Update README.md

README.md

@@ -1,19 +1,25 @@
 # Counting_With_Adaptive_Auxiliary
 
--------Object Counting with Adaptive Auxiliary Tasks--------
-------------------------------------------------------------
+# Object Counting with Adaptive Auxiliary Tasks
 
-**SHA**
+# Data Prepare
 
--- Prepare your data  
+# SHA
 
---Download [checkpoint_best.pth](https://drive.google.com/file/d/1HaRTgBhW1Evr5NBOCduaDY2h2Xdkb4l5/view?usp=sharing) and put it into ./checkpoints/  
+- Download [checkpoint_best.pth](https://drive.google.com/file/d/1HaRTgBhW1Evr5NBOCduaDY2h2Xdkb4l5/view?usp=sharing) and put it into ./checkpoints/  
 
--- Run test_SHA.py  
+# prediction
+```
+python test_SHA.py  
+```
 
-TODO:  
+# QNRF
+- TODO
 
--------------------------------------------------------------
+# JHU-Crowd
+- TODO
+
+# NWPU-Crowd
 
 **Our method achieved 76.4 MAE and 327.4 MSE on [NWPU-Crowd counting benchmark](https://www.crowdbenchmark.com/index.html), which is currently the largest crowd counting benckmark**