Installation
This project uses uv for Python environment management. Make sure you have it installed before proceeding.
https://github.com/benluks/quick-convert
cd quick-convert
uv sync
CLI Usage
This project provides simplified CLI entrypoints for running different pipelines (e.g., anonymization, ASV training, evaluation) using Hydra configs under configs/run/.
Basic Pattern
All commands follow the same structure:
uv run <command> <config-alias> [hydra overrides...]
<command> → the pipeline you want to run (e.g., anonymize, train_asv, eval_asv)
<config-alias> → the suffix of a config file in configs/run/
[hydra overrides...] → optional Hydra overrides (key=value)
Examples
Anonymization
uv run anonymize knnvc_clac target_id=6081
Uses config:
configs/run/anonymization_knnvc_clac.yaml
Train ASV Model
uv run train_asv clac asv.overrides.batch_size=32
Uses config:
configs/run/train_asv_clac.yaml
Evaluate ASV
uv run eval_asv clac
Uses config:
configs/run/eval_asv_clac.yaml
How Config Resolution Works
Each command maps to:
configs/run/<prefix>_<config-alias>.yaml
Where:
- prefix is usually the same as the command
- exception:
anonymize→anonymization_*
So:
uv run anonymize knnvc_clac
→
configs/run/anonymization_knnvc_clac.yaml
Hydra Overrides
You can pass any Hydra override directly:
uv run anonymize knnvc_clac \
target_id=6081 \
pipeline.out_dir=foo \
hydra.job.chdir=false
Getting Help
Running a command without arguments will show available config aliases:
uv run anonymize
Notes
- Commands are thin wrappers around modules in
quick_convert/cli/ - All heavy lifting is still handled by Hydra + your existing pipeline code
- This interface is just a cleaner alternative to:
uv run python -m quick_convert.cli.<module> \ --config-name run/<full_config_name> ...