Update app.py

app.py

@@ -226,9 +226,15 @@ footer, .gradio-footer, .svelte-1ipelgc, .gradio-logo, .gradio-app__settings {
 with gr.Blocks(css=custom_css) as demo:
     gr.Markdown("<h1 style='font-size:2.5em; color:#007dc3; margin-bottom:0;'>SR100 Model Compiler</h1>", elem_id="main_title")
     gr.Markdown("<h3 style='margin-top:0; color:#000;'>Bring a TFlite INT8 model and compile it for Synaptics Astra SR100. Learn more at <a href='https://developer.synaptics.com/docs/sr/sr100/quick-start?utm_source=hf' target='_blank' style='color:#007dc3; text-decoration:underline;'>Synaptics AI Developer Zone</a></h3>", elem_id="subtitle")
-    gr.Markdown(
-        "<p style='margin-top:0; color:#000; font-style:italic;'>VMEM is fast memory used for high-speed operations, while LPMEM is slower and mainly for less performance-critical data.<br>Keep the Arena cache size smaller than VMEM to ensure the model fits and runs correctly.</p>", 
-        elem_id="memory_note"
+    gr.Markdown("""
+        SR100 includes the following on-chip SRAM memories:
+        - 1536 kB of Virtual Memory SRAM (VMEM) for high-speed operations.  
+        - 1536 kB of Low Power SRAM (LPMEM) for images, audio, and other less-performance-critical data.  
+        
+        The amount of memory allocated to the model is customizable. Any memory not allocated to the model is usable by the application.  
+        
+        For the best performance, ensure that the Arena cache size is smaller than the available VMEM to ensure it fits and runs optimally.
+        """, elem_id="memory_note"
     )
 
     with gr.Row():