Please post the first picture directly here in the Model Gallery. This is the best way to support creators :)
Hello ♥
for whatever reason you want to show me appreciation, you can: ❤️Ko-Fi❤️
_________________
created this Lora to help for the 📖Text-Tacular Showdown Contest 📖
try the text improver Clip Model in addition
https://civarchive.com/models/793605/clip-l-fine-tune-by-zer0int-for-flux-and-sd
You will get more variance in text and more often the right text.
If you mix it with (too many) other loras, their is a high chance this will not work
The Lora is pretty strong. Please check the strength comparison
Settings
0.3 - 1.0
Trigger
aidmaTextImprovercan help
with text:"YOUR TEXT"Hope you get a lot of beautiful results ♥
Description
v0.4
try the text improver Clip Model for better text
https://civitai.com/models/793605/clip-l-fine-tune-by-zer0int-for-flux-and-sd
added ~ 25 new images
you can add this lora without trigger to improve the overall quality of the image
FAQ
Comments (18)
civit wont let me use the latest version in gen XD, welp guess i gotta wait o3o
it takes some time :)
Hey there, Love your resources! I just have one request - please add a tag we can use that isn't normal English words. For this one sometimes it will generate text that says the tag, for others saying words like halloween skews the output. AIdm4txt or something weird like that please!
pretty good most of the times, decent at worst
is there a way to get turkish charachters working with this lora?
no, this is just trained with english text sorry
@AIDigitalMediaAgency is there a way to train this lora with + turkish letters like "ş,ü,ö,ğ" etc.?
@adsazamet679 yeah if you train just the turkish letters, it should be possible.
A few questions if you don't mind:
Does this Lora work without the trigger word?
If the trigger word is mandatory, where in the prompt should it be placed for maximum effect?
Does the use of quotation marks or other punctuation improve the generation accuracy?
Thank you.
aidmatextimprover
@Doctor_Faustus oops, sorry for the late answer, missed the notification.
You can set the trigger at start oder end. It will work better with trigger, all training images are tagt with the trigger.
I use the format “with Text: ‘YOUR TEXT’”
iam not sure if this work "better"
For me, I use the - with Text: ‘YOUR TEXT’ where I want the text to go in the image and it works great 9/10 times.
Using the LORA here it includes the trigger word, if I use comfyUI from my pc I have to type the trigger word
Hi, first of all, many thanks for this amazingly helpful LORA, I can't thank you enough.
Let me share my results of using your LORA and follow your as token of gratitude and praise.
https://civitai.com/images/58556485
thank you, glad its helpful for you! ♥
https://civitai.com/posts/13155511
I tried some more, hope you don't mind.
In my country people call AI based genration "gacha" as it's luck-dependent.
My PC isn't decent enough for local generation...
I think I finally understood how to display futuristic fonts?
"sci-fi font" might be the key.
I used to type "futuristic font".
How to use this in ComfyUI
Pure awesomeness! This adds so much to some of my projects, two thumbs and one tentacle up.
Details
Files
Textimprover-FLUX-V0.4.safetensors
Mirrors
Textimprover-FLUX-V0.4.safetensors
Textimprover-FLUX-V0.4.safetensors
Textimprover-FLUX-V0.4.safetensors
Textimprover-FLUX-V0.4.safetensors
flux-Textimprover04.safetensors
Textimprover-FLUX-V0.4.safetensors
Textimprover-FLUX-V0.4.safetensors
Textimprover-FLUX-V0.4.safetensors
Text_Improver_-FLUX-_-_FLUX_v0-4.safetensors
Textimprover-FLUX-V0.4.safetensors
TN-FLUX-V0.4.safetensors
Text_improver-V0.4.safetensors
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.

















