GUI

Paper

See our report paper at NLP_Project_Report___On_the_Application_of_LLM_as_Dungeon_Master_in_Tabletop_RPGs.pdf in the repo!

Environment

First, apply for an LLM SPI key and place it in dnd-storytelling-game-main\config.py.

The environment requirements are listed in the requirements.txt file. You can install them by running the following command:

pip install -r requirements.txt

Models

DeepSpeech

Download manually from Releases · mozilla/DeepSpeech:

• deepspeech-0.9.3-models.pbmm
• deepspeech-0.9.3-models.scorer
• deepspeech-0.9.3-models-zh-CN.pbmm
• deepspeech-0.9.3-models-zh-CN.scorer

Then rename them and arrange models like this:

DeepSpeech/
├── models/
│   ├── chinese.pbmm
│   ├── chinese.scorer
│   ├── english.pbmm
│   └── english.scorer
├── audio/
├── chinese.sh
└── english.sh

pyttsx3

pip install pyttsx3

GPT2 and LlaMA3.2

I will complete this part if necessary.

Stable Diffusion

We deployed SD3 on autodl servers, and here are instructions on how to use it.

1. Log in autodl server and launch SD3. You shall contact Xiang Ji to turn on the server. You are also recommended to configure ssh-key authorization on the server.

account: ssh -p 14058 root@connect.cqa1.seetacloud.com
password: le2LAenxPLNH

2. Launch the stable diffusion webui

cd autodl-tmp/stable-diffusion-webui
conda activate sd
./webui.sh

Later, you will see an ngrok link in the terminal, like https://81d9-58-144-141-64.ngrok-free.app. You shall copy it to dnd-storytelling-game-main/config.py.

3. Now the SD is done, you will see progress bar in the server terminal when some image is being generated.

APIs

Config

• Directly fill up in config.py.

• Stable Diffusion: SD - Api

• Deepseek: Follow the instructions in DeepSeek.

Usage

python gui.py