From 46ca7c6dfdb36c75ee21adf54013267c4a933cee Mon Sep 17 00:00:00 2001
From: Anton Lydike <me@antonlydike.de>
Date: Fri, 22 Nov 2024 09:26:24 +0000
Subject: [PATCH] final changes

---
 report/.gitignore                             |   2 +
 .../VGG38_default/result_outputs/summary.csv  | 101 ++++++++++++++++++
 .../result_outputs/test_summary.csv           |   2 +
 report/figures/gradplot_38.pdf                | Bin 19455 -> 17627 bytes
 report/mlp-cw2-questions.tex                  |  44 +++++---
 report/refs.bib                               |  21 ++++
 6 files changed, 157 insertions(+), 13 deletions(-)
 create mode 100644 report/VGG38_default/result_outputs/summary.csv
 create mode 100644 report/VGG38_default/result_outputs/test_summary.csv

diff --git a/report/.gitignore b/report/.gitignore
index cdee849..b0253fa 100644
--- a/report/.gitignore
+++ b/report/.gitignore
@@ -1,2 +1,4 @@
 *.fls
 *.fdb_latexmk
+s2759177/
+*.zip
diff --git a/report/VGG38_default/result_outputs/summary.csv b/report/VGG38_default/result_outputs/summary.csv
new file mode 100644
index 0000000..4c27ce8
--- /dev/null
+++ b/report/VGG38_default/result_outputs/summary.csv
@@ -0,0 +1,101 @@
+train_acc,train_loss,val_acc,val_loss
+0.009600000000000001,4.609349,0.0104,4.6072426
+0.009326315789473684,4.6068563,0.0092,4.606588
+0.009747368421052631,4.6062207,0.0084,4.606326
+0.009621052631578947,4.6059957,0.0076,4.6067405
+0.009873684210526314,4.605887,0.0076,4.6068487
+0.009136842105263157,4.605854,0.008,4.6074386
+0.009536842105263158,4.605795,0.007200000000000001,4.6064863
+0.009578947368421051,4.6057415,0.006400000000000001,4.6065035
+0.009410526315789473,4.6058245,0.0076,4.606772
+0.009094736842105263,4.6057224,0.007600000000000001,4.6064925
+0.00911578947368421,4.605707,0.007200000000000001,4.6067533
+0.009852631578947368,4.605685,0.007200000000000001,4.6068745
+0.01031578947368421,4.6056952,0.0072,4.6067533
+0.009789473684210527,4.6057863,0.0072,4.6070247
+0.01031578947368421,4.6056023,0.0064,4.607134
+0.010189473684210526,4.605698,0.0064,4.606934
+0.009957894736842107,4.605643,0.006400000000000001,4.6068535
+0.009452631578947369,4.605595,0.0064,4.6070676
+0.009368421052631578,4.6057224,0.008,4.6070356
+0.010210526315789474,4.6056094,0.009600000000000001,4.6070833
+0.009557894736842105,4.6056895,0.0076,4.6069493
+0.009600000000000001,4.605709,0.008400000000000001,4.60693
+0.00985263157894737,4.6055284,0.0084,4.6068263
+0.009200000000000002,4.60564,0.0076,4.6071053
+0.009031578947368422,4.6056323,0.008400000000000001,4.606731
+0.009663157894736842,4.60559,0.0068,4.6069546
+0.008484210526315789,4.605676,0.009600000000000001,4.6063976
+0.0096,4.605595,0.011200000000000002,4.6067076
+0.00951578947368421,4.605619,0.0096,4.6068506
+0.009242105263157895,4.6056657,0.0072,4.6067576
+0.009326315789473684,4.6055913,0.012,4.6070724
+0.01023157894736842,4.605646,0.012000000000000002,4.6066885
+0.009494736842105262,4.605563,0.0072,4.6067305
+0.009810526315789474,4.6055746,0.007200000000000001,4.6067824
+0.010147368421052632,4.605596,0.0072,4.607214
+0.009536842105263156,4.6055007,0.007200000000000001,4.607186
+0.009452631578947369,4.605547,0.0072,4.607297
+0.009578947368421055,4.6055694,0.0072,4.607313
+0.009410526315789475,4.6055374,0.0072,4.60726
+0.00985263157894737,4.605587,0.0072,4.6072307
+0.009389473684210526,4.605559,0.0072,4.607227
+0.009852631578947368,4.6055884,0.008,4.6070976
+0.008968421052631579,4.6055803,0.008,4.607156
+0.009536842105263158,4.605502,0.0076,4.6073594
+0.009410526315789473,4.6055517,0.008,4.607176
+0.01,4.6055126,0.006400000000000001,4.606937
+0.009915789473684213,4.6055126,0.008,4.607185
+0.009305263157894737,4.605594,0.0064,4.606834
+0.009326315789473684,4.6054907,0.008,4.6070714
+0.009094736842105263,4.6055007,0.0076,4.6068645
+0.009052631578947368,4.6055903,0.008400000000000001,4.606755
+0.010294736842105263,4.605449,0.008,4.6068816
+0.009578947368421055,4.6054883,0.0064,4.6067166
+0.009452631578947369,4.60552,0.01,4.6066008
+0.008821052631578948,4.6054573,0.009600000000000001,4.6065955
+0.008968421052631579,4.605544,0.008,4.6063676
+0.010147368421052632,4.605516,0.0064,4.6068606
+0.009600000000000001,4.6054597,0.0096,4.6072354
+0.01008421052631579,4.605526,0.0076,4.6074166
+0.010126315789473685,4.6054554,0.0076,4.6074657
+0.009705263157894736,4.6054635,0.0088,4.607237
+0.009726315789473684,4.605516,0.007200000000000001,4.606978
+0.009894736842105262,4.6054883,0.0072,4.607135
+0.009663157894736842,4.605501,0.007200000000000001,4.607015
+0.00976842105263158,4.605536,0.008,4.6073785
+0.009473684210526316,4.6055303,0.009600000000000001,4.6070166
+0.009347368421052632,4.6054993,0.0076,4.607084
+0.009178947368421054,4.6054535,0.0084,4.6070604
+0.008842105263157892,4.605507,0.0076,4.6069884
+0.009726315789473684,4.6055107,0.007599999999999999,4.6069903
+0.009536842105263156,4.6054244,0.0084,4.6070695
+0.009452631578947369,4.605474,0.0072,4.607035
+0.009621052631578949,4.605444,0.0076,4.6071277
+0.010084210526315791,4.6054263,0.0076,4.6071534
+0.009326315789473686,4.605477,0.0088,4.607115
+0.009010526315789472,4.60548,0.0076,4.6072206
+0.010042105263157897,4.605475,0.0076,4.607185
+0.00976842105263158,4.6054463,0.008400000000000001,4.6071196
+0.01,4.605421,0.008,4.6069384
+0.009536842105263156,4.605482,0.008,4.607035
+0.009915789473684213,4.6054354,0.008,4.6071534
+0.010042105263157894,4.6054177,0.007200000000000001,4.607074
+0.009242105263157895,4.605473,0.0072,4.606825
+0.009726315789473684,4.6054006,0.0072,4.606701
+0.009684210526315788,4.6054583,0.0104,4.606925
+0.009642105263157895,4.6054606,0.0104,4.6068645
+0.00936842105263158,4.605405,0.0076,4.606976
+0.009263157894736843,4.605455,0.0076,4.606981
+0.00905263157894737,4.6054463,0.0092,4.6070757
+0.009915789473684213,4.605465,0.0068000000000000005,4.607151
+0.009389473684210526,4.605481,0.008400000000000001,4.606995
+0.009789473684210527,4.605436,0.0068000000000000005,4.6071105
+0.010273684210526315,4.605466,0.007200000000000001,4.606909
+0.009789473684210527,4.605443,0.0072,4.6066866
+0.009957894736842107,4.6053886,0.0076,4.606541
+0.010168421052631578,4.605481,0.006400000000000001,4.606732
+0.009242105263157894,4.605444,0.006400000000000001,4.606939
+0.009621052631578949,4.6054454,0.008,4.606915
+0.00976842105263158,4.60547,0.0076,4.6068935
+0.009873684210526316,4.6055245,0.0064,4.6072345
diff --git a/report/VGG38_default/result_outputs/test_summary.csv b/report/VGG38_default/result_outputs/test_summary.csv
new file mode 100644
index 0000000..ee4a638
--- /dev/null
+++ b/report/VGG38_default/result_outputs/test_summary.csv
@@ -0,0 +1,2 @@
+test_acc,test_loss
+0.01,4.6053004
diff --git a/report/figures/gradplot_38.pdf b/report/figures/gradplot_38.pdf
index 98383ffd50e972fe156073362dd8c2801f4bb6fb..b909a61e2d3848cb6ad2c94d0c62abce2d40d876 100644
GIT binary patch
delta 7127
zcmZuzc|4SD_fM!WsVo_Uv1SQ#&zNBlMOntar0mNu3Q^W<jf5<tFt)6ZeGd(#gtAPw
zXe32iWXqN%d+2wMdf(@*e)G?q&$+I%eb4ut>;92K_b!UADu#}rM;d-`Bx3%X)Wk&r
z$jJB6i^m_7T0K%2EIno=lXRkjQ}jMZNA$UkwM%hiA+6l%Gj%$Z*h<rc+Bt5>BI|lp
zJ#`QN_QvK~O7FcNo9pS{vUhv~<hGY*ACeS*kanIZbbV8u>b*r)R4r7M)lay-x#HNs
z5x=uCy(EjvTv#S;5!NdS8Tw5b{K0x28?iQP>thKA7R!m9KXUeE9PRwLwKYBAoELjw
zV{Wudbw{CQbyeB5RYV%HotUV#xb<~&Ywg?2x4cT9oio8ZS@)9fW!#&i-<~6xc61$<
zw(fkA7sFy1v@qpSB+%pGk(MN)MLnUdmK?x>)y(ng)&CfGTY#YNq0Q^FgQY4xVFOY>
z^`LEb7H8kQl0RRh`LV=H*>)RoF=%CLd2}NtrSivi_L@Ofga7vT^^~optyde<-(NjE
zUnCI0Qh(@d5F1spWV^qpqrX6xTo9plL($IdK#!%5@98shqZ)UwUk@f~2}X(3db>;y
zIZm^c3>vHFYw>9jWUv8IVo5~9dxF&7$`8ul->_>p-Qm@We>@Tkv7a7vobD?IuL@7!
z^3IAd=qt1o_6gI#Vg*s}kKm^*>}QOpk{|FWQV&WbGE=7(li!IaDqew0oaKwUt!U)!
zet(7OwR@RpqEL<+QIPeabCKb5qPTSUeH)DBuR{WOI-r)Gm>)GT^FnP(STiwRt?M5(
zNk#d=%`lgTuh}{co%3}S2ceI{74c`xrz3lX8@~)Y&ajuhH`5W<$y}6t6m-X(tMeEp
zH-kgtP;Xj^7pcOa*w<1zr#<C?MUgoD$+ZBQsZqaOd)w7;=L>It`6U@ru~WAQlcQ!i
zv+=@`TK3QBRDJ%FvFXcZ^W%Egzxc5DbVo)T8>$jd2u%IQaqdx@fT^WS=N?6`7ii%p
zezsY)oq6!4>6n?xAO5R5<LPe<+(neH<cm(*!I#UvEQ8rrf!Us#Ca#^<Kt5Tz&7kba
z;IdTHP)*?^$egmDVOsz}-0QM5cJnrt!cqSIT3M;OW}4XsL5ZD(>z}HhAaeLG35q#s
zoPS_ajMLya_ygFFSeeT!F(@?Ck&bVVxuf4VDz`r({j4HFMnoj0I26@3+ElgD(GTL!
z;lHRY!D-z%PeVd3K1J`U>upM5kKO~79Kz7$+Atwm6~>Ar9`~0OYJA*WE=%`G9U~X0
zk2(g9o7Pt(8;ZorE!$&4W{*y%9vM%H^Sr`Tv};Q<T@o?9KQtp4IxN{Ae>I1-|46oG
z(7%@UaT?E41^gS-_>j4|Y}nu0lY7L%KaWC)ivAfipefqx$HIT=VT<Bu&oO8rSDWOk
zYmnlx7nLBZcY>wdtP$_~%+$q&W8q>v^PZC4ep+7n9!`mD8a>;T=7HmOV7ejE47$s$
zrs`^wAkKJ(J}0O_Q}lC!{jVl$LT3ujjY>UyU%L4+^`#-f)$Gtl=f9*O9?_R_U4d7>
zmFNeAJ11`N-IV125&nx&Czv>ESM0Cg9y0rg*JrTxuxiKsYpKN(nxrSTKVrAo_O4aQ
zUQUk8bhlu&+IKEYXKjMw21?|{E`X+vR$B3?3phD6jctV80WobC7|p0{(Ho*gjeas~
z7jZxFU(aB3hw5Rsw6}MWO3)Ej;~8%LpJUST&-HQ|ahYm-h};wYL)3%pV<t=^jx%i2
zAg$EQm&T_2CsFV7(3g)+x7WQ)(9G;#D8Kt_zzDC|u;zkJOK8X|PS61|x7|yB^u+uA
z|297@jJa<=QdwXv0<EcXu!5?VHlmTLcye%)!3~_wmlsZZEK0EYp#4NX0m3hh_rHQ$
z!+iq4xql1t?jKP)5^F%6NiBQN!ccJC^^MO0^Vv@VzI_(m-d<W+dF5wdLpZrTzfx~p
z`LJmC>-i0K&-<!h4kwAFXfU7s;ueG?bVzUehApR{D>(E6MlT!1sSG}jZt5z`=ym9|
zd`#WWlaq|EJWl5CWQgvG+JhbEA}_UPu%z4?ewTu7V5hnqSI_Bnoarlao6hYmFGyCg
zi4l)|(G*mtC*b#w3Pz77*N+gcETCsS&2iAWRpQ;HOYFiUvX7vl^?E!$pI%ZsS&B|j
z1N#cARw9wEE~d^OjMe!oh*kMSqm3hh=_>nK`7V>2BuEnO70m+$27~=)v(0s+gLOHK
zO0?W9{&Ci`6bK$$U%ZiZ^z>1ka50k<#k1;MS37+zhOPy`hs4h<#4E=TENd{HvxYc*
zUcIbwksikxV=!5v-DCkLS*-tE<Nk+VaID9aG#GidJJvsckaRn_*`3C-a`$2X%jbp=
z4%Esp4>zw(Z}FY<K*O^>buj-96F!zv$5%HtH@nC&p)xXo51MS70yjTzZmqw6H7Q$Z
z+w_C5EMKsLtvj6^M7Tt-mbCS{8J%QPAZ1#9S@K1jXI7t;jP2dQnAbeWz-k41ysXQj
z^J?yL5MEZY0u-vQ;bo^Uf`XejUba+FC(wLMp3H^1jQnoehC*_k)j<Xt<sr#jq7tBR
z10ZuT?ned!O)6vqWM{#nj8mC|G5*h}UmbnJJE!guV%|IJCy|x7>a_=Lbn#ZRMoxk8
z^-XWBuk$CCOsB7U9WcY=WqnW9n1~2@sQA_G2+f<>7~I7_eoL-xFByJf;T#yRR`)1v
zIDnhmG8B#p@vJhSv%TI<wI;betAEkT_jQ$p#p6!;lFNY0TZZNkc=lm?-IwQ@U&fk=
z61fwQ1lMM01YB%d(q3k+`~WVrf{nspBC31+%bNpU)s;}>s_P4=i~qCK*#l)`5xvsf
zq;hyfV^L2aKamcna-MP~0+1NsCV3|TzFKJ&+&HIFq*%BjE};@B_DB*}zU)JRys@?K
zUU=k6xlocqT;wQZl&`AN4W0Cv;t1tk2kZzCC^;6W?u0&8DkE0=G>b>Tht4tK&T+zp
z;R7MubcS@@3)QYqFPQr|$gDEkb+f_KzVejpn{*p#RT$6PZ?^}!2XFb0<FtPFmtoVr
z1?>#j+f0ej<6>!#+1Y1GQguE(U`-Ta7Y9_jBZ)p494S}XR$2p$uiTA^%DVTqFPxqL
zoPr!CGXjlSbEvE+Qp9sVMnL8Dkx}ke;#}ns<j|W)?m5@i^?(O{LBZibR!19gH385N
zcgnLIb0B*9SBp2$fk-?})w{|z+7_nhpfR7<60=C;GCAJ;rQ9?l^{E0-s-Fs!)t`K)
zRHG_PJlbC#%W|AD&a9TimI2+fj}S6A(Hf>`4}E$~Ot<v7gN9ezO>N@-mW6_MH%Hd*
zQ*TtqyT3cUB1A2Veev<OZ+lCK?5D0qScl&}ZHT?{EhxTu2uae;YY2tn5tnMthl=#V
zDHYs#OUA|xcEn!#6Bs6eePDvCK!le=YHliZhY23>4Y!p5+{kD47M6t)aPO<MtM%+4
zP<o16nt`!yMl0tB;JR+PFg}C7wEsPT0&bj!So)pMzR&-z<SZ8A(A*cBb1zQ!x|UW_
zWX>YgDRLYOh3$)^bKr4@kY*3Aq@~6*B*awvP>8F{IJp++3wXq7BK~k5cPn%s&?Vuz
z<e+BEHD7p&$Zt@{gC{KCpjkiS1CpZ>kFiT&#Ecy#NI?w?UmYsXU~fkyP8T$!oWP12
zJ0u@k4;Z&T$xF&ThK+XSwk2`OA033k_Yogu6XykzVn?7}=Lb{A^L<pjO-ruEe{mZa
zV!>r<PVs@2DaR3IO}M<Ai6dL7N_NC(iNW#m1lEn3+DHOXmR7kV(HZ*6@<ZT9<Qu>7
z=dBgQ>i)L;@%x*%ZRT%DzkEsy<A1oi&w3o^RwVE${Z6$G<>F)KSj+RDsPdBPXBhcG
zC$rrfA2@M!7pCrcHf9L40C+FPEYAFg=etT=6T<FwNu4m+3lgPQSitK%<|8Mr=79HT
zz7wJR486w8*xreI&!F&$>&cc#ousdBk9_Q)E&b1l8%fMKwM-yACdlWg^5dE_kcK21
zKEoUPTlWBmluQFIvM@=y6Z7Vg71)J~cEw|dCFp=t-P@+0%=ZBeoY#_16P*s@xLYsm
z1Cmd7PY`8&+H4;nWDUZx{?h1!fnZH^IKnDG@kLM&#Ogytz*h-p+HScnGYs7<{?xG4
z_DpV<7jn>7>5Jo~xKzgY#_RnZ1(~<qAY76Fi%@+T&TI<0DdrH}rPq9JNyt1vun(Y_
zsNz!u)IS#I%?DOjs^3m(ucx*@522lq<;@a$B%7?XPYEOy-zk1dPXHm=Z#*9Ok@H%^
zxyE#<1uZtt2C|N7#AE&hZ04FWh-=UTTHAUrJKw^VCnltndFy`Id#FtC8P@`1&|pVO
z(6I{hav?+D<Mn5l-IC&B2Jg<V^g$b<NeOYj%B_qbH$$K>{2>R^17&S-eNjQ>rt>Ec
zQ9^)hDI9?k>?QSm-BYu7S`kOt4hCgJV7Om)Y3s3*UFRs?{G~P6)y^j-*DUz2CD4nC
zni#&_VGX0R=hr|K`epV7bA9xIi^Un@J5o6=p4to>QUb|*`kxx4p{JmRNk<#QB*gbE
z0}domVGK}fs7+O{(6e~;)Hec755~9a&LK?pdd?))S>sY2A?y-Yoh4t>wah_Vy3H%m
z3oml}-aN2UIrZ>De!Nq)1n0uuCvGh|#iQXbRuFen7gzkQ;CV-zt0vWWBH+V0mor$%
z4!R$>BgE?<BnJ))OgaA^u<kS3)<%6U1*n6Nx&F@BD1GrFlQR#vc6V08pv^pCAjeov
zt#8sC+{IvfWKBQce%P3IABt~*x_72YMH}h<)7q-tJu-0aI3>L}1vwzdTp~hSTp>5u
z$seO~<SA`}3>dCa(O9Kj@SmnO%<TV?u2bd|XQYtwg>|>~eZ;|f4ebKQWFj#3Ts(0t
zbT?KuV3ANq(4oYARuL<&-A4lnPf&?aB&V5n(i0=lVC!iRxl`KnXZ=x?XCkSE-?H}8
zMkpD`vWNIBxH&*ujw$m_M~+NX_+-L|nL~U0U2n@Kp%`oerOta{HCnX@xPx@-dmDq&
z>M_cxCeBmo-&}&bwGRl1Ub<qzdzik}ji9buDl7(%Wa^gG!XI&Tv`Q6>u@KaCJivY~
z6K?c|iG3y{s)RL_%9W-B3}97fsY>)juC&GjF-)R|_l046GkxqbpkI~S+ZBM}LcjQ{
z4nm`?@EZxvP!s8e7vWFJ&Ojvew+}5i(29G*9J**zc@6MsCA#LwPujn|j1nfOL-CBy
zGLIA{L^J6Zj%#0De*Nl--;kZAB<TK`ci)aE1MW;sWMXJF+v^7KsiJS4wQ<pFX^}pn
zPA}Y%U%k8DLV!SSl2xP*<(jlx9P^`}>v&a&qTmW{#9__~cmV^rF-~>wXmwO(K-FzC
zNRgyd(8G9Mdj~=$Jcl*)>gQraPHVS0Ieo2E(M}_<pe98khDq0)=sfM9@%=$_{THni
zr;*IH<u)f-+Nuvpwe|rZxW~N$pYRodF@bki=x2>C50|gtDRKAFLsy?Y6Vmr|A3l7D
z#xp1bwtlbZX7M<lRi)`1HpY<YC;Nzker+;@UKg*83$7AILA0fIs@taZ0{4oihaMMW
zu+Fm3E~UPP0}ZYf@9A+6FA>LBz-ZGn-Sj@7eXEbQm2U`;U)^DniM^#Nvjr_`82Gs|
zK;GyUY{`T<ZDb;wi*=~9*DVBih>S>oJh57^1i{rw^YITgnr7J(x?Obk>!r*-u`LXm
z>hr!L`RoquIsZP`cltFqJ5)8|Bu9eOvr*qSW7q|&OrRl9v9*?$)E3{BS(1@=l`D^t
zwpu3M&Y@0@Qm#nYKCc6AU&S38rNF`O20n?87D<);RM|D}D>mXKB4(1b^<16)5l{*v
zpKa{sAIVZ845vvg>8)`}P@h&U-WQy7m>7zPd@&EwW#UG!)yE`WFONYe5A>O9ZN?*C
z4H3=R8v?Xap7%&@8@&dsiXD~IMJ?@#Wc%7}K7b+>-cmt(PubsmKYp*Gugr(R?B(`s
z)dbHj;_q)+c6HsQXOY*?6u6Sh24bQdNAQL`hsyAM(+tNZnTt-U_azxdsfaVq&g=oB
zEZ|ltSjEBDBLn&@Iv%aLt{~u;Ytw1(vP;*7(eBe}n1`|NvP%6}p1jV%hOx~Q*z{OU
zUd4K>T)}~;>~rt_uo4$=Z3bLBotk&NX-!a`_3kcu3N<f#O4>Lj?+ULVFo1F1K}Wg4
zRjXPy6FFs5YI+=E@|hR<l*m%>CWOhVQU2VX(VG4;#y2v4R#*n?9>^ZAS!{tpL90BR
zSdzHDFW~)jm}|=ZCkepr6LZtC`$L}KNv0+{qF%qx46w`nL&99&N>0?>gHdjcq}#y6
z4hSSYYou)<0=+z59%l;iH2>_H7n!9s3O>WTA50P0K4~&AMo}0yeVWJy&fOm1`@==+
zy%g#Sn?E>EPxrvH*Am%MVn%#pOHP<D$L~DCt_28QV)(RfPU9szSykUSpw<<#0Z4oq
zgI_)f)@j%Mb=29ZrVGV@jub!F|8|>|NZ9>0vt;hb^y{lPTED3}@0j{<z#io>_a$Bb
zLILfIy`3`qf^ztr8L#5oR9g>r{mb+qL4$b}!Ef)99_ZJX4_aYjK%}-$>y@rV-Awo_
zLR3tjM=ZxV%*cOz-#PQb{M_>jfk(kN3w#(eW2%SVrg4#;v;5kwgs(v!o%9~H9Eb}^
zazdhD=8v+skJXW#@47+ibJY2O%<mX@(8y}TC~KDWT%hJnA=d1g1^e<iIi9Nyyx?k8
zj(e4;;v_&(@R0NYqCX$MgTjKa-c-2cCqwC|DV?vCtI9Mhj-l-VI9?W)7)z1i&YM#4
z8eL=n@6i)LSiWyT^OC=>SqPWcd=UJ|@)M{aqDY%0QLs(=SE64pXj7_&UOi);ZcHG2
zdeT==t9zKow_=r|v$?XAN7WdfEpo~{+G$W(w`JAnNTqD?wMog$mpl;shRU+#I7zq~
z5NN#ou@bQ$*U(Ru$4atuF4bpk2bgV6OfP>?^3iY3uDp(MLfD*C-M+T-d9l8z*Vd-#
zI$>x0b^0OOEP|qKFky3bdYOKEM<7hmfXM9R?&#(1>12PICV4oxFarn}97YpRC<L6$
zc5L|oX-KS*^jwUc)G91W&@<AGriVl7CpkAy3EGgJ6sg};pZxKrlQkT}z*Lz)N>07N
zatH_$IB{}4UeHT^oA-L|V7+-jLdW2b)JY)|qjlqA83#4%NZ#I@BMT>UT-nU48x}Dq
z^6q*y?iUEJV?gH6;exdhkBTRf1ljc=KF3%s@|){i9a=tHP?ckc3GUxD@J?c32k$tl
z4MRgBVKaO0Lr)~4(d-EG9{Rb19S1G!-jwhf#e_;F4lsG=nzfd8e<Du|_{ytGFMTQw
zUfjHBmSPu9)mgc4%{E*mwdAhv*95hI5j`cLZhq!`gp#yWZD8xP`;AT2F5`vmj<t2Z
z%0n-<tTKK9MJ68aWg(5pX_BaN93)lLW0Dk#oeV)o(~&knp+p{_BPGkTljao0NJ9$j
zq%n*uiHcz-IVgsZ3>4W(D~h*CpTYlRrDjs75(n8$*=#SHEdG+O6AX=l$)1ME129=l
zX3{4WWmXsfz+tlPFj)_n>=kCRq-q!=D;|bGf&DRt?b?tU&WN+XWG}&xaGH{tv~UJT
zxMGiUg30YFV6xsYIpF7g?~9&Jpq|!UV`f=zKM$B3^5=;UOiu2X<nHL?>4o#~1TX#y
z;|r5R|2*`A$;tne{9$qmG>KU@z{wNzgV|N9sO(aR{Nbz=9m!l7LgLi)<-;HVKwb`R
zA;J9X7$nv;PbWMx9EJi&#hS7g(MSLcK>NYmVa&S%fQBo8S^uE}05IP_bO_Y{qmu`#
z`STtIMN-vLU_fC|a8jU_$X*nNB&fqqK-19q{jQuG@-KTN=I=NtH0IBHa1=%!#Qisa
za5RAYGe<aD?x#P({}c^JqZIxG0*=O@|I*1L6#jB3k5T|T@_S%;1q|5i-*pNI;D3__
z<AXm1ImBNM7366Ip#A>a6}SQh_!BKS21(=N@Aoh$^k0ENVE!ruqd-9a?GU5zCxieT
z@i#@_C^*Q*-vR@0^k0DiIOcCU(3HEjgnz38*+m<o-yH(-$iIRB3W&dS2>4$;1`u+z
z!TH@ELjHGtxPiw25)S<35DCzR^LJw;0`pgbNIBG>2m(knxJmx7MI+_Wq)MG*N5So4
j?BU@J1FdLFMB5$j0Ru;3M2DRPjY7yXi;A8#&|v;Q6RkNf

delta 8911
zcmZu$2RxPk_jm0rt4Q2iwyZmFt!w6zQC3RG-kYrO2pN%eMRu}FSs7Vb*B+tFuVjUi
zm5eg}55C{;|M#ok^Lk$H=bn3>^Eu~z-shap^D-F;IZB3<MnM4MiVI>1VQaq(wz!R8
zTk+kA!e-uK<mN6<+}cbuG|sfM#M(T%MU8Ab{1u=d;D4372ww!TuS=$G`F!Dxp|PL3
zpMdVwJj~%v@{#|+>_Qgi*VeuZfC;ES+@1}`NguHTCK=!M)^@)Sk7vlWW*P8CO*hKY
z-^Q&Er-f$`UTFn(`!I)F)x*na!yQ?lIORUo4rnDlN;h^@(b@CgYTx)ZJzak!yK9>3
z@Au~CMs?n+wSl|LmJJv6;xf9H_g5G8<OW;@pulX;=-&6m8vVb-omY7G`^IZ~Z!pYY
zs$M#-+JAppAeYsM296qasDz$h%_hC7lI4wIy))tYlZo&U@chZ~tmZqGb`)!r9MXp5
z#^vs>&rS{(3}``*1~Z2D#wDU^1{yI33;WGqtB($1p>O>+rX~TC!*Xcmkze5U-ZI^Q
z)*Zp>A8ZlyZgVRY^OaW|28D0si)j)o-l2@g^KRVl*Vfof&*U36aG|4*WR+m{DCYQO
z_*y;JP7@gqBy<T!F<;WE=1)*&)akxIn*Z`jopS03ie9bl`rw0%n?Ou+4O)BYg~W8j
zZi$G{p|oRZo~gSsx0}-#XycikCSxQ6n|gf8St0MM^R#h%_ZZn%V@AS2D_o$Jcs`t2
z+7rUsK^(Jt$K6Xd8D*E67A8ucJ0`2h>5eLM+W^_o=<T4Yyap-uL0o!YSXchimnzw$
zZ}064M>TlEdx)rcY$S7ezs}RGT%(n6QKHlz=7`!ZLcOrdkFO@Zp_o{|Rhx01eJEK*
ziBzi!;rzTjK%nny>K9_h1gGdFkL~3xTC(lDpq<wuv1Rs?hE>-Je2tMAgwznIdCkql
zUzUJWuC%r&|2>(F=XMU`n3T_V`z(V;lNhlE(E|lvmC3w#eIl6LJRYa$KK?drXELf$
zpvt4u4)hRA6%AiJh}`J*!ry;BQI&N1BX1a@AIB60OQ(8ocBL;p(HolC5YpFZ^rN|_
z<Y;5IV|2RVV6!2(;|H>zCm-)nII$1htkQjJXN){(V@*;WlooL}_^F0?<&$6O`x><A
z0&e)07*E*7C!X@Fa!cBBzeOm-<*mZ5#USQ~wU}mQh@Bbxc>TNe%-21GKDVO3R(63w
z`~ZX4>M<u|QnyQ4^_!4xVV6`fdyvL-wqU!=Nq<~8BQXC}q}Iol&+7~<fS{*Yi57N?
zw`nxA7jB?bC@@(tT>mOXD~8&Njl4BJQ;KFLCOvH3RJrkHjKkzed4LoM?{a8K8JrmK
zm0o$j9CtXsvATn4zh3>V%&|agxYk-~#y#x5bt^(?S1*G$J*+8h@)Mq2$x1@s>X&Eg
zsMv61v0FlbqQ}zm1vythpoNhwAK!9o@^1|F3M@nG7dOrQr8?-+A2vmZetbOs@wpQi
zRt6ZB=fhGe$E}NNPqe(S{a1W^oZN5SF#P9H4c-_Yx6RU-p3pD4ccol~Tc|!xn)$Mo
zINk2qUPhHT<l~xz$!R5Sg-*Nxb4Ouvvnt6SaQa^Sw_E(aIHV+G-S`lH*c<Pq-(6^Z
zORs2R|DOcCd&q*nVkj7OZ%AZLcFnhfQX6Chm>}i1re;{5uN*ut0PcermbcC2kydwW
z$}^sq8~;4eFxEXtm?3`mSgqm?CfOr5P&l~DYR1fI^E82duK^EpQ~bDN&!>P{=@zKu
zxw6hrXAvcupvOfg9*6fq9-~EW{H2vayR6>KCPL9!^e@{A<ar{hxq2v@z-Z(4e9}^s
zdy4laswy3`w7!E`pm&Tpv*CX8*&DAfo~nRihG(|ohpLY93@J@8flSTI>BT}sTR>4&
zqnx|&r6z|S!P{G}{>jkGvL>Hr_=&$sE=<>>vHFkUe5vLeLE(Kp!8^KEQmz7oXCupe
zmscEyL24jCYIJd#8=0f8s1u*Kti{g~lcxWm(l+aAMXL4rJljzB?In-HqCvyMPP>S>
zjc%Y(L+jPQ;p&E|1px!+lmZJtN^%Dik@n=v;KlR+p~-XWawQJWbnP?}o+u}gWZn7F
zeqooAD^2q$cYm?{cZ+eYgM5(m9`7#YGMxi{WzL7tC4EC+tHphVCV8hpKKozqHB43<
zObHY^pk-~$r0H@Wbx0d64*-51cZ*~V%d5hIm?7EabEXVNv!x^AC04(-(m<|?G|dN%
zCLWZpk^<6UUndT>Hh;W$d$7BZwr0}MK3y{1=zqA{BQbEaxKPKr^~}`V<a6qXI0)m`
zG0|VF<n-KVGoF<nuZb%-`1WFS0N7>4H_APJE+=qjc0&fO<_>U-mEzBD;cAa^-fy+i
z&x+``bi^bPsm#~vy9!NKh`H!8vQ5ALZnBNNbpi7Ob~W;XHO0sq_YHkJ${TdPj|BC3
zj9ASzUg=m1P3H9G=U++Dlto&7l2!e@q|@(VysusB6E^~<a9+|*e(o49Jem4Xc+!uj
z)+afrhR;=zf87XRmC|4pMhaTatQ2<_Tir_>fB%}VqR1v+LnG|cf@(T`;PFgOVsoc>
z7tx}de!jqDc5hYwv)-!A6oh29vFOaY;cL5B8WYdzDL+b-nshgXwzfMKa2?IJA~{ry
zOt_)@-xul@OH4B5?(OV<9Tk~A+!f)2+Am~8Mxhq^58vAZ7y|c7-kN;(N~$X1SK9$2
z_zxx>wKWsd2IoelC#n{{(!SE%tF^P!HF6pnPc%r=5nuQq9IfeF`PEd{<jS~9&*;*I
zrY|#d1OFYOKnc7)%)!`%%-Om7g*+Y<KdBbFLo_J9p#SMZemx~>BGKT!SlXsQoTaV_
ziIyX%bRTfqt0f5J=QZHFBeK%?wo&wzNojqehwrT$>TYW#!j<pUudVe&l^KXsg6eq>
z!Z#PxU!-hl{`k<lk)LXx@2P8|N#oo-I$o-I`^9+H&rQFooW{oOS=I}RSz*|9>$%J%
z(%{ar>t$Yb#6BLgo9+P%T1}+ww_CI8>WliUV{Zeks#K5pci$(s1s9P89u+Ti*%qzs
zZEcn4NeQqY$qm9IjIudtmRUBLWrj{~mRPz4S?Dt5-?v>awzlQ?w%;XhhG?q2S>Rl;
zdpPd0ar45N-LqzJ9tj`iw@rm^f6bQ3gtsYd6#D4Na717mtz}G_EhrxX9UQS7iZl8i
zWenAx(I$L(XJZGi!#ap$TKX!f$Q)#n*us)EK5)I-)ZQ!f-~ZOvQ4<ww3EcT?QQy)r
zwjQsJEg!kNTB(HXjx>}fZ9$i&?>7a8vNiOYIj&%X(^fa#8whnEML{yuan1!0wYQh$
zG2=n6+Y5Re7reINxVOE4+53uEHBm|<!<A;zSjZG#Wku}Lhg+W>@#1r4wKsfnfd-GC
zZe<e8u{0dUZ;-LgnhgiG8y3AZunql$6?}2=UX82At(=LQvAdVTU=Ok2QybT1@-s$q
zwB%xnvC(uY)vAb6D0e}RyZVBObj{TLV9#yNP$4(*sV|kSAvwU(i>1pXxuBE29zlA)
z+-&7|Z2bB5GX3Q96Ua0De#LW#Eg8xuY+xcC7NLU>K4P1@klV)rF~!w0#=R{eN#&Iu
zbA6)y61qP4EVA#XhFVP&Mo2~!7NDqGOXBG1;MP_f^wM7s^AQF8(8{qvPZ#mBLrfNS
z1$~DaKM`aJ%yqSfuF$9^t>`!sq^NgLE%{NZg8poFlB{HKc-@evC0K4qeEdW?hhJ4O
zoZ_~861iJ`LUDdIOmgPFIEVOogC-KCV}$kO?IU&@AVWz=xcO@$c}hPGdoPxxIQM~-
zB-#B(BzCg%PSmQTfJOALn|UAw1UTqSvbAMsf%T>JPXrz4h#;m7xEKUeqf7qyTjyJ^
zq;6X$9vHl~;chNuwi~-Jr-_wM*mP^cmep7!55p8l;#d?JFRiMt_N&6F_Z?&ayJ{9v
z1Lyl!yx;M`nB%G&6mu?*amqi-vgV-(IRovtwQ*253&;?}zb9nKNz+|-dPvEi66Bf<
zXvqtJ`VvkG_SEyt@(E8ZywwB<R9OPEZ)ca0X3f0A?+P-ebZI*%B&oM7_?>2o1-8Ro
zR5Nd@hY|PB4L#HuQls0o%`l?IJ5lF~uzFyrP?F|G;-r?sSA7{4azN_(3|}7W?tnlo
z%v3ycG{ZDfjO&8fpE9j4kplN@1ylKQ$e|A}>2KaURlY}C)h}-Xu}l<9zPT<0vnXdu
zGajQ2o<nxsy(K_+2GUVy+?vF|KYwM1MF9FZRw7*~y;w^=X7@>+pe*AfbAJ{dOKC_X
znUhnr^SmkV=!1x46ow(#LJW+PvQr82W+2cgSR+u*zdnhp8d%^=Mw&ZTg;1!D#bUf}
zwaKnhi|M$?&4!q{<x0?2CF;%yUsDHbPY=U^?`XJ)DMrR|%Fo`PW#|sk9HURIwDFvg
z+`S-p_hKA3YRGqToqjfjYt{-{oEUqYHIaGFZ}Xg&p;&+79-n%=Xp3~C(H*FHg!E<J
zwjKZ8!3w~snuW&|c?<iYs1JM~XLF-^F-JxI@VvhzKOO3OXLH!WlUj?j?&BIuwYs%3
zVJUsz2*s(foZ61>NX{Rt=YIpYc^A0VZdHyjY*KmfSNi4`u0+3Fx%M+*=IdO>F_bFd
zp2_5w{>hxTO2mth0Gc>Sb=N8Wu`0kbI^2<|5<;>4pvTeqtEoq$Bex@=KcYR{fb2FH
zcD)@@LSif@Mruy<=5R!;IDI9l<))R3R{{hk85;f)eVa(f{2jUKY2vc{1=R#p7~{5Q
zwjO-@sUckMrPklQ9F|Zdoqw4%tWT}B_4)~tA332!!Ysn3lZlRijohqXJKu5@i@Cog
z)ZIaja@PU*6^=SBLSYC6FM9Z0P;XZyW86|)u)6U0#Wd#BO}f`4mg6Ch%m~bw;!ZU=
zApCYd4B`YSZ)7RhS`p9CG~1?go2y(ao3IEJvCiqXyilQEgZ#$$W?IbzBi_)NW+E9N
zDKPn`Md_e%{Rbw@Y5T%#L?|qGjO2I=arErxMA3!wqo`2V%X%(mUiH;TRWdrY#SIOJ
zX$w(0Ybs}Lr<>0F2?F;*A4ODwstXFL<Tnl+P5}=O5|s3<_s$_XAFVK5Y<e$Ew1(e3
zd5(IHv&_SZVWS)%V*`Xf>75iIW#=zKF438aJ`}O#xmn$qX0O3^QG(nGW#b4R!)=y{
zr(ly2fQfaDA3zdB!;D!nEXa}YT4A*-MoH!>xBLl7DBj<W(~J3~bMCl9>QQYT5>VJi
z;N@t9I?vpu8j_q-H0S75m<{w;Az*cNfHc;ET^Bg+TpusOl$7aciR5P^e|p7}()DMz
z<|<?wrR}<Z>a|YGMk=#hvsXB76O^6zh;wbW>NYO*!5&>BUbjV)*=f6YYn_wX<<g?_
z)5jdwqjwkfVM>?-N<7Nbzjx`aiAr5*$ecAr-3mN9PtN`<6n`>U03X3XICVV`4}3F~
zt6F4{qJk*V5*?*-+1Oe5yO%+yxYgrMMi=PbIam*|*%|37D761hVTIn<rv8$;ENtVt
zlQR#jCxjdxo(q=e$=01%PpYkuc!*uV$}V5nhO+M`Q|Z6L*Us7|L~;q9G9@2Uv?xdo
zVZAz0&JunRz!Xhds<s=h6&2)Yaub*5<%s#%Mt*3x0giu{)EIk=2^^}%P&Wu4+r3}Q
zg<(xL9iuMS8;i^9(SgYGP{|q3-2*AunMRtMY2iZ>#}}LY;4w$s8zhW~DgKj`JfTF)
zC7PjKf<laaDiUPGCb;`qu+K`#XmJoBT-iPWoXyuuld!^7cdMj3&(M({a65#?9(PPG
zs1=msO0_B72D!q8J1cd8KL_)L*c-?s%|ja`MWIgH*|#&x7RqjkL78J8V#T0qaj|F8
z(XqIQU+QB%R<h6P=E_a$);Vm@1?pAnnm*UEubYpPT|TK(u&PNgNi-jt_soI-^u#p5
zkj0ZLT5hNV+b${lZykF5#S|utdVI&NwGV5Y#kxr!mKEx+ra6L_`+Ufxgm4^Qc-QK6
zAVoaNO}`U17^!wz5#WFgtZ%`-4r{eN9JFSeuqY*AmzP<%A0o8SOe4<1u)ms1A@Da?
z=i>|rD5&#&@IX**e~@6(rMg3w^Fc|Gd777mM~KtL_)o|NqG}kYBtOd#W5zPg+2%y+
zMorYeaF&$RQtjf$wjE+DPcr+*)2?|P)1Kkav@kao6k3zBA1f|B?evs?D$3RtV;zXD
z$QOq)a(I*%Lyk*)PBSg`%bh#T7FSW2&%g)CX%+d}Xs8F*G9q14sae|j;A?qs2G?Z3
zYn_v-w3qux3VV=5CXpIs`PERKfwX#El$asks!mZisnpG(IqHO^^;T2~OmP<S&j#*E
z_iM*bP^fB=_Ob1ba>YA(x4$rTjd+wgt+8C_sVDUj%6`#@5@Vuvo&NSwD$wQA%}+C8
zvwp$f-s|0NdQTS>-0tBqRlpNbY><ZFZ1`nVm)B(!iSi16G=Z9=XSiuH0d%IA^%zQ3
zl6iJQhQrGF%a>6jUi@utWRVOkk_5T)z27(-4cz4uET@!mjtd+GyG!wl5*CuPDUXF?
zE_%W-mPsVZ_eqw4_3_oX(P+Q*r#QnCdzwT#12<U+hmxeD5Lb2XLzT{P#>AE<qf22g
ze~eTIS-LRD-0hn^F)9xGfNv!1w9H|1M!Zg_<$7Zq`D3lb<(%Dhx8(dwEKnz=M+^oh
z819g(@|+uTjsJX6))J3RP$ZmHpw6x6!ru=E`sP0b9T%@ldaM+QZ(mTTq3>%IBBF!|
z6vj3(R*%$s%}o8FF^3y}fnaI=hW3d$Z%Wm2x<!Y(LDb4ZFm-e6qEj+?dGVm?l2-Ob
zPlx34^Kqx75Z+sIYD5t650UOEi(edarI`#{Qe@*1CdjRU`MrB_3^6bq^j>V`a!~Iq
zn%0~3D&;qE$Je1UZ)C=7Z5#4>4pW^nDZA1uoKeS}*X4Yiy#g(SVqk1WjV|^J1FsNl
z5aLgps6v|CKr9U&25b4tPR{A|b{O!i<|xmj1%E`N+UQtPM58KWTR!NGy*txvh8|3c
zwK_-B{-hH05;z{=dxf;y%)jU(dgcT%F}J_d4<MtwId{%)=TOtOoL$UWg&)6+9CssG
zlQNE;5&3%yPaX)XU&#%|V5iQemNWAjPCkjnH^d~}eD8lcN<hHl8Mb7JEnTNLT5pD3
z{;~Kj@or{YqZn+kMABvm_2!@X3Upa*(OmOe2!yZ+QFkG|IlR>Px~3#cDEtybs*XDg
z*sV8cH~6xPb?LrgzYxKQINFrHJPeDek)96QA)ndwQG4;?r2D!g#h(9&oX<gG7GX*l
z>(5@{W&++TwRnf<;Dkr@<3Yx>+3_qQQFEsztv3ok&oRqP088`Thb@kEKT<Y17qbtR
z7`+$YeS&b4r+NO=J~>r?gu-6VAhJ}I<ZFfTtcDTx2)Y)i?&cIfm1J_XZT!hG3&ikn
z_ofy+)ItUY!>z^ml0<Z4QvTvDyeaNZwtUu>d@k1Z-xJwf&dS`}0{IZTf3(3a3sABS
zH<>wpoAP-OyFm$^dvfwucW;U2w!7C8rTrP#(DuN#odSh@&f<=!Nn`9zqBf4tZuew}
z4S&uN9-!VJ-#I@|HMf-MIJw}!pd|b`evgkrVhEC;{t<cZec*9M=twO8qUg*Gx&RqU
z2`xiLuQii2QUOdMiU4oUUPn2u^hb%A!qxsu=LwGI@pBSiAjopg>@&?CAKh2+$vtd)
zp&LpR7Fi@4oZl<Qv-PYH`|j`ltb>W!UD6c3X6s2zBWydoa~*NP{55!*-B`P3Y0EIL
z;((JHJFNx~B8Ex#?pD^$3+8MPHSQ1}u4!%{NYLiu1X+j&8zeH^KAE%Ww5#+H>M*81
zB4C|?7l_i*caXg5fn~ybpFyn(&-`Wbg%&QjWIA|q>f5-E*~DNT&bnF3k;V;>aggrL
zW17gOCQM?~mx{5A{@(KEKK1M+Xd0aMzTzK+t3S@fS{akDyIeT3RjLDyQi+AI9c%<T
z-F&MAksHk;&-V&X&N*w0am^1lNY9sfmC7?ZF)hb52?uQrWtl#qPIu^=&N+F9TBDEU
zcdvC1%pM5Yw|<kJD=*7|qw<QI)(Ga+IXMt?cifxjly>OwZxizV%qhj;DlsM(qMedw
zoNgAqx+L;F0XV@1oVPCiOpC}^x~FBv!AQ@b6~qLFz;gCX>h5~a-h=UOjN!>dQk|0X
zU<(qQQf#O7qBbT~buo3qSRx#RRPAk`J$pv!3Bm(>9AMRSNVjpiCiCSEZnLIkIE|JJ
zP7fMz(j?iSEnuickH=*3>WaoR>&MLuV2;}%Q{9@@u&a6d?keRJAx(30I}y9&n0Oxd
zum$dEutoGO1KZQ7$;A)D1_@ps)80fkb?B0HmOq=DGW|s@9WzBiWg9_PsCzPp1)yv3
zvD(ERt<2|rhG;lu;;4jk=%GFJ^P*Q?4ew#|a)vhYm%%K=#h60UmjD^hl3^G?$I23m
ziSfEK8D~6al1$dz)_*wtS>j${**e`dXPw1`SB(w(Ri07dom_+j7YxoB1x_24zcl-P
z|3`56HEM^_g~R@m=Sgp+0s!~jBP_`Wj_9b1Io_Xi-}ryQR~KB%%D^1vNQXRJw3$w)
z9L@>|_}Q};GVoFM=;sfao@%+wx4z$}uKnk~Q$D>tB1xmh1W;NzTY7l9TbVf>uUyRS
zDd7m1DC~G4E+LMp5I?s?PK1;c!ws@>0y^r>czN>4MJA)CL+fb+zB>9-^}k#klyr&&
z>LcOpq5<7;Zx+;z8QdOpRhzuJEtBsreA_U^H=-`;5jUplujJK>RmPseWoUNnCKVz>
zK)JB*`C#_0V$uV$IIZ<;s+IuB!#R<l=Fw7`ni*MyLMhj0e#LfYZPru<1$wi^c!0Su
z_j=Z?mTB$PuR915dza`uDvJwS6t77XomAThzsM6PuO#ZdjEk}1VtmUyWn+H9baAj8
z@`_bYIqBX?M?_J7VRXH68|IqkPrf7#p&RYax~zBh6Qkezdy)tgp*9mY-|1Jpqq-UC
z%djoEue!T9;&?Gk%6Vinasi6OBN%{wQ*}wx?BqG@r^#D)>0hzGsp<E;4}Co1s`2`3
zVCRo(hn|y$eRu2K$)CwDXezAN(Y{hCYimPbG9K~!JrgWFpIWbiY&JVXKeE%>Np$|e
zz5WcUz~R#Co0rlm&6OWLbo7vIARf_J!%L2|GPXFbWsGE(Bas3ZgC}ak?7zx;QB^d4
zcdg2jLoPjY!kJxQ#4W)XaRRW$3V}<lR76OGD43*6fEbu15;*zf0Dp~q-ukuv@6Vq{
zHG)}UY+%+9HJ#%ZxizHT_?~+*uXX-5bcef4@nh__)LtF@&e;m40ihD?0DZL(C9yW6
z?9CJr=1cDfww9y{U9V={Hz2QOVyh2PBM7nICF96L!mcmMx7V2?bclkv#ysrY$Rpi(
zB7t0>K69G)Wp4JC?5&iZX`FiAJ9bU2B?!J7ulCefZSQHyldH-J$zKD?6t}{f5@aw~
zcvsMhc%;)a5}rklJnr5$CaKa6!g}k}5Bf`{Ega$e(FRl%g&dBXw;3PjvcH^T9T)xB
zBdTkjJyS~Z{<FDN@an>xSV>F7H*K78tLti#e4jV@pzR^s+3rsk%n3o*f-(jgJ9JY-
zQ>F<TZ}Oc{lC~+rA|q9h+QOEnz%M4U>0M?;o>_YbhvAR!bb8DF;kb*!ytq{%6&#@m
zBMvKEhszRWtY8*-2*J(487s;~;Sk&ef)Te3pU2%qGUC1=9B^O3PX}ZWE*iy%J4D9d
z%*7aSS*RA=t{hE;vzRd<hZqVb;$i0l6BCDtSlQXwdQysu!9)~b5^$IZ7KWRWkfEi7
z!$n~t&M*-dn20N^f)^c1LU*DHLx_O|sRI+S#=Vi{qoNeCgCRvvB)An>3jnkV&aD##
z4_7k_D;NrX`ppxDI<90LN)b<6cPmhi{9Eqr0z;ur*LPv4zm4B@wzP8huyAn)6~#|~
z^n#%zPFLPA6#95YDdG!5NuJ95tlV8-Vxko-7zqdjj{7NhpB)WHAw-elhWwNm%s<`4
z$KA@BQWPeJz&(){u|>lX;LgVv%o#>`vVcpV(V)rSWReIlivO0OMaBPz3=PKkn;u*g
zi4##k6HCA)Q8)($E<ge<0mlB{TH+`q80K#>1mX`*#3cUIlR*99v8V(B{(EGi5=aRU
zKv4kP<==2cCB!9<h5s!>A7}KxWoQJ*hu>vlh(Df1qmjQuASwy}Pp6Of2QNvK7?{N4
z>;IjRltlc2f+!pba^k=AKtx6Vz+O}W{->U#*zeB3;ee>*ADh4tVt*KfBPD*17!DT~
z{qtG$9|*z`qJM6TK#HHF1dRXRh(Z3Le}@o`kU;#gF%p#ijsjBbf9Q$-ktR424gXK4
z$&RrEO9L0LzzB$m{Xs<(8g%eCz)}ArF=8O0e%BL2Ab+PFTnu$wt;g5@hmD4#PxNrH
x$}C`85YcgQ@q~dHIPNQI&emWhpMC)g*Td7y-Sebu;b<gMjFOxC$~6_r{{uj<vw;8r

diff --git a/report/mlp-cw2-questions.tex b/report/mlp-cw2-questions.tex
index fe466ca..0a4a80f 100644
--- a/report/mlp-cw2-questions.tex
+++ b/report/mlp-cw2-questions.tex
@@ -62,22 +62,29 @@ The difference between these two methods is that the first approach using a $1\t
 % 
 % The average length for an answer to this question is approximately 1 of the columns in a 2-column page
 \newcommand{\questionFour} {
-\youranswer{test1
+\youranswer{
+Our results demonstrate the effectiveness of batch normalization and residual connection as proposed by \cite{he2016deep}, enabling effective training of deep convolutional networks as shown by the significant improvement in training and validation performance for VGG38 when incorporating these techniques. Table~\ref{tab:CIFAR_results} highlights that adding BN alone (VGG38 BN) reduces both training and validation losses compared to the baseline VGG38, with validation accuracy increasing from near-zero to $47.68\%$ at a learning rate (LR) of $1\mathrm{e}{-3}$. Adding RC further enhances performance, as seen in VGG38 RC achieving $52.32\%$ validation accuracy under the same conditions. The combination of BN and RC (VGG38 BN + RC) yields the best results, achieving $53.76\%$ validation accuracy with LR $1\mathrm{e}{-3}$. BN+RC appears to benefit greatly from a higher learning rate, as it improves further to $58.20\%$ a LR of $1\mathrm{e}{-2}$. BN alone however deteriorates at higher learning rates, as evidenced by lower validation accuracy, emphasizing the stabilizing role of RC. \autoref{fig:training_curves_bestModel} confirms the synergy of BN and RC, with the VGG38 BN + RC model reaching $74\%$ training accuracy and plateauing near $60\%$ validation accuracy. \autoref{fig:avg_grad_flow_bestModel} illustrates stable gradient flow, with BN mitigating vanishing gradients and RC maintaining gradient propagation through deeper layers, particularly in the later stages of the network.
+
+While this work did not evaluate residual connections on downsampling layers, a thorough evaluation of both methods put forth earlier would be required to complete the picture, highlighting how exactly residual connections in downsampling layers affect gradient flow, feature learning, and overall performance. Such an evaluation would clarify whether the additional computational cost of using $1\times 1$ convolutions for matching dimensions is justified by improved accuracy or if the simpler pooling-based approach suffices, particularly for tasks where computational efficiency is crucial.
 }
 }
 
 
 %% Question 5:
+% Briefly draw your conclusions based on the results from the previous sections (what are the take-away messages?) and conclude your report with a recommendation for future work. 
+% 
+% Good recommendations for future work also draw on the broader literature (the papers already referenced are good starting points). Great recommendations for future work are not just incremental (an example of an incremental suggestion would be: ``we could also train with different learning rates'') but instead also identify meaningful questions or, in other words, questions with answers that might be somewhat more generally applicable. 
+% 
+% For example, \citep{huang2017densely} end with \begin{quote}``Because of their compact internal representations and reduced feature redundancy, DenseNets may be good feature extractors for various computer vision tasks that build on convolutional features, e.g.,  [4,5].''\end{quote} 
+% 
+% while \cite{bengio1993problem} state in their conclusions that \begin{quote}``There remains theoretical questions to be considered,  such as whether the problem with simple gradient descent  discussed in this paper would be observed with  chaotic attractors that are not  hyperbolic.''\\\end{quote}
+% 
+% The length of this question description is indicative of the average length of a conclusion section
 \newcommand{\questionFive} {
-\youranswer{Question 5 - Briefly draw your conclusions based on the results from the previous sections (what are the take-away messages?) and conclude your report with a recommendation for future work. 
+\youranswer{
+The results presented showcase a clear solution to the vanishing gradient problem. With batch normalization and Residual Connections, we are able to train much deeper neural networks effectively, as evidenced by the improved performance of VGG38 with these modifications. The combination of BN and RC not only stabilizes gradient flow but also enhances both training and validation accuracy, particularly when paired with an appropriate learning rate. These findings reinforce the utility of architectural innovations like those proposed in \cite{he2016deep} and \cite{ioffe2015batch}, which have become foundational in modern deep learning.
 
-Good recommendations for future work also draw on the broader literature (the papers already referenced are good starting points). Great recommendations for future work are not just incremental (an example of an incremental suggestion would be: ``we could also train with different learning rates'') but instead also identify meaningful questions or, in other words, questions with answers that might be somewhat more generally applicable. 
-
-For example, \citep{huang2017densely} end with \begin{quote}``Because of their compact internal representations and reduced feature redundancy, DenseNets may be good feature extractors for various computer vision tasks that build on convolutional features, e.g.,  [4,5].''\end{quote} 
-
-while \cite{bengio1993problem} state in their conclusions that \begin{quote}``There remains theoretical questions to be considered,  such as whether the problem with simple gradient descent  discussed in this paper would be observed with  chaotic attractors that are not  hyperbolic.''\\\end{quote}
-
-The length of this question description is indicative of the average length of a conclusion section}
+While these methods appear to enable training of deeper neural networks, the critical question of how these architectural enhancements generalize across different datasets and tasks remains open. Future work could investigate the effectiveness of BN and RC in scenarios involving large-scale datasets, such as ImageNet, or in domains like natural language processing and generative models, where deep architectures also face optimization challenges. Additionally, exploring the interplay between residual connections and emerging techniques like attention mechanisms \citep{vaswani2017attention} might uncover further synergies. Beyond this, understanding the theoretical underpinnings of how residual connections influence optimization landscapes and gradient flow could yield insights applicable to designing novel architectures.}
 }
 
 
@@ -102,9 +109,20 @@ The length of this question description is indicative of the average length of a
 \newcommand{\questionFigureFour} {
 \youranswer{
 \begin{figure}[t]
-    \centering
-    \includegraphics[width=\linewidth]{figures/VGG38_BN_RC_accuracy_performance.pdf}
-    \caption{Training curves for 38 layer CNN with batch normalisation and residual connections, trained with LR of $0.01$}
+    \begin{subfigure}{\linewidth}
+        \centering
+        \includegraphics[width=\linewidth]{figures/VGG38_BN_RC_loss_performance.pdf}
+        \caption{Cross entropy error per epoch}
+        \label{fig:vgg38_loss_curves}
+    \end{subfigure}
+
+    \begin{subfigure}{\linewidth}
+        \centering
+        \includegraphics[width=\linewidth]{figures/VGG38_BN_RC_accuracy_performance.pdf}
+        \caption{Classification accuracy per epoch}
+        \label{fig:vgg38_acc_curves}
+    \end{subfigure}
+    \caption{Training curves for the 38 layer CNN with batch normalization and residual connections, trained with LR of $0.01$}
     \label{fig:training_curves_bestModel}
 \end{figure}
 }
@@ -117,7 +135,7 @@ The length of this question description is indicative of the average length of a
 \begin{figure}[t]
     \centering
     \includegraphics[width=\linewidth]{figures/gradplot_38_bn_rc.pdf}
-    \caption{Gradient Flow for 38 layer CNN with batch normalisation and residual connections, trained with LR of $0.01$}
+    \caption{Gradient Flow for the 38 layer CNN with batch normalization and residual connections, trained with LR of $0.01$}
     \label{fig:avg_grad_flow_bestModel}
 \end{figure}
 }
diff --git a/report/refs.bib b/report/refs.bib
index 1863eff..3ce0ad9 100644
--- a/report/refs.bib
+++ b/report/refs.bib
@@ -161,3 +161,24 @@
   year={1995},
   publisher={Oxford university press}
 }
+
+@article{vaswani2017attention,
+  author       = {Ashish Vaswani and
+                  Noam Shazeer and
+                  Niki Parmar and
+                  Jakob Uszkoreit and
+                  Llion Jones and
+                  Aidan N. Gomez and
+                  Lukasz Kaiser and
+                  Illia Polosukhin},
+  title        = {Attention Is All You Need},
+  journal      = {CoRR},
+  volume       = {abs/1706.03762},
+  year         = {2017},
+  url          = {http://arxiv.org/abs/1706.03762},
+  eprinttype    = {arXiv},
+  eprint       = {1706.03762},
+  timestamp    = {Sat, 23 Jan 2021 01:20:40 +0100},
+  biburl       = {https://dblp.org/rec/journals/corr/VaswaniSPUJGKP17.bib},
+  bibsource    = {dblp computer science bibliography, https://dblp.org}
+}