This is the ✨Merge of 4 Trained 🦄 "Pony" models:
1 - Everclear PNY by Zovya: https://civarchive.com/models/341433/everclear-pny-by-zovya?modelVersionId=399640
2 - Proteus-RunDiffusion: https://civarchive.com/models/351450/proteus-rundiffusion
3 - Pony FaeTality: https://civarchive.com/models/331116/pony-faetality?modelVersionId=370979
4 - The (OG) = Pony Diffusion V6 XL: https://civarchive.com/models/257749?modelVersionId=290640
Prompting is the same as for the OG Pony💦 Diffusion! Use your favorite LORAs for anime, artistic or cinematographic photo style/look.
ATTENTION! To use it in Automatic 1111 or Forge - go to Settings - VAE - Set: BOTH "VAE type for decode" and "VAE type for encode" to "TAESD" instead of "Full" or get this SDXL VAE: https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl.vae.safetensors (added link to this VAE to the description on top).
CLIP SKIP = 1 Test is here: https://civarchive.com/images/8252103
This is how this model was created in ComfyUI:
This model works both in Automatic 1111 and Forge (with TAESD as VAE):
Description
the initial model release that very well responds to almost all promps
FAQ
Comments (8)
Hi thx a lot.
Can you add the project workflow in a higher resolution?
Getting lots of error messages when loading this model
yes, but it works afterwards ... just ignore them
@OliviaRossi It always said "Failed to load model" and something else and then show a small error message under the preview window
@AJFire which webui are you using?
TypeError: 'NoneType' object is not iterable
@AJFire I'm always overcoming such errors with enabling TiledVAE in settings and using VAE-ON