doc: Initial readme.

README.md

@@ -1,3 +1,14 @@
----
-license: mit
----
+---
+license: mit
+task_categories:
+- image-classification
+language:
+- en
+---
+
+This is a mirror dataset prepared for the Summer Workshop, School of Computing, National University of Singapore.
+There're 2 datasets: 
+dataset1:
+(Chinese Traffic Sign Database)[https://nlpr.ia.ac.cn/pal/trafficdata/recognition.html], which  includes 6164 traffic sign images containing 58 sign categories. 
+dataset2:
+(German Traffic Sign Recognition Benchmark)[https://benchmark.ini.rub.de/]