Maybe some of you were expecting this, others maybe not, but I want to share with you the template I use to generate my embeds ^.^
Currently this is the model that gives me the most satisfaction, but I'm working to create an even better one
Model created by combining different types of NSFW checkpoints.
If you appreciate my work and want to ✨SUPPORT IT✨, you can buy me a coffee ☕ KO-fi!
Check out all HDA embeds
Instructions:
recommended parameters
Sampling steps: 20-30
Sampling method: Euler A
Upscaler: Latent (recommended)
Noise: 0,25 / 0,4 if use non-latent upscalers / 0,7 if use LATENT upscaler
CFG: 7 - 9
Description
FAQ
Comments (10)
The source of everything 😍
Can't wait to try this out. :)
can u upload HDA_NegativeXL embe? thanks u so much
I'm getting inverted colors if i use the embedded vae. But it looks fine if i use external sdxl vae.
I haven't incorporated any VAE, so you are obliged to use external vae sdxl.
Thanks for this comment. Every other model I use now has a built-in VAE and looks worse when I add the external SDXL VAE so I had completely turned it off and didn't know what was happening with this model.
Is this different from the Ko-fi version?
The model on Ko-fi is the first version. This version is an update and generates better results, in my opinion
A nice alternative to Autism for animated characters that holds better likeness in my experience.