The first approach is based on a plain implementation of Detectron2 [9] and it is pretrained on PubLayNet dataset [11] containing modern scientific papers ...
確定! 回上一頁