DeepFaceLab Training Error "00M when allocating tensor with shape and Type float"

[sagaface]

Hi. I am using the DeepFaceLabCUDA10.1AVX, the prebuild one for windows. I've done all the steps such as extracting the images and also faces. Then whenever I tried to train it, either using h64.bat or 128.bat or even the DF.bat, I get this error. PS.I am running with everything default settings.

I have also tried multiple batch sizes and I still get the same error.

[claudetheboof]

man im keep on getting the same error does anyone know how to fix it?