String to Vector

Description

Converts string values into their corresponding vector embeddings. Given an embedding model and a pooling method, the input strings are encoded into vectors, which can then be used for vector operations like vector similarity search.

Input

• SOURCE [STRING]: a list of strings

Output

• RESULT [VECTOR]: the vector embeddings corresponding to the input strings

Parameters

• Embedding model: the embedding model used for creating the vector embeddings

  • all-MiniLM-L6-v2

  • snowflake-arctic-embed-l-v2.0

• Pooling mode: how the embedded parts/tokens of each input string are combined into one vector

  • ‘MEAN’: the average value of each dimension across all tokens is taken; captures the overall meaning

  • MAX: the highest value for each dimension across all tokens is taken; highlights the most prominent features