Update Readme.md

Readme.md

@@ -34,7 +34,7 @@ OpenVtuber: An application of real-time face and gaze analyzation via deep nerua
 
 ## Face Detection
 
-[RetinaFace: Single-stage Dense Face Localisation in the Wild](https://openaccess.thecvf.com/content_CVPR_2020/html/Deng_RetinaFace_Single-Shot_Multi-Level_Face_Localisation_in_the_Wild_CVPR_2020_paper.html) of **CVPR 2020**, is a practical single-stage [SOTA](http://shuoyang1213.me/WIDERFACE/WiderFace_Results.html) face detector. It is highly recommended to read the official repo [RetinaFace (mxnet version)](https://github.com/deepinsight/insightface/tree/master/RetinaFace).
+[RetinaFace: Single-stage Dense Face Localisation in the Wild](https://openaccess.thecvf.com/content_CVPR_2020/html/Deng_RetinaFace_Single-Shot_Multi-Level_Face_Localisation_in_the_Wild_CVPR_2020_paper.html) of **CVPR 2020**, is a practical single-stage [SOTA](http://shuoyang1213.me/WIDERFACE/WiderFace_Results.html) face detector. It is highly recommended to read the official repo [RetinaFace](https://github.com/deepinsight/insightface/tree/master/detection/retinaface).
 
 However, since the detection target of the face capture system is in the middle-close range, there is no need for complex pyramid scaling. We designed and published [Faster RetinaFace](https://github.com/1996scarlet/faster-mobile-retinaface) to trade off between speed and accuracy, which can reach 500~1000 fps on normal laptops.