luddite478 / noise-gen-model

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Noise Gen Model

Dependencies

  • Docker
  • MLflow server
  • S3 server (required for MLflow)
  • Prefect server (used for orchestrating scripts within flows/)
  • Docker registry (To handle remote Prefect flow runs)

Initialization

To initialize the project, run:

make init

Setting Up Environment Variables

Fill values in the .local.env file

Building Docker Image

make docker-build

Pushing Docker Image

make docker-push

Data Download Options

You have two options for setting up input data:

  1. Run the command:
    make download
    
  2. Alternatively, manually place audio files in the directory:
    data/input_files
    

Data Preprocessing

make preprocess

Model Training

make train

Data Generation

make generate

Prefect Deployment

make prefect-deploy

Run Prefect Deployment on the infrastructure

make run-training-pipeline-infra

About


Languages

Language:Python 51.1%Language:Jupyter Notebook 41.2%Language:Makefile 6.5%Language:Dockerfile 1.2%