Skip to content

This project use PANNs for audio tagging and sound event detection, and finally get audio embeddings. Then Milvus is used to search the similarity audio items.

License

Notifications You must be signed in to change notification settings

zilliz-bootcamp/audio_search

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

❗❗ This repo will no longer be maintained, please visit https://github.com/milvus-io/bootcamp ❗ ❗

Audio search system with Milvus

This project uses PANNs(Large-Scale Pretrained Audio Neural Networks) for Audio Pattern Recognition to perform audio tagging and sound event detection, finally obtaining audio embeddings. Then this project uses Milvus to search for similar audio clips.

Local Deployment

Requirements

Run Server

  1. Install python requirements

    $ cd bootcamp/solutions/audio_search/webserver/
    $ pip install -r audio_requirements.txt
  2. Modify configuration parameters

    Before running the script, please modify the parameters in webserver/audio/common/config.py:

    Parameter Description Default setting
    MILVUS_HOST milvus service ip address 127.0.0.1
    MILVUS_PORT milvus service port 19530
    MYSQL_HOST mysql service ip 127.0.0.1
    MYSQL_PORT mysql service port 3306
    MYSQL_USER mysql user name root
    MYSQL_PWD mysql password 123456
    MYSQL_DB mysql datebase name mysql
    MILVUS_TABLE default table name milvus_audio
  3. Star server

    $ cd webserver
    $ python main.py

System Usage

Type 127.0.0.1:8002/docs in your browser to see all the APIs.

  • Insert data.

    Download the sample game_sound.zip and upload it into the system.

    The sound data in the zip archive must be in wav format.

  • Search for similar audio clips.

    You can upload test.wav to search for the most similar sound clips.

Please refer to https://zilliz.com/demos/ to take a try in the front-end interface.

About

This project use PANNs for audio tagging and sound event detection, and finally get audio embeddings. Then Milvus is used to search the similarity audio items.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages