AI text embedding

Jina Embeddings V2 Base is an English text embedding model that supports 8192 sequence lengths. It is based on the Bert architecture (JinaBert) and supports symmetric bidirectional variants of ALiBi to allow longer sequence lengths. The model was pre-trained on the C4 dataset and further trained on Jina AI’s collection of over 400 million sentence pairs and negative samples. This model is suitable for a variety of use cases for processing long documents, including long document retrieval, semantic text similarity, text reordering, recommendation, RAG and LLM based generative search, etc. The model has 137 million parameters and is recommended for inference on a single GPU.