From the GitHub page:

Memory requirements for GPU acceleration

If you want to use GPU acceleration, you will need at least 3GB of RAM on your GPU for demucs. However, about 7GB of RAM will be required if you use the default arguments. Add --segment SEGMENT to change size of each split. If you only have 3GB memory, set SEGMENT to 8 (though quality may be worse if this argument is too small). Creating an environment variable PYTORCH_NO_CUDA_MEMORY_CACHING=1 can help users with even smaller RAM such as 2GB (I separated a track that is 4 minutes but only 1.5GB is used), but this would make the separation slower.

If you do not have enough memory on your GPU, simply add -d cpu to the command line to use the CPU. With Demucs, processing time should be roughly equal to 1.5 times the duration of the track."

So let's make this happen so users don't have to upgrade unless they REALLY NEED to and can AFFORD it.

The most important thing for me would be to use my existing video card even though it takes up two slots, or get a one slot card, and yet I also would like to know how much of a difference in terms of percentage each card contributes.

It looks for the amount of RAM. However, it also needs to have a ‘compute capability’ of at least 6.1 (which all the 8GB cards seem to have).

Unfortunately, the GTX 980 has a compute capability of 5.2 according to this table:

https://developer.nvidia.com/cuda-gpus

Does the software look for the amount of ram the video card has or do you guys have black/white list of GPUs that the installer looks for?

It is a requirement of RipX DAW that the GPU have at least 8GB of VRAM.

Was thinking of getting a 1 slot workstation card but you guys don't have support for it?