Better Images / Prompt Slider
Recommended range short prompts: 2.0 to 5.5
Recommended range detailed prompts: 3.0 to 9.0
Lower when combined with heavy LoRAs
This is in some ways a detail slider - but I don't want to call it that, because it is so much more than that.
It's basically a prompt enhancer in slider form. It makes all images better - that's why I named it the way I did.
In my opinion, this LoRA is pure magic to use. It can turn one-word prompts into masterpieces, and it can take a long, detailed prompt and push it even further.
The strength will vary a lot based on the prompt. A one-word prompt can handle up to 5-6, while a long detailed prompt can some times take 10 or more.
It combines really well with all other LoRAs, since this is a slider LoRA at the lowest possible impact rank - Rank 1. Stack it with anything without conflict.
I can't wait to see the images you guys manage to make with this one.
I'll give out 250 Buzz to the 10 best images once the feed starts filling up with gems. <3
Giveaway is done
Give it a thumbs up if you like it.
Description
FAQ
Comments (23)
you are doing great work but i want to use these onsite also so please keep its extreme strength to -2, +2 , otherwise its useless to onsite generations on any site civitai or tensorart🥲
I have increased the range on my loras so there is no problem using them onsite.
The range its limited to is set by me when i publish the models. If you encounter any model that have limited range its the publishers fault for not using the correct settings.
And i dont care about tensorart. That is hands down the worst website i have ever been on.
@Loraholic maybe but for a user with no good device , that works for sfw. i know there is no issue with your slider lora if used locally in comfy, but these sites offer strength only to -2,+2.
@TaiLong NO. This is NOT limited to 2. You are clearly not knowing what you are talking about.
Just click the create button on THIS site, and you can see the lora is set to -2 to +12
lol , never knew that, thanks but i dont use civitai for generations.. its beyond fucked up
@Loraholic They might be saying that using certain Lora nodes in COMFYUI only has a maximum strength of 10, but that's enough...
@wyxzddsjj919 I know LoraManager have a limit of 10 when using the slider. But it you want higher you can just write the value manualy. Never heard of a node that has a har limit.
In that case, the problem is not the LoRAs range, it is the node.
I like my ranges, and i will keep them that way. When a LoRA i make have over 10 as a good range out of the box i will normalize it to be below 10. But the 0.0 to 1.0 range is missleading for my LoRAs aince some times they need double strength, and some times it can only handle half strength, so it would never align with other LoRAs anyways.
Slider lora is all great. I'm impressed. Would you be interested in producing a lora that can control the wide angle and close-up?
I have tried, but with no results im willing to put my name on. It seems impossible to get this working with the quality people have come to expect from my work, so i have put that away for now. Maybe i'll pick it up again at a later point.
This is incredible, thank you!
Woah! Your loras are amazing! Some of the only ones on this site that do not change the original image!
Any chance in a tan lines slider? The only ones that are currently available make the image horrible!
I will put it on my list and try. I have tried it before without luck. But i've gotten better since my last attempt, so maybe i'll have more luck next time.
@Loraholic Cheers man, you're the GOAT!
What on earth is this magic? It is truly amazing. With this slider combined with other LoRAs, I can now do almost every infinite thing. 😀
Another great lora, thanks! Do you think you could do a pubic hair one?
Done
@Loraholic too fast, can you do one for tan lines like this image Pin on Powerfully Pretty
@TaiLong That is not a job for a Slider-LoRA. A slider could make tanlines in general, but a spesific concept like that is not really possible without using a normal LoRA that train on images.
@Loraholic ohh, i really hope you do such lora too , using pinterest girls images
This is terrifying! How did they manage to do that? Theoretically, refining cue words requires using an external LLM to enhance complex cue words. Lora can actually do something like this?!
In a sense, it's like a mix of random and detail sliders. It's very, very, very great