Description:

TL;DR: Lets the bot answer you with a picture!

Stable Diffusion API pictures for TextGen, v.1.2.0

An extension to oobabooga's textgen-webui allowing you to receive pics generated by Automatic1111's SD-WebUI API

History & Roadmap

Note
Consider the version included with oobabooga's repository to be STABLE,
experimental developments and untested features are to be pushed in Brawlence/SD_api_pics

Lastest changes:
1.2.0 → 1.2.1 Implementation of recently added state_modifier(state) function to override the streaming parameter while generating picture response 1.1.1 → 1.2.0 @GuizzyQC 's fix for mode on load, addition of Hires fix with relevant settings; UI refinements
1.1.0 → 1.1.1 Fixed omitting Auto1111's metadata in received images
1.0.0 → 1.1.0 Lots of new features and fixes; introduction of Connection check & VRAM shuffler
0.0.0 → 1.0.0 Initial reliase

Planned features & changes

Installation & Usage

Prerequisites

One needs an available instance of Automatic1111's webui running with an --api flag. Ain't tested with a notebook / cloud hosted one but should still be possible.
To run it locally in parallel on the same machine, specify custom --listen-port for either Auto1111's or ooba's webUIs.

Warning
Attentions WSL users! Windows and WSL exibit troubles interacting via loopback.
Use your network address (ex. 10.10.10.69) instead of localhost / 127.0.0.1

Installation

The stable version is already included with your TextGen-webUI. Load it in the --chat mode with --extension sd_api_pictures alongside send_pictures (it's not really required, but completes the picture, pun intended).

To test the experimental version, you can clone this repository into the extensions subfolder inside your text-generation-webui installation and change the parameters to include --extension SD_api_pics.
Or you can simply copy script.py and any other *.py over the files in extensions/sd_api_pictures subdirectory instead. Up to you.

Details

The image generation is triggered:

• manually through the 'Force the picture response' button while in Manual or Immersive/Interactive modes OR
• automatically in Immersive/Interactive mode if the words 'send|main|message|me' are followed by 'image|pic|picture|photo|snap|snapshot|selfie|meme' in the user's prompt
• always on in Picturebook/Adventure mode (if not currently suppressed by 'Suppress the picture response')

Features overview

• Connection to API check (press enter in the address box)
• VRAM management (model shuffling)
• Three different operation modes (manual, interactive, always-on)
• User-defined persistent settings via settings.json

Connection check

Insert the Automatic1111's WebUI address and press Enter:

Green mark confirms the ability to communicate with Auto1111's API on this address. Red cross means something's not right (the ext won't work).

Persistents settings

Create or modify the settings.json in the text-generation-webui root directory to override the defaults present in script.py, ex:

{
    "sd_api_pictures-manage_VRAM": 1,
    "sd_api_pictures-save_img": 1,
    "sd_api_pictures-prompt_prefix": "(Masterpiece:1.1), detailed, intricate, colorful, (solo:1.1)",
    "sd_api_pictures-sampler_name": "DPM++ 2M Karras"
}

will automatically set the Manage VRAM & Keep original images checkboxes and change the texts in Prompt Prefix and Sampler name on load.

Demonstrations:

Those are examples of the version 1.0.0, but the core functionality is still the same