But don't get me wrong, I look forward to playing with ggml SD and its development.
The fastest implementation on my 2060 laptop is AITemplate, being about 2x faster than pure optimized HF diffusers.
But don't get me wrong, I look forward to playing with ggml SD and its development.