It is important to note that the training worked when I used the defualt coco dataset with 91 classes. ... It's generally faster than Faster RCNN.
確定! 回上一頁