Trained on r-mk's (NSFW) furry art.
Remember to support them if you enjoy their art!
Weight: 0.9-1
Triggerwords: furry, if you are not using furry models.
Notes:
Is capable of doing humanoid characters as well (As seen in the second example picture).
It is very prone to create NSFW artwork, but can create safe images as well by negativeprompting nude etc. and adjusting the weight.
Sadly the mouths sometimes get a bit warped, especially when prompting things like smile, light smile and smirk.
Feel free to share some of the pictures you create with me and post them here!
Description
FAQ
Comments (10)
Based. I have some ideas for merges / styles with this one.
nice!!! definitely need more furry artist style loras <3
I don't know that many that I find appealing, otherwise I would've done more already. But yeah I agree! x)
@Beelze I like Fluff-Kevlar's style, but the current versions of fluffyrock seem to handle it well enough. I tried training a lora in Teranen's style, but it was the first one I've ever made so it was so overbaked that I had to set it to 0.2 to get any good results lol. might try again another time but I think fluffyrock can attempt her style at the very least
Another great one! I apologize for asking again, but do you have the training data for this one?
Thanks!
And sorry, sadly I don't usually keep the files once I have a LoRA I am satisfied with. That's why some of my LoRAs release with a v2 or similiar in the name, I simply don't publish them until I am at least somewhat satisfied with them.
I delete all the previous epochs and training data usually. What were you going to use them for if I may ask? I'm curious x)
@Beelze Apologies for the late reply, I wanted to see how to tag well, because I believe I am lacking in that aspect when it comes to making LoRAs and I thought a good example would help.
@jegdjjfgsdfus It's a bit of a trial and error thing (at least in my experience).
I didn't read a lot of things before I started out either and am by no means an expert! I however noticed my results improved a lot more through adjusting the images I actually pick to train on rather than the tagging, that only rarely needs manual cleanup after using WD 1.4!
There's a lot of great resources on this site (and on other sites as well, I would avoid most youtube videos though) under articles if you want to go more indepth and don't be afraid of asking people (that are more knowledgeable than me x) )
This guide is rather helpful, but the writing is really cringey/immature:https://rentry.org/lora-tag-faq
This has some general dataset preparation tips: https://spainias.notion.site/How-to-make-LORAs-with-very-few-images-Advice-and-Tips-b30262a7d1b14604bd7f9c85124f6ab7
@Beelze Thank you again for the help!
Awesome model! Well done.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.