Trained on a nice mix of nsfw and sfw images
Reccomended positive:
a beautiful and detailed portrait of muffet, monster girl,((purple body:1.3)),humanoid, arachnid, anthro,((fangs)),pigtails,hair bows,5 eyes,spider girl,6 arms,
The good: a relatively sizeable dataset lora that can produce okayish images of Muffet. You can use the activation word "muffetwear" to get a more accurate representation of her usual garb.
The bad: loves to produce way too many arms and sometimes throws in an extra eye to spice things up a bit
Make sure you use "open mouth" or "mouth closed" to specify mouth states, and be sure to always neg the inverse state to get the best results.
All images were generated with the anythingV4.5 checkpoint
biases: spiderweb background, pigtails, hair bows, chest bow,gold border,simple background
Ensure you use alot of emphasis when adding these biases to the negative if you want to not see them e.g.:(tag:1.5)
for best results use one of the popular anime models such as "anything"
Description
Edit:changed from v2 to v1.1
So in this version i added a few "from behind" images. I also aliased the tag arthropod to arachnid but i dont think it makes much of a difference in the grand scheme of things. The big thing in this version is that i cleaned up the body colors alittle more; the current body colors are blue purple and pink with purple being the dominant one. Please note that purple body will still give you lavender and blueish bodies rather than strongly purple bodies.
do note that ive had alot worse problems with extra arms in this version.