Powered by LLM, TabLLM-Copilot is able to automatically identify forms and generate corresponding data analysis and trend predictions based on users' questions, which helps you gain valuable insights from your data with ease.
You can easily and directly experience the project demo online on HuggingFace now. Click here for Online Experience 👉 Lesion-Cells DET - a Hugging Face Space by Tsumugii (just for placeholding right now)
- Complete the Gradio Interface for multi-input and multi-output of the first OCR processing stage
- Add Dr. Yue Wu's brief introduction
- Add a gif demonstration
- Deploy the demo on
HuggingFace - Finish the LLMs interface and prompt design
- Finetune OpenSource models for more powerful data analysis
- Try multimodal LLM such as
LLava,GPT-4-turbo
First of all, please make sure that you have already installed conda as Python runtime environment. And miniconda is strongly recommended.
1. create a virtual conda environment for the demo 😆
$ conda create -n table python==3.10 # table is the name of your environment
$ conda activate table 2. Install essential requirements by run the following command in the CLI 😊
$ git clone https://github.com/Tsumugii24/TabLLM-Copilot
$ cd TabLLM-Copilot
$ pip install -r requirements.txt 1. open .env.example and fill your own api keys in the corresponding place if you want to use certain LLM, then rename the file to .env
# 智谱AI https://open.bigmodel.cn/usercenter/apikeys
ZHIPU_API_KEY =
# 阿里灵积平台 https://dashscope.console.aliyun.com/apiKey
DASHSCOPE_API_KEY =
# 讯飞星火 https://console.xfyun.cn/services/bm35
SPARKCHAT_APPID =
SPARKCHAT_APISECRET =
SPARKCHAT_APIKEY =
# 百度千帆 https://console.bce.baidu.com/qianfan/ais/console/applicationConsole/application
BAIDU_API_KEY =
BAIDU_SECRET_KEY =
BAIDU_ACCESS_TOKEN =
# OpenAI https://platform.openai.com/api-keys
OPENAI_API_KEY =
# Anthropic https://www.anthropic.com/api
CLAUDE_API_KEY =
2. Open Source LLM
- planning to support ChatGLM, Baichuan, Qwen, LLama, InterLM... Coming soon~😄
I would like to express my sincere gratitude to Dr. Yue Wu for his invaluable guidance and supports throughout the development of this project. His expertise and insightful feedback played a crucial role in shaping the direction of the project.
Dr. Yue Wu's Google Scholar Homepage
Feel free to open GitHub issues or directly send me a mail if you have any questions about the project. 👻