Update app.py

app.py

@@ -35,19 +35,32 @@ resource constraints. To train models that are robust to truncated thinking, we
 introduce a lightweight `budget-constrained rollout` strategy, integrated into GRPO,
 which teaches the model to reason adaptively when the thinking process is cut
 short and generalizes effectively to unseen budget constraints without additional
-training. 
+training.
+    """)
+    gr.HTML("""
 <p align="center">
   <img src="figs/framework.png" width="80%" />
 </p>
-
-
+    """)
+gr.Markdown(
+    """
 **Main Takeaways**
 1. ✂️ Thinking + Solution are explicitly separated with independent budgets — boosting reliability under tight compute constraints.
 2. 🧠 Budget-Constrained Rollout: We train models to handle truncated reasoning using GRPO.
 3. 📈 Flexible scalability: Robust performance across diverse inference budgets on reasoning benchmarks like AIME and LiveCodeBench.
 4. ⚙️ Better performance with fewer tokens: Our trained model generates outputs that are 30% shorter while maintaining (or even improving) accuracy.
 
+<p align="center">
+  <img src="figs/aime.png" width="46%" />
+    <img src="figs/livecode.png" width="48%" />
+</p>
 
+<p align="center">
+  <img src="figs/codetable.png" width="90%" />
+</p>
+    """)
+gr.Markdown(
+    """
 ## Citation