NEW MODEL RELESED
Disclaimer: despite having "yiffymix' in the name, the model itself has nothing to do with yiffymix.
Hello and welcome. This is my custom furry model mix based on yiffy-e18.
It's able to produce sfw/nsfw furry anthro artworks of different styles with consistant quality, while maintaining details on stuff like clothes, background, etc. with simpler prompts.
I personally use it with these settings:
CFG: 6-8
steps: 23-150
Size: 512 x 704 or 578 x 768 (then upscale it. Some of the example images here are upscaled.)
Sampler: DPM++ 2M Karras or Euler a
(Clip skip 1 is recommended. 2 also works but makes the style different)
For better results, it's recommended to use it with anything v4 vae file or you can use versions with baked-in VAE.
Starting the positive prompt with "uploaded on e621" and negative prompt with"(worst quality, low quality:1.4)" help too.
PS: Consider using bad-artist embedding or boring_e621 as well. it's optional to use bad-artist embedding and you can try other textual inversions as well. If you see this, don't be an idiot like me. the way to use bad artist embedding is by adding bad-artist instead of bad_artist and you'll see it doesn't nessesarily improve the quality of the result. Boring_e621 is overall more recommended.
Description
VAE baked in the model. It doesn't need any vae to produce images with correct colors. I screwed up the this version's file size. Now it's fixed.
FAQ
Comments (18)
Do you have any tutorial or tips to train models like this? has good results
sadly it's a merge and I don't remember the formula anymore
About to see if I can make anime furries with this model. (Usually use “inspired by Posuke Demizu and Aenami Alena”— anyone got better suggestions for modern anime style prompt keywords?)
Looks rather promising. Keep up the good work!
Try out the v2 version then
https://civitai.com/models/12979/lawlass-yiffymix-20-furry-model
Will give it a shot, thanks.
Mine are looking to blurry and filled with light, I'm using the 2 GB version and I forgot how to really use it, I could use some assistance.
Can one use both Bad_artist AND Boring_e621 embedding?
Is it a bad thing to use both?
I encourage you to experiment with them both
@Lawlas also, Lora's. I noticed some people use Lora's at partial weights on some prompts. is that good to do too?
And lastly, is describing the image in a prompt such as (A Anthro Furry shows off while leaning on a window:1.2) alongside other prompts or E621 tags be viable?
How in depth could one describe the image desired?
I'm also curious about the depth of detail one can use when describing a prompt. It appears E621 tags work great with your model
But would you also recommend an in depth description of the image desired? Such as (An anthro furry female in uniform gives a salute) or can one in depth describe the setting too, such as ("prior description" while on a military base with a vehicle behind her "perhaps even describing the vehicle")
If one were to use a Lora with the model, is it possible to create images such as what I'm describing? Or is it best to use only E621 or Danbooru tags?
Lastly, I am using the 7gb version
@Lira512859 Lora config is located in "additional network" where there are a few sliders you can adjust
@Lira512859 I prefer not making a tag two long, so maybe yeah, describe the character and the car seperately\
@Lawlas I will keep all that in mind when prompting!
@Lira512859 Depends on the LoRA. 1.00 or higher can do things like warp the art style you're going with or start inserting stuff you don't want into the renders. Sometimes even just 0.1 is enough.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.