Introducing concept LORA: Viewer Holding Leash!
Key Features:
Concept-Specific Fine-Tuning: The Viewer Holding Leash model has been meticulously fine-tuned on a diverse dataset of leash-holding templates and variations. This ensures that the generated content remains true to the concept's unique style and aesthetic.
Getting Started:
Recommended Weight: The recommended weight for this model is 0.5-0.7. If the leash start to not look accurate make sure your prompting is correct ( check image examples for reference) weights vary as checkpoint model weights may vary.
Clip Skip: Clip Skip should be set to 1 or 2 for optimal results.
Additional Details: If you want to add more elements to the image, like ( specific poses ), just include them in your prompt after the trigger words you can also make the leash a chain by typing (chain leash) after the trigger words.
Backgrounds: You can normally add any backgrounds you want to the image preferably before the the . Check the images in the showcase for reference.
For further support, feel free to DM me on Discord @.gamakichi_
Happy creating with Viewer Holding Leash!
Description
clip skip 1-2
weight 0.5-0.7
FAQ
Comments (4)
Works very bad on a1111, style bleeding is horrible.
That's unfortunate to hear. However, you can clearly see what methods I use in generating I never really used a1111 in any of my tests and even if I am sure it is a prompting problem.
Now to discuss the style bleeding to proof this is statement make no sense, I will be uploading another showcase in the gallery with the ENTIRE STYLE PONY XL models I created and show how smoothly it blends without bleeding the style. Once again if you don't follow the guide I write on every model I release and check the showcase images I post (which i always post with prompts and negative prompts and steps and everything) for reference I don't think it is the model's fault at all if you can't follow the instructions clearly...
Will be uploading the showcase today with 5+ different styles
@Ayanna_01 yeah yeah, don't even sweat it. I left comment to deter people with bad internet connection like me from downloading it to use in a1111, yet it's a pitty i could not make it work, concept is brilliant
Greetings uploaded more images in the gallery showcase for more referance to use and to debunk that it ruins style blending. Prompts and negative prompts and everything is provided except the seed because I normally keep it at -1 or random.


