realworld

src/pages/index.mdx

@@ -165,7 +165,7 @@ We randomly sample frame triplets triplet (<LaTeX inline formula="(\mathcal{o}_{
 We compare PROGRESSOR with R3M and VIP by freezing the pre-trained models and using them as reward prediction models to train RWR-ACT on downstream robotic learning tasks. 
 
 <Figure caption="real-world experiments">
-    <Image source={kitchen} altText="real-world experiments" />
+    <Image source={human} altText="real-world experiments" />
 </Figure>
 <Figure caption="real-world experiments">
     <Image source={drawer_close} altText="real-world experiments" />