llm-rp icon indicating copy to clipboard operation
llm-rp copied to clipboard

✨ Your Custom Offline Role Play with LLM and Stable Diffusion on Mac and Linux (for now) 🧙‍♂️

Upgraded project here ==> ImpAI

The project is not deprecated but i just recode all the app in a new repository !





Code style: black

✨ Your Custom Offline Role Play with AI on Mac and Linux (for now) 🧙‍♂️

LLM RP

📎 Requirement

For now, Mac & Linux:

  • Python >= 3.8
  • pip install -r requirements.txt && python3 -m spacy download en_core_web_sm
  • Install git lfs

🛠️ Config

You can custom global variables at the top of app/run.py as you want.

For increase performance, you can reduce the generated images size, the number of iterations etc...

🖍️ Note: the output images folder is in app/images/

👉🏻 Start

Each time you want to run the game:

python3 app/run.py

The first time you load the app you will wait a while, because the program will download, export and quantize the better llama model for your config and install stable diffusion.

Next open your browser at http://127.0.0.1:4242

Click on the 🗑️ for reset le Role Play

You can customize the prompt here: prompts/RolePlay.txt

📝 Todo

  • 💾 Create a persistent role play (with save system)
  • 🖼️ Adding Quantized Stable Diffusion
  • 🛠️ Compile for GPU Linux / GPU Windows
  • 🎤 Adding whisper.cpp
  • 🔉 Adding Bark or an other Text-Prompted Generative Audio Model
  • 🔥 Doing a better interface
  • 🔒 Lock user input when model generating response

🔎 Resources