PyTorch_YOLOF icon indicating copy to clipboard operation
PyTorch_YOLOF copied to clipboard

A PyTorch version of You Only Look at One-level Feature object detector

Results 6 PyTorch_YOLOF issues
Sort by recently updated
recently updated
newest added

resize() got an unexpected keyword argument 'max_size'

大佬你好,我自己也复现了一下YOLOF,但我是在tf上写的,然后在pascal voc07上训练,然后测试的ap0.5只有0.66左右,这个结果是远远低于retinanet(0.7左右)的,甚至比FCOS还低(0.68左右),所以想问下你有在voc上做过实验吗,然后结果咋样呢

源码中的标签生成逻辑是: 1.利用预测框与gt的l1距离筛选出topk个锚点,再利用锚点与gt的l1距离筛选出topk个锚点,将之作为预选正例锚点。 2.将预选正例锚点依据iou与gt匹配,滤除与锚点iou小于0.15的预选正例锚点 3.将gt与预测框iou

我按照yolov5的提示,将数据集转换为了COCO格式,但是是这样的目录: RSOD - images - train2017 - val2017 - test2017 - lables - train2017 - val2017 - test2017

if I want to do custom training. should I mark each and every human in a frame with some class label? in my case (suppose try to build a running...

![image](https://github.com/yjh0410/PyTorch_YOLOF/assets/72019624/d58242a5-9d71-4a38-ab7a-449f82c1e812) 可视化文件夹图像时,只能检测一张,便卡顿无反应了,应当如何解决?