ZP ZHANG

Results 44 comments of ZP ZHANG

> Hi, Thanks. We are checking. However, it could be also due to differences in the VOT toolkit. Thanks, keep in touch.

> channel_selection 虽然为nn.Parameter,但是训练过程中不会update,因为这个parameter只是作为feature map tensor 的index来使用的,所以backpropogation的时候不会有gradient。 请问有没有遇到过channel selection layer加上去后训练速度变慢问题 (不是imagenet训练,是训练其他结构)

> Hi, Zhipeng > I noticed that the tpe result extractor seems do not match your tpe toolkit. The TPE toolkit worked for about 1000 times as you suggested, and...

一般情况下还是可以使用的,不过有资源重新对修改后的backbone预训练下当然会更好。 huchenjie339 ***@***.***> 于2021年10月19日周二 上午10:46写道: > > 你好,作者,我想请教下,如果我想使用resnet50(共5个stage)的前4个stage,并且我想修改第四个stage(步长为16的那个stage),那前3个stage的预训练参数还能使用吗?是不是需要重新预训练修改后的整个backbone?希望作者或者其它知道的小伙伴能帮忙解答一下,谢谢! > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view...

> Thank you for your contribution. I trained the SiamRPNRes22W model using GOT10K and TrackingNet dataset. And I notice that the Parameter penalty_k, window_influence, and lr affect the tracker a...

> Hello, thanks for sharing the great work. > I'd like to ask some questions about how to training network if I add small sub-networks to an existing model (e....

> 是test中的pytorch model下载不了 : siamdw_res22w.pth Pls refer to [here](https://github.com/JudasDie/SOTS/blob/master/lib/tutorial/SiamDW/siamdw.md).

> python siamese_tracking/run_video.py --arch SiamRPNRes22 --resume snapshot/CIResNet22_RPN.pth --video videos/bag.mp4 > > I run the above command for testing my own videos but i got the bad results. In videos Occlusion...

> 首先在Ocean下问siamban训练的问题有点冒昧,但确实经验不足,忘见谅! 由于整个实验室共用两个卡,而其他同门的任务默认卡0,导致出现卡0显存已满,卡1还有很多显存的情况。而算法在第11个epoch开始解冻backbone时,显存溢出,所以请问siamban单卡训练时的训练命令是怎样的? > > 原多卡训练命令: CUDA_VISIBLE_DEVICES=0,1 python -m torch.distributed.launch --nproc_per_node=2 --master_port=2333 ../../tools/train.py --cfg config.yaml 抱歉没有用过siamban的code

> @JudasDie Could you help us with this issue? Hi, as for the data, you could use the provided files in our new [code](https://github.com/researchmm/TracKit). As for the bugs, it may...