ThePengg

Results 10 issues of


                                            ThePengg

how to deal with No module named 'prroi_pool'

2

comment

The link to the pre-trained model is broken.

Can the author post it again?

PreciseRoIPooling/pytorch/prroi_pool/src/prroi_pooling_gpu.c:12:10: fatal error: torch/extension.h12 | #include <torch/extension.h> | ^~~~~~~~~~~~~~~~~~~ compilation terminated. ninja: build stopped: subcommand failed.

How can I solve this problem ![图片](https://github.com/vacancy/PreciseRoIPooling/assets/152894708/9cea62d4-72b5-4e5e-8b24-71d270a1fbb0)

Doubts about TrackingNet training

Hi author, I would like to ask if TrackingNet training uses all the datasets, I see that DiMP is using the first four

Questions about threads

Hello, the author. I want to ask about the impact of the next thread on the test. I think it is very difficult to set a single GPU above 0....

作者想问下您的centernet的FLOPs是如何测出来的呢，为什么我测的是几十G

作者您在使用预训练模型训练时是将灰度图转变成三通道图片进行训练，还是修改模型的网络结构进行训练？

作者，您在目标检测训练时说使用的在imagenet上的预训练模型，是指类似mobilenetv2+agca这种的全部分的预训练模型，还是只使用了mobilenetv2这一部分的权重

4

comment

数据集如何处理的能分享一下吗作者本人小白不会操作

3

comment

YOLOX有多个模型作者您使用的是哪一个？

4

comment