iMindRobot

🔧 Quick Start

• Prepare virtual environment

conda create -n imindrobot python=3.8
conda activate imindrobot

• Clone our repository

git clone https://github.com/LePanda026/iMindRobot.git
cd iMindRobot

• Install required packages:

cd requirements
pip install -r requirements_tts.txt
pip install -r requirements.txt
pip install -r requirements_api.txt
pip install -r requirements_webui.txt
pip uninstall torch
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
pip install streamlit_webrtc
pip install streamlit==1.26.0
pip install --upgrade langchain

• Download the model: All checkpoints can be availible here, and you should arrange them as following picture: M3E: 🤗Huggfacing;
  ChatGLM2-6B: 🤗Huggfacing;
  Bert_Japanese: 🤗Huggfacing;
  Bert_Chinese: 🤗Huggfacing;
  TTS_MODEL(character): Baidu Disk;

• Start serve and webui

cd ..

Run following command to init database

python init_database.py

Run following command to start webui and api serve

python startup.py -a

API

How to use api? You can reference to http://127.0.0.1:7861 after starting serve and webui.
For using api in Python, you can reference files under iMindRobot/tests/

💖 Acknowledgement

• ChatGLM ChatGLM2-6B: An Open Bilingual Chat.
• Langchain-Chatchat Langchain-Chatchat (formerly langchain-ChatGLM), local knowledge based LLM (like ChatGLM) QA app with langchain.
• Bert-Vits2 vits2 backbone with multilingual-bert.