Not sure, I don’t really dabble on the NSFW content, so I won’t know lmao.

No problem! It was actually an oversight on my part lmaooo

Here is an updated code in the AI Artist.

It seems that there was a problem in setting the properties in the character itself, in which those properties are already existing in the AI Artist character.

But since you’ve transferred the code to another character, those properties are undefined and cannot be accessed so it cannot change it. I’ve managed to fix it by just setting default values.

You can share the character link like this:

This won’t share any threads, just the character.

I’m not really sure, have you added a new style to it, maybe some syntax errors that doesn’t allow the custom code to run? Can you link a ‘share link’ of the character?

I’ve added a fix here to not say ‘AI Artist (style)’ but ‘Character Name (style)’ and the styles should be added after clicking the button on the /change-style, although you still need to edit/create a new image to apply the style.

Here is a video on applying the styles.

You need to create a new image or edit the current one to apply the style.

Lemmy.world is separate from Perchance, and isn’t affiliated with Perchance. It is a Federated Forum Platform, see this post for more Information about Lemmy.world.

Perchance.org on the other hand is a platform for sharing and creating random text generators. Logging in to Perchance would allow you to save/create generators. The AI tools is just an addition to the page.

You can use different login details on both pages.

The custom code is written on the character itself. You can open it up by clicking the pencil icon on the thread. Here is a step by step guide to add custom codes for your character. You do not need to create a new page for it.

Here is AI Artist v2.2 which has my vionet20-t2i-styles-importable. I’ve also commented on the custom code there the formatting on adding your own styles on the styles object.

I’ll also link some reddit posts that were having troubles with gibberish outputs, mainly on the ai-chat though some are on ai-character-chat:

• AI Chat Gibberish - 1
• AI Character Chat Gibberish - 1
• AI Character Chat Gibberish - 2

I haven’t encountered one myself, and I’m don’t really use it much.

By editing the syntax, you mean this input fields on the character settings?

Currently yes, after every message, it would reapply the current styles, which would remove all of the changes you have made on those inputs.

Here is the updated character (AI Artist V2.1) that should handle those user inputs to stay. However, if you change styles, it would be overwritten by the new ones.

If I misunderstood the request, please feel free to point me to the right direction.

Edit: Here is the updated AI Artist V2.

There’s the T2I-Artist and the AI Artist V2 in this ai-character-chat Docs. Although it is not optimal. I’m going to try refactor it and I’ll update this comment after I’m done.

I think if it doesn’t currently break, then keep it the same way XD

I’ve just added the functionality to use hashes, thanks for the suggestion. The tabs were accessible through URL params before, e.g. page=learn and for the chat rooms e.g. chat=general. Hash now takes precedence to set the tabs on load. Although the URL params would still work if no hash is provided.

I’ve also added an Id to each tab so the hash would be re-added on the URL for the specific tab opened.

It seems to me you are accessing the other inputs with :? It should be [input.Lighting] not [input:Lighting] if you have an input with:

Lighting
  label = Lighting
  type = select
  options 
    ...

Also, you can have the lists in the prompt like so:

Line-Art style
  prompt
    [input.description], (line drawing:1.8), (no gradients:1.8), clean lines, no color, black and white, visible hatching, high contrast, high resolution, high detail, intricate details, 4k, wallpaper, concept art, pen on paper, [input.Lighting != ""] ^[input.Lighting]
    [input.description], (line drawing:1.8), (no gradients:1.8), clean lines, no color, black and white, visible hatching, high contrast, high resolution, high detail, intricate details, 4k, wallpaper, concept art, pen on paper,  ^[input.Lighting == ""]
  negative = [input.negative], low-quality, deformed, text, poorly drawn, 3D, color, gradients, greyscale, shading

So that you can use Dynamic Odds in selecting which prompt to use if input.Lighting was added/selected.

You could also link the generator so we can directly see any problems.

Is this regarding an Image Generator with the t2i-framework-plugin? You can with:

Btype
  label = building type
  type = select
  options 
    Default
      , 
    House
      Generate a house with garden in the style of [input.Environment], House with 2 floors and slanted roof ,

Probably a ‘report a generator’ to report/flag generators with NSFW images?

Please see this relevant document about the AI in Perchance.

Btw as I’m remember I was told, the ones that start with $ are essentially private properties and there are methods that are used to get them. (Reference Post)

Ex.

• $allKeys are accessed with .getPropertyKeys
• $children are accessed with getChildNames

No, I don’t think there is. The public ones that I’ve seen are:

• api/getGeneratorScreenshot?generatorName=name - for generator screenshots
• api/getGeneratorList?max=num - for the general generator list in the /generators page
• api/getGeneratorStats?names=name1,name2 - for individual stats as seen from https://perchance.org/generator-list-with-stats-2-template#edit
• api/downloadGenerator?generatorName=name - from the download-button-plugin

The endpoint for getting the generator stats/metadata is https://perchance.org/api/getGeneratorStats?names={names}, it would return a list of data based on the generator names inputted and it would give a metadata property per generator that you can access.

Here is an example of using it: https://perchance.org/generator-list-with-stats-2-template#edit