Nan Xue

you may need to enter the directory first and then run the ``make`` command for the compilation.

Thanks for your good suggestion. We will do it later

Hi Felix, Thank you very much for your interest and efforts. It is great to create a docker branch for this project. I have created a new branch named ``docker``,...

> CC: Hello @cherubicXN could you please fix access to the branch "docker" Hi @korvindest, did you fork my project to your own? After reading the guideline for the pull...

Thanks for your contribution for the exporting of onnx

I think this is because of the version. What is your version of pytorch?

After checking the official documents between 1.4.0 and 1.8.1, I found that ``torch.hub.hub_dir()`` was changed to ``torch.hub.get_dir()``

Let me fix the bug tomorrow ylxbyy ***@***.***>于2021年8月31日 周二下午10:10写道： > After checking the official documents between 1.4.0 and 1.8.1, I found > that torch.hub.hub_dir() was changed to torch.hub.get_dir() > >...

Hi, I think the problem is in the multiprocessing of the data loader. How about changing the num_workers in the data loader to 0?

Yes. In this work, we did not focus on it. It would be a good direction.