tencent-wedata-auto-ml 0.2.16

AutoML SDK for Tencent Cloud WeData using FLAML with MLflow integration.

wedata-automl

AutoML SDK for Tencent Cloud WeData, powered by FLAML and integrated with MLflow for experiment tracking and model registry.

Features

• FLAML-based AutoML with graceful fallback to RandomForest
• MLflow integration: experiment tracking, model logging, and Model Registry registration
• Quiet, production-friendly logging helpers
• Simple pipeline API and CLI demo

Installation

pip install wedata-automl
# Optional extras
pip install "wedata-automl[xgboost]"
pip install "wedata-automl[lightgbm]"

Quickstart (Python API)

import mlflow
from wedata_automl import run_pipeline

# IMPORTANT: Set MLflow tracking URI before running AutoML
mlflow.set_tracking_uri("http://your-mlflow-server:5000")

# Uses a demo dataset by default, and creates/uses the specified MLflow experiment
result = run_pipeline(experiment_name="blueszzhang-test-automl")
print(result)

Quickstart (CLI)

wedata-automl-demo

Documentation

• 🚀 快速启动检查清单 - 新手必读！
• 故障排除指南 - 常见问题和解决方案
• MLflow 版本兼容性 - MLflow 版本要求和兼容性
• 项目 ID 配置 - 项目 ID 配置说明

Notes

• Ensure MLflow Tracking/Registry is configured in your environment (MLFLOW_TRACKING_URI, credentials, etc.)
• XGBoost/LightGBM are optional; install via extras if you want those estimators considered
• Python >= 3.8
• Project ID is required - Set via project_id parameter or WEDATA_PROJECT_ID environment variable

Troubleshooting

遇到问题？查看 故障排除指南。

最常见问题：

• ⚠️ MLflow Tracking URI 配置错误 - 必读！
• MLflow 实验创建失败
• MLflow 版本兼容性问题
• 项目 ID 相关问题

License

MIT