From b00bb8e1f74d5e361118fabc01fb77595f8e7056 Mon Sep 17 00:00:00 2001 From: msohaildanish Date: Wed, 29 Nov 2023 14:42:20 +0400 Subject: [PATCH 1/8] add website pages --- README.md | 2 +- index.html | 873 +++++++++++++++++++++++++++++++++++++++++++ static/css/index.css | 160 ++++++++ 3 files changed, 1034 insertions(+), 1 deletion(-) create mode 100644 index.html create mode 100644 static/css/index.css diff --git a/README.md b/README.md index 1158dfe..964c80d 100644 --- a/README.md +++ b/README.md @@ -3,7 +3,7 @@ Oryx Video-ChatGPT

-#### [Kartik Kuckreja]()\*, [Muhammad Sohail Danish]()\*, [Muzammal Naseer](), [Abhijit Das](), [Salman Khan](https://salman-h-khan.github.io/) and [Fahad Khan](https://sites.google.com/view/fahadkhans/home) +#### [Kartik Kuckreja](https://www.linkedin.com/in/kartik-kuckreja-930531221/)\*, [Muhammad Sohail Danish](https://www.linkedin.com/in/muhammad-sohail-danish/)\*, [Muzammal Naseer](https://muzammal-naseer.com/), [Abhijit Das](), [Salman Khan](https://salman-h-khan.github.io/) and [Fahad Khan](https://sites.google.com/view/fahadkhans/home) \* Equally contributing first authors #### **Mohamed bin Zayed University of AI, Birla Institute of Technology & Science, Australian National University, Linkoping University** diff --git a/index.html b/index.html new file mode 100644 index 0000000..2aec955 --- /dev/null +++ b/index.html @@ -0,0 +1,873 @@ + + + + + + + + + GeoChat: Grounded Large Vision-Language Model for Remote Sensing + + + + + + + + + + + + + + + + + + + + +
+
+
+
+
+ geochat_logo +

GeoChat: Grounded Large Vision-Language Model for + Remote Sensing

+
+ +
+ + Kartik + Kuckreja*, + + + Muhammad + Sohail Danish*, + +
+ + +
+ + Muzammal Naseer, + + + Salman Khan, + + + Fahad S. Khan + +
+
+
+ Mohamed bin Zayed University of AI, Australian National University, Aalto University,
+
+
+ *Equally contributing first authors +
+
+
+ + + + + + arXiv + + + + + + + + Code + + + + + + + + Dataset + + +
+
+
+
+
+
+
+ +
+
+
+

+ RGeoChat is the first grounded Large Vision Language Model, specifically tailored to Remote + Sensing(RS) scenarios. Unlike general-domain models, GeoChat excels in handling high-resolution RS + imagery, employing region-level reasoning for comprehensive scene interpretation. Leveraging a newly + created RS multimodal dataset, GeoChat is fine-tuned using the LLaVA-1.5 architecture. This results + in robust zero-shot performance across various RS tasks, including image and region captioning, + visual question answering, scene classification, visually grounded conversations, and referring + object detection. +

+
+
+
+ + +
+ +
+ + +
+
+ +
+
+

🏆 Contributions

+
+
    +
  1. RS multimodal instruction following dataset. We present a novel data generation + pipeline, to leverage existing object detection dataset to create short descriptions of + the images, followed by using Vicuna-v1.5 to create conversations using the generated + text alone. Further, we add visual question-answering and scene classification abilities + using their corresponding datasets. This results in a total of 318k instruction pairs + for RS domain.
    2. +
    +
  2. GeoChat. Leveraging our dataset, we finetune LLaVA-1.5 to create the remote + sensing-domain vision-language model - GeoChat. Our LoRA fine-tuning is efficient and + avoids forgetting the necessary context embedded in fully-tuned LLaVA model, whose MLP + projection is trained to align images into the word embedding space of the LLM + (Vicuna-v1.5). This allows GeoChat to retain the conversation and instruction following + abilities of LLaVA and extend its domain-knowledge to remote sensing tasks.
    3. +
    +
  3. Evaluation Benchmark. We also address the lack of evaluation benchmarks to assess + the capability of existing VLMs on remote-sensing conversations. To this end, we setup + evaluation protocols for conversation grounding in RS, as well as a setup a suite of + tasks to allow comparisons with future efforts in this direction. We show various + supervised as well as zero-shot evaluations for different remote sensing tasks, + including image captioning, visual question answering and scene classification to + demonstrate the generalisability of GeoChat conversational VLM.
    4. +
+
+
+
+
+
+ + + + +
+
+
+

geochat GeoChat: A unified framework

+
+
+ +
+ +
+
+
+ + +
+
+
+
+
+
+

+ GeoChat can accomplish multiple tasks for remote-sensing (RS) image comprehension in a + unified framework. Given suitable task tokens and user queries, the model can generate + visually grounded responses (text with corresponding object locations - shown on top), + visual question answering on images and regions (top left and bottom right, respectively) as + well as scene classification (top right) and normal natural language conversations (bottom). + This makes it the first RS VLM with grounding capability. +

+
+
+
+
+
+ +
+
+
+

geochat GeoChat: Architecture

+
+
+ +
+ +
+
+
+ +
+
+
+
+
+
+

+ An overview of GeoChat - the first grounded large vision-language model for remote sensing. + Given an image input together with a user query, a visual backbone is first used to encode + patch-level tokens at a higher resolution via interpolating positional encodings. A + multi-layer perceptron (MLP) is used to adapt vision-tokens to language space suitable for + input to a Large Language Model (Vicuna 1.5). Besides visual inputs, region locations can + also be input to the model together with task-specific prompts that specify the desired task + required by the user. Given this context, the LLM can generate natural language responses + interleaved with corresponding object locations. GeoChat can perform multiple tasks as shown + on top e.g., scene classification, image/region captioning, VQA and grounded conversations. +

+
+
+
+
+
+ + +
+
+
+

geochat_logo RS Multimodal Instruction Dataset

+
+
+ +
+
+
+
+

+ Types of annotations available in the GeoChat instruction-set. For a given RS image, we + obtain object attribute and relationship information, referring expressions and region + captions along with their corresponding region annotations (shown over the image). This + structured information is used to create the rich instruction-set with a total of 318k + image-instruction pairs. +

+
+
+
+ +
+
+
+ +
+
+
+
+
+
+ + +
+ + +
+
+

geochat_logo Qualitative Results

+
+
+ +
+
+
+
+ +
+
+
+ +
+
+
+

+ Qualitative results of GeoChat. (left-right) Results are shown on grounding, referring + object detection, and disaster/damage detection. The user can provide task-specific tokens + (e.g., [grounding]) to shape model responses according to the desired behavior. The model + can generate textual responses (right), only visual grounding (center) and both text and + object groundings interleaved together (left). The model can also specify object types, + object counts, object attributes and object relationships. +

+
+
+
+ +
+ + + + +
+ +
+
+
+

Visual Question Answering (VQA)

+
+

+ GeoChat is able to hold multi-turn conversations, based on various types of questions, + including presence, count, complex comparisons and so on. It is able to detect objects + and hold conversations against low resolution images as well. +

+
+
+
+ +
+
+
+ +
+ Qualitative results of GeoChat's performance in Visual Question Answering. +
+
+
+
+
+ + +
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
MethodPresenceComparisonRural/UrbanAvg. Accuracy
LLaVA-1.555.4668.2059.0062.77
Qwen-vl-Chat38.5767.5961.0055.35
MiniGPTv255.1655.2239.0054.96
RSVQA87.4781.5090.0086.32
EasyToHard90.6687.4991.6789.94
Bi-Modal91.0691.1692.6691.63
SHRNet91.0390.4894.0091.84
RSGPT91.1791.7094.0092.29
GeoChat91.0990.3394.0090.70
Comparisons with general zero-shot (top) and RS-VQA + specialized (middle) models on RSVQA-LRBEN dataset for + VQA task. LLaVA-1.5, Qwen-vl-Chat and MiniGPTv2 are evaluated in zero-shot setting. GeoChat + outperforms other zero-shot models and performs competitively to + SoTA-supervised models like RSGPT which are specifically finetuned on target dataset (while ours + is a generic model not specifically finetuned on target dataset).
+
+
+ + + +
+
+
+
+

Scene Classification

+ +
+

+ For scene classification, the model is presented with a list of dataset classes in the + input prompt and tasked with selecting a single class from the provided options. + +

+
+
+
+ +
+
+
+ +
+ Qualitative results of GeoChat's performance in scene classification. +
+
+
+
+
+ + +
+ + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelUCMercedAID
Qwen-VL 62.9052.60
MiniGPTv2 4.7612.90
LLaVA-1.5 68.0051.00
GeoChat84.4372.03
Zero-shot scene classification accuracy comparison on + AID and UCMerced datasets. In comparison to other + generic VLMs, GeoChat performs favorably well.
+
+
+
+ +
+
+
+

Region-Level Caption

+
+

+ Given a bounding box, GeoChat is able to provide brief descriptions about the area or + the object covered by the bounding box. +

+
+
+
+ +
+
+
+ +
+ Qualitative results of GeoChat's performance in region-level caption.. +
+
+
+
+
+ + +
+ + + + + + + + + + + + + + + + + + + + +
ModelROUGE-1ROUGE-LMETEOR
MiniGPTv232.131.210.0
GeoChat87.387.283.9
Region level captioning performance.
+
+
+ +
+ + +
+
+
+

Grounded Description

+ +
+

+ When asked to describe the image with the special token '[grounding]', GeoChat outputs + both the description of the image as well as the bounding boxes for all the objects + detected. +

+
+
+
+ +
+
+
+ +
+ Qualitative results of GeoChat's performance in grounded description. +
+
+
+
+
+ + +
+ + + + + + + + + + + + + + + + + + + + +
Modelacc@0.5acc@.25METEOR
MiniGPTv210.830.916.4
GeoChat11.733.948.9
Results on grounding description task.
+
+
+ +
+
+
+
+

Referring Expression

+
+

+ When asked about an object as a referred expression, GeoChat is able to locate by + predicting bounding boxes around it correspondingly. +

+
+
+
+ +
+
+
+ + +
+ Qualitative results of GeoChat's performance in referring expression. +
+
+
+
+
+ + +
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelSmallMediumLargeSingle-object groundingMulti-object grounding[refer][grounding]Overall
MiniGPTv21.79.921.99.13.68.22.67.6
GeoChat2.913.621.716.04.310.511.810.6
Performance (acc@0.5%) comparison of GeoChat on our benchmark. Small, medium and large + refer to the size of the objects based on the bounding box area. Overall, GeoChat + outperforms the baseline, but there is still significant room for further improvement on + this complex task.
+
+
+ +
+ + + + + + + +
+
+

BibTeX

+ 

+    @misc{kuckreja2023geochat,
+        title={GeoChat: Grounded Large Vision-Language Model for Remote Sensing}, 
+        author={Kartik Kuckreja and Muhammad Sohail Danish and Muzammal Naseer and 
+            Abhijit Das and Salman Khan and Fahad Shahbaz Khan},
+        year={2023},
+        eprint={2311.15826},
+        archivePrefix={arXiv},
+        primaryClass={cs.CV}
+    }  
+
+
+
+
+
+

Acknowledgement

+

+ This website is adapted from Nerfies, + licensed under + a Creative + Commons Attribution-ShareAlike 4.0 International License. We are thankful to LLaVA, GPT4ROI, and + LISA for + releasing their models and code as open-source contributions. +

+
+
+ + + + + +
+ + IVAL Logo + + + Oryx Logo + + + MBZUAI Logo + +
\ No newline at end of file diff --git a/static/css/index.css b/static/css/index.css new file mode 100644 index 0000000..ea3e115 --- /dev/null +++ b/static/css/index.css @@ -0,0 +1,160 @@ +body { + font-family: 'Noto Sans', sans-serif; + } + + + .footer .icon-link { + font-size: 25px; + color: #000; + } + + .link-block a { + margin-top: 5px; + margin-bottom: 5px; + } + + .dnerf { + font-variant: small-caps; + } + + + .teaser .hero-body { + padding-top: 0; + padding-bottom: 3rem; + } + + .teaser { + font-family: 'Google Sans', sans-serif; + } + + + .publication-title { + } + + .publication-banner { + max-height: parent; + + } + + .publication-banner video { + position: relative; + left: auto; + top: auto; + transform: none; + object-fit: fit; + } + + .publication-header .hero-body { + } + + .publication-title { + font-family: 'Google Sans', sans-serif; + } + + .publication-authors { + font-family: 'Google Sans', sans-serif; + } + + .publication-venue { + color: #555; + width: fit-content; + font-weight: bold; + } + + .publication-awards { + color: #ff3860; + /* width: fit-content; */ + font-weight: bolder; + } + + .title + .publication-authors, + .subtitle + .publication-authors { + margin-top: -1.25rem; + } + + .publication-authors a { + color: hsl(204, 86%, 53%) !important; + } + + .publication-authors a:hover { + text-decoration: underline; + } + + .author-block { + display: inline-block; + } + + .publication-banner img { + } + + .publication-authors { + /*color: #4286f4;*/ + } + + .publication-video { + position: relative; + width: 100%; + height: 0; + padding-bottom: 56.25%; + + overflow: hidden; + border-radius: 10px !important; + } + + .publication-video iframe { + position: absolute; + top: 0; + left: 0; + width: 100%; + height: 100%; + } + + .publication-body img { + } + + .results-carousel { + overflow: hidden; + } + + .results-carousel .item { + margin: 5px; + overflow: hidden; + border: 1px solid #bbb; + border-radius: 10px; + padding: 0; + font-size: 0; + } + + .results-carousel video { + margin: 0; + } + + + .interpolation-panel { + background: #f5f5f5; + border-radius: 10px; + } + + .interpolation-panel .interpolation-image { + width: 100%; + border-radius: 5px; + } + + .interpolation-video-column { + } + + .interpolation-panel .slider { + margin: 0 !important; + } + + .interpolation-panel .slider { + margin: 0 !important; + } + + #interpolation-image-wrapper { + width: 100%; + } + #interpolation-image-wrapper img { + border-radius: 5px; + } + \ No newline at end of file From 5686f0f781293b3686f622a3ab66a6839fe265f0 Mon Sep 17 00:00:00 2001 From: msohaildanish Date: Wed, 29 Nov 2023 15:01:02 +0400 Subject: [PATCH 2/8] add images to website --- images/geochat_fav.png | Bin 0 -> 49950 bytes index.html | 14 +++++++------- 2 files changed, 7 insertions(+), 7 deletions(-) create mode 100644 images/geochat_fav.png diff --git a/images/geochat_fav.png b/images/geochat_fav.png new file mode 100644 index 0000000000000000000000000000000000000000..899d4383a6ac601c836f08076db0f753a504f22e GIT binary patch literal 49950 zcmV)eK&HQmP)PyA07*naRCr$Oy$76KWtl(zeNVsV_PNtbdLto$m?eAt!r7)1s8W+ zb=N5NzPbvFiU=z$0xAk3EfS;|S|C7rnaT7!x1N5^`~E)9d!vgL0zvqH{&OeunaRw# z_nvvr_dM@YzR&kzn11}x3W&8(?%s6N$5T+wPn)I@YkFBQ_G#am{L#vbABg_{x?K0Z z@%Z_|^{x$kcd;6AHHkSQO@UU*-%#x7{J>!sts7>5c%j#az0fD{VtRkEW{rA}IWe0z zo$uOybeydjEM|&O%=Ck5&OUd+B{%%`#cl8>`6INPW@?HS8L!c+*& zh>cwO^GyfTUp`kc}C6ftPw-r>37KY(| zYB3_l*z;x`pFHiftRrey`{l}OG%)5yAtsU#&@^N%gXd=U{@5~dpIPyPbvy6^Tj^;) z?+3I1*d1>^_6Q>p{~3(ek2e#pXDtIv*AXQUhH3vyv2ga1b?ci4+X{V42y=KM7zjM= zkVH>j#-S_&*r9R4GMuqgD1`1quqa_7#CZYBqn)!l;d`^N7>LKF>uM#^}!HG5gSYj~Z6){G%@U$t}}>KG^7Zmaj+K z>TmaKI9mzAPgZu0&eame3aJz{T}LbgRKpqxh5m}8Z(hIsS-!=eKTQaB-cY@$B$;>m zNZuA}%3MunDOG{0LNHTRq*Gc?1R4q8MFL?Y5OE;kKumX`5L^+6(gNcJhPw(XyU43+ z_guV^9Tb26zSqCHBm)29xappbI8rK}k4=?2>Yj(RqM@^x!>n1|BX%iw)+;W*9s;A z&JOQo5ws6Sy7F3PF?5z1WbhjPQ>9xK*$-AM1Yg6rhi<8$~laL zC3#?1jB_Zh7h(j@!4;vH;@;Zs4C`q=8LZ@=jv?+ynH>uEs$-;%KnZ$9b>#|bXowqs-=4>?O~ItKE2lyW&FTp)tx zB|Li1So_SY&seuEIpEda|AaL{s0;Q6Cs$+kMPaZ7x}|Em&Y-CRmd0Q*23rxxDgYM@ z`$CSf5G_C_fFCOo&`B&n7mz0mg1(OdokJ8jnq3Ot(2+Ji^zRGU&S!g`x36BgP=~mh ziR7>#im;&UCd>K=<2LsvSBMQtVMs)|@JFokX8iSiGxyz|~Cp3RjLLaFN*N%-NSgS;vlDKm}95VZKj~9Dhx{#Z3Edax>5Zz?_9szKhX6!*vC&>=%=jc z&rVx1Pl5Ukf2{G^TFqm_eu&{9L`u252hxlKB0ZN^t4Q2cCJcW$i9u z>3a0Tx`H!&4T`EW7`g(Bz^pRp01O5g3PUG2x)q=%7@mj)cKZy?gu{yy#0istZj~_@ z5^1Fg)BjXxSL=7k~6g=xsY>FjKT<;#wKy= za0Q6~`b-mZGZ~apmUMTNG?MnN+*(#lzI?{I$HRjS&;xzFX+S?^#jaViDpO#!^Mc9X z6U|B2ob-L{bvz7t5&CT#t9v`qku4yBK*ZF&vB0}m-FDY)&!pHwL11ApIK@?QwWBf1 z*8p8pV98Gg1Aqzvl`#qGPI5UIFvWq#VqmvVL1BzQ05L_1oPiL=7=w^M6NssLM?s?> zNl+IOV5V-M*HHH6m^k~~WvM%#1|aN?laB23n%-3^XGbS-$hK8$xOs2_lVO5(L&f}b z2L0&_D87IuV5If9rzd~@P%r=P`(6Z<)M-FJ8+%`Fv!5BJq`T1yA&f=irv%32dyn$Y}}|rc;=m z&Y)oEFm(fp3ZycIXz$H^yrU!j+Vfp>db%}k?N`%}KditVt5#{%!sx4bP5f_Ii8{O} zLD_48df4;Ohy{)=6|rh|FEnNViV}Jd7doN(=~M2ytMa#&^1Fk-`z392ziJu*-iCVXzA!t}De!*s zvIUFGM88TA?64q=G2}S-`S3Wl`XMq3!(mw)OR`y+sh+BY21!y4iyUgL+wVFEXbQ4Ru3cp-K)JnZuV%*)t#=b^J<>ju1lgG0R& zEc(ae?s{bFU$Rg?u8LwM39ksCotTJ-Wt1r_Qa55DyBj5pLQ}ZRrhwT{U;#2bROALq zRuE`{@dl815S+m613Umv9FWs2;t)=l#c;3y5Al~{LvjoTlTHM_0!<|7(KXD-806aQ zwwbx+LyOqPr=I7x=E3YME&MMn(ccq=v04cuY?z$H9p$>bsY;rLqYGKg%Vnf>wsZqr zB#0C)I%aqN%GOe69l!3@+6!%+rvd#JuD~^Gpyq?+YlE`?uYnsIK5h3&f~pr{&~-8B z_%c9xRc||9J?4I6XvI*JDKEx|tVPlu$wAx%GkwX2}bugpBnQmo6EP||b*KR8vzuO}*) zh!fZf!+g`kG3`b4q-|J+*7}>12qLBWoijQvNN3cqufBBsGs>Vm-=Lp%IsRmYDbBj} z^q2f|T#G*MH^Y=4#_%~u)ekY`c-ZGS*z3noR0VHYG=RnVJfc`Y#N1(8^7i95Y`*rX zFDb-Y<@T|c{kxe?e|m|wS3Ar3TL_(Ij1?(guYaJp3Q)qCGjH2N1jE!U4 z+obQTs~S{to*6@kh-(?`?w@? zkKslU#=H=ZRBG7k2AGNyC``p0=XK-o_B>1>^TMH;VwcT*Mf-iB8h$QCNsZjNL&NCOChBnlk2B26 zrEu1QK6K~?Se$^Hd^p{od;7w(_T6puJ3sNJlGhXl7BFkzjjgI)JdQX~;3O%G`6-mq ziOHl5Lym=UH;Y8c!P5+Q%tRnGSx5=v5Ippt#7(5)CfY>}y;cqVsVNkbA@s0)=r)I8 zV+v|i0}>Z;5W@F3+@?U&(UC+Z%v2VSK2VEAn?L74vK9RQhDHytO#N9TQco zFIQznkjn2b$=O)d-j0lsg04{6pCRTkB1Fi}ZhJCU%AIrSkM6kbKtcC>9zPA}zZe)U zHWVMHD_-(--P8Jb6iE&;6*bD9i@QcE*xc|C3QAouoH@4}D?1Cw2!fZWdd||0 z4wsYLLnUwC^4Ajvl?c%yf*HE7f(km~A+*Q4F{5!W4z(UcueKX$HGt>FsMif?rER-_ zjrUg}0)O-1Z&Op*hW!##Q`5B53?yFe1i8cz4kPq|H_JAvu1IIHE_Px5Fb-wZc3hnP~ayK2CEI+ zI5LIFI6;F0iwbF+IkyLWmWhb`R}t5x!M-=bm!q`tpm(-Iz|>{cZEXt~znaGPRj} zQwhW?Wqp~daYCRG2G~5&#NCw|hLS{D)z|cwa6(Tz6mn0Z0Q!oK`p`VnblMi1d)i+C zw&ayCibZI;D!Cez!ypWkPf8hQ$|$q~Mfz-kl$dn%8I!BBzr0OfH83dTiU@%q7Lu@3 z4_Y)*0y}{}3KN9md%+*yfWFFoSl~a3V+R`8^VmLgbO`al28F4^S!5LzZk*uGY7>t- zF4{B|Cv+9D(9R$uGcObxB=F+|E{}zg);D%_<JW+w$KN-|E*mMW$~>` zPKdndds>q8F_kgS6hxd#;sG{I*755~6Al;XwM?Az@|7sgnTK%3Y-qf7=rF2A|EjYTv5Q70#nnV z8wO0vfN5H?qs9k=VVlTgQZQ{3re(oMnbM63!T{m;81$_VLVx%MXxs0^#G?}^xEbVB z4b?Eh4U<*5D;(FE#i3RjY1JTht8DY~;t-8EM!KNw%6H^Hrqs>r-**5k!vkFQ=XChA z%kebJcg^C(MpRUfOFaMcN)j$(s>(QH@HsGcb{Cg**@yvC0q`8Wh_`s#tPeS>aQ?R|pF*Q1gW@Qp_vjHCXV37wNhZ496 zk?@lQJdPn^$`MBhJVydNqT|R{NT`59uxx>vE=(f0{26@^fayl!#Ofz+n`uyW9g40) zRSo&QrkS$fLqMcJPpB&BCCRr=2=uBBC7pxT-i^|X8R+Y4LsxAC2F=KX+S>%K+_?2tXgI1C7d}lG4aW< zsp=f9e}U3`=ST3X`?o`_l%Z19jsQ&M>Ci{$x@gpEaGfUHMjf8pL>#!_A-NNw1iJ(; z$RfA==oXw?p&@;xxUcp7e#)q<9JYQ!g=>1w?~iMNT!o^lP&6GDfjnhFGcD*=3c6`S z(F`!vLZUKRUS~!QnYr_D$cp6{aK@0i;c}ScLo$<5;}L2}5ZQU-!5KYWpX_UU;*K-c z$y0h!Kc)eF0GCIsU8UXeBs=-Zy@Q{gP>Y8Rp8QU>VbWDXjzO$1A8$ zO~S2?!LN;g1$CI<$S{U9V=$Fg-{9}NO9*%YQ9~uH3G568Wl_KpvzE(2qMF|4RD~p!tBS0z?_)*RmLI!2z zFKYyH$%$@(RRUVR5|^nv`l(Fc$zZZ&(MprO_?2@(XA9&hMoM*9o&+AjB0sSnO_l~uzw5hSw%Dq+o zna6p@v5)Mlq$*n5dNA{c~~XOOC9}8{2&5XTJaaTVITM=BK+{(|~@u6?&TA z-+cB7=^-ZGS8wLXI1(O9C$?pg$%iE;K}8V9h=QeEMePh2#(r_xn-6ZlaCD!>q^OjDJlc-d4+sW%GVVvZn85QGuDIEKzzmXw4L%YhG7 z(UHpIMBtxv|7`4+_cIeXNH46uR)vt@QXMXX*zu6;s5SHzEd5LL2KgVUQ ztzFZ$efPc}?A|kZYNZ;gVZ>yam;yd#8c;G>WV^ay+ctQ;1~FNN7KN?iiV*Ok7f=c5_LM&F;ngyS66ml8N>1c}vdV9Z>(f9>(FJAYe zXsbGqOEeAW16i#)5mWGlqLQ!TiqK9qAscy=C*g#1~m^R)uC z6axy3qdkm}-MJarB$B?hOTMu}Zbm5Gfs8DaC>U>0QCpy|y#U+LrCTvntxCy)G*e|e zl&Tr1nQ4r)bz#EHq7ui_Iww@fkq!q1oD3;VLtX)TBM+TU1Gd|crFIg}$){6)An(r@ zXnQ!H&i?aDe)RLF)A4u?tM~jobsEsmVO@88aZSEMSO53n`*)taW!t#o>lsYVT7*Hf zh=%JUA%B)CeVqCNd8><0lIj@rY!R~{aOj>*=mg0W1k%4As?=neyN!epvl%o39~x7U zOX;vp0~O!Lw(&_R^P8rCQB}~;3=GX)hP~9(>$x&0BwmOlImuS97RgEGL?~EQHT0x( zEDc@sI5otfFWsY_juLFz1=^G!`(gAi$6a~b*mGE(7v!nafPR+CPWk8)Fub5kIM;9A zKCx)Gt73b)gngt8O4*7iLKwDs{ix(k0U2dH6pb<*)cv7BxAT~3Xn5)P7ARH9kO>Ue z9VB)JY1>97laoOr<)5`AM1~2}oCcoUyBB#Y1xqu~u+!MnH4nRk7@q4$ASVfzRY|hG zQZtO+nBLd0pc^Uqf4cA_aI=s)!Xp=U=p&;k=*n6s70d_J414d(et7eP&vJQQu%}G} z`T&=HEuidq^XvK)WS4an>^G!&^w!Cco3)T~yG+tH<2XcAucJ9J3_*bx1#YzY6S+uL zVAy$Na_y3bVWcvcle5vY^#RyPh-6>^{GJ&YZn|iCK4L-Kj}S`2VXKOi_Kl_{U=Hs@ z+^A!6?sDu5xU5b_zK@_@hgY2fOCqGRHZpcvW; z!$NO1jhQAe(;apD(&}ZyN#UI5A{%^wOM6fbp9b^+F5_EITGbCDe^IPktA!AhT~ifB zgRbZ@R~$GFY9oWFj0|JZoVi$b_~B6Olyp17I7Y>(V{qRv#ws<~xuQ(#8Fm`I85>HH zNc|6oh)ZE0S$JeiiJ(N)LxqvQtinv$$f`PY-$z-|P$^F!Z2D-HC&77yfw}Wg?C3<1 znp+Dwq)dYd(kKrN;;!Frz~I;fa$Pf!DYiqUw4&TPSj)#xrEp9d!Kz$%pRbBHo&3X_ zCJu0kUeLp*0eyhWf6iH_E-DAcFA}i15(rE{1u@< zNul98czpXdJi28os;-0f?mlGGHa4tZ4>M1se@dn?L;*F#k^rN!pX9%v=)a%uK@~vZK~eToq@<6}Faw@=Loq7P?zTRB#P@E$YCkpP09WdTIeZ$>2e|I< zUUS;KddPk*pdF$pt#UJEd5HqbcWsBLjKfqly!TU|#Zf1p1YxSUef@fT<6Gax=FLwa z@B_T&l+*E!xBmk!{QT!p3q54pdnA1{@JOP7xX>IdaG@t5bWo9|rQ9EzX&^N{uWq;%mw)$580hcCD^5HK|Mbz1VqoSB_(6cH zF1i@kf9132$``P=eKrDxUXN}f4yvc;LKkOraPO$OonPOR_!pe~?VJ9pz~ccf-}8R> zG@uV~iQjeF@x4=#b{$jABdPdDmX!>3+NooHeUchbp{6r<+eP0;&*DR{bJs51f8R#j zz2UdG>hjAZiNB}67Yd2nxees<8Fb8B48{eVY87#962-v7io*`YlI1Ir%484)9!3WD z;^9r3uyuF>Nwx$nmxFFt2&-l2dv>5MVl?awUVqlR5hflkx#&U!P6L^A8jBV$#>sDZ z1KxkmIkJ;w+vdk{`T6I;sFkr@Nnx9kX}LC3B4+~a2yj{|;ICLZ^Lob+?_T+(|L^3Z zrvZI{fL^m^jp__-IY&*keL=HwDJp4`e%!pq6iUr9TrWmx{sO%1l1pWm$aPmNRGaMV?!?l=m*Eqi z{wxk%whV5gi65Q&VU)LQLL*l2pizKN4FPe4Z0upNl3-=tz#5#IH~;`307*naR7;o7 zz708k#*($`rVenKUa-Tb0sSnNe$}c~+DlWlFYZZ7=W-*ZrlT11s$)n;E>sSb=FGvX zFZc{hJB=$Yy8_q$>_!ak-UB}j!AakliXGG_t1z?`bLd&HOai<*v>T_deg)q3@lT?C z=8XLWWC7tt@GDg`hIeDnV-MlV>wb>M#+zW44QpZuoqipoOu-lv7#pw1e#A&N*;4LX zpslTlg^Lftu}2++m%ri_D0H;rp8x(9>RX?{gdgE|dI1fgKpWbQ*@=&(9c^e&+n7DG zG}b%2Q>Fp^ES6ndvqru1#y#J;ZlC_1AtSGLgdr9;rjVlcsw9G0 zDq+QkK7jVVe%yM;2K?rphw%6#kD}hFONlPB(v-2VOa?oXMzMd6OuH2n;Pc=4E*30b zF7ws0pO9E|^t!GGuRM;*o?X~*$L+Z4&imji77CSN%tDAgUVus_jmfDRe3#UdA|yNk zPdGZdIx%~}e7yG6uf~#v3lMl79=rA@2!{tT9>usDw)E2nhIXJG0<%kPXwPTS)6PhnHecrGXC5!4l2G-l+!m&`pp|V`{uCiz-x^RCN`7 zv*)0vrwix)FIb3brHXYw`5AWZ9ft3FaO-tBio$hjaBCG9g%0@8@QPz!j`y8+F0zHZbVuaW z3CVG8sYFv2LMGQZ;FhacclnQT$+~ONrmA?U?!t3DY%NcrmdnG)<`FoaOewatw?j2d z9RKp8apJ0%!w)=c-?RyLU3Mi5-^U)!#11QirJ;|7v5z1kiYiBnCFsmqqPshDKQr~S zkGkT{Z3n(ufB$DpyBr6&=7XPI(^uX%dClam`q8mw;|9e>P!G^QF)aHBNjZsVnXi1u zJFw!EQ=!`i%GC;Py!jSvdgM_AjXFZ70VVR$(K&$OL_y!IIrz{=&&R;*S%_mP8ZChw zg{?gF@5N_`37xhS7>bVScoqNhk#n(UcmzjS5qfojiE;&dY7LB1fS1j}&gNwy=BO3R zaq6p9$>4RWJc+w*yak(XUk{BZ7;W!Gn9kzZz(Kz%B#V#4%!$?4UC3fapT5IRr{B2b ze{SCNVpB=_OP6jM&<6jEqCm1evqDRCyiY2g#Ga>~U6L?X8Y|6r{{?5m9 zJ()MI`0s1COauA=!SZ51eDj($M(Hs2OLxqB^~knCHab~Fu73bObxjxocun;k0T5M%$qwOYu@mBESNP5U;XM=aQlWku&lj^rFsD7 zIe_mYpHE@d?0HC~GqMOnMU#Z4XvHx+^7Y*aw@!K3IyHe6VSss-CWBAeL|_Yi|zvm9rnl-FkdN|H}^Sdx_=?ZLma3ik0_Q!Z~%O=#FCIPW}^vvU{giGKdTH`X2SkrYN=3mkBp=yJXzY)I!KQ~N8xDr1 zrm%O{9(?#eKZVzwc?Ryf{T@92@D}XZvlmxicPU)Tc${%I`ex3>y&Ham^>^P6Ck)Wv zHy5*JFF-b5M84RD5B>88VWjf7{np#@-`}_Zj#tN=o__Ra(&%CdW*H29Z3X1AX>g&S z?#0*_$9SYt$F|XZFl!YoF?IB()9A2lWDQd`*CxbIk)xVpwCB}i)~wuhI+FJfeB&Z58AhMr6gLIElf zi8Dqr58bFuVPbRwOI99%um9*$w3SM5>P=M2HSFCzg8%)&g}Cd^+p%Q90_3tqY}>vC z+jj55iWNs->B?hdk0HHRHebM+GtQLY9vv&=iXUB!>wj_?+-41Jr4AI@I#9B6$eSiK zwI#>f^g~S6r%;(1K`{z3H zI+dnQJPUFyBAQ1mg%laKk5}>FWEFd(2o2J{NNBoiEJs9mzK5#kqkG0IeC>yq;3Y>M zC7E?HVN5au3CF&@L%8jxU*U$UuSV!Pn4FrxVMiW|51oHLy83$IdI4NA`HCU+o{3uO zcs`zd{0Xf8)z7fum)D{^vKLGt{XLTs(em6B()i#UDIt(2UbSJMWa=P&=)7s5J)K5p zCWk^c4?hgxMgc5cN10Ra zX!=_huUh5&ep=F>zRAX0uE}Ro>9aIexNhDt$BaLhlX&WL=|n&4kqd6S#4NDKLECQy1VJAO0{5odyH_(O)nB<9PwT@%b;{iXU8z z8M9{NOW(N=OAcQq^V0zhG4dd9k|ZVx-dNsH)k+P+yLaHBTYiCiesVd+r>3MRjifr0 z;0?_bS5>4mnpdd9Hh_$#%GRqCajO*tMcc+esY`OS$7@rvFL7pfCl)N~+$ePAEsMW$ z#fv@{{RcpAyYWUdDdfH>YL&OdJgUT8xnH-;dvfjRow}t}HLD0!)yvTAx~a$U#K_K> zK>U53DGzaB{8&*lqnZ)dd8{T@zBD6SDz!nV9GaF0Awo^lTQ&rjg&x4#FkdEF_{4P6G9 zVGIOOOVC_S3m{qEgywtuXjIA=-?;@_Z@M1a?z|bpQ#C1*K=;0;YEoC)&{QdrklO#1 zh5=xf9(g^91KC%H;EbqFJ$#EYfYL>xGBSOiMDhKmqQ0rLbm`!8u)=@!shig>HoEz|FOTi4p5rzm z)i6`2)#}KkvPjvsB+mz7h&owl(uMXzL)82yJGA$tm2bp8K( z$fnj<$SU8+-re~5wO8ZKZ+$n4ogI<`&1rlDXpReOS!NPZkE!@VeT0duR5m7tQGWa( zOx}1c%7Z)5B=pN3!j^yyNsdv(C%eU|VOB7tbOR}sh7@t>Ze&s!=$eYr@+4RkqOVkp z4_(yvrAg*|_Qf(_^RHc~Kk}ik`Oeym65&FY_-AUE(p8O?T#@R4#Hmr$6d0o%rG_YS z;MVrZi^%jGGMOTTXlb~~GOcDp7-`w)Nx=m*(nnJFhji9jI>*$%N&a27Py&q=Gc^+y zZG$A(P1xeEkp0aaCy&E=&oa>DtPBE;D}%a-5xd;ogaJ62h# zzSR30W8tSJ2kV8HbC_ldDN@R}ED0KFjP-*EjvvT&D|!$ulq&+X|5&|!SpKQcKreJa zF|)EFNEbjhWJ?kfN8tK!Yjuq6+l||Ay%BGE)7xaxg`UrK4O~?bbDAQ+rK`as4ncEQ zD9`OS;7pF8KDZO*yMB$-zO6E4C$*w!Ohv-w%|%u_h$GohOJ_hDOB%85NP!6|8lf-S zu{_U3R-q91BTC^UG+Fp$xiob2Q#WMx_kn$)Es zrL9QtgoaAfg_ZOF09|iMXc0S&3KNt`ps@}@-7jmE#PY@}v~?q$Ye!dCSE6d#rM%&M zXyE9h4|=|o)@I+ozIIh}(z$kW)G1m9je;;?rRXMaou@AK=(rRnmsWBA>e7-#qES?U8bEkeAM4&y zX=!7!&QXFS4B^&mD31-JF*Oe6)Zjhw2-=3XqL4{LRa7~DB?=M=W-4z}&_}!t$^PSz z>P@t5TB?o!0u#*!bjHwar%_DnNO$Qw)SURQa!tGB#LIu}JRe5mV7%s^rR^xwF;i`R z)HM00O)W5!$d#QDQtgryY<~HNz+NgkcK@9s38qXBMVOL_d*JIlX#&VyAVh4qX3K;YM}#_ z_`bAtTRD{4`_S3dUrm|XX>*P|cEiC~oIi30) zCajbR!zAs`);W`flMm9ZBj1+;b*?~J{VLQVAM=+Th0fXYB&d@hLQEaoz9WUvDH~EN zPvE}$?!~HC9S7Ua$n}Jxfz#j4sQRc#mnQNsz>w(z04L!9n zbmX(L`j^l+a8f*tX$OfwGt6lhNupS)H_=W=%1tV{AeM!O@1x;3Xt&dt(UIr9bF&xL z-O>M8`rW&#FXkS@znF*q&PU&Ic*-)a$>h!XLby^_O)3@>u;Pd!H^hMt+p@rjmyk{& z4m>m}lSp;U2Gg_9H4BPrfg2`N&4Q-4^w|_@%O?Y7=H!S33Z#gOOTBI^^pN2#3r+rH zOw2E>rTZh9XLK_$zzb;98A;7jjOof2rCh#*j_z4JoA12Hbli{3Kl$W?-Ul~&{t4+u zBfepBUt?9u$}uyQlGQ(w1IT45FqD$!(mE3Wf6r&89f>UCr0Zk6s|ybe?m?=p18;fv z2Tm``_stM#z`z^d<3;IaK%QVXQU6U>uG?1t=p3IMM--_jbOq-!c0V3 zQMw_dO5Y@<%^*OAJFrwP=SowcLq2KC)#fds6it&qYxtL_?;CZL?@i@GvHW3blq>M;P%?U?3okBK~hOX=Krjyp0x8w&}u|fhT zj4;+{V32d{<{{L!0%k2-hWYarpx9P|M{`$#K!UkR8HV9;+;ji$@T%3TQR?V`LM{hc zdpUPSmHu#}QODlByRh+rhv7N_Y(+s&DT8)Hg|&A#<`i=1%IBcV`P?mch~^x65s{CG zTarj6)F=v}YPysx3*$&iGGxqDYll@BGPWT)I?VsckK^3=H@&zUX8&rX@QTm>`y!UI zu7d~{A+9l#`zW-Rqz}zV@0fxVvYaw#R$B9=eFvOe0;?dUV!?E%W*VxIk^V0A($NB$ zN(5Dj5W_^o(ufrkf-DdL*Y}{uO&|$mDohScCHIAfQBrDz7MVoAlNMm8ttkq9*@Z!A zsm{)sj&5fD`6VZva>>C?F%5p~w0Q<&Kd$VpuW0(Q>P7-4(~(XY$Y#@W#F|vnkbtII z2~j^=J54k2F;cBzPqhZGvxuYK^k(o(5qGV>9iHPNolVO)!f_ppPgd~IlRNRqw%s^n zW*-*Mn}d$F0u0jEpgg7HW3pU<*YKpy^Md6oVNh#Vy#nD*K^+@HvDw6owj#P{3YP5V zZOQ4n6klc?bflp#KR2ZV*p8|5zfEUtO8Y9trdtrYEtB`U%v6!@l1P=! zNFhL{r4^CCPdunW10Kz`mg9I^14%SujP0M2qwXL6P6bWh4bZ4o;5M5GV;{v*zsQ#c zeo^=N*~g!8#z9rjBrQ8_-1~tu+Y=<;uFov#FS(wbsL`C{@( zhbuMgE>9vT<#Fo2d;)Fr7h~gH>v8ihe~C>`ZpHBUB+9-oOVo}>V(471fkDL&QiurbQc6Q>(V_%7x^XI@Tmu1KJ#J0^ywmybVreKC_rJ`ykxhu4j(u7)49TlWI zgMj8HhOq>+?2u8I1TzJZuIC|5lfHBvmd4OGVBMJR)!(}Oo9myYZuA;;;lk}*q36CT z0Y3sr4rfV{3wRtvG2d$e53<<$?8zQE(=JMVW2qmANOFf0#X_WI zWpJ2PI1HKslC*>isbVD*5<7$pe!{?&`B(B$D4}Q>Cb+60;4uO}g2DtEwJLJ?e#fwK zYx8+_(UR3?G@sqQ{^!rS`@A)&0bcuJV=VkY971=4mKbAaXS>`@C`Benw1om)rnR&> z_E+;7o`-!?RqULcL_^o`hX4Ew<{W(-sVQIR!mbhY(Jjg40xcRc1VULx1UFOQ=!K7=RlybboyE_9`B%xW*8m?eHjD+VDz z)97PLk#d?J9#TU>B!fe88_9PjRntnfE_)H9Btmzog!c9#+DgiY2?zGs-W zV{6u&XLWa5nfo5xJhNG^oE3)6*ZZE+7Y43D5^28|vZUTeikdXltR|6nixR8siqYk_VoUuW4 z^vr~r>y#rF=;HW+BQHue*CoNNu>JG<TLhaMMY7eYs5KGNDGjwiaNrD@y&)KE&?VF$RIr} z+wQXeT00}HTes=p^PGuYD7I$+eSaFx(xB*;OsP?kgZx{%0Pe&jY*UBX-UCMSqUl|T zYa0hLe`=+Q(CLCx@;-3o+`eKqCAkfh#HRr_vZx|uVQCmVjXP_(C=?lG^o)$kDQL7z zQ~L4b?lh$)91XQpk;-&ZcI7e)#y=f);)%~WZ%Wn8%EwQBZP<)2-?6(=sMZ2xEE8>o z4Dz|GEJYIlGTB^f5*4X^Dv-jBK@h|DB23mD?3|p!-l-|{ylf>-KmR|F>zM_Ogv%K) zIl2!UZoUQg-hB^-heu_WK^8w$9Yj^l=5tuF@+d4{xk7fVC_KXO@CXL?>_mBVA0D~? zK1^-A7YoxCX1BMYyOf7z=};*5Pk|tTS2d($80~0`D{@8)&GL5r0L`W+Z_f5~R%-N- zJ4eAj1-#w;rC;URvu|E`@w)ojwP0)4{?V8>cG5{F8ST0DIl6A2WSQx=WIB2mGBcIg zvu`hUlU%~+9*mCck$Lu(w6ENz|L~A9zq*B#m6iH@imFcujxW_3>^s2NGroRL3uwwq z&(HBwEhBYt)R^crSsiwFkF>~kIt!C>)RZrl(Y+$e(v&Y|P6b-A6HpBZn$stXF=SA< z%vz9tZKw0#yfux_HWH~@Ouv!-rySN!DJ8lP0YV-?5jxEMIc>^6lVzsHbti0;Fw8o6G98n-?qWkvj!szHI>WwCbcJ08Hd+tH8K7~U%iWnewBcG9h zGNvF9QOKR!?5e)&QnciWHvQtOVNTD}hkh>lk2urij(c88y-JL!0 zO266xrVW%u`%l-cTbI1|z3B5jCXWQi#1m(9SmGm{}OU{NAL4qReQ1;&3XM(%SH6OTV5OZ~KfruV(^#&4O) zVDLSjgl`R}b|{%rL5^BT87Y{SEdx9Pvmy&BOjh%FZ4w&e(AxVXn8_z=mAnJQR8G}W zWXGT|D$T}7r&?SM0)k{zf&fdF!$IAz%CwTLwr(KOl%rV+iC{8~N9B2H+Ydrl`my!8 z4{r9Nt<>hLf_=1q(bDhyHP+w(0G$oK`@|wEu(wx6yiZJw)Ji!TH<7XAc0OL7lGmoM zuUp=)1GQ3+LgOH@pt%LQ%>Yn06MabU}7$)A5q4 zA*rdF9G*!8B&ShDwLFH(WEr(u1!2TdZ#J>zkw*~k+Jku+0}FdPQOf6Fr){V-7CEB4 zb0kN(IRRB(8OAD8lGs!#7G;Kr0#7{qouXM*K9#YVRqBG7X~V?CC`Lv{@!MOj$MC-GEpsj_>-$IC z(M`I)Lc_uYoH!?R!|u=FXVQ6SWUO>!5?p)h_|&dnKa*?qi~vn9bkh&M{`w?}zwg&} zwKw7qBE%$asOqRx90cef5n;%D|OhJ3{*{#`FT=BA_`~5 zrcn>Xa%+*1ydcI%t&S&#M=>6H=v}!S3y(YkhM^+}d|70XGreh&7P0aP1&Z1#eo&se z>A-C?W3pQ@EU6@pCET?f@h z4ef13Vu`^ZR)=6xP`w_9FjXBrZ5_zlSp>}{+Hxqj4TKln=BV$RXq!=}E4@3Jl2$aq zW*fSp8Z?5LWPW=F_S#l|rnYT!Zj>R?ibn1VP89*LYzWywpv?*ITG07*naRBg&nw*!d_UG-b808c=$ zzYB#glqHa9uh#ekpWGMG5)y2U0A|wPixJ3SH=1I~)RJ^%C{L{_s7%%2^ETv*9h3>F z8k+vjo&^hjb`Y#I_MUU{%E?|RdN=Z#dq7EgR@b8Y&!KWv=kCcf zN=P7q$Vq@fMkIp`##m06XJZ?K12zWxU1Vd7FeVw33n4}?k}e?k$RtiK_g11@m5R3=G@gO{K3#HQv= zmy^uoO<|M5$(9A^^g2kDX_TuoaP1On;@wK8_ty^vY3bmrpY;SSRo=CBVt8&AB#Q8i zZ40Gx34tHsz_DYfSE{n=UTJS)_rk(t<*%!4?0%2i|~2hQ{~ zT8neAoXJd(MzKJfd?Pqdu&4#jOLFu^9+r=-Bkv7R&I;_Qy4X9@LaS0fc^gciP2$Sb z2_D|uKu`b{RMe9UHOocIQE+C-hNY+&`3a5}8jco_w!^g@nNent;Ta7BZXV%Cl;A&h zJPPZ3&K3~Bmhdh--iJJp_q2Ioiysk;hu=aG zOtGaXH7ckz8p1eZ|HZb#B>2xwB0VMY1o@yV@w+-Hma@Q3pm9+nOq^0FjomV`IHdl}Cte^9$JO_Hkrm9W%9>K#qfu zSiCkVOepA%d`XpM;dAXz*+ZBCoXoCQLgv=zwE83kDD`0$55>S=-{ zOGi!9uuyfeP_Ln0Dj`ZDYzhol#esHL3>m#B6k*|`O&w2ZlrdFx;g(ETnu)D(jH_1L zSSbM4b#O+>Mcq_TgfZsS3^OJqed}}96!t&z88GWDNj;QIaoK#aJql57G%@r?*x1~_ z^*3LSyAJF}&Q&9U0mEWEE@eh2`0s38_8B}1S;&-!@np_BJ)~O$Ik_n@YgA`gS4EoY zUl{qxZ{F8%_<#!f(n~Kgoss*YH0)of<*DlE3A&z#rQN5%vP|IxFw&Prt?~&Z8+5P% zt2P7OwM9+rm@ZV647`&o^MZPY;dVi;J)g6TR^WqrjJC#lZ70jAGcjfsNr1KRCFIK{9D42tWQI$Yz6>3)CiK>dnRmoGgQESvNN;6!w*2ez8N2$>8)J7GLny!n{`dAB*O*=@| zJ@{oC|K2za!|BseDOX_1vWE98Q3_((V^nH2Y<0R=ZExbHwc|Lx>l7rr_5pMAK&3iq zGe&(V_uLBX{}GBjhLWaGLuMytk|U$|90_cCp^;p>CJ3H5_22-;*nt!w;c6s#arm3T5 zlNhF<;w9Lw>S#HRkj?gYJJ=uks45DcHCxBFW(|5?$b+gsYZ|4!HRO?wLt6ztQB&V$0HavU;kr|^@u4Ajch0bVz+oxJM{DNOaZc(UvvgP3a8kz(eVfg@X^ACZw z6{sU0Y7|4`3O)#s-*XUZlEAjBm}xG-vPxuBw!0zO!za+M8@8 ze^qC78N;<>nA)?1#_SR*wvDuirI)2BGx0r)HrKGQXD>{rB&+gCaJzHTU1)OB606-D ztY9!|qwGvW(+k)J`O~%(HJ2~u= zDH6C&DOXMNpBDF?`MUcm*!%kXJAe1_&BQN0c48}hp3T9DCt~1ZX=)bDYE{B6iTM&< zQB4D5-I&`&b7(-@s7!yryjGbXv;!ze|h zsMuF4$w|!bL42P*nr)is40>|XHyA!7r@n@56&EF9;sq5eh5T+tkzjYJ2{(bP;ctz{ zqGowcql7(E6KVq8nGTJ5aL(V3!qJc^25#R_aaDH-no~j5G!R&tWQ`MT;pd}dA#A6L zt=<;KX@Ipe&%oeM|57NzM8U>EEH@|T05}sO6pkK(^8If>-PnXW8bDzrhZmx_;|>^p z0NbizwzUH#w@lBGwjYLP9X+~#{UPh2PktliijTf~ucGANNP@vme>6ZIv{7GNM7>^v z!K+O*!7iC$HiV6)w(pd7K-EmBrX__70kavgwjQG}!ur_9#(0D^FF?a_P%}6rXXyIA;L&?qE!0g@2DTuH#fQ!{G-P>-L{TKH zQ{i4aO5l)B9!+@Tq>otu7E3jhR2?h5K5p{H2#W%zYbs8!l`vByV^9NmC)aIsvG#~L zfYTD9ob8KAk zJRBH|k#RUAnlA&zyr zh^;xa98*q{i5@K97s@71^HTYR3$`?X(RntUoG(tc0B@^5lhcKc9CKUO@(aY6!Oi_-N0Y|!h zTsIoa#`Y{j!Jd+fxw;F-;e(apP#qS09DusM2Cdtc zB3Ev`4SF1-QJ%$Ya|yO(px5qwJ=*HMl@JBMeJ{)LG19KG>aPV0(OMn+GLejj7c6D{J> zjvYutZZ{|7tKrzk(D!ky*TW!7amv&zX3Axerfm%SqVa2$N~o7yDN+h?SN)Of02b;^ z4EsH_I~^<;7FvZOy*=9-ZY(W<*9!DK4+jSx?v5j{F?pDwVY_Q#zE*-^Y49W`F~sfN zF9c-@RN<#f2tFjAHWSC{GhR}4d0j-czqNGYT%EMMjcdIRn*EZ91ejZz${RnnUPa! zLSmxwVzM!0ESYZF6V9~8m#iX=Nyr(&P%~M5l9)q~e}Yp}pvyOu&pVN6US`CwCu$BB zjTa1XbgK>BoR$wt7B+>P<5}*GcM}Q=Xrj zg<+bAeDX;WX$gA$fxw7q5Mmqz(8(yvfpVn=!*$T>bwt0%zH)VQ6TPu7A@+@G8H?41 zoPGn}!#c5i&A_&LU8a93Ti6e8g#kLghk3h%c};;CC#V}1N+yHvHG#v(#n>AAI570l zO;c1g6^n+7Im^aW$wArDMUJ%@cvuS~tc4kxhJsxsT~zw*IKe3&9bJa9 zwE=b5N4B~uo6_3hWz1L&%(dnahW_%sx9xvX+h4oSsOF!#pyibQ@LSHYH0SF@UQCl4 z91PZw=L0zPlHi%PYM{i;)Z7#*jp>u?L(cu#WC7#w)Hb2&rfh~LDsP&n#G;fCN~WCJ z-1Fz6*dWr8E8HHHP)I?uq=^_%GnXMrLmb^ajx=?J%WIqTp5yWdML6`|*Ej7ydoX?U zJ->I}RIKTrbQAMg$CuZzywb+bg?a4Q!D(E_pxc8{a^ZV{EJh&S85!(S8N>cS9&m;U zt!=fX@T;y1KMJw3wF&4JO16pFYC~?gjcy0s@kn}XqfoFgHI2ehu{9XT;ASK-YEdG^ zS$8KbgQ$lDLiDlL?&07#z*?50Zm8I4>uBkQu<2Sf%E?o#^g7rG1GM8DqqM-ZrDJE= zpo0oF(i}I^epgYj(X6A>tO*iBx~D@=Kd}wwu3&OwMiFD7iI*pc3tCX5h@S8Z*z=cf zKx2Lu`=bmS5jBk=yv`=Z$BrXAejLV$HQ867y9sSv$ZK3WbQD$1M7`E1I-SnHUw_k8 zzegU`eH}&*s-Q2u=pwVSqw)q-vHvEC6OC{3Xt0Vn-opHj?eZboMF!RD!){K)s@CKQ zW?>S&Q0W;V1EO*A&Cw>>IhK4H$f>MqmJ~Es!d&Lay3{B^p#V1{x?w?4CZYkE{)!A6 zy%luE1s0nNuuSJ9*OuQK`;}BVzqM~4fzF@mpF3ZChUxCy`LbqKdhh0uwtMT5s=9AQ0R@c~s^$2Qjz*;N$h7-TWlvlO>N!2!cW?3Ce^EkQGQLf0xKOw}uL+p^W; z_oPOVT1)Co{Wt=3Frj83c+N{v|IG{0btZE}IP4=_TSd6OiF9)vMSBxUuY>xD9;)Q) z3_EZWAI)+DvF9zX9AEzJyN}*+<^B1<9#lc|2Yl^=Z=P43+RZ31o2Pt15!%OZM18i7 z*5VErs)=mSL9}@SZoL6dwb7WFMX6C2PBc$pQ@7a+Csmf_TkyUb^3KTr8 zg9O3Q6QOK(^qQ1L8WUb9L^x;y}f9GVOOeqY*YoBlO}J0~*v6K*P4MtyYqD zh4g{V@fh7O#1?6!gd9y3-*-xAJ5?A~1)5!jZdY<68r_wSxAvDzZFaqI*8NwlJm`YvXYsjzz4&34VSTzNinAnIFB%~Am*Li@;fV9j1e#0w zLnM`kP)1D>QQ{az=0jQ9jS~B54j77xF&KTOp`^Je%TPD@FiFEqsjv_lD(R#aEWIj2 zBYQZqJkl&eoQ3ec4FuyHrSf*vYb<(EmFGp4fAp3kM_&1~r#)@r)%8>UbLgc{c#ggq zecaWphEu7)w8=+Nq=;j3FDQB9_Dd|!bKztV(vw$0@oYBHK(gm0EJ+Acd^}E)-${|6 zQxvsQd6XduQz1VEC*@J&D8-P}TJ9-I7N)91K+13pLmnFDB~s|x_eW@Z9yYkhNl6n` zFz>pUtyNIwK$=FPNZ$&5834z5fniZ#Qv>cMaNRV}sJdv@D)2iAh9|%# zW<-e`!$OgU(Wuk@Q8enksedf^!O8>F6+R>dO^!z;sJy;VwKwENVP>>5@Q=c(HeuSX zm~L~MmB%Rt!vU&GJ78t1a68iWGSuk>*tIF>rVcOj(TyjoVOz5??KI%%uE1lWY7k!a z1oaSW1dW`piSk_37kPv<3lT>njC(_ zkF~!UtabPBR&hz$l%dRk!jq9FDmOkHj)n=FHh$%c)8$G%efe}28FIH>x` zkVc8bNTlL1onsWm7=?*Y1r$wzV`-RgwJ=j}fP7>wmSS;W3NQ#p*cgwo?gi+@lq3U7 zmVv#EI%=dTFc@ewL_Zj#mnG=ssrXrsvjQtdfg?8Hn+lp`3)8IRA#TuVi120jj$e1Mix2YJq1Y?3sOS0bvW+9_fpuEdC~e6d_xUw#Wd_k)V6JV zQ1J@y{~s5h|Mc}cK0dX4&veiaq!ps%>c?UBTs5=p_^}XF7NX+SL4hhIM?ASFC$h8@ zvTdWpsvy89nm-PD6twy*Zm~>LQMOEJQ)(p#)n*+A{ZWW2L?PlZ!eBhYy60mv2+)pF zK`Rz273^x(P_ej%Rwjm!;TXd>fLCO4DvygCWc{pYIX0_6WUAuM#Mz@>b+LO}1>0t( zaOD^7K^mfP>h8Dut?9q~%r&>({J&pvEbOhFJB`73vzZ`SP&D&xeJ?$aTT>GMwJe9D zE5|YzpD$EHRSiQg)S{E;`kh`lymcJMhuDDqlo#iZvr2#3|NQ^`@7PO!>5_N8R#Egf z1^&>=1_w}S)&ZR~MGixXT!m?O6Y9(&^ioAu*EEz-+Q(2V>=n+mY3oSy2t&VzTvOpX zRm_&Ah4U;L!!%=KfdaAM*qUBN!HZPkhXsUrL`4`(GS$33dfhG*qY2lo3p`k;d1R@| zsS9U3a{0l$>R)~RC7*uo@Ya_+?yf6UwLg}E;$mTv0z2`Yo2acP)@Eb~sw`G1THdyk zjT=v5$u^{=0QHcBEqP$G!l_F%G}K%h&9VchM9JHP+=Z0l)*SITKYpAlyeP1Fr1hlCN{HCC+Y`F&vniEt6kmhJCz5~M?e41K6KLy zhqs(_+MVA+cWoU?oWNByd41G5vPI%;Lj0Vn>Vk^~S^(4PfPJ#22;w1)5(N&rD%()P znOuqp4o{p{oT^BHlFSR)%F#nQ_5ygrfk?1MT#ZMG7^`!To_1|4m0VQZs@#xq6u=8c z;+-c(GklLx$mzV9W>`@QY*lh-Q-ErlxINyBW$Z$3mJk!U;X+GQ;I@~scl}PZ@(?JO z(9xXWrlYIB*U7R^{x`*_KJDyNuF&kt1r+A?Js)8ZAojy+4OM;Ky{oJ32XmR%q;{J694zD7BA^tcf3 z$Nbqj=O}#wo#_fJg;ZKYNJ#?Wxh_hMB?yMNlyX?4Fjyy=Q;5Lfe;J{3^Olmwq6$Bp zcDpg#1)t7JA%=mEE$R$2MO;GXD)iX1g-6bZ-sfJxqXFa?VH%NO7&vn{8sgLawYh((^;lDfQ>6#+;J`-qvwMMHzZ|8k$)`!LSjKfJT8XHIQy>2b)i- zUh~;JS>JfFztQW)VY;DcXy~Q#J#|}C`Xm1daZo(6R4UUk;xma}9B8QJLB|b@VQ;Lz zyZVezp4;C437Q_)8O4ydI5jrmH+{A+W`d;5(yBDn}q?|ku#m0eMPwg`KFb?xD`KU@nE zjTQ+sh%|IPZ`2x}O%>%e7hZT_=c0=)y3c^B2Xcx(}tNPo8<53LZ9 z;G*-j7Rr^16kg`{5m8J!JBQ}db}3dR4iRr2L%(dIFilk53Q8t9Ar<+1LV6-brQC;C zw^5RK2JyTTGQml~0IjT6jzSyex8VQ)AOJ~3K~$~oghJAjoUfB4RT74Bol-Czr32r- zCcowzHFXg0&c<<+r&*8}xs?C|_Q6QAnJ~$nB+0WN%G#=m6OL^TP0Mn6!-4zV*~jeB z6R-5#-koTKKrr@kyt^f^Sc%uOX~J@BVW`nSoLfPfy=YFa?XkO@@m9C;5)F7_YL#}{)$$YlOeqii0Bai?==8;o zqL^LSey{D+f8^NiamOlOa{1*Sx#xit?N9pJhrFQq!#?rWKVH&Q>(hx^Jc)AHxQJ1+ zHPjtP48k)#g%$)b0-$x;X>z)-SIou(1npH6b2AfClEx36R8(r~RAhxF(Fx=^a9tau zq?lTA#gb7cOt3t|aPtJ>)P~#KCeHT>PdtLq{9tnhN}rGnQ-5=2n#DvBTgtp+$Bu+6heHO$T~BKgLL(7xk~w;7Ih!M^&Q zFgDardnles&bmVTW7Cpoi{WsH@pv@Ro}z$La*5L9wZ`m+&wBW2fACP9)IXDg{_aH= z>6_K)uYlr>v6Ab7l1Lv$*r@EUfZn#XiCz+mkd&SmY+vb`%Ks(DnOF$7bF{;aQ$9SzsV8bBSJP0kEs;%e zn^!AYD6u&C;O9(FNaW;ZOEKm)&X3R!BP2;Ky?Vvw$*&{Pbd2%>cNQA1bf$2m+(f1` z9>l{NpS+3l9$Q9Z&Vga(7;Fak(Dgg;+0x5l)#}K?2+^ny)LIh+X)uDt8!q$_lUUVT zg;SkEZGI2@8~^8#wR>*))uqC}Lshfi&s6+Up%|woc`?V=ME?-J_X$56Mu|W{xl%4> zXJ?Z%%D(gcTW@&oNu=XJAJjjSg63~^`TPFp2ErH`P0&)wpaQN$C>m_01Fd;>a*3gfqc z$tj<-O-Z#X(DnixEmZ7p?LzddU&ndN-^WuAUyGTD`aMG;hP}wg(XCCK7!A=OQHYdfU`Ws8WadLmyV{DTqiPvwmTffL3M#H6svH@D#l|2*5GF#PXA$#H#|y9#rWhre zg!{B?0~LqbO%+j5;GVd^cMBW)7xp4(ElKt`(TZJe8}B@S1&=t(hVJOXX&#&y;qAAa zfp4^a7v)kJ(P$*96q5`;KY`}?&^-_1!}ntSJDYHb#kM?5!-RpYg*U$4IZ zJX~ZB=1`e}`^TngyunMUwF0WThE5t|80YXAa-`+A+FI%DVI2JE$Y}Kc@0Nov;JY7v zXFXN%SuYqp%TGOwQcntmhT5gdgr5P8SCWHTmZO#yu&QMgwWfG==(LoL#z?m|q0BTU z-fv>SO_xualsFR}F2Bu5J6~R{T%MRofoR;9pEn9@lxBCrwwy^KK$;*3b3t3ID?bm9 z-Vp)aL1V=6#4>^XE?GPLSzK4MXXmrSBrfXym_NAc4{`6APeuQNmt)`h)p+9me?xvW z!*pa~(KIn#uEL~+7u`qcazqx;$+!wxwou_L+CZi@PKsR6JlTX2?fY@4W{Myp(I>#j zqa8wsewbjuc7uEjU6(?x*}AwjF>7!m&2Ueu;G6C=R+dgfHnR=6;RwQ^xg|W(J&w0O z`BrS-Wk9hwk!KjJ4DjY#AA=hfejBc3V>BMY+`e5n=?Eg2qY$G*2eAItOVGRdyW%%E zNvH)1w^8W(o^n7jr=QpU(TC_xs{GO3=@+}I_VI>ln6&n-s|vQ#468|kJCh(D7WvtaOnA5Z-|$jMt`K$WtckPZrgcs3pXM9D|G#rMa2Z04)O@GyulXgx+_W$p30kt zPC#jGN3}U4ICRW@Sp)$Jylbz-`e!~2T~^I>3Z0|`5b|PG(ddv54fTQ6 zVTiStyb6PJo{oi$n{fW-dvM~$qZlr$*rB-CO^KDRqog|%o(Hv!V(%!tw3BP`LTtFi zkK&|0kAocTcS%DXQo-&;7|_%`PBBgkcsT(`1v9penUaO7Wy2D}eJ-2717VEs6b^2# z&%xV%CiIyl6jljpsRFanfKjR9Z2x+^^w#Cbj%<_vC+oCFZ?^=?|uTMI1tmvVVa6Fx$M5}YM#wL>@~w93-2#*S^gjAKkP-d zO!<_XC?>NMIUfYMg0_<5NI8tAW}G)-I(gTt-*f+1boaNYe`W>EU*PlaylAK4+LxQA zc2yRbyhEm8R%#wxKjQ96Gd=63z$JA`}#aCsuL$TmEmvWPYU? zT+0%EG4+d7GjsS9nHZEwgo8MtawrV2N?UZgBSZBfPeR`xcBA}y6Xj; zM8_UO$AYelN`$>F$?$|I#pX=tmW6ON*!%i)Po+a?LU*WKAm@VW(*B8 zjwM<`4u)Z#35<7d5aNbR!L_AntS_DlV{QpOHdL2CK%(1{1I92z_-A{FAMko}=6J+dW8Ln%(t zBuU)jb)P_4UqQ{Wq3_s%Q9^oQbt2UvM|gGyC{?i;#j>5N>C~@LQQU*q-aR~K zE@Z1QqRy?&iSO}fWd*mq@13wI{~_Gx1p@q6EaqWPR@3}_Mzot{&(703%h8Er9B~|E zuYWJ{))L&|GA=AW3TNd8EZ^P1iGwNbJse@JlVfpK$HTX&$io8TevW>(Kve@Q?!Pnk z9d*kP^?FHYs2m_!@Hr0qG42i{QI_0gRdI0FBcN>C z2e&p0yF^oX6Ha3am8B(EB;(UVBuTJu<2!iE-v7Yt_L5AhB<=TBJ^bx0Psa5P<{~<< zcPzmvm*9keHH>ik2j7bMAASnwR7zOjjYHIks z#b>i3o#iS!gpWX_Pp}SK)kQ=brNLxfOmaWd0dfUZhSnGvV&b8?33Fo=rZ-0B)Fii0 zt2BW5ok&&=!d{q#m(nlIM3H7tjvs}R`p6p#P^RX^6;)t3uFh#fP6l$Epu4sL@7fLU|;thsFf*{gLVA2_Hj%N zZ^rVy>sUV=VgGG0?mgn+QBRu2)6a9^jT78{&j6QSImXUu6;%s}yotLgHFVN9CqVf` zfjx$bny$e$bU{ddEJdSPJkzv}MKSir8IDjDqp4^aI&PZYgOx`;6Yk6$T)QGmkZQGt z=B_=^t0ihB#PT4HGVEIaKHjq963p$cqM#jH;-SA9;*B>v5!cn8i*mC9ch{0wEYt#^ zx4e#(fBZu{W^@GmoRaWz2%nP9%NCK$ze8gmH-w|nQ)jBzJ-S|cY8Hh?!67z?(GN2m z8T)W-2Rm8~Y#@X>rDE^)*{@Ek)t5i)eP8*BUK{u8Xg>KxKl6hA;U#aINmcvuAd4S6 zQESKWiXjSn;_Occ4-(U%q(TdIY|5*27=8#xO+}vg z{k7eA{4<&YwVgQL$2Y#chBNk6P_3%+@DaHfjRDVRa$br%zBj~dT%f`hg|`X~Dsz&q z3XF;ZeIhx4xPDHREX>n3*EaCE+G$vS)H6|Pwqy}yyH!l@+5x9tL&o?IX1y>nBuX%K z>}tGq@q^fPDn+-XsCsCxM);d+9*Y}W&p~Z^3e|1%;#O4}#JJ~+pThLl-iN2vZB(R{ z%VfT&xOBB5rGYeFV*Lq9T2_oBUtt~$UtceZuoc8Y^xr!(g_&|?k`7Eb({p<&|WVk*JfZ9y}$2Zkw_5{nfEWT_>{dTm6jE3i&EP1MwsWf`_D z8aYbC#O2b|vDx;Ks59b(uaA1@e&JIB-)S_9{PC)pu0-lY<0x4tRC#2;RzzB+kBu`Q z1@qN!gW)((vk({LU&o6Y-$D084|m_OiQBJ^;oqC%7k^_1PXC3PIQX~M2DtvZ4V-<> z3>wq4Kp-}rVskUVi4`AP>jA#{tsd5n1+tXcVVTId59JL(6*rAA4Gj)ozh=O7TvRJA z##xNd&Yp|$qn{1UcA=L`s4vW;I#r*w>_#`P1oSvDfK zZmr27#bMP;Vyt_ikjKuNrmgbHMN-D7Y|hDgk?KZl7|LBv&3?@VU-{k-{wwax{XM0B z<^@e!=~qAYw$~e`{r6dxk=cnT9wW%d$ZbuoEJNWUz(R@H*8&oGE}>hG5biZiAT#nRbZ9u?jnGSD^d9{r*spvuH13SE3}>Y` z;nh>0fg1QYaML=jy<&*&a)D?4#u6TWep7z#U~7Q8_qTD{nX{<1$|AvHd)4iXarAH> zYik4C_QMX=?}{;P+t@ulJ27pEVpN=xL{k&hAtR5yFe$2)dKE{!2%p>Ybl{Qa37S!z zpGLV^hh>|BpiBnjA1ks`luq1+zpwri&N!8xCp3Wc(c6e{@ztl{%IYtpJ~NNT?2PbF z)~@^}{7e2GzdT*WF0%jV*+)(aje`uT=}Ej|?CAtU_3{K?IJP2N&{>OqC@SV%(paZp zivDnt63k*IuZH~`c_VbNE!XS=rlr4r)}FfJVV7O@Q*oky<^?U7^ri3p70b3iJsOU6 z?ukh045J|mm&`X_^P=|?Zib`*FbpV+QRRlL*^+p;iTy4|NDYQnm*65chDjK~?{C7a z))8#2V|e%wYG*tWR%wz<#nWxlKc_IM(NlrJkPu%Ds#%i{y#4)e@ zi5f{wEJC&yu_vo>UqS~m5pEe76Jod4&+{fC95$?XZjYHRXIBRYJ+h=Fd z?zDxVRd!0^6&aW z(nQUrbcw}76LcD&>qy&uw8~{vXXr&yLXfMd(Dp!^M5S?fq|K@~>|=HRRj4nW1}*L( z+}MJ<=UliIdd*GZiwpM9Oh!x-X+`+(i~=!|Q*6`1 zHzfanr=K~HzBgk$#ioxBT(f}7ieEu(ej9477F-oryW|~s%F$c!=(2&Fp-g-@Gz~kO zO*wJNJ&7_pkrq%)1EB_d*^6=Yi8a*X7_UBaH)d$HKs-O0WK(i+#o311&`>Dcs+f0D zg6G&~F*{W~YFFL!&b{P{WB0e!`0>~MTnhSIU;NOlk!1U`tZ1?)q&sI`F!0ewtR^Ux zZInz}02DCEk5&vh&D=~yV`&=}ePSKwK4A*addYSe zrY3|RCf+KFj&hmNb`zVkAPmrJ`*K1bJ2b)_-)v*2Jqw*PM7f0dsg@A@X%if$Ii{x? zGThy}at!nHb2uC%_~PR)g!1U8pgdIJ~ocJrCviU8ppt;84u`xqrs9w(i0EdX5hL^$MW&xm`!55`mBAZbJEKu=8_>Fu$-T@?06?x@h5%LX{E*ov08QV*=FU< z+mz;8wq17F0}9{#xfJyB_rKIw?QUE<_PkT*P|2+kJt7Jv!8quHHPx!w(hg`!BKpBF zHIZ(1Fu!vb+;T$_W7&$R45U1A55`+vw6-lsXnV(;fibfS4xKeEOJqhojo6}uL!Lq* z^4>aT8YYa^5(ROxrR)b9#sPil$~gSpui?&1{uNUkbmA06N)mm5VH%+zG^J=5C$zhb z(2XL@{Q29EJ?fdlNug?=2JYLVd+_I#f5m(?KzDVBPkrc+aI}Bp5BI^TI7p(2RRJ#! z@=r|3l_?Me0R{sP?R6jPCqmrzjScM37KHCnt&}m{sv`DdtZ#Ncj zH#>uNp5t>*{4H2d`9-+(>SQ5AI8e+hSK-_!uf#j;)a=EWSO>2I`h22S%T5T2bK0V2Fch zievQ|49j)&mycq17UB0!-w8cqn?V6{Ap<&j?GYhVA(Tx>S;WZMhZ_ck1_lhsUOEwJ zd16Gy2`GAqyt)&)S;i>f?$E+`?F6oS`0UgEYaO zi$8&Q*I7bIEetMU6qM2}yi)x)oZCDgTfpnSa~wBaKgP>mz7NfrrdZyx)nZ?(k;lTJ zP$|ULYdi?h-te)x5#X*L_AoP=Ma8ki+H9uTM1L@nqMx0ahFdD3-5=oY)ir3QfeoXA z>t6QPXrA*};b5!AB+-fm&4Ysl9dR&DV`S`&=^r3kfj5~=1Hxcq=Nz&bkPhu zXqIfatworPX%yBZn#_Pp9{$^j3R`i0iwvC$eoy-m=BSK zV_7z!yd9C^$Z+}9x4s#Nulpe?g@#6{B-iVfA0if2tBQUcBS>@jnu&+K^HQYqyM(<~ zNNTJKt%z`b_yxSQ@l|ML6`d13T=lgM9{=Pe%x#|*CJz&D$=N0qmgJ{N!fpY76k>C2 zh^@^KYX=h)cUcku#28^lw9%@U;#ejn5Lq~~x`wskK#-4{n>#T4qqm{9ZAp4jNt)tT zQ*OdZmN$EECJoR=!5?^d%=%~X`aPGUQrBdm*_u3e5*?%@h(`VCa0Hw{Nc-E00 z;1MZX9U3jtF;>NC&9Wd<5XZ-3+%-;cmubV^eLC!N1*i13@KiH`B@fcXl8$o> z`Q$zo^2ADLJuJQSB<_v*mIjjvY!l6zb8oXz|CJ|P_N@ox7W;E4Xug;mzxA1SjfUQ< zDct9#Q^X%Inix`a%^fsVQ?M6{*;(M1~im(04 zyJWTN>IQm@3+97Is~Gl>X(oTtN)xBQ_0z!AoP^>q{DuMr4q>~3JMbFqy=YA(NDB|Q ze6NjCxs1JM%!tz}*>_w3@qw1~T~bps4a1?2trLB8Ix%{i1;*E_Xj&x<#{&r^s+T8f zA4V`U&v3chg-+R)TgErfem@e;^J$UaIpMnu`N4w;|R{%yammOxsOa0(S%IZFm;)C#z`!^akdXJ zUH)|os|&k?Sk_oQj;CrNsL3api^la-+K2}pU%O=HFt(2(Vv2+(wicGf*b)uRS@*8# zR`rEX`owo15ViE@R?xS6>$86sC)w+hG}qZ~Oqi5(8se46a7ZpMI5Z3pYY`NKG*J?J z%BYwIrd@{Kq|i$(XoiI`V<(Jx7&OFo$y+#?s6^GZ#sGPm3Q~D(%Y+?xBY9{D8H!#n zO=JMkl@tg=WKJCi4jjgZ|MIoM(dI8G5SK!zR1t+VKUDUM#o`|9x#-`3N=@1ko|2;4 zrYmRM!>{(>{~>0dbRHDl5c3|!-$|CC zV9%6ur4lK}pPSJxVH~2|z7=n(z8(7(C;r+&Kg6*kKHh(I72ixAgG#A}J^fWYyK@JU z(GabA85M<&WpsvBFi0XnedIkN=}Oaan_0s0%uZ31msXGC1x^kp2R8k&K!`#epfZS- z6cat2kj0pA%h+bNb>#QZ%yVk7d`r1reZey?`PKtMI(}{ieZyBj{pvI?F3tf*aN#@? z{>DTA8bVbvnYh)0sj`zHD_{YON9sfRQMm{?%lq zU?8h$-5@_eh0V|*rb{Iu;v|5hsFXxi3FAGi+J#Fm8h@X~WQv_#xK6`T@M_+-q>=={3>h9lm=LQ`1$n z=BlzJioyutcm&rq@2$Nt{ei@ zS_$h>gir1{4@X15+IFbdHfxY#Rg!`vg1N`0fXTW!MLr(&nH9UrG(I(DoxX{BGKIsC|Vre2^pkRUkVA2wlnpTC@%=%v0uyb7(;)6 zJc%&dY(Wjj*z#=Tt-UC_WmKzWxTYdxC0gf7I)K=_=CEAmi5udiG$Gbc96`yW=J3c! zxbb)_3mAspTO>kx-FrUy7ysACqRys2JyUloNM2?s_)0I#zPYkK+Stmbj}2aN(Mz-G z#Sg-=6$VOD0L#pA%hccv$GGJq|A@an<0ic9`~|3biuL0id98EXT7YiCABI>xzJZ0M zx~z~%2N?Hbbl1mdwYe7do&Z(xtt z#dBIEs0Bs8baNtAm=hDT&zWeJBNPv`iNzZ~B-^p_YT5izURIv-im%`KfbH6TZUuey zmp=Z4tSG*iCCL&a3+rj0=WHG>&gE@vYVDyAzp?D*7_3fz7 zZAYzI6BUZEhDgNZ`jig^T`Vb^;;Bn`O#G!M+B^y&O*+k?2PTr@Eyyy&8f)C2R6e|=!vZz>F1VGbX)j{Mz-vUE4r zaO=DN2EW|88?XC=y=XO547($7x1U*PieiU|#_cy8#qLw4(U__UADabYvXZdhDSFXAz?9mlm*S$KQXQs z4hXYZm@6(tT`bPGuC`0Y@Bhjb*WY=6Pv!sp>wazp{q0LX{20^Hu22+Zh9|Uc*b?=| zZ4-fYwiRQ41GcHcX8Kh@#A4!JzMa)Asaq%^q0oSc9>C@y9KZwzc^<~t9^>(|jZF`0g8gP3u)%2veZLUU)JPK%%jZ(=jIG1k3i2VSe>+&CG zIKUgdkKzq8N1@APn@2&H!RH@vaQhD?VTnyBX@m*E#+%W{6w}B^8j$A=u<6-w2xt=q z2vrSV)Fu(iq%#T7URlAAIEIdeYqTt`si@>MAy6u0Qwfm;XW0>EnCCqv#MkO<{ zkRuC=r#oqMv9w~TE$5^g3lkHE=`H*AB46oNYKo*^l9h8m5z4jQkG=YxSFY4I8e}RN z(aj?g_AAG!((wns3Fkxa!L$(Kl`pH{|9L$L?n5|(0EMcKyfMPwo;tk3`SPA?rWmu} znl9Q46wp+W0G{|>AHgXB1%|q)7;n^=_k8HYSuS&J9INp>;Js5q11f%CZGv6dblx^=e4`;` zZg7LN>BWga`VSsH7KBuha-pUOCn^=~ohR=*abVy7!Jz)%TDdP;KtK7wJ-a-|`%;p` zGx@y2831b;sQpQdWU`vDTJVA%`wvo1OR0yPC4)y7F=Ez1XN3fv)d9lBWl;3Os0&4O zyNtlpdR;3)QD{g&93f^&$iy2^qA!i~$J`bdGYF`;LIE9vS49eF-6#&rmCB3$zsoN_ zQ7+h5<+Ab{R3US-BIGt@Qt0jKr10eH-UYMVL@MME(v*ytMV_RczKaw4KLP#j+p*D5 zuxqA(8(yYi`{h(djF4Als0QWSBCNp>2OsO;vON=!4H@phLwmtOYc+uB1IM2-VXPFO z=M)a0O0|rkW#Z)AGN)2a@-UP|e7$}N`fvY3sO18JbAz@ccUKB#Qnif&nyVW`B_)R{ zmz>5%^?9Lv8b9OTftO7!A}82LGZ9w%DIRzN_{39V5HcF{oPe#0fLDY==y3qrFjML# z%Mzm0@tBs+z zz&Zyqtf#36nISO|_xUA(EGvN4@u!g4yccSr!jD;6@yShJqSq{MQL}0$hZ0L_>oVtFCrajn%z2j__cTZ`rB@nGPz$#W1%4>f*=y1Gfc5G zu@S?c{!Ii1<--!9oQyO>H(n0jpocRbe?O|<-j5eFs@PsGAYb=!?W+qYSBa@dkVZf* z5>mm(sY6}V#tJA@HF(1it@$CAS3LCX7_H?P@s4 z->q#$>lfb#wOVE+9GzF1OnF+=Pv?MwI`YYQi1KGD-BC?C;5~#r7+~}A*YU3Ux1i;0 zh#B%82bNkf{^4N_M+OzB%E%$$dNIaUzXh6eA#1smSbS7+3j;hVC^%MVa@B_>;|Kj7 ztU(WknoBm%&i<}YYW&F^cix%&r@auzdQ)S$oYK|h!0}hF0Y_+UP zf>NF)aAI73gMsNwv(OBxTgT8U1f&EOPPDj=-k2!Cw_>#B2bf=S(YJwNKa1g$5jK^_ zxgJgx8k%lzcLyktGgekaKXCDQWft9E`Y%vxqnn$m5LDq{;u8lrRXIrC60$c7RVpIS z``?gs*BK&e&EwU>yYb?&IcPbu{Gd^-3=Yl-xa)8cfoPyqu3&p}412mqQH@FbMWU(z zS_Ww667L~L&pU;hkYw}R5;uoKc;*o0qLy5`ZQDC;zw2|iKik!JEt|xPwnCqL{2SYa zICGDb%WvapVWRDaqaHu4ummV$#5D40D8 z=75brsnHbq4m@W7cQAmu`~=>t-j7R)1E_hh^|I-Sxc8`pN4s^D$EUDu^Cn!nauAiU z!-Eq}-$YH%nomX9Z#~q>y&_*M=BUnYlGti!!nbYIDtdCs?ABlU$lZ5;Y%Kx#Y>&BU z0o^(F(9R&rd?87aS+>^Y_HV#}M}mH0wG9Uyr12{AI*5%i@>CI#%gT@>0?ULfA6+{9 z1k$Z9<&;BL6&~H9T{l@;a5ET1ha==bKEn=~bXgV3=MS{D348{4hN@JjstWKyk_?J-+!akeek$_WnMEx8ppD{0G4SYbZs=ptL~<$iMM?8 zt}lFcEdlv#kGXgOJ%9K;kstg)9K{k@jy>(xL_&LNRg+*{AcyF%!&HB6q7W0Bv>ar~ zkRdAuA925hc0WWqdks{z07F%I0$C$(apF3Y0ViW?sc=8~_1JZa&;<671Dg6QRF9_B zU*=J*1m?hmqUk)@_*;MZF+6mpi(I)5rQYPceEvJ%!0KQKqqzySMibSITOc*-NG001 zW)YDU@&1oOx%bn!+)!}i){WRyEy3;f;n)LAT&rQz%`%h%k(&%lu*y-F+r;|6>>>Kk&~F&ykwd9fgR!GYu|y&E8c)e1YCBr zrFKQi-_3U%oIQ3N-kIZ=eBq0cj+hKcDWGHOnnp-z2!OhZ#OU_=&^jk@eft~O6&;3@ z4dDwh!bHZBvKzf_il%uIleHW+&6MJ;mu~#}pWU;v=g6YppeGs47FvWZvxf*h7c&!M zo$= z6>v0dLiJ;6ua1I~G`=ZftBa*hgv^HBP-Gp3oQJODVJMV4)3N6Q`&{9fS7=5u2k?eH z1`|>{oT8i>jO?42CUA!|WT3(JY`pUWpTvnlj9j6Ne7S;rHiN@ox(i9UiClF8x>`Va z^A_amRWASKGZ_T^E?S@ZbBsNHAFi8d;`%Kcp&`Jq)5Xv=F?+L&sXg@F1YXrii%lg& z+OXM91apw0*9&oa&cVrJ9#+mqXykNEXe!Di$D|FbM%V$MNK6w5SnT$3f2)TlZ+;VI zf9@TK@>#Yd_Nbyk<;?tO^Sf~92v)!UZA@Q(19Iibb5-8aN=;9$lvZg9YN~utjoRs< zGCYeN2S1I;Y!|8&Le&Ls@h7RofFRl%Y}hciJii(rzvB^qYs?2&Ge zO_R+tlVcPA?r-kC``}sv^4T79Q3ASi>{ua8-G7Qb=W1#-(@)q zz6lW)LmJtbk_z~Qzam@X9F^}0t@dcYaC$Mov2z%MW}$$lszN2NZJUJtWIZ9Htg9T9G-G(E)5TYI z+>GWg{w@SvWw8*MGLb4SC+Fe2E>@mA#`$cL6JShjfJ9S|B&g?V2sM@hB&>0v@+nnA z9M+2aU-}p}xhF8&6fma7s1`&N3t3bOdCbnvT4^@=;eFrjOg*xoPc9Axk#s4n9P|R% zME?(c%uY3rPLzwU{OISucIw#%>ssD_Q3BdrT)e~$?e9g7rH8)F8$uqHq6R=bnkl2{ z6+h@fkwSoEaj&Z_$1GyKut*kI=_leehT`;M$`-B<1IoXunB87$Tf<_RL=!8 zcf3a$4$?3L{wQ~+jT}iDX+uaU1?8{AoBqf9zJtI2?r~(<6fz5EWf7;p^?69uCW<>= z$Qo!d5m4H^1-XLGtCbkX7@m9*-tGIaTL^H&>@+TE)F7o)DR8km7+@HfC~xyIxhn(B z(D>)teg;}jfL4&86m$%%2#-C|#eJV&!nC5Js;k^)CMZ1R_%vf!D;Cfw7U22;7TaA0 zntyk4C+xTU9x}B$=g^)2Wvm z!jxk;5MiS|~y&&eimo~o* zqsRexd}S400<4V9;P_kK2VrW8D-#sZwl%tONov|Ydv%b)qim&qaag1;_!K(`JbHbgo5 z@xU{GlA0PkdPI3GYYnLXH4-6(5!nm(fz%y2chlS-&6*KoP8?D0sS-jB40~X~EbKt4 z7O=5iK~W~{2Z@9zcvG%&@JwNb{&{XwdCqIwC(k~|J_k#uHC$k841z%t#pr>1_v4Qb z9EFqB*i$8>-5VJbmFp~~vD-_?mK)HTlaTW{tdnU#()2KkA zNpZ(UyWfT7xmf92xc1gEcHcb4v?Qvlm&#NJmpPzaGsN+uZ5({Khrz)Zmu=X{)t;4= zMF!+3gn;k@ZaWhVm9PR?S%D{s`2MTkiR6VZK}`13G>Gc>Eb7k*!0O=>XjU}tZk{|f zhwkD$%8f~=l?Ffc2*yr>yL1D@_5kM6DLB0rw)YM}J$gTasmsvWa1+Y)G0cuP@uJ}o zC{Ntu?z*J>nFmiS|JwV%bZm1`%e@zp@;W#01ol*q)0JG3empN{KRw%Wet73dd3wQL z7d)WNrDJu^4!;z*=2c{>Zd(>NRA_}JqjwsaqSb*+g#er8g1I6=B?yhgL}s6TMLJf3 zSZ|_Gtnvs|syCRrIbEqjC1XoYIkF(S=%}eeOE5KMXfrt~c@Rc`CctTf$gWgR10Sgt zER8g~-}>Tb@PUU{VUukfyvd;SI+jy{ZQ zXU4I$Ud04;2qB=|ZR7OPB1=rY*vBpJ97l6=34`va3Q%j5+3l5N+Kw4v{;Y{pCk8lt zV1Q}8hH{|*&l)nEl2R)grJ^!tsZd0hc-Axu;Bp?drP`cqfHnZ6#S zQW4XQI$j}m5Pj|A!PKvdJ-YFEhqvJ*#zBtpiElJCvC{+~S!_`l$h z{J*dA1^!7ZM-J`|Joj^m{SBmSW7rC93MoaP)swr19(uzSgwc?pjA@!CscR=0&r4#t zNJ(z0=(RMG8-XaR64jj3G!|s?>nqm;Ng_Nr3B-xc$+@vWaw4t?a+f_rd9*-{8J~F1 zthS?#5$*d3wL6ko(yI61{!gYKdeRbxS#qBxb0Z09PJ%%5a}d!!@d5um2>#~>u%|zV>o-iJmY2{VIcFH))Z$sbu}Hh^WK7)h8!=3Vf6D*jyIo$vJ4g%gj4TSCJ$F&YYdYa4@_zK-R@x)cW<$N}d z!WnA2M!q`|+jB)Nlf-EnMPX=lyJuUzV|ppk0>b#X-(G$!m(MNeGuzT!E<2Dj*$`SG z)EiT|a#CqyPRuu?m~K>YW@Q0w%f$BSS*C0Tx``WqrHQPZ=aON+-9fckX2oGh zHWWi2i)U>N%>-wEXhUh`(J1P0U5~q)g;EKtD;=1YgQ>A5Q#&6!ehN?79+swdAphof z!`OZ)s6|D2_h4YM1%NdJb^GljVa|{Fb-@GrOJDkne9f=i zpc5<&@I7bfhe?p+$p}c-6G4)L&=)%^r=Oa~nSs^Xxifw0si)HSzW2Q;iN6WQOaJ`0 z&J!K;BmVyB|M#!`j#R9sKMK)dNuKREPWUTteB;L9_G|asIi*he3LX~{BqU@#3<2Hr z)#M!j03ZNKL_t);PawZ(2Qn&YMkI(-^6Dx!#yHkvQ*;HC~DJ7ISIYK z1=sd4QEzbWdt`nAKWz1JrrbpIhIgU*!kZCiGYF_2JaicHk_?B)b~_NW3i5`ELb-r} z?O|xToR5z-g`@H&rGhk`MvbTZ(fhH*J%dB)ZireDiXy|%b?hBiF!s3*1Ga@7TPMEX zD8#pH{mU;8f6TFbF0p(*e2eGNcev%&+h&xlTMzg`c9wm_sq9DFR8>br!=Go*AU8eB zTxbfg3Hfnh!+HX>~2&sVHDtaHe0!E-MR&^T&%X*FfNI4?HeZ8K!wsD&$f8g zqmm&33|%L{%DfH7CR~G$&f~zCQDOKal@4hLq222+mKxZxrL1=t)mgOryaU7F_1g)AjStr zaopGqNiDIZgsLjIajJm&7e9iu-NN>X%2H!m-@E67UwTH!t_yX^FZc?5p}zQq`rMbl z``wzVOn=)GGP^0SX8Q!Hbf`M4UI$LM19^4_i=e2DP31x?o<;EKKgTUfgq`&YG^%cR zE{u|aX0w68@+ubEZOq=RW6vuaBU&keYdn{gfwb}y5{m|9J1f-up;mm1g@;m9as~q^ zDd1OnLoByD*iavbOc()8MQ1R;L#NKb%I0x=*Uc#W?7N^;t4w>c91nw`g~Yay>2{FM z%gE#<7}W}fro$p8|7mmkQACe`T{JX7_HlEMqrUhEju&@B(#q^#NbI~9PL6Ewh))T5KH-2&;TwO&J#H)$>q^Ga}u#@ zp?bx2$f_!{>1b9j(_ew`4}XfAb3S&DH%DfhQHaJw1EqqFg_CnQ)mp)Iw~k}S^);ko z#APy?W~bbkEcX>4tS`b#;i9AzWLP-R(mF{+`O5)qdm)JkMF~) z9Sjp-xjBpE?e9Tx>m?(7iW8vMA0i-xZ?6NHzB4&|6-%%kcQg%7zKzsWITy|}tV%zh zR_M&~AyiI3fH?|iqsBJ`xdL20ZlHeG2O;)*7%x>&9oPO|HsarS@!uWjU#Nvg)`HX<_)i**e6nHb`^h}6W2f-)*1UF`U zTs=Jw$8lJIU!QDnl79U787y0Uy!>q&u=$D-8zgbAnozZ6ROz9BcFYiimdjmKT9F_7 zr#^Idj03vcA7ZZ6<@WqF+cxnu&+&y7EOvY75ax@BOmYEBKleM(uf2%_+H(SQ`+Y>F z4fH!ubbS=iMv2{N=?OtVbk3d6Z3-#0sjEtOiFD~O>ZcySQt2{e^cobEhL9Cp)le{Y z*GG`F+8EOf=wta+V=VWETmJ4__pdeZF4S@B0sSM+?5l74gKI@ud|w!cdx9h`#;GjX zo-Oo`9fVq+gmLK=kO_LjjR+6n+);#|{V;AZvY0JsFf;|aMh2l7IF5&t3yVnAEME5W zO>Eqy6G4Xqh9Ej5ait&){e)8vyBDy&g4|;se{cZn1X)XnSnl`HF>UOaXkzQcIA;e> z%rD{C;wsXtfRZv=eGlLAW<)Q$l|PdjEB(F+r`P4xR4bR*ct|Z6JpSfcHcz!s%0nGn zmIyFbVzem~7LOrw=-U{Uw?L{)GQd$)Wn7vyG4sv+fMubK47O~qoGmr+wmpCKt#4ka zKwazSt_SpwI8(Grf9J_T$;o!6A}=imu6@1JTY49=MpJ5TM1FE3L|&mqsCfx14}1mk zfqQUMS;mB>;F66~jPj*=ai=$gnMTO!dED@mI;MAN%&I1~HFM$U3Q)JvPvHy$xI$ zH?Ro7)tXxXI;*HuOAuuRc}*Kpi%5jeu2ur;n3-%ut21>*sOd6)6wV`Ggq1arpS=QV zV*|9jgqOkOQErbW5jVET_g=%RPeC(v9xfy;|BW@;5&vSAVl=_$ySiy9n+9fOY-y{?39 zR~1GY5^6P5$CJFixw!*)Hm#a8nnnRF;ouKi_{KdgsA&!gbwe{E>bi~^dlSm|&XJQC z1_5R&71RukSKZTk8Ha!N-4Lg?aRr0)GOhUq*k?|mlue*lYse@n5|R)MO=L4E4<0k} zfQX_ry^@5;I3|qV8F-I<6AtNO%1vk!JJ8S!%pUzdn${3TUPiT2N+-6|KD;fTzI)>K z&l9%w$NE|i=rx}HTW`HpELiiGUlK+4F6Ro>Lrw-Bl_r+vpe>)pq@3fy*vW|oqi4N= z#Zx71+eXW_5G6itcwHS=y`(;x#NkJLiX^18V`LsgO^(n_;JGQHRK(GPU3~4X6;vgI zS7AT&p=5xmdX3${jw~&)RFgJlgkB~}e;*kP~Bor}KgxwjR*W;M9Kimv7hxGkI`sVYPa6p-YM~SeDBzlrkKC^kvO*J((1*}Q?3_7f;Kp#A7y~oG zk%L`)@0%7T)FNXV2olq%=oqh7v9Q`j%d)vnE=w|^l%{+}6D!|qY(xJwzksydV59-f zG`3G3g*$%=LVF%!^3lZ#0v<0(pir9A?$_X zM(7V+oIGRW@rQ@-mh)&DI?FnXc?nx4$C)X8dU2I`)T;xB^I_5`5$JO0T4;$fPF?#d z1iNlRoXxS%jJyoVp^cn;D@Grc`Nb}M? zOQ#2WiAGO`H(``=`qAcg3HnJF@!olBA&1AF_tI)9In!Y_NQl8Zi{i;*sPBrG7U|^xT?XTsTrggD;ONyJyW4~5`pj;gs@7N*y4vqU*LB7M4}BMl{Vq@D zP*Z{`6qY>_4Uw8}Y3N|t@z4$el$1OsRE=xPWjTwwqOwk#Pz`c6i-F~$tdogS9<4zi z<&wdVf6Mo=)NZ3$E-~P8zTJZr`CKO^ntsc4uwvVs6I1z*x`L4)U`W&639$u92c1F7 zb$RZR#8pLA#j0&$V7n-(1}dt?pd0$Vo0`U@LuwGsZk~QHTh(6nn&&~ujxObTK(8^N zPyG5zueLhgmj~^j+UoQWXGE&OqSfuA-M27aDZ`LtwEBGxV)}1F%@P)i%6n9*8?Tm7 z(q*jnx@=M52N6zo`ZznF3F#OUG{K(9U}-SG#&Q8QRfFyO$WujvYSx(qTPCNO(zMuG zW&c95PF`qrdDxX!Xhq6mu|LF9iwm48=(55tvR%hStyttKb=pLlQH0axfSG;Nupyx3 zg6Z(ka-*POqF7+6>dBQA?i@A>C8)B--(AUy80YD89kbguepV`F-*Ww*Ju?gAHJW8$c`S&~+8dtv(N~HV6P2M_B4~;6?#9j*VeJ z!?;S@$ zs}@TzR1GWb7E4NLZn8ZXqFE{;Ckku_F$_YS?GhkHMzcUh_b!~s!$h^rzK=9P-lvs%CJYuX-L#am^c(!KC z_yqHzWl3W6neA9u8JOsL9twF0)0GnZDB@h2a(iMl(}-JHCK{B7SV+()8O*UZJs*pv ziB*R>g;~_}BHQP60+-8-WVc7{a2k{AI1bmH%bEd0Rgj_2_FWo*OSezf9=Ys_jlaC- z&-Xw6e2|W=2lN^P`r+Gd%fUbRKLoMxPIIM~SzYO}J8ah+qBFFxd18#`4Hs6IQD$aF z$oa7y1ke=?#k|b=XKEFRusWbAcmdP( zF$T;m^?Izmm@HR0xL2(qhOUn-V>K=p(#=7poQ+bMIe&f2MJ5^PM(DqFO=eQ-s%PWq z(h~ByJl`NiMTZv$XbnwxVZxK(w1Fgc_Rw?T2OgTmDpygc{a#R1s0BF|BI_^TcHoKM zT;plHD2J`59BVkuciakb+xSaw3e4n=`C}_JYEhHh7HyiSQbBak4*7`(0mjQU)-2v{5pcq2=n>z{L>rL>B2{Dx7$#vS7=B=&!J`Nt1@at){O zq8+jx&}%%^-~WXdZ*ML3?pt2yHdt;cW-xR-7AO@JnO96zR#1>fekxQpbNmnk%i_ie ziRh`qvDE3I@A#OY?qN2|k8x^yn~uY+X!?J&5v2b|_GvUFVtO7u%X+h7A zoVV0Gu}6TSMir5SRe4$R0TgpM-RofM)FfmfgSpjJ$f}Hzs$yw4#6o|FZR1TAG!2M< zXPG==ML9P$`=uN99zVQ#cfaSAs7>vVQ*DOaXR}=LB0)W^gajz0p7>a~g1%{U+CY*_I?+`5 zSm_Vg9CW%?W$zQpu?b|}F^AAK6$+L5D7b?FMMYtRfff4fXI?f`EUd0_UR_splr%UYDN+C0*yn4LNMdnH)CPs>#%J{WE1n{nZ=);fuXBp2CZA*m^*(@iag7 z%h&IjUogMob)zvF|BthPoXfK(A8|KmbtUH9aL|XQD;O^ovD!CrxuGz8yL(@dZw3zirfXtrjqFB&TRaKa-hco>FxALc}^${94 ziQohwI*!Ly%e4ZTg95AO0C|YGWMXVMHBo(&ovnZQjd$L8QHV%uyiDtL=(V1xum1ey zS^@eUE2lfJUhUgZXeA3q=1rvIAkZUCjr4{Co+E4+I?i@`IKJ9LRZ}o#=&X<6q1q(j zTy(Whz9}(OLW)T_a^Pb3C|dQ2syeiWoGUkL4b~{r%`w;MW2#(+md!$w@*}=UR$ze? zd7F4K<<%mmHO*=T8%hPtuaa|C9&OvjLT`vV)tM`0e*Z$Rk5iovlUu2Q5>T656xrD- z&E{YQA!Qw0NEJ0dfV&|o6U(=-t@`4mezXyF4l4DEA$%A^f%wOSI)TI{k^vT z;+17fA|?iv#i;z|_|)|yQNDcq0jK}hKawYcfkwJn%rGeO>B zDKJqka^0D5z6HIAg<%i27qY;JEItg?z?ps@t%1p%K`pOhso!IKFW)HS7VG=04lET5 z{QdKpOmgYwnY*XP&7ZyM_D7xL9E-=dXjDqeW*Ci21XZe7 zxV{ZJn`eCq*}sX|JWh8zY;amr3asZi*%N)u>mo7ffxUe$j4_51I;ot*Q|^Hg52x2>lf&+ycK z>+M%hTRroAXHS`Ljp9_G&Km_GkwGbvNvpg_v``^Ks~O6wikgyVwg+v#@=^wdoMSyS zX(MQ=Wxm(v+dcok- zhpaD#0h!f@Oh(-_Hpanzw$nx5u`pFGq9*H1Ev1&bF9@8vkfjCzb%>SSc07XkbN!hh z=!z_1kc6-t7qjDaOpetLmrJ>ydCh0O`S>$Dbrl10&jO|m84Xoq&4(5EINj}Y9zJapP*w79 zqX-M8iJGCJp{nfdJ=YuJNN0$OqN1S}`R5J-59NXi+0b$9^c-f&x>v82-jnh3fAK4i zJpVR?bSc*Z`aA-f-txIOT>YlOiusYng|?OmS&S9*5v|{Hu%TXuETr6lYdJ0ySw&e^ zaCX?k@$LXQiP+dc*Nf00f5K>F3!$Vyl}v;(8J^-M*(oXLD0PVvz$NZ;oFJbSpyfrr z5iH_pK!6jC5c#DXP|syiBmYJz$G9*viZNbLVNj_rCvo$Io(?oLkWC`^Y08b1Ln}iR zRT-0|D*N6n*%q6iQia3vEPL~Y@`u%2_OIXkKIYHs0ezmqOmDmERoh>(FlXK| z*R`vpqBEo{I+lgJn8j?l%({$W=)>|yT|ycFm^b_AdLa~iIq+%t(*H{^YI)iPD@QgrUL?sO2(P3YPWTc4?q1QOm>95fOue(S#Y^* zFOb~^2)EO$D}a4K#^0sy2t)j1yW?N(Ckdtv4bxQCLkc_axrxIbG~{6_0EbrxAiE$+ zN0_&V6e^8=Az@?F)09^mO0aWh^ACeX{OR*GcYgXZt_Spa1~k3xj$5x&Qq#X_ZmIpT zZqFL0Du5~~IkS%XXB{<89gdcAVnuYC29=UuD&U!JG+fIi>IIrqLhZr!Upo%Sb&UHjF8 zp)HXgIa$n`z6UED3HGVu#(Ix9Wdy*IWn;zm;l>GDQRsOgDe9TW^0~|(+t%hOAb`CYHO`h|m)!H4D-JC{=)O)u&e zGSLVEj9eZ;lJMj;H7~3w9wS%Cb5(upsKY#ljdw1%+(EI()iv1s5$Ik?5gqc!_=kjPoij;VS>ax?AaRsIqa?3fB zLsV-T4{15|gLWJD+`hQ@oY6b??IcZ6`L828`T0zo3Yj$FRypZ9j;-`zg()TrGHObW zH>pmVVuiHSvWO%S&<#Uqh8k>}uKwz~zkU3p&*AC1pdYoqLO%zmm%4}bZ|r!X*Bk!N zZr5s(VvfLu7snZdO$zB(=6=WyEmV=dz)p)-qc9E55-8;ve%siw!RdXY6#F@U(JbW2Q_YwAcJPtd4yQ@dO19HPEG4-s-9L5`Vbuwd z$z`GB^4wst6CkgM?&jJ0|Gz<+`J=ZVIN&|!U+w~Z;Ceux=ZPg=w7F_tQ&jW+(F@I+ zwL(dAZ6~*~a9R-BizLy-OuaxM!i@|Z;(fbd{hEIDqv8hD2^Id@lnV*C4Uf1b9u|Fz4u z9?<6*&@aE~<+F`SO08t&Nn0)(-MerE;!ZVQoTGwqW= zE!iZT(=*9S9v%9}&-2GWqqptbv#~Vj_5LOoCqE@;GD7xTVnfp+oW?c)00NLnL_t*R zQkFbkM7YQx5i;n9q1W-;pUvpOSMQw12(9{m(Q7@R&+}rv@`jhbt6nbr_jb1}ZQ6Mm zT1)fjcUl;V)0&ErKkyp+M^?N$q?PMv^d_c@RWXFz`E zh=N2x^8Y@%)c)(|dL#IwzWFuRTr=A!H$It3lj{S*eUZOmr-fZRE8H)o3DHEZPOd3-yhgR*>`;mEE|qx z;`*Ckf?VFf>dF!pmd-F_O|taKF?D>TPppUm_$D# zD@r*lWj7>ABDrpWOfH8FQ%x~RGXlY9dYx_>#sSGgsko8ENf@PZ^0*k{jVE3A?0Nq9 zXY#h}jW4+N)AdH}^}cTl^n~b|7NkrDd#=150co-aef0Y+3{4xgMiYa6$3!Z;u?q59qa=r5|z3MnJYTtyB=C8ieSoY%W)_d|wEKv{ccJR{@v`13&Qt-;Uz6 zlg(sL3mM^H5Id_vn)%_f8$N=O(CC7`cJACMmBid#hA#aC(fgw~M7wV?5^T%1%Mb(( zT-SkR_0jJ1QEN`a54});^v4f;_r5>87*CMr@&x=ZF~-55u|)9i00000NkvXXu0mjf DRHK&9 literal 0 HcmV?d00001 diff --git a/index.html b/index.html index 2aec955..4e4858c 100644 --- a/index.html +++ b/index.html @@ -309,7 +309,7 @@

geochat
- + @@ -468,7 +468,7 @@

Visual Question Answering (VQA)

- +
Qualitative results of GeoChat's performance in Visual Question Answering.
@@ -582,7 +582,7 @@

Scene Classification

- +
Qualitative results of GeoChat's performance in scene classification.
@@ -644,7 +644,7 @@

Region-Level Caption

- +
Qualitative results of GeoChat's performance in region-level caption..
@@ -701,7 +701,7 @@

Grounded Description

- +
Qualitative results of GeoChat's performance in grounded description.
@@ -754,8 +754,8 @@

Referring Expression

- - + +
Qualitative results of GeoChat's performance in referring expression.
From 7ca158edc9e2b4485c899749be16a5b1dc2e4028 Mon Sep 17 00:00:00 2001 From: msohaildanish Date: Wed, 29 Nov 2023 15:54:32 +0400 Subject: [PATCH 3/8] fix website mobile responsive --- index.html | 66 +++++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 56 insertions(+), 10 deletions(-) diff --git a/index.html b/index.html index 4e4858c..968b27c 100644 --- a/index.html +++ b/index.html @@ -111,14 +111,19 @@ background-color: transparent; } + .table-con { + overflow-x: auto; + } + table { - /* width: 100%; */ + width: 100%; border-collapse: collapse; text-align: center; margin: 0 auto; margin-top: 2rem; } - table caption{ + + table caption { text-align: center; } @@ -143,9 +148,36 @@ .qual .title { margin-top: 2.5rem; } - .title img{ + + .title img { width: 3.5rem; } + .mobile-yt{ + display: none; + } + .desktop-yt{ + display: block; + } + + @media only screen and (max-width: 600px) { + table td { + padding: 0.5rem 0.5rem; + /* Adjust padding for smaller screens */ + } + + img { + width: 100%; + } + .hero img{ + width: 7rem; + } + .mobile-yt{ + display: block; + } + .desktop-yt{ + display: none; + } + } @@ -235,7 +267,7 @@

GeoChat: Grounded Large Vision-Language

- RGeoChat is the first grounded Large Vision Language Model, specifically tailored to Remote + GeoChat is the first grounded Large Vision Language Model, specifically tailored to Remote Sensing(RS) scenarios. Unlike general-domain models, GeoChat excels in handling high-resolution RS imagery, employing region-level reasoning for comprehensive scene interpretation. Leveraging a newly created RS multimodal dataset, GeoChat is fine-tuned using the LLaVA-1.5 architecture. This results @@ -247,14 +279,27 @@

- -
+
+
+
+ + +
+
+ +
@@ -646,7 +691,8 @@

Region-Level Caption

- Qualitative results of GeoChat's performance in region-level caption.. + Qualitative results of GeoChat's performance in region-level caption. + .
@@ -743,7 +789,7 @@

Grounded Description

Referring Expression

- When asked about an object as a referred expression, GeoChat is able to locate by + When asked about an object as a referred expression, GeoChat is able to locate object by predicting bounding boxes around it correspondingly.

@@ -849,8 +895,8 @@

Acknowledgement

This website is adapted from Nerfies, licensed under a Creative - Commons Attribution-ShareAlike 4.0 International License. We are thankful to LLaVA, GPT4ROI, and - LISA for + Commons Attribution-ShareAlike 4.0 International License. We are thankful to LLaVA and + Vicuna for releasing their models and code as open-source contributions.

From c93f56c3901d3c09f3ff88a86e02797759a4bbf0 Mon Sep 17 00:00:00 2001 From: msohaildanish Date: Wed, 29 Nov 2023 16:01:23 +0400 Subject: [PATCH 4/8] add website link to readme --- README.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 964c80d..99f5f3a 100644 --- a/README.md +++ b/README.md @@ -8,8 +8,9 @@ #### **Mohamed bin Zayed University of AI, Birla Institute of Technology & Science, Australian National University, Linkoping University** +[![Website](https://img.shields.io/badge/Project-Website-87CEEB)](https://mbzuai-oryx.github.io/GeoChat) [![paper](https://img.shields.io/badge/arXiv-Paper-.svg)](https://arxiv.org/abs/2311.15826) - +[![video](https://img.shields.io/badge/Video-Presentation-F9D371)](https://youtu.be/KOKtkkKpNDk) --- ## 📢 Latest Updates From ad9c853ef0bf5902c8a9c0e4eaa3cf9b1c26696f Mon Sep 17 00:00:00 2001 From: msohaildanish Date: Wed, 29 Nov 2023 16:31:22 +0400 Subject: [PATCH 5/8] update index.html --- index.html | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/index.html b/index.html index 968b27c..fa20dec 100644 --- a/index.html +++ b/index.html @@ -210,6 +210,10 @@

GeoChat: Grounded Large Vision-Language Muzammal Naseer, + + Abhijit Das, + Salman Khan, From 2e7a9fe0a9f08884d57268f437bb7e8830bc84ed Mon Sep 17 00:00:00 2001 From: msohaildanish Date: Wed, 29 Nov 2023 16:32:50 +0400 Subject: [PATCH 6/8] update1 index.html --- index.html | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/index.html b/index.html index fa20dec..cc7ed60 100644 --- a/index.html +++ b/index.html @@ -225,7 +225,7 @@

GeoChat: Grounded Large Vision-Language

- Mohamed bin Zayed University of AI, Australian National University, Aalto University,
+ Mohamed bin Zayed University of AI, Birla Institute of Technology & Science, Australian National University, Aalto University,
*Equally contributing first authors From 64e33a7bf266d6742ca86c5fd5416bf76aaf6fa6 Mon Sep 17 00:00:00 2001 From: msohaildanish Date: Wed, 29 Nov 2023 21:30:47 +0400 Subject: [PATCH 7/8] update2 index.html --- index.html | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/index.html b/index.html index cc7ed60..e3d9323 100644 --- a/index.html +++ b/index.html @@ -225,7 +225,7 @@

GeoChat: Grounded Large Vision-Language

- Mohamed bin Zayed University of AI, Birla Institute of Technology & Science, Australian National University, Aalto University,
+ Mohamed bin Zayed University of AI, Birla Institute of Technology & Science, Australian National University, Linkoping University
*Equally contributing first authors From 03cdd15ad9d36d786b7f7b68e70c0d6d7f057110 Mon Sep 17 00:00:00 2001 From: kartik kuckreja <131653487+KjAeRsTuIsK@users.noreply.github.com> Date: Thu, 29 Feb 2024 12:55:47 +0400 Subject: [PATCH 8/8] Update index.html --- index.html | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/index.html b/index.html index e3d9323..6581fe9 100644 --- a/index.html +++ b/index.html @@ -233,7 +233,7 @@

GeoChat: Grounded Large Vision-Language
- @@ -920,4 +920,4 @@

Acknowledgement

 MBZUAI Logo -
\ No newline at end of file +