I think it doesnt take the second 4GB vram you have into account. I really dont know how to solve it
I think it doesnt take the second 4GB vram you have into account. I really dont know how to solve it but would suggest using google colab with your google drive instead. It works a lot faster then it would on 970s and it is free for the most part. Like even if you could get it to work on your computer, training new models or textual inversion embeddings would take so long, and inference is already a lot faster on the free version of colab. There are a couple of videos on it on Gozukara's channel.
