Every result you get is a result of a different Christmas and gifts you got from it!

Also merry early Christmas! 🎁 💝 🎅 ☃️

When I try to make a new button and click generate using that button it says (text-to-image-plugin: Currently, the only valid resolutions are 512x512, 512x768 and 768x512). Anyway to bypass this?

This is reported by Lamatiel on Discord.

Using the external lore file (lore book) doesn't properly get saved/used. The manual lore input works just fine.

Here are some fixes that I made in this copy of the page

• Line 7777 @ HTML Panel - m.text to m.text.split('\n').join('\n\n') to have proper empty lines per lore entry on the textarea display.
• Line 7783 @ HTML Panel - e.text to e.text.split('\n').join('\n\n'), same reason as above, so even without reloading the URLs the formatting would be good.
• Added let newLoreBookEntries = result.loreBookUrlEntriesText.replace(/\r/g, "").split(/\n{2,}/).map(e => e.trim()).filter(e => e); after Line 7790 @ HTML Panel to parse the lore from the lorebook URL files.
• Added newLoreEntries = newLoreEntries.concat(newLoreBookEntries) to combine the lore from manual entry and lorebook URL entries, added after the previous line.
• Line 7816 @ HTML Panel - originalLoreEntriesText !== result.loreEntriesText to originalLoreEntriesText !== newLoreEntries.join("\n\n"), this would compare the new lore entries to the original one better since the old comparison only compares the manual lore entries.

Some notes:

• Upon saving the character after writing the lorebook URLs on the Character Edit, the lore doesn't apply immediately. We need to go to /lore, show the character specific lore, then reload the URLs, then save, for the lorebook URLs to apply.
• Removing the lorebook URLs doesn't remove the lore from that URL from the current lore list, we need to manually remove it.

This morning I can't seem to find the public gallery/comments on the text-to-image generator anymore. Did it get removed?

https://perchance.org/ai-text-to-image-generator

Is there any other way to access to it?

perchance is loading forever - ive tried clearing the cache and using a private window, but it doesnt work. i=it loads on chrome but not on firefox.

More info:

Browser Version: Firefox 120.0.1 (64-bit) OS: Linux Mint 21.2 Cinnamon

The generator is frequently updated but this is what we know for now (based on all the different topics here about it):

-the "style" add some prompts and negative prompts to your own prompt, so you have to add it on your prompt on automatic1111, you can find them here: https://perchance.org/t2i-styles#edit

-the model is different depending on the prompts/style, for example realistic vision (it is confirmed), and maybe deliberate v2 are alternatively use for photo related images (mainly realistic vision since an update made one month) ,and I think deliberatev2 is use for other kind of styles too (it is a very versatile model). But it was used for sure because someone find it on previous jpeg files properties . I think it is still used. We don't know if other models are used for other styles

-sampling method : it seems it is Euler A, that is what someone find on previous jpeg files properties too

-sampling steps: no more than 20

-"no loras, no embeddings, no fancy sampler stuff, no hi-res fix, no hidden ‘quality’ prompts" (from @perchance@lemmy.world )

-BUT (and this is the "problem" lol) "there is hidden stuff to reduce suggestive/nudity/etc." (this is the words of the owner of @perchance@lemmy.world)

And THIS (the hidden stuff) is why we just can not reproduce the perchance text2img generator locally on automatic1111 I think, even if we use realistic vison or deliberatev2 with the right sampling method/steps, the exact same prompt and style prompts. Stable diffusion is very very sensitive to very small change (just change one word on a prompt and the images can be very different). Even the resolution change the result (portrait vs landscape). So without this "hidden stuff" and other settings we don't know, it is for now impossible to reproduce it ... I tried to add negative prompts like "nsfw" or "porn" and things like that but it doesn't repoduce the perchance generator.

So we really hope he will give us all settings (in fact it can be explained in two minutes ^^) and so we will stop to bore him with that lol

to upvote the topic/comment can help ^^

here is the generator https://perchance.org/rainwing-appearance-image-format#edit

i have each body part of the dragon have organized sub-lists of color, and i want the colors of some parts to reference the colors of other parts. is there ANY way i can do that? i don't exactly need the sublists to reference the exact image in their dynamic odds, just which sublist theyre in

Could we self-host it?

So I'm working on an item roller that has two outputs. (The generator linked is just a quick example)

The only problem is: I can't figure out how to make it so the random boxes can have separate amounts without affecting each other. For example: I put the number "2" in the "Item Drops" section and the number "4" in the Puzzle drop. Both outputs roll 4 items.

This isn't a big deal as I can roll one and then enter the amount and roll the other, but I would like to know if there's any way to make it so both amounts work independently.

Can you help me understand how to create a character that you, and not the AI, will roleplay? I can't find character sheet for USER (only name edit and pic). And how can you give the AI clarification to write something special in the chat, like “detailed descriptions; more communication” etc.? I scrolled through the tips, but I’m completely dumb in this one and looks like i can’t figure out something simple because in the old AI chat I create myself and my interlocutor, and everything is simple. But here, the user seems to be insanely limited and represents something amorphous. And how to create world in this? Cz i saw mention about it, but don't know how to start. Help please because I'm so excited about the new AI chat in 90% of things and I really want everything to be perfect with this project!

There seems to be a bug with creating a consumable list, then using .evaluateItem from it doesn't consume the item.

Demo:

items
  a
  b
  c
  
output
  $output = [this.getRawListText.split('\n').slice(2).join('\n')]
  A: [x = items.consumableList, ''] [x.evaluateItem] [x.evaluateItem] [x.evaluateItem] [x.evaluateItem]<br>
  B: [x = items.consumableList, ''] [x] [x] [x] [x]<br>
  C: [x = items.consumableList, ''] [x.selectOne] [x.selectOne] [x.selectOne] [x.selectOne]

• A doesn't work
• B and C Works

I think it is because .evaluateItem essentially converts the consumable list to text, which will not consume it, although it should do a .selectOne (a reference post) behind the scenes which should consume from the list, but it is not working.

i haven't been able to figure out how to dynamically change the background in response to things happening in the generator. i've tried both changing the background color based on a random generation, and using the background-image plugin to add a background image and changing the image to be used based on a random generation. i haven't gotten either to work yet.
if there's any way to make either or both of those work it would be a big relief. thank you in advance for answering! :0

Hi!

I have a question about the privacy of the prompts and responses from the AI chat. Are the prompts and responses stored in any way? Does anyone have access to them? Does anyone review them?

I'm wondering how safe it is to provide personal information in the chat.

Thanks in advance!

I'm trying to use the remember-plugin, but I would like to use it like this generator did as a scratchpad: https://perchance.org/ai-character-generator I want to add it to the Story Outline Generator and Story Writing Helper to keep prompts or story snippets.

scratchpad // Note: the 'remember' option storage is 'based on' this 'scratchpad' name, so if you change that name to something else like 'notepad', people will lose all their saved scratchpad text. But it's safe to change the 'label', below.
  label = 🗒️ Scratchpad
  tip = [t2i.defaultText.scratchpadTip] // this loads the default text, but you can delete it and add your own tip. Just change this line to "tip = You own cool tips."
  type = paragraph
  remember = true // <-- the text that the user inputs will be remembered so it's still there even if they close/refresh the page
  width = min(750px, 100%)
  height = 30vh
  takesUpFullRow = true
  foldToggleState = hidden // or 'shown' if you want it to start shown, or delete this line if you don't want it to be show-able/hide-able
  visible() =>
    return Number(localStorage.generateClickCount) > 1; // so the scratchpad only shows up after they've clicked generate a couple of times (to reduce confusion/clutter for newbies)
  examples
    [t2i.defaultText.scratchpadPlaceholder]

I am new to AI. I am using the Character generators. I have used several and they are all great. My problem is I am unable to add my own characters using photos of people from my own photo collection. I tried the plugin, but it doesn't work. I tried lists. That doesn't work either. From what I can deduce, there is something in the text to image generator prohibits this. I like that fact it works with celebrity names for the most part, but I would also like the option to add my own personal people. Is this possible? Again, I am brand new to this, so I ask your indulgence.

Thanks.

Recently, a couple of people started losing images in their private galleries from the text-to-image-plugin.

Here are some posts about it:

• Reddit Post about Gallery being Purged
• Tsuyoi KaiZoKu mentioned from Discord that some of their images in a private gallery disappeared after saving it into the gallery after a certain amount of time (I wasn't able to replicate it and they mentioned it might be depending on the number of images on the gallery).

Please comment below any information/problems you also encountered with the same issue.

Some helpful information you can provide:

• Generation date of the image saved into the gallery that has disappeared.
• When the purge/disappearance of images happened
• Generator that you have your gallery on
• Any test you tried to replicate the problem

Thanks!

@perchance@lemmy.world

Hey there!

I need help using AI Text Plugin outside of perchance, I tried to ask @perchance@lemmy.world in private message, but I don't think he/she gets time to read those. I am trying to create a discord bot which will use the API and I find the parameters and everything in the plugin perfect for what I need. You can even make it subscription based which is reasonable since it's hosted on a GPU Based server most likely.

It'd be great help if I can get more guide on how to use these specific plugins outside of perchance for other type of programs.

Plugins I'm most interested in are ai-text-plugin and text-to-image-plugin.

Thank you

Note that the 'auto' checkbox doesn't actually clear variables - it just fully re-executes all the square/curly blocks in the HTML. So if you're doing stuff where your code needs variables to be "cleared"/fresh before the first generation, then you'll have noticed that you actually need to click the 'reload' button (or refresh the whole browser tab) to get your generator back to a completely "fresh" state. But, as the annoying message would inform you, clicking the reload button required saving your generator first. That's not required any more. I should have fixed this about 6 years ago. Took about 15 mins 😐

I have some questions and want to learn more about how to use something I've been utilizing on the site.

I started out trying to split this giant list (nearly a thousand entries!) by gender. When I got down to the C's, I found Cobra Starship, which is a band. Continuing on, I found Fleetwood Mac, Florida-Georgia Line, and Foo Fighters. Eventually I found Jersey Shore and The Bachelorette, then Dsquared2, a fashion brand.

The comments in the source suggest the list was made by a machine, and it looks like it.

https://perchance.org/yu4wj171gf#edit Can anyone fix it? I can't figure out what's wrong with it.

i've toyed around with the interactive terminal template for a bit, but the commands there seem to be isolated from each other, and can't contain spaces. is it possible to make commands that work like, say, a "look" command that on its own displays the list of subcommands connected to itself ("look" might display "look left", "look right", and "look up"), and then each of those subcommands can have its own output?
thank you in advance for answering! :0

I need help debugging my Warhammer 40k generator. My generator works the first time, always, but then after a few clicks of randomize, it slowly breaks down and it suddenly will only generate one feature for each of the categories of Inner Cauldron, Primary Biosphere, and Outer Reaches. The chances of all getting only 1 result each are over multiple generations is nil. So I know that a variable somewhere is being overridden and stored upon a new press of the randomize button. I cant find where the issue is. The only thing that works is to refresh the page cause it always works the first time, just not on a soft reset. Please help.

You will need to press randomize a good 2 to 10 times if not more to really see the issue.

Hi!

I need a consumable list where two items are picked, but only the first is a random selection, and the second item is not at random but always a specific item based on the random first item.

Say, if item one is "A", then item two always has to be "X", and if item one is "B", then item two always has to be "Y".

To complicate things: "X" and "Y" should not be eligible for the random first item, they can only be the second, non-random item.

To complicate things further: none of the items are numbers, I'm sorry.

What I tried was to do something like a consumable list with dynamic odds, where I used == instead of != in mad hopes it'd work, but of course it didn't. (Is this something that could work if I changed it up a bit?)

A friend suggested I could use key/value pairs, but that would be so incredibly unwieldy and I'd rather not. I want to use this for long lists (plural) and for multiple possible outcomes. If key/value pairs are the most reasonable way to do this, how would I do this? Could I define the pairs in my list?

I'd really, really appreciate if someone could point me in the right direction.

Sorry if my question and/or explanation is confusing, I had too much coffee and I've been staring at code >24 hours now. Please let me know if I need to rephrase my question to make it make sense to people who are not in my brain.

Thank you!

(Not linking the generator in question because it's almost 3,000 lines of code and what I want to do is currently not in it anyway lol.)

This post is based on messages from the___messenger in the discord server.

The abstract-noun generator sometimes outputs "time)" and "naã¯vetã©" and this shouldnt happen