add framework

Author: PangzeCheung

.DS_Store

@@ -1827,6 +1827,29 @@ <h2 style="font-size: 3.2rem; font-weight: 400;  color: #f7fcf9; ">All-in-one Fr
 
     </section>
 
+    <section id="Framework">
+        <!-- <div class="title" style="font-family: 'Alegreya'; font-size: 1.8em; text-align: center;">
+            Framework of OmniTransfer
+        </div> -->
+        <div class="title" style="font-family: 'Alegreya'; font-size: 1.6em;">
+            Our Framework
+        </div>
+        <div class="description">
+            <strong>OmniTransfer comprises three key components: 1) Task-aware Positional Bias: exploits the model's inherent spatial and temporal context capabilities for diverse tasks. 2) Reference-decoupled Causal Learning: separates reference and target branches for causal and efficient transfer. 3) Task-adaptive Multimodal Alignment: leverages an MLLM to unify and enhance semantic understanding across tasks.</strong>
+        </div>
+        <div style="text-align: center;">
+            <img src="assets/imgs/architecture.png"
+                alt="InsertPipe"
+                style="width: auto; max-width: 70%; height: auto; border-radius: 8px; display: inline-block;" />
+        </div>
+        <!-- <p style="margin-top: 15px; color: #999; font-size: 1em; text-align: center; line-height: 1.5;">
+            OmniTransfer comprises three key components: 1- Task-aware Positional Bias: exploits the model’s inherent spatial and temporal context capabilities for diverse tasks. 2- Reference-decoupled Causal Learning: separates reference and target branches for causal and efficient transfer. 3- Task-adaptive Multimodal Alignment: leverages an MLLM to unify and enhance semantic understanding across tasks.
+        </p> -->
+        <!-- <div class="image-caption" style="text-align: center;">
+            OmniTransfer comprises three key components: 1- Task-aware Positional Bias: exploits the model’s inherent spatial and temporal context capabilities for diverse tasks. 2- Reference-decoupled Causal Learning: separates reference and target branches for causal and efficient transfer. 3- Task-adaptive Multimodal Alignment: leverages an MLLM to unify and enhance semantic understanding across tasks.
+        </div> -->
+    </section>
+
     <section id="Video Transfer Comparisons">
         <div class="title" style="font-family: 'Alegreya'; font-size: 1.6em;">
             Video Transfer Comparisons