Skip to main content ↑ ↓
Topic: Welcome to AI Image Central! (Read 11905 times) previous topic - next topic

Re: Welcome to AI Image Central!

Reply #60 –
New update:

- Added Age verification pop-up to the NSFW section.  Once Yes is selected, the pop-up will no longer appear for that browser
unless you clear the cookies.  If you access from another browser, you will be asked again as it's stored in the browser.

Re: Welcome to AI Image Central!

Reply #61 –
New update:

- Added some basic LORA support for SDXL models via a drop down menu.

Right now the selection is limited as I'm just making sure it works and I still have to change the other languages.
The trigger word is automatically added to the end of the prompt.

Current selection includes:
- Glass sculptures
- Stained glass
- Blacklight
- Renaissance
- Fine nude
- Greg Rutkowski style
- Wowifier

If you have any requests for any specific SDXL LORA's then reply here with your requests.

Re: Welcome to AI Image Central!

Reply #62 –
Added a bunch more SDXL LORA's... Now up to 37!

Re: Welcome to AI Image Central!

Reply #63 –
New update:

- Added a new model called Lumina 2.  The model takes 2 minutes to generate an image.  Does not do nudity or famous people.
- Added Babes by Stable Yogi Pony SDXL Lightning model.  This is the third SDXL model added that is trained for nudity.

Re: Welcome to AI Image Central!

Reply #64 –
In order for the LORA's to work better and possibly provide more detail to the images, I have increased all the SDXL models to 12 steps.
If the model is already loaded, it still only takes like 10 seconds so no big increase in time but should give better results.

Re: Welcome to AI Image Central!

Reply #65 –
The site is now back up and everything is back to normal.
Added a new 2TB SSD and switched ISP's.

Re: Welcome to AI Image Central!

Reply #66 –
I added another model to the Stable Diffusion 3.5 section.  The new model is from TensorArt and is called TurboX.
The model is 8 steps and takes about 72 seconds when not already loaded and around 40 seconds when already loaded. 
The results look promising.
 

Re: Welcome to AI Image Central!

Reply #67 –
I have removed the SD3.5 Medium model as the quality is quite frankly the worst of any model I have ever seen.

Also I have been having issues with the 24GB GPU overheating as it's a P40 without any self cooling and I had to improvise
a method to try and help cool it down.  But the constant use of the FLUX 20/25 step model is causing it to overheat and stay
overheated causing the Flux generation to take 7-9 minutes instead of 2-3.  So I have also removed the Flux 20/25 step model
until I can get enough donations to purchase a better card with cooling that can produce a result in a much shorter time.

Re: Welcome to AI Image Central!

Reply #68 –
- New update

I have fixed the issue with Unicode characters causing the generation to not start.
I have also fixed the issue with saving Unicode characters in the image.
Feel free to use Unicode characters if you want from now on.

Re: Welcome to AI Image Central!

Reply #69 –
-New update

I have fixed the issue with the search engines not always updating the search results.

Re: Welcome to AI Image Central!

Reply #70 –
New update:

- Fixed issue with Song Player sending too many bytes when sending a requested song. Only about 1 second to load a song now.
- Cleaned up logging of song requests.
- Cosmetic changes to player like removing brackets and file extension from songs displayed, bolded text to read easier.
- Added song name, play/pause button and seek bar below volume bar for current song playing.

Click the purple icon in the top left to open the Song Player.

In case you were not aware, you can drag and drop your own MP3's or OGG's into the Group Chat and they will get added to the Song Player.
The Song Player currently has 269 song selections.  See Rules/Help for more info on the Song Player.

You can also drag and drop images into the Group Chat to share with others. PNG/JPG/GIF/WEBP
 

Re: Welcome to AI Image Central!

Reply #71 –
New Update

- Added a Search feature to the Song Player to make it easier to search songs or artists.
- Fixed Remember function which somehow broke.
- Added Auto scroll to SHUFFLE feature and removed the global song display.

If you only want to shuffle certain songs, you can click the checkbox next to each song you want to shuffle and the
shuffle will be limited to those songs.  If you click remember when some checkboxes are checked then it will SAVE
your selection. Next time you start the player (With no boxes checked as any checked boxes causes clicking
remember to SAVE) click remember to automatically bring back your previously saved selection.

Re: Welcome to AI Image Central!

Reply #72 –
New Update

I got sick and tired of the slowness from running the website under IIS 10 and pretty much got everything fixed so I decided
to upgrade the website to a better webserver. For about the last 4 days I have been working on converting over to Nginx.
It's a lot of work converting everything.  I have almost 200 files in the root directory and several folders to go through.
The directories changed for everything so it's a major undertaking to switch but the website runs faster with the upgrade.

What I have fixed so far:
- SDXL generation with image auto populating. (All models and Loras)
- SD3.5 generation with image auto populating. (Turbo TEXT/NO TEXT and TurboX options)
- Did 3 days of work alone on the Song Player before switching.  If you have not checked it out, you should. Top left music note.
- I also got bored before deciding to switch webservers and have been working on a video player. 
  Not sure what I'm going to use it for yet, but it works great.  If anybody wants to try it out give me a shout. 
  I could use more people for testing.

Still to do (That I can think of...):

- Flux generation (All models)
- Lumina generation
- Fix the buttons for Face Fix, Upscaler, and Remove Background.  Also add the same options to the NSFW category.
- Fix the GPU percentage indicators.  Seems Nginx has a different way of handling web sockets so it's been a pain to get working again!
-  And after all that, I have to do it for all the other languages.

Update: Fixed the percentage and Loading... indicators when generating.

Re: Welcome to AI Image Central!

Reply #73 –
New update:

All the image generation models have been fixed.  The only thing still broken are the buttons for face fix, upscale, and remove background. 
But these were actually broken before the upgrade.  Don't know how the heck it happened.  I have a ton of
backups but during one of the major conversions using the AI models they broke these features somehow and I have been
having a hard time getting a working code solution.  It really messed it up good, but there were so many good changes I
could not revert back. For some reason the remove background no longer works on the selected image, it always goes to the
last image generated and I have been a hell of a time trying to fix it.  You would think it would be an easy fix but I have
tried using the top 3 AI programming bots multiple times and they always screw it up.  I will keep working on it though
and hopefully will figure it out shortly and also add the same features to the NSFW folder.

Re: Welcome to AI Image Central!

Reply #74 –
New update:

Realized switching to Nginx also broke the search engines.  They have now been fixed.
Fixed a configuration issue.
Removed ads from search engine pages.
 
AI Image Central - Welcome to our forum!