Xiaohan Ding

I understand the np.float32 part, thanks for the suggestion, but why is the cuda part deleted? That causes a gpu-cpu mismatch during inference or training (RuntimeError: expected device cpu but...

We are working on a journal extension of RepVGGplus, which will be put on arxiv in two or three weeks.

Hi, we have not tried Windows, but it seems a cuda-version-related error. We used both cuda 10.2 and 11.2 and it worked fine.

Where is your "after reparam"? Please read the README or tools/verify.py

> In fact, in my opinion, if you get an exact same result, that's what it shouldn't be. Your opinion is untrue. Please read the README or tools/verify.py.

Hi, in MMSegmentation, a training task is defined by a config file. Here some_config.py refers to any config file we have provided. You need to download the weights of a...

Thank you for your interest in this work. We have not tried it on other tasks, but as the method makes no assumptions on the network structure and loss function,...

Hi, I checked the code and found no "synchronized()" so that the time recorded may not be the actual running time on GPU. I would suggest you follow the speed...

Thank you for sharing the results. As explained by @xiaocenxiaocen , our implementation is designed to pursue high throughput. Larger the batch size, higher the throughput.

I have refactored the code to improve the readability. Please check the latest version.