sqlflow
🏗️ Frameworksql-machine-learning
Extend SQL syntax to train and predict ML models directly from your database.
SQLFlow is a transpiler that compiles SQL statements with ML extensions into training and prediction jobs on various AI engines. It supports multiple backends including TensorFlow, PyTorch, and XGBoost, and integrates with databases such as MySQL, MariaDB, and Apache Hive. Users can write simple SQL commands like 'SELECT ... TRAIN ...' to kick off model training, or 'SELECT ... PREDICT ...' for inference. SQLFlow abstracts away the complexity of defining ML pipelines, making AI accessible to SQL developers and data analysts. It also manages model versioning and metadata, ensuring reproducibility and governance. The project emphasizes a seamless user experience, allowing users to leverage their existing SQL skills for advanced analytics.
💡Highlights
- ├─SQL extension for ML training & prediction
- ├─Supports TensorFlow, PyTorch, XGBoost
- └─Works with MySQL, Hive, and more
🎯For
- ├─Data Analysts
- ├─SQL Developers
- └─Data Scientists