yolo_person_detect

yolo_person_detect copied to clipboard

I would like to train YOLOv3 for detecting person only on Ubuntu 18.04 without GPU. But the following error message occurred. ./darknet detector train cfg/voc.data cfg/yolov3-voc.cfg ./darknet53.conv.74 -gpus 0,1 |...

snowuyl

[yolov3_train_loss_too_big.zip](https://github.com/pascal1129/yolo_person_detect/files/4043780/yolov3_train_loss_too_big.zip) I follows procedure described at the following link https://zhuanlan.zhihu.com/p/47196727. But training loss is too big. The command I use is "darknet detector train data\obj.data cfg\yolo-obj.cfg darknet53.conv.74". Could you help...

snowuyl

我想要在一个个人项目上使用行人检测，目前是在学习过程中，想多跑一些demo，因为没有太好的计算设备，所以希望能找到一个训练后的权重，想问下您是否方便提供呢？

windzu

请问一下在人这个类别的测试结果中，训练时只训练人这一个类别会比所有类别都训练要好吗

nationalflag