Don't forget the negative embeddings or your images won't match the examples
The negative embeddings go in your embeddings folder inside your stable diffusion webui folder
(if you have a different install I'm not sure but someone will know online)
To use the negative embeddings properly they need to be added to your negative Prompt only
Also this is trained on 768 resolution images so it won't do well with 512 resolutions, different ratios do work but are hit or miss.
4 months, 700,000 images, and a very high electric bill and the model is finally in the "Good Enough" stage.
My goal of this model was for it to generate highly detailed and realistic images without the need of those tags in the prompt and it works most of the time, but not every time so that's were the negative embeddings came in making the model about 50% better in most cases.
There is another version of this model the pushes the realistic style even harder, but loses some coherency for it as they're just not enough differing images of that type that at least I could find, so I haven't released it. If you guys want it I will release it in it's current state but I'll probably never get to finishing it as it is now Summer time and I'm tired of it being 105 degrees in my room all day long, so this might be my last model at least until SD LX comes out. Might do some Loras if people want them, but for this model there not really need.
Description
artistic and holyquality tags for great detailed images.
This model is very similar to the last one as the only difference is I baked in the kl-f8-Anime2.vae
suggested negative prompts: widescreen, intersex female, intersex male, fine art parody, full-package futanari, simple background, text, grey background ("mawshot" if you use any other mouth tags and don't want just a maw portrait)
The Trigger Words aren't always needed.
FAQ
Comments (7)
Unable to get this to work, everything seems to be off and drastically not 'Furry' or anime. Even the model seems to be off. The example images included in the page when downloaded and PNG info within use produce oddities. I validated this on 3 systems thinking it was something with my Automatic1111 install. All different releases. Same issue. At least it consistent across all installs.
According to your images meta data your using the "v2_512-base-ema" not my model.......
If you tried switching the active model to Furtastic, but for some reason it didn't load and stayed stuck on the SD 2.0 model (or whatever you had loaded previously), check the console output in the command window for errors.
One possible problem is not enough VRAM and/or add-on scripts for the WebUI that don't release all links to the current model, so that the internal "Garbage Collection" won't clear the memory it takes up. If the console output indicates that it is a memory problem, there are a few things you can try:
• Check if you're starting the webUI with the "--lowram" option. This loads a new model directly into VRAM instead of regular RAM first. This can speed up the loading process by a little or a lot, depending on the transfer bandwidth between RAM and VRAM, but it also requires enough VRAM to hold both the previous model and the loading model at the same time.
• Load a smaller model that takes up less VRAM, before loading the Furtastic model.
• Try disabling any extensions you have installed, just in case one of them accidentally holds on to a link to the previously loaded model for too long, preventing it from being properly removed by GC.
• Edit either the "sd_model_checkpoint" in config.json or start the webUI with the "--ckpt" option to make it load Furtastic at start-up, instead of switching to it after another model has already been loaded.
Does this work with nsfw sex tags? I am trying doggystyle, missionary, penetration and get too many artifaces and incoherent generations. Any tips or is this planned for the future?
this model works best with portrait style images as those are what I mostly used in finetuning it, you can try some of the Loras on this site or wait as I've been working toward a newer version that is more varied but It well still be a little bit as training with a couple 100,000 images takes forever.
@sopenit494 great to hear, this is the best furry themed model i have tested out, and i've tested nearly all of them. Hoping for a good hardcore themed version if possible :)
Would absolutely love a 2.1 version. The other e621 model that's up for 2.1 doesn't include vore or hyper pregnancy.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.