This dataset is distributed as tar archives for efficient storage and transfer.
After downloading the files, extract them with:
mkdir -p slakh_preprocessed
tar -xf slakh2100_yourmt3_16k.tar -C slakh_preprocessed
tar -xf yourmt3_indexes.tar -C slakh_preprocessed
After extraction, the directory structure will be:
slakh_preprocessed/
├── slakh2100_yourmt3_16k/
└── yourmt3_indexes/
The slakh2100_yourmt3_16k/ directory contains the preprocessed audio/data… See the full description on the dataset page:
https://huggingface.co/datasets/choihy/slakh2100_yourmt3_16k.
