1.1 OML4Py 2.1 Updates

The following updates have been made in OML4Py User’s Guide Release 2.1 in 26ai.

Features	Description
Vector series data class	You can use oml.vector vector type to manipulate database objects using the OML4Py transparency layer. To learn more, see Oracle Machine Learning for Python API Reference.
New types of ONNX pipeline	Import Pretrained Models in ONNX Format provide methods for : Text Embedding - ONNX Pipeline Models: Text Embedding Image Embedding - ONNX Pipeline Models: Image Embedding Multi-modal Embedding - ONNX Pipeline Models: CLIP Multi-Modal Embedding Text Classification - ONNX Pipeline Models: Text Classification Reranking - ONNX Pipeline Models: Reranking Pipeline
ONNX algorithm	Allows you to bring your own ONNX-format model, load it in the database and score using existing data in the database. To learn more see, ONNX.
Export or Import Python scripts	Allows you to export or import your Python scripts. To learn more, see Export a User-Defined Python Function or Import a User-Defined Python Function.