The ${\displaystyle \underset{n\to \mathrm{\infty }}{lim}}$ best audio dataset formatter!

This is currently compatible with datasets from Ai Hobbyist's datasets, but in the future will be compatible with as many formats as I care about.
This currently supports exporting to parquet (hence the shit name), with LJSpeech coming soon!!

Deprecated in favour of dataset-creator-neo

1. Download the src
2. run uv sync to download dependencies for your GPU
3. run uv run src/main.py to start!

1. Download the binary
2. Pray
3. Run the .exe
4. pray
5. ???
6. profit?

Please just do method 1. Also method 2 is already outdated 💀
Method 2 is also completely untested on any device other than my own soooooo your mileage may vary. I can guarantee method 1 works though..

• Manual quick sorting ui
• Parquet exporter
• Audio reformatter
• ML audio processing (its just uvr lmao)

1. Import dataset directory
2. Press 1
3. Go through using W, E, R to process the dataset
4. Finish sorting
5. Press 3 to export or 2 to process the poor quality files before exporting (doesn't guarantee the files become good)
6. Export the parquet file to HF or something idk