Eric Lee

Here is what I thought. When you use large amounts of dataset **_or_** set a high value of batch_size , your cpu will have many things to process at first...

有解决方案吗

你用的什么显卡？如果是游戏卡必须是30系以上

It seems that you will have the same problem on linux. I guess the vram size detected have something to do with tensorflow version. Try _**DeepFaceLab_DirectX12 build**_ and you may...

-p后面直接接端口啊，不需要port

### +1 plz