aluds123
aluds123
Hi, you can convert VOC dataset by this: YOLOv3_TensorFlow/misc/parse_voc_xml.py /
It looks like u don't have trai.txt?
Maybe the label is out of the image's width or height.
我在做量化訓練也是遇到這個情況,,而且也還沒摸清
Hi, do u have QQ, maybe I can send it to you.
@yifeichenNJ 你好,我剛剛發送出去了,在麻煩你查收
@yifeichenNJ 抱歉我在這方面也還算新手,或許你可以試試slim.assign_from_checkpoint_fn() 1. https://www.aiuai.cn/aifarm317.html 2. https://www.jianshu.com/p/0237ebbee5d5
Hi, you can use the pretrained weight which provide by wizyoung.
可以試試train跟test改成絕對路徑試試,還有路徑的名稱在檢查看看,我前幾天也有碰到這個問題
也許可以檢查看看是不是label的框比實際圖片大