論文筆記：Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering

論文鏈接：Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering Bottom-Up Attention Model 本文的bottom up attention 模型在後面的image caption部分和VQA部分都會被用到。 這裏用的是object detection領域的Fast

>>阅读原文<<