Data set of high quality is the cornerstone of the current data-driven machine learning models, and plays an important role in promoting the development of various application areas. At present, image analysis and processing techniques have intensively involved into the tasks of inheriting and protecting culture resources. However, currently there are few effective image data sets about the traditional Chinese Tibetan culture. In this work, we provided a small data set referred as CYTKv1(Chomo Yarlung Tibet version 1) which includes 1700+ Thangka images (an important and representative carrier of Chinese Tibetan culture), and the main objects in each image are manually labeled and bounding-boxed with semantic words. In addition, we shared a list of tasks of processing and analyzing the Thangkas to enlighten researchers about the challenges and potential applications on this data set. At last, we tested several famous deep learning models for the purpose of validating the annotation task on the new data set and presented the results of them, and finally selected the best one as the baseline for the annotation task.