🚀HarrlogosXL - Bringing Custom Text Generation to SDXL!🚀
Teaching Stable Diffsuion to spell, one LoRA at a time!
Harrlogos is an SDXL LoRA trained on my own art and graphic design work, unlocking the ability to generate custom text.
It does not require any control net, any img2img, or characters from fonts on the canvas.
All you need is the LoRA, and you're ready to go!
Versatile, Responsive, Powerful
The model comes with a plethora of styles, modifiers, and extra content built in, but that's only where it begins.
One of the greatest aspects of the model is how well and how seemlessly it works with almost any checkpoint, other LoRA, or even things like Stable Video Diffusion!
Using Harrlogos:
Harrlogos works with individual terms, separated by commas, in a specific order. You do not need all of these in every prompt, but this is the general format:
YOURTEXT text logo
Text Color: blue, teal, gold, rainbow, red, orange, white, cyan, purple, green, yellow, grey, silver, black
Accent Color
Background color
Style Modifiers: dripping, colorful, graffiti, tattoo, anime, pixel art, 8-bit, 16-bit, 32-bit, metal, metallic, spikey, stone, splattered, comic book, 80s, neon, 3D
Accent Modifiers: smoke, fire, flames, tentacles, hell, glow, horns, wings, halo, roots, embossed, blood, digital, ice, frozen, japanese, chrome, pastel, robotic, hearts, cute, egyptian, viking
Additional Content: cat, sword, owl, cat ears, cthulhu, sun, roses, clouds, space, stars, skeletons, demons, fog, trees, moon, skulls, bones, planet, earth, cherry blossom, pentagram, crosses, lightning, bolts, crown, circle, moth
Example of a prompt with the resulting image:
HarroweD text logo, white, grey, red, spikey, splattered, dripping, blood, hell, crown
Helpful Tips/FAQ:
Harrlogos is trained on a 1024x1024 data set, so it works best at resolutions: 1024x1024, 1024x768, 768x1024, 768x768
You will not get the correct spelling of every word, every time. Some text may take multiple tries to get right
The model works best with single words, multiple words are very achievable but it may be more difficult
Try weighting the text in the prompt, ex: ("Harrowed":1.5) text logo, red, black....
There's two schools of thought in achieving the perfect Harrlogo - A. Brute force spam through seeds B. Find a good starting point and lock the seed in, then slightly asjust the other paramters up or down, which will create slight changes to the image.
With a fixed seed, raising or lowering the CFG, Steps, or even resolution will modify the image, potentially fixing things or correcting spelling.
In the negative prompt, consider things like: Double letters, repeating letters, malformed letters, and more than X letters. If your renders keep coming out with a specific double letter, also try putting that in the negatives.
This is a work in progress!
Description
FAQ
Comments (32)
The bomb. Took 40-50 tries including some seed variations to get exactly what I wanted, but hey that's what computers are for.
I think the time you put into these kinds of generations mostly always pays off in the end tho. Some results are just awe-inspiring. :)
Tried with Brightprotonuke and it pretty much hit a seed right off. cfg around 5ish or 12ish for retro looks.
Just way too much fun to play with. Thank you for making this, can't wait for new versions! 🥰
Thank you zer0, means a lot
Downloaded and saved to LoRa folder as usual. But it does not appear in the list. What could be the problem?
First it did not appear for me either, but Refreshing the lora tab solved it.
@Metaphysix I've tried a few variations, of course. The most obvious ones with refreshing the tab, rebooting SD or re-downloading the LoRa file - didn't help me.
Are you trying to use it on a sd 1.5 model instead of a xl one?
@ILG_UA same shitt hpnd here :( stpd lora
@VojtisekTomas same problem like ILG and i am usingA1111 with have models delibrate , ghostmix , shattter , edge of relism and resluistic vision and maybe my A1111 is SD 1.5 :O HTH i make it appear and activate for work all things i did but not appear nor work???
@AaliShine With what model you exactly trying to use the lora?
@VojtisekTomas @Metaphysix @AaliShine @ILG_UA Since A1111 1.6.0 you have to go into the options, additional networks, always show all networks no matter the model version (something like that)
Because.... absolutely no reason.
Then as usual when downloaded in models/Lora, refresh Loras for it to show up.
Hope this helps.
@VojtisekTomas realvisionXL and realistic vision .......... also i have delibrate
@ze_thriller no no no man that nonsens i do it already i know it already then u here... but but but my Q is that option says ( Always show all networks on the Lora page
(otherwise, those detected as for incompatible version of Stable Diffusion will be hidden)
but my thoughts are its says incompatible lora are shown but did this mean these incompitable lora are working wrong or useless in term of desire result bcz they are incompitable ???
@AaliShine Have you tried it in ComfyUI instead of A1111? IDK what is the problem, but for me it works better in comfy.
@ze_thriller This Helped me thanks
@Metaphysix naaaaaaaaah sory buddy comfy UI is a crap poor tool in front of A1111 in the earth there is only 1 tools exist which has many features then others also option or tools or Ext or anything is A1111 comfy foocus all are limited boring version i saw then all so why i choose less beneficial tool so i have A1111 the boss okh ?? but now my problm solved and anothjer arrives
THIS IS AN XL LORA. YOU HAVE TO USE XL MODEL!
@sirvz i do dear i do sdxl i used but still not generate texts on 1024.... i want bing like result
amazing goooood job。would you please share how to train this Lora? what images should use and tag, training parameters. thx so much
This LoRA was actually trained with all of my own art and graphic design work.
WTH i did all tries but not appear in lora nor working
its not working like dalle3 text its not generating text................. ? fake lora /???
if it's not appearing in your lora section, try refreshing it. if it's not working, are you sure you're using an SDXL checkpoint and that your generation size (latent image) is at least 1025x1025?
@Avakinz i have A1111 install with SD 1.5 base ok but my models or ckpt has only 4 1 is TDxl (which is SDXL model) 2nd is SDXL original with 1024x1024 3rd is realvisV2 which is also sdxl lastly my chpt is MJ shatter which is only i have SD 1.5 despite others so now whats you think ??? more
@AaliShine I'm sorry, I don't really understand your English very well, but have you tried going into the LoRA section (looks like a little orange book in the middle of the top right under the generate button iirc) and then there's a section that says [Textual Inversion] [Hypernetworks] [Checkpoints] [Lora] and if you click Lora it's not there? Have you hit the [Refresh] button to the right of that?
same. it doesnt even show up
Finally!
Hope you can make a better version that does longer words!
Thank you!
Well done LORA ! The only problem i run into, is to write a text with letter and number (like "Leet speak style "l33t5p34k" ) but it'doesn't seems to respect the text . I hope it could be improved in newer version.
As of v2.0, it's not currently trained to do numbers, trying to get the letters as good as I can before taking those on!
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.