https://github.com/basharast/a2iprompt
Automatic1111 to InvokeAI prompt resolver
https://github.com/basharast/a2iprompt
ai automatic1111 invokeai prompt
Last synced: 11 months ago
JSON representation
Automatic1111 to InvokeAI prompt resolver
- Host: GitHub
- URL: https://github.com/basharast/a2iprompt
- Owner: basharast
- License: mit
- Created: 2023-07-07T16:25:12.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2024-06-29T11:01:18.000Z (over 1 year ago)
- Last Synced: 2025-03-22T21:12:27.354Z (12 months ago)
- Topics: ai, automatic1111, invokeai, prompt
- Homepage: https://basharast.github.io/A2IPrompt/
- Size: 4.06 MB
- Stars: 16
- Watchers: 3
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# About
This tool will help you to convert [Automatic1111](https://github.com/AUTOMATIC1111/stable-diffusion-webui) prompt to [InvokeAI](https://github.com/invoke-ai/InvokeAI) prompt
## Usage
Online test [click here](https://basharast.github.io/A2IPrompt/) and use it directly
or download this repo and open `index.html` it will work locally.
## Features
- Support InvokeAI (V2 and V3)
- Support wide conversion cases
- Support translation from invokeai to automatic1111
- Fetch prompt from online image
- Copy shortcuts
- Tokens counter
- Prompt syntax highlights
- Auto prompt copy
- Weight limiters
- Weight randomizer
## Standalone (Developers)
If you want to use the translation engine in your project use `converter_standalone.js` or `converter_standalone.ts` for `TypeScript`
note that `.ts` version will report error for missing `encoder` function, you need to import that (`GPT-Tokenizer`) manually or disable the line if you're not intersted in tokens counter
```js
//Create instance first
var invokeaiResolver = new InvokeAIPromptResolver();
//Prepare options, all keys are optional, you don't have to add all of them
var options = {
invokeaiVersion: 3,
rawNegative: false,
limitWeightPositive: "$1",
limitWeightNegative: "$1",
randomWeight: false,
usePowValueAlways: false,
dynamicPrompts: true
};
//1- From auto1111 to invokeai
var output1 = invokeaiResolver.convertAuto1111ToInvokeAI(positive, negative, options);
//2- From invokeai to auto1111
var output2 = invokeaiResolver.convertInvokeAIToAuto1111(positive, negative, options);;
//Both functions (1 & 2) returns object as below:
/*
{
from: {
positive: { text: , tokens: },
negative: { text: , tokens: },
},
to: {
positive: { text: , tokens: },
negative: { text: , tokens: },
}
}
*/
//Calculate tokens without translation
//Require `encoders/cl100k_base.js` to be included
var tokensOutput = invokeaiResolver.calculateInvokeAITokens(positive, negative);
//This function returns object as below:
/*
{
positive: { text: , tokens: },
negative: { text: , tokens: }
};
*/
```
## Contribution
Feel free to submit any request or fix
if you want to make your own fork just don't remove copyright section.
## Privacy
This tool works locally and doesn't send or collect any usage data or any texts
## Credits
Bashar Astifan (Developer)
GPT-Tokenizer thanks to [Bazyli Brzóska](https://github.com/niieani)
Syntax highlights and code editor [CodeMirror](https://codemirror.net/)
Tooltips [TippyJS](https://atomiks.github.io/tippyjs/)
Notifications [Ryan Morr](https://codepen.io/ryanmorr/pen/MyVvLg)
Checkbox switch [Edgar](https://codepen.io/BuiltByEdgar/pen/jWOVYQ)
HTML page has some elements from [Navneet](https://codepen.io/heynavneet/details/yXjPLw) and [Trent Dec](https://codepen.io/Trentdec/pen/YBEQKm)