From 8eb76ef35d6538e27c59b7dd9df2c259bebe7632 Mon Sep 17 00:00:00 2001
From: hubenyi <867320491@qq.com>
Date: Sat, 25 Apr 2020 12:13:51 +0800
Subject: [PATCH] upload

---
 docs/INSTALL.md | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/docs/INSTALL.md b/docs/INSTALL.md
index c015f85..d824348 100644
--- a/docs/INSTALL.md
+++ b/docs/INSTALL.md
@@ -57,12 +57,12 @@ To reproduce our experimental results, you need to first download these datasets
 
 For image retrieval task, the dataset should be divided into two subset: query set and gallery set. If your dataset has been divided already, you can skip this step.
 
-In order to help you to reproduce our results conventionally, we provide four txt files, each of which is the division protocol used in our experiments. 
+In order to help you to reproduce our results conventionally, we provide four txt files, each of which is the division protocol used in our experiments. Before you use our script to split the dataset,  you should construct the four datasets like this:
 
 ```shell
 data
 ├── caltech101
-│   ├── 101_ObjectCategories
+│   └── 101_ObjectCategories
 │       ├── accordion
 │       ├── airplanes
 │       └── ···  
@@ -82,7 +82,7 @@ data
 
 ```
 
-These txt files can be found in [split_file](../main/split_file) and you can use the following command to split the dataset mentioned above:
+Splitting txt files can be found in [split_file](../main/split_file) and you can use the following command to split the dataset mentioned above:
 
 ```shell
 python3 main/split_dataset.py [-d ${dataset}] [-sf ${split_file}]