KaggleGo

Fall in Love with DeepSeek Hackathon Project

This is a project for the Fall in Love with DeepSeek Hackathon. The project is built using the Camel framework, which is a framework for building large language models.

Installation

To run the project, you will need to have the following installed:

• Python 3.12
• Camel 0.2.19
• Kaggle
• Streamlit

You can install the required packages using the following commands:

pip install camel
pip install "camel-ai[all]==0.2.19"#for all the dependencies
pip install kaggle
pip install streamlit   # 安装最新版本

Usage

To run the project, use the following command:

streamlit run KaggleGo_app.py

Demo Video

You can watch the demo video showcasing the features of KaggleGo below:

Demo Video 1 Setup

Demo Video 2 RuningOutput

Demo picture:

（to be updated）

Taggle the button KaggleGo! to start the demo.

Todo list

• 1. 使用critic_agent引入human用户协作 (Introduce human user collaboration using critic_agent)
• 2. 实现KaggleGo的核心功能，包括数据获取、数据清洗、数据预处理、模型训练、模型评估、模型部署等 (Implement core functionalities of KaggleGo, including data acquisition, data cleaning, data preprocessing, model training, model evaluation, and model deployment)
  • 已测试可执行比赛查询，比赛下载，排行下载
  • 分别给两个角色用不同的后端模型，给ai用户用deepseek-r1专注理解，ai assistant用擅长代码且能调用工具的模型（Assign different backend models to the two roles: use deepseek-r1 for the AI user to focus on understanding, and use a model proficient in coding and capable of calling tools for the AI assistant.）
• 3. 实现一个简单的streamlit app，用于展示KaggleGo的功能 (Create a simple Streamlit app to showcase the features of KaggleGo)

Dev log