v0.2.0

What's New

• Our paper is available on arXiv.
• We enable most of the models to train with minbatch, see model list for supported models. @kayzliu @xyvivian @aha12345678
• Add new models CoLA (beta) and ANEMONE (beta) by @harvardchen
• The first community contributor @zhiming-xu add a new model CONAD.
• Add new metric eval_average_precision by @YingtongDou.
• Improved device setting by @yzhao062