leon

Results 13 comments of leon

> Dear author, > When evaluating the mAP of VOC dataset, the performance is much better than darknet YOLOv3 version. However, there are much more extra predicted bounding boxes. The...

No such problem has ever occurred,please check your code and environment.

> > I have the same problem, did you solve it?Thanks. > > It's a function problem in OpenCV. The network will fall into the function in opencv and can't...

> > I have the same problem, did you solve it?Thanks. > > In my case, I changed NUMBER_WORKERS from 4 to 0 in the config file and the problem...

> 但今天在自己电脑上测试的时候,发现又没有问题了。经过后续测试,发现是train_dataloader的num_workers数量问题,如果置为0,作者的multi_scale代码正常运行,但是如果num_workers置为cpu总核数,代码就会出现我上面描述的问题。 你的pytorch版本是啥

> > > 但今天在自己电脑上测试的时候,发现又没有问题了。经过后续测试,发现是train_dataloader的num_workers数量问题,如果置为0,作者的multi_scale代码正常运行,但是如果num_workers置为cpu总核数,代码就会出现我上面描述的问题。 > > > > > > 你的pytorch版本是啥 > > 我的torch版本是1.7.1 我的是1.0.0的,使用前请看一下开源代码软件硬件环境,这是最基本的。你可以换成1.0.0再试一下这个多尺度训练是否有效果

> > > > > 但今天在自己电脑上测试的时候,发现又没有问题了。经过后续测试,发现是train_dataloader的num_workers数量问题,如果置为0,作者的multi_scale代码正常运行,但是如果num_workers置为cpu总核数,代码就会出现我上面描述的问题。 > > > > > > > > > > > > 你的pytorch版本是啥 > > > > > > > > > 我的torch版本是1.7.1 >...

3、in the finetune.py, ` if total_test_loss / len(TestImgLoader) * 100 > max_acc: max_acc = total_test_loss / len(TestImgLoader) * 100 max_epo = epoch` ![image](https://user-images.githubusercontent.com/31852119/80908465-a5664880-8d52-11ea-9455-b5643956aaf5.png) you test return the error between the...

4、why in the main, you set the mask below the args.maxdisp; but in the finetune, you set the mask up the 0? why?

NOTE: about question 1 and 4, i have known the reason.