You know when you look at an image and you go "WOW! That looks great!" and then you look a bit more closely and go "Wait! That doesn't look right!"? Well, this happened to me recently when I created an image with a daisy and it came out great, then my wife saw it and said "It looks great but you know the leaves are wrong, those are rose leaves!" and I was like **** now it doesn't look that great anymore ;-)))

This is how this model came to life, I tried to train a lora that can actually make proper leaves/foliage for the actual flower that is being rendered! The model still needs a lot of work and at the moment there are only 8 flowers included. In no particular order:

Chrysanthemum, daisy, geranium, snowdrop, hydrangea, lotus, orchid, tulip

This was quite hard to train as I had to look for specific images that included the flower with the leaves with a good quality. I had to go through thousands of images and it still fails a lot of the time, I'd say ~50% of the time but don't take my word for it, it's just a rough guess as this is what it looked like to me.

I had prepared 180 images for this set and it went through 3620 steps.

The next version? Well... when and if this model can generate enough buzz to train additional flowers

You can buy me a coffee ˗ˏˋ☕ˎˊ˗ here ❤️