Work-in-Progress: Furion: Alleviating Overheads for Deep Learning Framework On Single Machine | IEEE Conference Publication | IEEE Xplore