Hello, I just love adobe firefly (beta) and have been using it quite a bit.
There is a very common feature in many other AI generators which isn't present, nor does it seem to be being worked on in Adobe firefly... that being an image to image generator. Rather than starting with a text prompt, I would like the ability to use my own image as a source of inspiration and then add text as to how that image should be modified. I'm kinda surprised this isn't on their line up of things being imagined. The generative fill can sorta do this, but not very well (only creating one image and being more geared to target a specific area of the image and not create variations. The text to image also almost does this in that one image can be used as a reference image, however that is only one created by adobe firefly and can easily be lost for future sessions.
Overall, Firefly is becoming a solid competitor to midjourney and is far easier to use and get good results over something like stable diffusion.
AdChoices