Paul
Paul
Hi, Vivian! Is it normal that the speed of Scaling down 1080*1080 to 600*600 without mask is really slow? And how long does it take to generate the examples of...
Line 145-148 in main.py I think content_image is obtained in RBG fomat. However, utils.subtract_imagenet_mean_preprocess_batch take img in BGR fomat. Is that right? ```python #main.py content_image = utils.tensor_load_rgbimage(args.content_image,ctx, size=args.content_size, keep_asp=True) content_image...
转coco格式,使用paddledetection训练,eval的map[0.5:0.95] 刷到0.74, 但把eval的结果导出来后发现:有一些预测框会框多或框少一点。 看看还有什么改进空间,有小伙伴一起讨论交流吗?