I have ~200k audio files with associated metadata and features.
These have been stored into ~100 datasets.
Is there a way that I can concatenate these datasets into one without loading them into memory?
You can concatenate these datasets without loading them into memory with concatenate_datasets
.