train_imgs, train_gts, 'train', train_index, data_transforms['train'], params['img_mix_enable']),
+	                      'val': ICLRDataset(train_imgs, train_gts, 'val', val_index, data_transforms['val'])}
+
+        dataloaders = {'train': torch.utils.data.DataLoader(image_datasets['train'], batch_size=16, shuffle=True, num_workers=0),
+                       'val': torch.utils.data.DataLoader(image_datasets['val'], batch_size=16, shuffle=False, num_workers=0)}
+
+        #create model instance
+        model_ft = params['arch'](pretrained=True)
+        try:
+            num_ftrs = model_ft.fc.in_features
+            model_ft.fc = nn.Linear(num_ftrs, 3)
+        except:
+            num_ftrs = model_ft.classifier.in_features
+            model_ft.classifier = nn.Linear(num_ftrs, 3)
+        model_ft = model_ft.to(device)
+
+        criterion = nn.CrossEntropyLoss()
+
+        dataset_sizes = {x:len(image_datasets[x]) for x in ['train', 'val']}
+        
+        model_ft_arr, ensemble_loss, _, fold_val_prob = train_model_snapshot(model_ft, criterion, params['lr'], dataloaders, dataset_sizes, device,
+                               num_cycles=params['num_cycles'], num_epochs_per_cycle=params['num_epochs_per_cycle'])
+        models_arr.extend(model_ft_arr)
+        fold += 1   #K折交叉验证
+        sc_arr.append(ensemble_loss)
+        trail_val_prob[val_index] = fold_val_prob
+    
+    #predict on test data using average of kfold models
+    image_datasets['test'] = ICLRDataset(test_imgs, test_gts, 'test', None, data_transforms['val'])
+    test_loader = torch.utils.data.DataLoader(image_datasets['test'], batch_size=4,shuffle=False, num_workers=0)
+    trail_test_prob = test(models_arr, test_loader, device)
+
+    print('mean val loss:', np.mean(sc_arr))
+
+    test_prob.append(trail_test_prob)
+    val_prob.append(trail_val_prob)
+
+    #save validation and test results for further processing 
+    np.save(os.path.join(args.library_path, 'val_prob_trail_%d'%(idx)), trail_val_prob.detach().cpu().numpy())
+    np.save(os.path.join(args.library_path, 'test_prob_trail_%d'%(idx)), trail_test_prob)
+    idx += 1
+    
+    trails_sc_arr.append(np.mean(sc_arr))
+
+    torch.cuda.empty_cache()
+    del models_arr
+
+    return np.mean(sc_arr)
+
+parser = argparse.ArgumentParser(description='Data preperation')
+parser.add_argument('--data_path', help='path to training and test numpy matrices of images', default='.', type=str)
+parser.add_argument('--library_size', help='number of models to be trained in the library of models', default=50, type=int)
+parser.add_argument('--library_path', help='save path for validation and test predictions of the library of models', default='trails', type=str)
+args = parser.parse_args()
+
+device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
+
+torch.manual_seed(0)
+np.random.seed(0)
+
+torch.backends.cudnn.deterministic = True
+torch.backends.cudnn.benchmark = False
+
+#read train data
+train_imgs = np.load(os.path.join('D:/datasets/savepath', 'unique_train_imgs_rot_fixed.npy'))
+train_gts = np.load(os.path.join('D:/datasets/savepath', 'unique_train_gts_rot_fixed.npy'))
+
+#read test data
+test_imgs = np.load(os.path.join('D:/datasets/savepath', 'test_imgs_rot_fixed.npy'))
+test_gts = np.load(os.path.join('D:/datasets/savepath', 'test_gts.npy'))
+ids = np.load(os.path.join('D:/datasets/savepath', 'ids.npy')).tolist()
+
+test_prob = []
+val_prob = []
+trails_sc_arr = []
+
+n_trails = args.library_size
+seed_arr = np.random.randint(low=0, high=1000000, size=n_trails)
+
+#create search space for hyperparameter optimization
+space = OrderedDict([('lr', hp.choice('lr', [i*0.001 for i in range(1,4)])),
+                     ('num_cycles', hp.choice('num_cycles', range(3, 6))),
+                     ('num_epochs_per_cycle', hp.choice('num_epochs_per_cycle', range(3, 6))),
+                     ('arch', hp.choice('arch', [models.densenet201,  models.densenet169,
+                                                 models.resnet152, models.resnet101,
+                                                 models.vgg])),
+                     ('img_mix_enable', hp.choice('img_mix_enable', [True, False])),
+                     ('v_flip', hp.choice('v_flip', [True, False])),
+                     ('h_flip', hp.choice('h_flip', [True, False])),
+                     ('degrees', hp.choice('degrees', range(1, 90))),
+                     ('contrast', hp.uniform('contrast', 0.0, 0.3)),
+                     ('hue', hp.uniform('hue', 0.0, 0.3)),
+                     ('brightness', hp.uniform('brightness', 0.0, 0.3)),
+                     ('val_img_size', hp.choice('val_img_size', range(224, 512, 24))),
+                     ])
+
+trials = Trials()
+
+idx = 0
+if not os.path.exists(args.library_path):
+    os.mkdir(args.library_path)
+
+#use tpe algorithm in hyperopt to generate a library of differnet models
+best = fmin(fn=score,space=space,algo=tpe.suggest,max_evals=n_trails,trials=trials)
+
+
+np.save(os.path.join(args.library_path, 'scores.npy'), np.array(trails_sc_arr))
+#np.save(os.path.join('D:/datasets', 'scores.npy'), np.array(trails_sc_arr))
+
+#from multiprocessing import Process
+#import os
+#if __name__=="__main__":
+
+#    best = fmin(fn=score, space=space, algo=tpe.suggest, max_evals=n_trails, trials=trials)
+#    p = Process(best)
+#    print(p)
+
+
+
+
+
+
+
+
+
+
+
+
+
+
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/prepare_dataset.py" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/prepare_dataset.py"
new file mode 100644
index 0000000..f88d16b
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/prepare_dataset.py"
@@ -0,0 +1,38 @@
+import numpy as np
+import os
+import argparse
+from utils import read_train_data, read_test_data
+
+parser = argparse.ArgumentParser(description='Data preperation')
+parser.add_argument('--train_data_path', help='path', default='C:/Users/x2/Desktop/pq/wheat rust/train', type=str)
+parser.add_argument('--test_data_path', help='path', default='C:/Users/x2/Desktop/pq/wheat rust/test', type=str)
+parser.add_argument('--save_path', help='save', default='C:/Users/x2/Desktop/pq/wheat rust/savepath', type=str)
+args = parser.parse_args()
+
+#read training data
+train_imgs, train_gts = read_train_data(args.train_data_path)
+
+#remove dublicate training imgs
+idx_to_rmv = []
+for i in range(len(train_imgs)-1):
+    for j in range(i+1, len(train_imgs)):
+        if np.all(train_imgs[i] == train_imgs[j]):
+            idx_to_rmv.append(i)
+            if train_gts[i] != train_gts[j]:
+                idx_to_rmv.append(j)
+
+idx = [i for i in range(len(train_imgs)) if not(i in idx_to_rmv)]
+print('unique train imgs:',len(idx))
+
+#save unique training imgs
+np.save(os.path.join(args.save_path, 'unique_train_imgs_rot_fixed'), np.array(train_imgs)[idx])
+np.save(os.path.join(args.save_path, 'unique_train_gts_rot_fixed'), np.array(train_gts)[idx])
+
+#read test data
+test_imgs, test_gts, ids = read_test_data(args.test_data_path)
+
+#save test data
+np.save(os.path.join(args.save_path, 'test_imgs_rot_fixed'), np.array(test_imgs))
+np.save(os.path.join(args.save_path, 'test_gts'), np.array(test_gts))
+np.save(os.path.join(args.save_path, 'ids'), np.array(ids))
+
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/requirements.txt" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/requirements.txt"
new file mode 100644
index 0000000..3ada6af
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/requirements.txt"
@@ -0,0 +1,8 @@
+hyperopt==0.2.3
+pandas==1.0.1
+torch==1.4.0
+scipy==1.4.1
+torchvision==0.5.0
+numpy==1.18.1
+Pillow==7.0.0
+scikit_learn==0.22.2.post1
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/utils.py" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/utils.py"
new file mode 100644
index 0000000..ec9737d
--- /dev/null
+++ "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/utils.py"
@@ -0,0 +1,163 @@
+import torch
+import torch.nn as nn
+import torch.optim as optim
+from torch.optim import lr_scheduler
+import numpy as np
+import torchvision
+from torchvision import datasets, models, transforms
+#import matplotlib.pyplot as plt
+import time
+import os
+import copy
+import torch.nn.functional as F
+from PIL import Image, ExifTags
+
+def train_model_snapshot(model, criterion, lr, dataloaders, dataset_sizes, device, num_cycles, num_epochs_per_cycle):
+    since = time.time()
+
+    best_model_wts = copy.deepcopy(model.state_dict())
+    best_acc = 0.0
+    best_loss = 1000000.0
+    model_w_arr = []
+    prob = torch.zeros((dataset_sizes['val'], 3), dtype = torch.float32).to(device)
+    lbl = torch.zeros((dataset_sizes['val'],), dtype = torch.long).to(device)
+    for cycle in range(num_cycles):
+        optimizer = optim.SGD(model.parameters(), lr=lr, momentum=0.9)#, weight_decay = 0.0005)
+        scheduler = lr_scheduler.CosineAnnealingLR(optimizer, num_epochs_per_cycle*len(dataloaders['train']))
+        for epoch in range(num_epochs_per_cycle):
+            #print('Cycle {}: Epoch {}/{}'.format(cycle, epoch, num_epochs_per_cycle - 1))
+            #print('-' * 10)
+
+            # Each epoch has a training and validation phase
+            for phase in ['train', 'val']:
+                if phase == 'train':
+                    model.train()  # Set model to training mode
+                else:
+                    model.eval()   # Set model to evaluate mode
+
+                running_loss = 0.0
+                running_corrects = 0
+                idx = 0
+                # Iterate over data.
+                for inputs, labels in dataloaders[phase]:
+                    inputs = inputs.to(device)
+                    labels = labels.to(device)
+
+                    # zero the parameter gradients
+                    optimizer.zero_grad()
+
+                    # forward
+                    # track history if only in train
+                    with torch.set_grad_enabled(phase == 'train'):
+                        outputs = model(inputs)
+                        _, preds = torch.max(outputs, 1)
+                        if (epoch == num_epochs_per_cycle-1) and (phase == 'val'):
+                            prob[idx:idx+inputs.shape[0]] += F.softmax(outputs, dim = 1)
+                            lbl[idx:idx+inputs.shape[0]] = labels
+                            idx += inputs.shape[0]
+                        loss = criterion(outputs, labels)
+                        # backward + optimize only if in training phase
+                        if phase == 'train':
+                            loss.backward()
+                            optimizer.step()
+                            scheduler.step()
+                            #print(optimizer.param_groups[0]['lr'])
+                    
+                    # statistics
+                    running_loss += loss.item() * inputs.size(0)
+                    running_corrects += torch.sum(preds == labels.data)
+
+                epoch_loss = running_loss / dataset_sizes[phase]
+                epoch_acc = running_corrects.double() / dataset_sizes[phase]
+
+                #print('{} Loss: {:.4f} Acc: {:.4f}'.format(
+                #    phase, epoch_loss, epoch_acc))
+
+                # deep copy the model
+                if phase == 'val' and epoch_loss < best_loss:
+                    best_loss = epoch_loss
+                    best_model_wts = copy.deepcopy(model.state_dict())
+            #print()
+        model_w_arr.append(copy.deepcopy(model.state_dict()))
+
+    prob /= num_cycles
+    ensemble_loss = F.nll_loss(torch.log(prob), lbl)  
+    ensemble_loss = ensemble_loss.item()
+    time_elapsed = time.time() - since
+    #print('Training complete in {:.0f}m {:.0f}s'.format(
+    #    time_elapsed // 60, time_elapsed % 60))
+    #print('Ensemble Loss : {:4f}, Best val Loss: {:4f}'.format(ensemble_loss, best_loss))
+
+    # load best model weights
+    model_arr =[]
+    for weights in model_w_arr:
+        model.load_state_dict(weights)   
+        model_arr.append(model) 
+    return model_arr, ensemble_loss, best_loss, prob
+
+def test(models_arr, loader, device):
+    res = np.zeros((610, 3), dtype = np.float32)
+    for model in models_arr:
+        model.eval()
+        res_arr = []
+        for inputs, _ in loader:
+            inputs = inputs.to(device)
+            # forward
+            # track history if only in train
+            with torch.set_grad_enabled(False):
+                outputs = F.softmax(model(inputs), dim = 1)    
+                res_arr.append(outputs.detach().cpu().numpy())
+        res_arr = np.concatenate(res_arr, axis = 0)
+        res += res_arr
+    return res / len(models_arr)
+
+def read_train_data(p):
+    imgs = []
+    labels = []
+    for i, lbl in enumerate(os.listdir(p)):
+        for fname in os.listdir(os.path.join(p, lbl)):
+            #read image
+            img = Image.open(os.path.join(p, lbl, fname))
+            #rotate image to original view
+            try:
+                exif=dict((ExifTags.TAGS[k], v) for k, v in img._getexif().items() if k in ExifTags.TAGS)
+                if exif['Orientation'] == 3:
+                    img=img.rotate(180, expand=True)
+                elif exif['Orientation'] == 6:
+                    img=img.rotate(270, expand=True)
+                elif exif['Orientation'] == 8:
+                    img=img.rotate(90, expand=True)
+            except:
+                pass
+            #resize all images to the same size
+            img = np.array(img.convert('RGB').resize((512,512), Image.ANTIALIAS))
+            imgs.append(img)
+            labels.append(i)
+    return imgs, labels
+
+def read_test_data(p):
+    imgs = []
+    labels = []
+    ids = []
+    for fname in os.listdir(p):
+        #read image
+        img = Image.open(os.path.join(p, fname))
+        #rotate image to original view
+        try:
+            if not('DMWVNR' in fname):
+                exif=dict((ExifTags.TAGS[k], v) for k, v in img._getexif().items() if k in ExifTags.TAGS)
+                if exif['Orientation'] == 3:
+                    img=img.rotate(180, expand=True)
+                elif exif['Orientation'] == 6:
+                    img=img.rotate(270, expand=True)
+                elif exif['Orientation'] == 8:
+                    img=img.rotate(90, expand=True)
+        except:
+            pass
+        #resize all images to the same size
+        img = img.convert('RGB').resize((512,512), Image.ANTIALIAS)
+        imgs.append(np.array(img.copy()))
+        labels.append(0)
+        ids.append(fname.split('.')[0])
+        img.close()
+    return imgs, labels, ids
diff --git "a/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/utils.pyc" "b/code/2022_autumn/\346\275\230\345\200\251-\346\275\230\345\200\251-\345\237\272\344\272\216WR-EL\346\250\241\345\236\213\347\232\204\345\260\217\351\272\246\351\224\210\347\227\205\350\257\206\345\210\253\347\240\224\347\251\266/utils.pyc"
new file mode 100644
index 0000000000000000000000000000000000000000..863636974968572fee27f88e1d9430217988b249
GIT binary patch
literal 5586
zcmcIoOK&5`5w0Fe)cgJ3r?iF>8Jk@=$zcJ*u&u1T#sao%&)Sy3NXv{kLu#nu3^hI6
zr5rK{;5{V=ACgNh$tAZSr{oXh5ag2lfgq<`f&e)=H(!$Pt0tvgV~<{uQ|#%k?y9b?
z`l?!`|ISQS?4KXjWbo(V_boj3pCBQ=MiP<AdD}?SlZrPSDoJxhDkJ*pNm7#5s8mM9
z3}%+4GB#u#m&&-dOOlLBYeFg$(wdaYq_n1_GIc&@S}N1SIWtn38MbGoGCOR~QH$oh
zR2H;nM3O~mElFibTFX*d##*ED1ZuBHSk}g>gk#!RlW<%c>)hId?Fno%{vZDKwhfxa
zG)R(HZS6P;J4uvpm)JWF5(Y(Jqrw(tytJ|~N^CL7!8}N9J<D5mo5m|fLGHJLMx?Xe
z?#A_xf`+YeiI?!q<GG8+{uV12B2AIcMf&4XsA+rXDY55sBmD{L*q@Zplm3)6z&0(#
zh%`r~2a}g%BOH-B&SF%1%a~)DW6~Sd$urNg*3FP_{8_$P>Hf+qCZxytvU~V*uRkZ)
zS#wIL%rupO6?>3_v*cv2KQG0MG-stZFTF7-=A>7Wa6&e)!1Cu%c0sx?>zZS3%{8yT
zDBW2PJW#vgL5z6)CFw13+KDImhM}-|3Nz`XN?y<xT3}hiN$KM-L06^QG!TcW8@q=O
z%>l%%Nq=2>YjUyqHR-KC^@`DXX)fx5G~iQ`q4F1h7fQzhH2Q|xZCQF75>5$rLkX9p
ze_48q+$}oSsqsr5wJ)jmXwOJ*RliU}HQiXNm1L_}l_r$IT5xJCXUv_^o6xNX<9(g~
zCVS4O-@#T>Xk|DXGicW<Pg$($^{z-0{jeB!x8Yg~!xrZ5BMd+Mw1>Tb^Ag9o{AJpa
zCv^vg?XBo`u`u=rlU?P0Fn9;%K7GsVbMlB9d!>Js+|6sc&am&AYUHf&mGr<eB_|@e
zw&AqT_3G2F1mFGBf60DtTAJ6N!vz)0=`S!2yO!m(KD-&0+*D<^1ifH@4QZ6?(z`y`
z9@^(2)f<=f{-#FFmUOoTl6DU_rGG=63-kupFug4*fa6z*?b1FvupYJnA@vbJ#0;W!
z55edDwDHp?w|9yXIvxc{%wLWAusCZ+1wkOmHTYXuD@uz_tMRwL^6x)yDjNVJPGhbX
zE97u~7`L`x!Y`Z?ByL68sAcUlhdoi$))xYfErKGdhH<TM%Vl}(SX-Y(d1i|-u+)<*
zD8Bza)uqUEtt3ku$_2OrHW8(=HYDjb9=x-wSk?~m0J5Xp`WTmD0pxCG<22ey(+GP_
z8}Ir$BZ*R9(T@s8Far=ogc?*Ec@P$aIiRJ!RHHOFN}^CFwt}t-x5ojDMFZSZm2Flp
zI7Xa23Th{hgFIBlZ4tG#5f@R*^~9e=x|Z#@4nniSa+q}>NSi^s9i^czmnKOy$*gsX
zXAR&LjwlPl>V+wt^=TD_cHK=ahsZq6QlCl{Qy`SYnXnKL@H{D0EmU%uA&efywMea*
zc3RciSuMfHGI;qKwX@o>t+u1wjW|=pDQ3)q<p``*RiOG9nMQ5FuAo*^jdWbMLS+Vw
z)gaH+F6}%!a(g^VbYT^)6ZNQ}W3F%K*-g1rCvCNlnc<iM^G=%Lu$|d>mcx!(mgi9o
zy3uRnu&cI?({^W&7?a?LiIGm2bqej~8CTu<c@$bzRrf&uQQ=Mu9mu_MyR>eU+NgyC
zb>q~Fp;3*Jpl#t!oZxIg3sB_aC~h1Vb{kDmV(+1K_qBtMEl%uUCu|||ECR5|@`K&T
zo)lU8VE;G@3cq8EyGdYeT#su(fkVI53gCMAL8pildmDksGt}4}oRCeS#&M?JWa|`Z
zro9zY*3V6|ZI;ZI*)&U|OJ>^4c$1!vcxz_P%#|ibR?)j|u9#b(%Vy4-Hk0@)ja)UG
z-mF<JuVbA}Z_}GJm%%w-E}I#A77VhA2)9o2-Nj@70YcbAhDU`}(l$Zd^h?tH1F>!Q
z@ORP&WFz&JG)(}-`lHgdOfS2KCz|8R(ns-7W<`f&4>;_XiEB+zP2*p@DF!atM5zD{
ztrk#!p!lGqah4rTByKPu%?%5OCV<*O>%Qiq)7NF6sv`T1vZ_Ey!PzKikkLnxbxLx2
z9jK2@ps3K9E3Qb}i;sY;3bsCv+Ghkf03`$k-Pp$X6)nOpc(>`iii2s}C!G0<B$r8+
zNI33a1yK*Z7bG_FH`%vE@(Kr7@P$zk0A^|X8oZi7fp~^jP>oua!e@YTMdV+?yH=G7
z0qUMsBb2K+=<vUbrY@C7*1yFU?g%er^m(TKJ(9OUUW_kmngOyLL4sWe5qRMY{I3kJ
zoh!}5=Uy=@X4cdgy~H7ENX=}99=eZF>Phz#gA+Z^^Y9ydcMspx%B)1U>%MB-Fsa@c
zaI)|rByVM%rvnT;?O6_%vT#^+#j{`2OehDRU^2WEl_(r(f^^pjrTc=zLtX3&l0PXU
z>SuLR92I@7Ea?a5d}S@){ypTwY-n=MhoWJy-njBD7z)OEC}W@IL#36Di7_TM@+t5W
zUQb3`SG3S-&awLA22tM*BMX7oTp$~CH;6aB=vK|?dnk==AP7+5c&zUSB}k1&Q5rm!
zM{x%PraZi*@e3<?jAsW0CuPYU6kr)68*Zsj%qc2$!IE5TPlLs$Gb<{8wDZmbci~|X
zbNf|I&nHSBy%0aN#Ie^KQ5tpI`5QmRi=|{B4#VpGJVyDXV61T;{}G)UjbIPbFlI=_
zv)D;Y4~=n4bz$Yd^MhEhki!DD+$%$0vqps2jgSRgwKgc45@d!j9;DKsI98K1vpDrN
zT1Ya~6xyxT7ll?)OGUWjDI%~!aaBN`u}E`9%{!6*Atlq!2=L%c!yTaU2>Dp|gV1M9
zLu>5peRO~4-TONa6nnNQLl0n31GO5~-yQbF{%dS<-FnKxjouNr8Y-7sy^urGgqKGs
zpIkO%O_W2h{T7I5_Pb$T_O`r|xnbskEAF{sN@m3SHc)28p$ZX&s^7(9&%TB#L>OeR
zePRm|n@a_|hxZKd1aJ-*;efjfKLsWLlC$3vBmgfbW0HSKe*q#q0T1F2*#@%|Ifj^l
zn=LrpU7FXr-zcL7IPw!f+~ud!FVFeL6bt?`<U`7RKA*eTVjcuzJ;27Ze7vAtzy`Ud
z+!c|R?htoyY!m4dHyor#pv5e6A|QmEdN+2cr{Dpo(JK*e#!Rj-%%X(2AE73g(R;b!
zv?cD}T!0I>>m`5zli;3u;_iF<pX~WB6l(kfNK=^c57F`;l2qA;=Nm8Eo`slSLc#!I
z{2!rxfh&H%*XJOLf5ZU?B4}wa-u^nWEMvqENg@(rg?~&^CxJ7I&+5c)lAMs>ziQ&Q
zNHUT($tej-4#f)}4kAA5V2yCwPGWcCtD$<1wi!M?<IZO;aG;>^A6mb|o<}5)L0*Vm
z^}gazfHxF^z|SDdeh0*XK-sK%*8l?3&%nTSg@J1*d}quh6d2nZ;MXcluRAErYsa_K
z)Rk<h<^k8IR~Ubl+!Wz|Orp5TT%kOVPy^z&#?<CBkNC8Q&wGPA8?_~?zM+)c^-fx2
zv5)&J%jt_(tpOEJKe+#{J5OSj|31kNNxCFI0l6@*w!X<L$2gXM9W^BW6Zr;rWv_uu
Z^8x`-oyO-1t_}Pw!qyWHCgvyB{}0h`YKQ;;

literal 0
HcmV?d00001

-- 
Gitee