Skip to main content
Topic: Welcome to AI Image Central! (Read 11878 times) previous topic - next topic

Re: Welcome to AI Image Central!

Reply #15
New update:

I added the Spanish language to the site so you can now switch to Spanish if you like.
I may add other languages but it's a lot of work... And I'm not sure what languages the models understand so will investigate.

Re: Welcome to AI Image Central!

Reply #16
New update:

I have added a translation to English that is automatically run now if you type in a prompt on any of the Spanish webpages.
So you can enter the prompts in your native tongue, and it will still work correctly.
The model works best with English and without the translation it won't match the same outputs as in English.
I will have to see how well it works before I decide if it's the best way to move forward but I believe so as this way I
can add whatever language and it does not matter because the prompts are converted to English for the models.

Re: Welcome to AI Image Central!

Reply #17
New update:

I have now also added Hindi as a language option.  Again, like Spanish, the prompt will be converted to English before
being put through the model for the best possible results.

Re: Welcome to AI Image Central!

Reply #18
New update:

I have now also added Traditional Chinese (ZH).  Like the other two languages, the prompts are converted to English in order to get the best possible results.

So the website now supports English, Spanish, Hindi, and Chinese.  About 3.5 billion people?

 

Re: Welcome to AI Image Central!

Reply #19
I have updated the rules, so please read them or read them again.
I have also provided some information about some different features on the site like the music player and group chat.

You can view the Rules here: https://aiimagecentral.com/forum/index.php?topic=4.0

Re: Welcome to AI Image Central!

Reply #20
New update:

- Added SD 3.5 Large Turbo model.
- Added Rules Tab to main screen. 
- Added SD 3.5 Medium model.
- Added multiple step options for both Flux and SD 3.5 Large Turbo (4,7,10 steps)



Re: Welcome to AI Image Central!

Reply #21
New update:

- Changed the dropdown menu for SDXL to make it look nicer.


Re: Welcome to AI Image Central!

Reply #22
New update:

- Fixed the Spanish and Hindi languages to now use the new layout. (Still 3 more languages to go...)
- Added Stable Diffusion 3.5 options to Spanish and Hindi languages.
- Created Rules for Spanish and Hindi languages.


Re: Welcome to AI Image Central!

Reply #23
New update:

- Fixed the Traditional Chinese and French languages to now use the new layout. (1 more language to go...)
- Added Stable Diffusion 3.5 options to the Traditional Chinese and French languages.
- Created Rules for the Traditional Chinese and French languages.

Re: Welcome to AI Image Central!

Reply #24
New update:

- Fixed the Arabic language to now use the new layout.
- Added Stable Diffusion 3.5 options to the Arabic language.
- Created Rules for the Arabic language.

Re: Welcome to AI Image Central!

Reply #25
New update:

- Added Image search to all other languages in that language.  Translates whatever language to English and then does the search.

Re: Welcome to AI Image Central!

Reply #26
New update:

- I see a lot of issues where people just simply don't READ.  I posted all over the website that you can't do text in an image with SDXL but people keep trying.
So I have changed the default model to Stable Diffusion 3.5 which is newer and CAN do text.  Using more steps will help with getting the text correct.
Text is still limited.

Re: Welcome to AI Image Central!

Reply #27
Stability AI has released a prompting guide for SD 3.5.  You can find it here: https://stability.ai/learning-hub/stable-diffusion-3-5-prompt-guide

I have also included another prompting guide previously in the forum which you can read here: https://stable-diffusion-art.com/prompt-guide/

You can find a guide to Flux prompting here: https://fluxaistudio.com/blog/prompt-guide

Re: Welcome to AI Image Central!

Reply #28
New update:

- Added Another set of options under SD3.5 for the Large Turbo model to also generate from the 8GB GPU.
  So you can run the Large Turbo model from both the 24GB and 8GB GPU's.  The 24GB GPU uses the fp16 T5XXL
  text encoder and the 8GB GPU uses the fp8 T5XXL text encoder.

I also changed the sampler for the SD3.5 models to Euler as it appears to give better results.  Thinking about possibly
adding more options like selecting the sampler and seed instead of being hardcoded.  Maybe add other schedulers...
Like Beta.  Any thoughts?

Re: Welcome to AI Image Central!

Reply #29
New update:

- Add sampler and scheduler options to the Stable Diffusion 3.5 models.

 
AI Image Central - Welcome to our forum!