https://github.com/blastre/mlcli
automated ML pipeline with LLM powered data analysis
Science Score: 26.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
○Academic publication links
-
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (8.3%) to scientific vocabulary
Repository
automated ML pipeline with LLM powered data analysis
Basic Info
- Host: GitHub
- Owner: blastre
- Language: Python
- Default Branch: main
- Size: 88.9 KB
Statistics
- Stars: 0
- Watchers: 1
- Forks: 1
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
🚀 MLCLI – Build ML Pipelines. Talk to Your Data.
MLCLI is a modular toolkit that lets you automate full machine learning workflows and interact with your data effortlessly — via either a command-line interface or a Streamlit web app.
Just provide a CSV and enjoy:
🔍 Smart Task Detection (classification/regression)
📊 Fast Data Handling with Dask
🧠 LLM-Assisted Pipeline Planning via LLaMA 3.3 70B
⚙️ Auto Preprocessing & Feature Engineering
🤖 AutoML Training powered by PyCaret
📉 Comprehensive Metric Reporting (Accuracy, R², F1...)
💾 Joblib Model + Pipeline Saving
💬 Interactive Data Chat Mode — Ask natural language questions about your dataset (powered by Groq LLM)
📂 Metadata Export for reproducibility
🌐 Streamlit Web App — Use MLCLI through a simple, intuitive browser UI
Demo Usage
CLI tool
type "python click.py start" to start
Streamlit Webapp
type "streamlit run app.py" to start
🧰 Tech Stack
Frameworks & Libraries: PyCaret, Scikit-learn, Dask, Click, Streamlit
AI & Automation: (LLaMA 3.3 70B)
Languages: Python
Data Processing: Pandas, CountVectorizer, datetime
CLI & Scripting: Click (modular command-line interface)
Model Handling: Joblib (pipeline & model saving)
🔁 Workflow
Owner
- Login: blastre
- Kind: user
- Repositories: 1
- Profile: https://github.com/blastre
GitHub Events
Total
- Push event: 3
- Create event: 3
Last Year
- Push event: 3
- Create event: 3