Change image to pdf

Author: ziyadsheeba

README.md

@@ -13,7 +13,7 @@ Our project focuses on developing a machine learning solution to detect noise in
 Our model architecture is based on a vision transformer (DINOv2 base) backbone with register tokens with a segmentation head. The frozen backbone encodes the input image which is then processed by the segmentation head to generate pixel-wise predictions. The following diagram illustrates our model architecture:
 
 <p align="center">
-    <img src="assets/images/model_architecture.png" width="100%"/>
+    <img src="assets/images/model_architecture.svg" width="100%"/>
 </p>