wdf
wdf
> how do you solve this error？ @HammerWang98
pip install sacrebleu==1.4.14 @liminn @HammerWang98
解决了吗，遇到了同样的问题，一个也检测不出来
> 这个项目不适合训练 不用预训练模型，训练完后也是啥也检测不到，有遇到吗？
默认情况下，训练的时候用的是DarkNet训练的，因为在cfg.py文件中默认Cfg.use_darknet_cfg = True，然而在models.py中测试时用的Yolov4在测试，改Cfg.use_darknet_cfg = False后训练作者提供的数据集，训练四十几个迭代就可以检测出来目标了； 亲测有效；
the cifar100 train image size:80x80, but when load train data only crop 32x32, transform_train = transforms.Compose([ #transforms.ToPILImage(), transforms.RandomCrop(32, padding=4), transforms.RandomHorizontalFlip(), transforms.RandomRotation(15), transforms.ToTensor(), transforms.Normalize(mean, std) ]) test image size:32x32, didn't need...
pygraphviz/graphviz_wrap.c:2711:10: fatal error: graphviz/cgraph.h: No such file or directory #include "graphviz/cgraph.h" ^~~~~~~~~~~~~~~~~~~ sudo apt install libgraphviz-dev pip install pygraphviz
do not solve it , but change torch version to 0.4.0 can avoid it