very nice.

This is one of my new favorites. I use it all the time now. Thank you!

Site stable deffusion not display lora on site.
Please, help me define reason of it

the lora model is not showing up to load please help

How do you make a lora that keeps the original image intact? I want to do that but don't know how, if I train a lora it ends up taking over the underlying model and producing very different output, but yours keeps the poses and general theme intact somehow.

Thoughts on using this during upscaling/img2img to enhance? Or would that potentially ruin the look of the original?

Thoughts on using this during upscaling/img2img to enhance? Or would that potentially ruin the look of the original?

What is the trigger word? I have tried with the base SDXL and MBBXL and this lora doesnt seem to have any effect. 1% change at best, but usually loras have a trigger word.

100 side by side images generated. With and without. 5 knowing. 3 found absolutely no difference, Two were sayng that results are worse and they were able to be correct with 70%. This Topic has some really amaziong images in the galery but they have nothing to do with the Lora - its a waste of time and resources.

this Lora doesn't even change anything, idk why people keep praise it like god

This Lora made just the subtle changes I needed to make my image acceptable : corrected an eye here, removed random details in the background there...
It's so frustrating when you love the idea of a picture but just can't get around little mistakes!
So thx...

To clear things up for people this adds detailing not details. For example, a simple plain white dress may end up embroidered. It does not actually make the fabric itself better and look more like fabric. Pretty much has a similar effect you'd get from adding "detailed" to your prompt. SD in general seems to treat any mention of "detail" as detailing which are not the same thing.

Great olora

noice

Nice!

Nice!

I like it, this model is good.

i though this lora doesn't do any effects, but i realized how beautiful this lora is, (i tried 1st generation without this lora, then i made another generation with the same seed but this lora in the prompt, it gave me a whole different detailed generation image with the same seed)
Keep up the good Work ! ♥

Great Lora, thx

help a lot! thanks

even now, still the best detail enhancer I've come across!

Great Job

this lora might increase the possibility of hands or legs mutating when on an anime model. idk whether others encounter this but at least on Kohaku xl epsilon rev2, it doesnt work better than that without the lora. actually, i have added some negative prompts to try to avoid mutation but it even didnt obey it

Muy buen recurso

unfortunately with sdxl ponyrealistic models this lora removes dick form the mouth.

Hello! Fellow nerd here. I hope you don't mind me asking some questions. This is posted as a SDXL 1.0 model, but the metadata lists it as "ss_base_model_version":"sdxl_base_v0-9"

The metadata model is listed as "ss_sd_model_name":"dreamshaperXL10_alpha2Xl10.safetensors" so perhaps the base model was mislabeled or it is the training script's error? I'm learning by testing and noting configurations from popular LoRA's to get an idea of settings that are most successful.

Do you mind telling me what you placed in the smaller directories, if you recall, and maybe your preferred toolchain for training? "ss_dataset_dirs":"{\"1_portrait\": {\"n_repeats\": 1, \"img_count\": 20}, \"1_all\": {\"n_repeats\": 1, \"img_count\": 188}, \"1_misc\": {\"n_repeats\": 1, \"img_count\": 4104}, \"1_anime\": {\"n_repeats\": 1, \"img_count\": 10}}"

If I am reading the time stamps correctly, you spent almost 11 hours just on the training alone. Thanks for doing that and all the other work that went into this.

amazing

Great one

good

Newbie here, does it work with SD1.5?

10/10

When using the site's generator, is it mandatory to add <lora:add-detail-xl:0> and such as a prompt?

Works well

Great detail tweaker, use it almost on every generation

For some reason, it doesn't change the image at all based on Pony

Does it work with Pony?

I have always like to use this lora but I have discovered one big issue, it makes all faces longer, you can even see it in the sample images.

How do people use this with pony?

Looks amazing!

Good

does it work with illustriousXL?

hey guys I have a problem with that , my base model is sdxl and Also every setting is ok

but the lora doesnt do any change in my image :)))

Despite tagging solo, 1boy, and solo focus at increasing weights, as well as trying to neg others, men, etc. if I use this Lora, it will insert the torso of another man into the render. Any suggestions on how I can overcome that other than to not use your lora?

👍

👍

bruh where to put weights, what trigger word to use if using text, its not working for me..

lighting in this generator is the best

wanted to remove all those unneccesary details but it cooked my image. Tried all weights. what am i doing wrong?

The description says you can use weights -3 to 3 but the slider on Civitai's on-site generator only goes from -1 to 2. How come?

very good to use often

works like a charm

This tweaker is insane. By far the best I've seen so far.

Multiple tests, it does nothing. Tried with Lora weight from 60% to 250%, tried with <lora:add-detail-xl:-3> to <lora:add-detail-xl:3>, same seed, exact same results.

yes

Very Nice Tweaker!

Did this just suddenly stop working with Pony?

why has this suddenly stopped working with illustrious too?

Hi, did you stop working with Pony?

Stopped working with Pony all of a sudden, was working fine a few minutes ago and then when i wanted to make a new generation it dissapeared entirely from the list, the same happened with other stuff

It seems to be working with Pony again!

Perfectooooo

Can I have an explanation on how this works?

One of my favorites to use.

Cool

Could use in illustrious？

Almost a most for many of the things i generate

Here's a converter from SD1.5 to SDXL, https://fooddesert-embedding-converter.hf.space/

could someone help me by telling me why i can't use this lora? i never find it when i try to add it to the images i'm trying to make. thanks

cool

Very useful thing

Good

it's not registering as a Lora?

can this work with ILL?

cool

awesome

Good

Are the weight ranges of IL the same?

how to use it in the COMFYUI ?

Good

awesome

like

Is this Lora useful for anime works?

Great

Wonderful

How to use? You need to add Prompt Detail.

GOOD

great!

Great stuff