From 6d3a6098669b9ebc496472c99e80a14695072557 Mon Sep 17 00:00:00 2001
From: maniast <maniast@hotmail.com>
Date: Fri, 20 Sep 2024 00:20:03 +0300
Subject: [PATCH] 1st draft

---
 index.html | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/index.html b/index.html
index 1bedffa..f52bd8c 100644
--- a/index.html
+++ b/index.html
@@ -48,7 +48,7 @@
     <div class="hero-body">
       <div class="container">
         <img src="./resources/ucl.png" width="150"></img>
-        <img src="./resources/rpl-logo.png" width="150"></img>
+        <img src="./resources/rpl-logo.png" width="100"></img>
         <div class="columns is-centered">
           <div class="column has-text-centered">
             <h1 class="title is-1 publication-subtitle">SDS : See it. Do it. Sorted</h1>
@@ -104,6 +104,7 @@ <h2>Quadruped Skill Synthesis from Single Video Demonstration</h2>
 
             <section class="section">
               <div class="container">
+                <img src="./resources/figures/method.png" width="900"></img>
                 <br>
                 <br>
                 <!-- Abstract. -->
@@ -115,7 +116,6 @@ <h3 class="title is-2">Abstract</h3>
                         we present SDS, a novel pipeline for intuitive quadrupedal skill learning from a single
                         demonstration video, leveraging the Visual capabilities of GPT-4o.
                       </p>
-                      <img src="./resources/figures/method.png" width="900"></img>
                       <p>
                         SDS processes input videos through our novel chain-of-thought promoting technique (SUS) and
                         generates executable reward functions (RFs) that drive the imitation of locomotion skills,