What’s the prompts I can’t get it to work right

rodsod said:
What’s the prompts I can’t get it to work right

I don't have the original prompt for this, but usually its something like "baby, pixie, cat, red bandanna around neck, feral, all fours, by pet foolery". You could also try lowering the Lora strength to 0.7/0.8.

Other than that, I have no clue why it wouldn't be working. Are you using a different model besides DomesticatedMix with its negative embeddings? I've found this lora doesn't cooperate with all models. It works with a lot of ones ive tried, but I've encountered some models where it fails miserably.

nickernacker said:
I don't have the original prompt for this, but usually its something like "baby, pixie, cat, red bandanna around neck, feral, all fours, by pet foolery". You could also try lowering the Lora strength to 0.7/0.8.

Other than that, I have no clue why it wouldn't be working. Are you using a different model besides DomesticatedMix with its negative embeddings? I've found this lora doesn't cooperate with all models. It works with a lot of ones ive tried, but I've encountered some models where it fails miserably.

I kinda new to it so might just be me I dident know you could even change the Lora strength but thxs I’ll try it

rodsod said:
I kinda new to it so might just be me I dident know you could even change the Lora strength but thxs I’ll try it

Here's a link to the base model that was used for training, DomesticatedMixv1.3. You'll probably get the most stable results if you use this. I think 1.4 is pretty stable as well.

https://civitai.com/models/90545?modelVersionId=116578

nickernacker said:
I don't have the original prompt for this, but usually its something like "baby, pixie, cat, red bandanna around neck, feral, all fours, by pet foolery". You could also try lowering the Lora strength to 0.7/0.8.

Other than that, I have no clue why it wouldn't be working. Are you using a different model besides DomesticatedMix with its negative embeddings? I've found this lora doesn't cooperate with all models. It works with a lot of ones ive tried, but I've encountered some models where it fails miserably.

there’s negative embedding?

nickernacker said:
Here's a link to the base model that was used for training, DomesticatedMixv1.3. You'll probably get the most stable results if you use this. I think 1.4 is pretty stable as well.

https://civitai.com/models/90545?modelVersionId=116578

im using it but it just seems to give me a random cat do i need to describe it for it to work?

rodsod said:
im using it but it just seems to give me a random cat do i need to describe it for it to work?

On the page I linked to, if you click the "show more" button in the description, you'll find a link to the negative embeddings. There's 4 of them in total. There are more negative embeddings you can get, but those 4 are the bare minimum to get good results out of this model.

If you're just getting random cats though, then I'm wondering if it's even using the Lora. Make sure that you have the full filename, minus extension, typed into the Lora portion of the prompt, and make sure that Pixie-LoCon.safetensors is in your "models/Lora" folder. You can also try increasing the wait weight of the pixie token to 1.2. Below is an example. I can't test it out right now, but the following should work, since it's what I would typically use.

Positive prompt: <lora:Pixie-LoCon:0.7>, baby, (pixie:1.2), cat, red bandanna around neck, feral, all fours, by pet foolery
Negative prompt: bwu, dfc, ubbp, updn
Cfg scale: 9
Dimensions:576x576
Sampling steps: 40
Sampling method: Euler

nickernacker said:
On the page I linked to, if you click the "show more" button in the description, you'll find a link to the negative embeddings. There's 4 of them in total. There are more negative embeddings you can get, but those 4 are the bare minimum to get good results out of this model.

If you're just getting random cats though, then I'm wondering if it's even using the Lora. Make sure that you have the full filename, minus extension, typed into the Lora portion of the prompt, and make sure that Pixie-LoCon.safetensors is in your "models/Lora" folder. You can also try increasing the wait of the pixie token to 1.2. Below is an example. I can't test it out right now, but the following should work, since it's what I would typically use.

Positive prompt: <lora:Pixie-LoCon:0.7>, baby, (pixie:1.2), cat, red bandanna around neck, feral, all fours, by pet foolery
Negative prompt: bwu, dfc, ubbp, updn
Cfg scale: 9
Dimensions:576x576
Sampling steps: 40
Sampling method: Euler

omg thank you i dident know i had to add "<lora:Pixie-LoCon:0.7>" thank you so much

What ai you used

coldheartsorry said:
What ai you used

If you mean what model, I'm not 100% sure. Maybe furtastic? Pretty sure I didn't use domesticated mix on this.