Most tags can be used with a color.
A note about this lora, generally I dislike using very generic words for triggers but in this case, I went a bit overboard and made a dataset of 250 images so you could have a lot of modularity with the outfit. I'll admit that that I can do the tagging better but it's a long dirty work which I'm not sure there is actually demand for.
So for now it has the more general keywords and if there would be demand or requests for it I will rework the lora but since this works well I thought I'd release it as is for now.
If you enjoy my work please consider leaving a like or a review. Pics of your work are also appreciated :)
Description
FAQ
Comments (6)
I think we could use a general "horse riding" lora, because pony kinda screws it up a lot...
The model was trained on some horse riding images, it is still far from perfect but I find that on average I do get less eldrich horrors with the lora than without when doing horse riding pics.
For some reason when I use this, Pony messes up the generation and the images go from "cartoony" to some weird realistic style
I can try to help you but I will need more information to do that, if you can send me a picture with the metadata I can see if I can figure it out. I do know there are some loras it doesn't play nice with.
@JaneB Not sure how to get the metadata, but I have two pictures with the exact same prompt saved if that somehow helps.
I'm using a LORA for an anime character, and their face kinda gets messed up when I turn on this LORA. So either get a messed up face or I get bad clothing without the riding LORA since it's very consistent.
@TheBooster You can upload the image to a image hosting site that doesn't srcrub the generation data like lensedump and send it to me and I will see if I can figure it out