CodeGeeX4

mirror of https://github.com/JasonYANG170/CodeGeeX4.git synced 2024-11-23 04:06:32 +00:00

History

Stanislas0 9e364eec5d Initial commit		2024-07-05 09:33:53 +08:00
..
models	Initial commit	2024-07-05 09:33:53 +08:00
resources	Initial commit	2024-07-05 09:33:53 +08:00
utils	Initial commit	2024-07-05 09:33:53 +08:00
chat.py	Initial commit	2024-07-05 09:33:53 +08:00
README_zh.md	Initial commit	2024-07-05 09:33:53 +08:00
README.md	Initial commit	2024-07-05 09:33:53 +08:00
requirements.txt	Initial commit	2024-07-05 09:33:53 +08:00
vectorize.py	Initial commit	2024-07-05 09:33:53 +08:00

README.md

English | 中文

RAG Functionality

CodeGeeX4 supports RAG functionality and is compatible with the Langchain framework to achieve project-level retrieval Q&A.

Tutorial

1. Install Dependencies

Navigate to the langchain_demo directory and install the required packages.

cd langchain_demo
pip install -r requirements.txt

2. Configure Embedding API Key

This project uses the Embedding API from the Zhipu Open Platform for vectorization. Please register and obtain an API Key first. Then, configure the API Key in models/embedding.py. For more details, refer to https://open.bigmodel.cn/dev/api#text_embedding.

3. Generate Vector Data

python vectorize.py --workspace . --output_path vectors
>>> File vectorization completed, saved to vectors

4. Run the Q&A Script

python chat.py --vector_path vectors
>>> Running on local URL:  http://127.0.0.1:8080