From 7c8e839642a0845b3ebd19a2b3f159f541a62937 Mon Sep 17 00:00:00 2001
From: ZhiZZhang <40889989+ZhiZZhang@users.noreply.github.com>
Date: Wed, 27 Apr 2022 10:51:47 +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 94a4dfa..de93953 100644
--- a/README.md
+++ b/README.md
@@ -4,7 +4,7 @@ This repo is the official implementation of "*[ActiveMLP: An MLP-like Architectu
 
 **ActiveMLP** is a general MLP-like visual backbone, which is applicable to image classification, object detection and semantic segmentation tasks. The core operator, Active Token Mixer (`ATM`), actively incorporates contextual information from other tokens in the global scope. It adaptively predicts where to capture useful contexts and learns how to fuse the captured contexts with the origianl information at channel levels. 
 
-The ActiveMLP variants achieves `79.7% ~ 83.8%` acc@top1 with model scaled from `15M ~ 76M` on ImageNet-1K. It also shows its superiority on downstream dense prediction tasks. `ActiveMLP-Large` achieves `51.1% mIoU` with UperNet on ADE20K semantic segmentation dataset.  
+The ActiveMLP variants achieve `79.7% ~ 83.8%` acc@top1 with models scaled from `15M ~ 76M` on ImageNet-1K. It also shows the superiority on downstream dense prediction tasks. `ActiveMLP-Large` achieves `51.1% mIoU` with UperNet on ADE20K semantic segmentation dataset.  
 
 
 ![ActiveMLP](assets/teaser.png)
@@ -21,7 +21,7 @@ The ActiveMLP variants achieves `79.7% ~ 83.8%` acc@top1 with model scaled from
 
 ## Usage
 
-The following is the guideline for ActiveMLP on image classification task, the usage on semantic segmentation can be found [here](segmentation/README.md).
+The following guideline of ActiveMLP is for image classification, the guideline for semantic segmentation can be found [here](segmentation/README.md).
 
 ### Install