- Docker
- MLflow server
- S3 server (required for MLflow)
- Prefect server (used for orchestrating scripts within
flows/) - Docker registry (To handle remote Prefect flow runs)
To initialize the project, run:
make init
Fill values in the .local.env file
make docker-build
make docker-push
You have two options for setting up input data:
- Run the command:
make download - Alternatively, manually place audio files in the directory:
data/input_files
make preprocess
make train
make generate
make prefect-deploy
make run-training-pipeline-infra