Audio Retrieval Based on Milvus

What this post added

This post details the application of Milvus for audio retrieval. It explains how audio data is processed by extracting feature vectors using deep learning models (specifically PANNs) and then how Milvus is used for vector similarity search on these extracted features. The post outlines the system overview with insert and search steps, including code snippets for embedding generation and insertion into Milvus, and subsequent search operations. It also highlights the use of a MySQL database for correlating Milvus IDs with audio file names.