Simplify OCR model performance statement

Removed specific accuracy percentages for clarity.

Author: preethamp0197

projects.html

@@ -62,7 +62,7 @@ <h3>Pet insurance Banner generation And Rakuten Optimism booth</h3>
             <article class="card">
               <h3>Image Translation</h3>
               <ul class="list">
-                <li>Outperformed Google Vision (81.05%) and Paddle OCR (75%) with an in-house trained Japanese OCR model with 83.07% accuracy for seamless Rakuten Ad banner localization.</li>
+                <li>Outperformed Google Vision and Paddle OCR with an in-house trained Japanese OCR model  for seamless Rakuten Ad banner localization.</li>
               </ul>
             </article>