90 detailed dynamic prompts with fantasy, sci-fi, video game, mythological and historical themes. Each prompt includes setting, background, pose and (mostly) gender-agnostic character details. Now with booru tags!
For ComfyUI, please use https://github.com/adieyal/comfyui-dynamicprompts, and https://github.com/GrvBdgr/comfyui-negativewildcardsprocessor.
The wildcards contain negative wildcards (text surrounded by <! .. !>). If you do not wish to use these, simply remove the <! .. !> text in the wildcards files with your favorite text editor. Use the following regex:
<!.*!>As of v3.0 the wildcards no longer require Loras or embeddings, for compatibility with all models.
Older versions of the wildcards are not recommended for use.
Usage:
Add a description of the character you'd like in the generated image (feel free to use character LoRas), plus any style/quality keywords necessary for the model you're using, then add __ffpc-booru__.
Example:
best quality, masterpiece, 1man, russian (male:1.1), solo, aquamarine eyes, long brown beard, __ffpc-booru__
Some example characters are provided in the character.txt in the downloaded archive, so if you just want to run a few tests, you can use the following prompt:
best quality, masterpiece, __character__ __full-prompt-fantasy__
More diverse character wildcards can be found here: https://civarchive.com/models/118508?modelVersionId=128565
Negative prompt should be whatever the model necessitates, or your preferred negative prompt. Some models have race or gender bias due to the material that was used for training, you can usually use negative prompts to compensate. eg. if your generations come out with female characters despite describing the character as male, add "woman, girl" to the negative prompt, use weights if necessary.
Additional usage notes:
Other than the above, feel free to experiment with additions to the prompt, though most of the wildcards are rather long, so if you want to have a significant impact, weights may be necessary. The prompts rarely use weights above 1.1, so if your additions seem to have no effect, try weights of 1.15 or more.
Description
Lite (Lora-less) version of v2.0. Please check description for details.
v2.0 (lite) notes
New extension required for usage: https://github.com/Klokinator/Umi-AI-Embeds
Textual Inversions (embeddings) used (needed for both lite and non-lite versions):
style-swirlmagic - https://civitai.com/models/25829/style-swirl-magic
style-rustmagic - https://civitai.com/models/10764/style-rust-magic
style-sylvamagic - https://civitai.com/models/7523/style-sylva-magic
space_style - https://civitai.com/models/7278/space-style-ti
DarkFantasy - https://civitai.com/models/30287/dark-fantasy-style
Style-SwampMagic - https://civitai.com/models/54966/style-swamp-magic
Style-GravityMagic - https://civitai.com/models/108174/style-gravity-magic
FAQ
Comments (17)
First of all, Insane work. For real. One of the most complete and beautiful post I have ever seen. I wanted to ask two questions...
1) I see some people put a script to download missing LORAs, ok, but could it be updated for 1.8?
2) This is more a really personal question but I'm dying to ask... are you planning something for XL? :O
1) I understand that relying on others to keep such a script up to date is not reasonable. I'll definitely make that a priority for the next version, and include it in the archive.
2) I'm not sure what you mean exactly. The wildcards supposedly work with XL, though A1111 doesn't have great support for SDXL, based on my experiments. Maybe I haven't taken enough time to research and mess around with it, but I found SDXL a bit more unwieldy for my purposes, even if I just consider the additional generation time. I did manage to set up a workflow in ComfyUI that employs wildcards, but all in all A1111 is still easier to use in my opinion. So the short answer is: no, not at the moment. Maybe if some extension, or tool comes along that makes me like SDXL all of a sudden.
@TxcTrtl About 1, perfect. About 2 I meant that because will need a re-start with LORA's cause 1.5 based LORAs are not usable in SDXL. That's why I was asking ;). Cheers and keep the good job!
@LDWorksDavid You're absolutely right about LoRas. Lite versions should still work with XL though.
Great job mate ! I use v1.7 and it's a lot of fun to use. Thank you !
I'm sorry but I didn't understand what do I have to do to use version 2 with Dynamic Prompts, that I'd like to keep.
Could you please explain the procedure in detail ?
Thank you so much !
P.S.
What do I loose if I don't use UMI AI ?
It's pretty new as I can see and I don't want problems using wildcards for other prompts and applications ...
I extended the model description with more instructions on how to transform your wildcard files for the dynamic prompts extension (scroll to the bottom of it).
The main benefit of UMI AI (for me) is the ability to use negative prompts in the wildcards.
An example to illustrate the benefits:
Some SD models tend to add umbrellas to the image, if you put "rain" or "raining" in the prompt. If you want to avoid this, you can add "umbrella" to the negative prompt, and avoid the umbrellas (for the most part). However, when you're using wildcards, and only some of them include "rain", you wouldn't want to keep "umbrella" in your negative prompt for all of them, because it would just add noise to the negative prompt, and potentially dilute more important parts of it. UMI AI allows you to do something like
{raining **(umbrella:1.1)** | sunshine | snowing **(snowman:1.1)**}
which will result in one of the following:
Pos: "raining"
Neg: "(umbrella:1.1)
Pos: "sunshine"
Neg: ""
Pos: "snowing"
Neg: "(snowman:1.1)"
This gives way more control over what the wildcards may include. Other than this, UMI AI does everything that the dynamic prompts extension does, so I don't see any downside to switching.
@TxcTrtl Thanks mate ! You're very kind ! I'll give a try to UMI AI ;-) You convinced me.
Well, I wasn't sure how UMI AI would handle old-style yaml wildcards like BillionOfAllInOne... It seems to use it correctly. Only thing is for some reason. I can't seem to get webui tagcomplete to work correctly anymore... not sure it the problem is related, though.
Anyway, you made me switch over to UMI AI so I'll explore it.
Keep making these prompts! They're incredible. The addition of negatives really helps with some images!
The Umi extension looks really great, and I wanted to switch to it months ago for my own wildcards. But then I saw that the built-in wildcard files only support female characters at present. And I'm lazy and don't want to have to be the one to add support for male characters. So I guess for now I'll be sticking with the previous version of your prompts.
Not sure what you mean. If you're talking about the built-in UMI AI wildcard files, you can absolutely just ignore them, and use the extension the same way as you would use dynamic prompts. The main upside of UMI AI from my perspective is the ability to add negative prompts to the wildcards.
If you still want to stick with dynamic prompts, you can still use the newer versions of the wildcards, just remove the **...** parts from the file, as mentioned in the description.
@TxcTrtl Yeah, sorry for the confusion. I was think of a totally different extension. One that seems pretty cool tbh if you primarily gen women. There was a reason I decided not to use Umi and stick with Dynamic Prompts, but I don't recall what it was. But thanks for the info on using it with DP!
What would be your 3-4 most recommended models for use with this, for both anime/cartoon and realistic images
My recommendations for non-realistic:
https://civitai.com/models/112857/hellocomic
https://civitai.com/models/54073/arthemy-comics
https://civitai.com/models/70692/lemontea-mix-painterly-25d
Realistic is a bit harder, as most photorealistic models aren't exactly trained with the themes of these wildcards in mind. The following produce decent results, and are more on the realistic side of the scale:
https://civitai.com/models/123777?modelVersionId=147577
https://civitai.com/models/84728?modelVersionId=90072
https://civitai.com/models/52298?modelVersionId=90072
I've been loving this for quite a while now, and the 2.0 update with UmiAI seems to be making better quality images. So thank you!
But there is one BIG thing that I wish hadn't changed. Previously, I could recrate an image simply by keeping the same seed. With the previous dynamic prompts, it would select the same randomized elements based on the seed. But with UmiAI, this is not the case. Using the same seed will not keep the same prompt in subsequent image generations.
Select 'Static wildcards' in the UMI AI settings.
@TxcTrtl I discovered this just the other day! I didn't realize that's what the option means.
txcturtle, been followin your stuff for a bit- love your stuff and I mimic your prompts to help me make d&d npcs for my games! I am running SD on an AMD graphics card and don't have a way to use this with it; as far as i can tell. (it's the plug in that breaks it, for browser. :c
If some techie AMD user has gotten this workin, could you please reply to this comment with how you did it
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.