A re-hosting on the Hub of the CAFE dataset — ~37 hours of spontaneous,
in-vivo human–human dialogue from 100+ speakers, capturing real-world code-switching
between Algerian Arabic (Darja), French, and English across topics like
science, technology, sports, and society. All audio is 16 kHz, 16-bit PCM, mono WAV.
Attribution (required by CC-BY-4.0)