Repository files navigation

MLE-Agent: Your intelligent companion for seamless AI engineering and research.

💌 Fathers' love for Kaia 💌

📚 Docs |🐞 Report Issues |👋 Join us onDiscord

MLE-Agent is designed as a pairing LLM agent for machine learning engineers and researchers. It is featured by:

• 🤖 Autonomous Baseline: Automatically builds ML/AI baselines and solutions based on your requirements.
• 🏅End-to-end ML Task: Participates in Kaggle competitions and completes tasks independently.
• 🔍Arxiv andPapers with Code Integration: Access best practicesand state-of-the-art methods.
• 🐛 Smart Debugging: Ensures high-quality code through automatic debugger-coder interactions.
• 📂 File System Integration: Organizes your project structure efficiently.
• 🧰 Comprehensive Tools Integration: Includes AI/ML functions and MLOps tools for a seamless workflow.
• ☕ Interactive CLI Chat: Enhances your projects with an easy-to-use chat interface.
• 🧠 Smart Advisor: Provides personalized suggestions and recommendations for your ML/AI project.
• 📊 Weekly Report: Automatically generates detailed summaries of your weekly works.

Workflow	Agent

• 🚀 09/24/2024: Release the0.4.2 with enhancedAuto-Kaggle mode to complete an end-to-end competition with minimal effort.
• 🚀 09/10/2024: Release the0.4.0 with new CLIs likeMLE report,MLE kaggle,MLE integration and many newmodels likeMistral.
• 🚀 07/25/2024: Release the0.3.0 with huge refactoring, many integrations, etc. (v0.3.0)
• 🚀 07/11/2024: Release the0.2.0 with multiple agents interaction (v0.2.0)
• 👨‍🍼07/03/2024: Kaia is born
• 🚀 06/01/2024: Release the first rule-based version of MLE agent (v0.1.0)

# With pip:pip install -U mle-agent# With uv:uv pip install -U mle-agent

1. Clone the repo

   git clone https://github.com/MLSysOps/MLE-agent.gitcd MLE-agent

2. Create & activate a virtual env

   uv venv .venvsource .venv/bin/activate# Linux/macOS

3. Editable install

   pip install -e.# or: pip install -e .

mle new<project name>

And a project directory will be created under the current path, you need to start the project under the projectdirectory.

cd<project name>mle start

You can also start an interactive chat in the terminal under the project directory:

mle chat

MLE agent can help you prototype an ML baseline with the given requirements, and test the model on the local machine.The requirements can be vague, such as "I want to predict the stock price based on the historical data".

cd<project name>mle start

MLE agent can help you summarize your weekly report, including development progress, communication notes, reference, andto-do lists.

cd<project name>mle report

Then, you can visithttp://localhost:3000/ to generate your report locally.

cd<project name>mle report-local --email=<git email> --start-date=YYYY-MM-DD --end-date=YYYY-MM-DD<path_to_git_repo>

• --start-date and--end-date are optional parameters. If omitted, the command will generate a report for the default date range of the last 7 days.
• Replace<git email> with your Git email and<path_to_git_repo> with the path to your local Git repository.

MLE agent can participate in Kaggle competitions and finish coding and debugging from data preparation to model trainingindependently. Here is the basic command to start a Kaggle competition:

cd<project name>mle kaggle

Or you can let the agents finish the Kaggle task without human interaction if you have the dataset and submission fileready:

cd<project name>mle kaggle --auto \--datasets"<path_to_dataset1>,<path_to_dataset2>,..." \--description"<description_file_path_or_text>" \--submission"<submission_file_path>" \--sub_example"<submission_example_file_path>"\--comp_id"<competition_id>"

Please make sure you have joined the competition before running the command. For more details, see theMLE-Agent Tutorials.

The following is a list of the tasks we plan to do, welcome to propose something new!

We welcome contributions from the community. We are looking for contributors to help us with the following tasks:

• Benchmark and Evaluate the agent
• Add more features to the agent
• Improve the documentation
• Write tests

Please check theCONTRIBUTING.md file if you want to contribute.

• Discord community. If you have any questions, please ask in the Discord community.

@misc{zhang2024mleagent,title ={MLE-Agent: Your Intelligent Companion for Seamless AI Engineering and Research},author ={Huaizheng Zhang*, Yizheng Huang*, Lei Zhang},year ={2024},note ={\url{https://github.com/MLSysOps/MLE-agent}},}

CheckMIT License file for more information.