I've been watching these and so far I think midjourney 5 is still best for plain text to image. If you want highly specific models (like one that only makes trains or whatever) then I think stablediffusion is better. For image-to-image I think controlnet in stablediffusion is also better. For some things it's possible that adobe firefly is better, like I think its outpainting is good.
I've been watching these and so far I think midjourney 5 is still best for plain text to image. If you want highly specific models (like one that only makes trains or whatever) then I think stablediffusion is better. For image-to-image I think controlnet in stablediffusion is also better. For some things it's possible that adobe firefly is better, like I think its outpainting is good.