Highlights

• Improved document website, including quickstarts for Orca, RayOnSpark, Zouwu and BigDL
• Orca library: unified API for running distributed deep learning (TensorFlow, PyTorch, Keras, BigDL, OpenVINO, etc.) on distributed Big Data (using Apache Spark and Ray)
• Experimental PPML support for privacy preserving big data analysis and machine learning (i.e., running unmodified Apache Spark, Apache Flink, BigDL and TF/PyTorch/OpenVINO inference in a secure fashion on cloud)
• Improved AutoML and Time Series (Zouwu) support, including AutoXGBoost, TCN, ONNX inference, etc.
• Improved Cluster Serving (including performance, stability, log information, etc.)