Steps to use the parser plugin:
- If it's missing, create a directory called 'server' inside the 'plugins' folder. For e.g.
D:\stable-diffusion-ui\plugins\server
. - Download the plugin from https://github.com/madrang/EasyDiffusion-Parser_Plugin/archive/refs/heads/main.zip.
- Unzip and extract the files inside a new folder
plugins\server\parser_plugin
so it's like the following:
D:\stable-diffusion-ui\plugins\server\parser_plugin\.gitignore
D:\stable-diffusion-ui\plugins\server\parser_plugin\__init__.py
D:\stable-diffusion-ui\plugins\server\parser_plugin\prompt_parser.py
D:\stable-diffusion-ui\plugins\server\parser_plugin\README.md
D:\stable-diffusion-ui\plugins\server\parser_plugin\tokenizer_conditionings.py
Restart the program. It'll start using the new parser. You'll still need to start your prompts with an exclamation mark, e.g. !photo of an (astronaut)
to use the advanced syntax. Without !
it will send the text literally to the model.