Cloudflare Docs
Constellation
Cloudflare Docs
Constellation
GitHub icon
Visit Constellation on GitHub
Set theme to dark (⇧+D)
  1. Products
  2. Constellation
  3. Platform
  4. Runtimes

Constellation runtimes

With innovation to AI, various machine learning (ML) and AI tools and frameworks are emerging and being improved over time. Constellation was designed to support multiple ML runtimes at its core.

Our vision is to support multiple runtimes. Currently, we only support the ONNX Runtime, and are exploring and researching others. XGBoost is in the final stages of testing.

Key
NameDescription
ONNXONNX RuntimeCross-platform machine-learning model accelerator. ONNX Runtime can be used with converted models from PyTorch, TensorFlow/Keras, TFLite, scikit-learn, and other frameworks.
XGBoostScalable and Flexible Gradient BoostingXGBoost is an optimized distributed gradient boosting library designed to be highly efficient, flexible, and portable.

Use Wrangler to list the models Cloudflare supports:


$ npx wrangler constellation runtime list


┌─────────┐
│ name    │
├─────────┤
│ ONNX    │
├─────────┤
│ XGBoost │
└─────────┘

​​ Supported runtimes

​​ ONNX

Constellation supports the ONNX Runtime. The Open Neural Network Exchange (ONNX) is an ecosystem supported by multiple technology companies and research organizations, defines open standards for machine learning, and makes open-source software for AI.

The Open Neural Network Exchange (ONNX) is an ecosystem supported by multiple technology companies and research organizations. ONNX defines open standards for machine learning, and makes open-source software for AI.

The two essential parts of ONNX are the open file format), made with interoperability and easy conversion from other formats in mind, and the ONNX Runtime, which we are running in Constellation.

​​ XGBoost

XGBoost is an optimized distributed gradient boosting library designed to be highly efficient, flexible, and portable, and it’s known for its performance in structured and tabular data tasks.

XGBoost provides a parallel tree boosting (also known as GBDT, GBM) that solve many data science problems in a fast and accurate way.