Text transformers have the ability to translate natural language text into a numerical vector representation also known as an embedding, you use such vectors for semantic similarity search or other Natural Language Processing (NLP) use cases.

Models such as BERT, sentence transformer models from Hugging Face, and other transformer-based models can be converted into ONNX format models. These models can be run within Oracle Database. These models can be used in AI vector search within Oracle Database, where documents are compared based on their mathematical distance between the vectors to determine the similarity.