From c01213690ea0bd43bfac3590e11a1557c32d70e9 Mon Sep 17 00:00:00 2001 From: zwzhao Date: Sun, 15 Dec 2024 18:08:11 +0800 Subject: [PATCH] Update README.md --- README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index ca1b918..af20169 100755 --- a/README.md +++ b/README.md @@ -5,7 +5,7 @@ This is the pytorch implementation code of the paper "[SlaugFL: Efficient Edge F ## Overview The following figure shows an overview of our SlaugFL. The SlaugFL consists of two parts. The left part, in the dotted box, is the preparation phase which aims to obtain a trained ACGAN model. The right part shows the process of our FL collaborative training. In the collaborative training phase, we propose a dual-calibration method on the device side to improve FL performance: during the local training phase of selected devices, the local model is calibrated on the augmented data and further corrected through aligning its local class prototypes with the received p-f global class prototypes. -![SlaugFL](./SlaugFL_Overview.png) +![image](./SlaugFL_Overview.png) We divide the implementation code into two parts: The Preparation Phase and The Collaborative Training Phase. @@ -40,4 +40,4 @@ If you find this code is useful to your research, please consider to cite our pa ## Acknowledgement Our code is built on the following projects. Thanks for their admiring contribution. 1. FedLab: https://github.com/SMILELab-FL/FedLab -2. FedGen: https://github.com/zhuangdizhu/FedGen \ No newline at end of file +2. FedGen: https://github.com/zhuangdizhu/FedGen