You need to agree to share your Hugging Face information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

SIP annotated calls

Private copy of mnawahda/sip_annotated for the SIP ASR benchmark. Arabic phone audio (mostly Levantine): data collection calls (AB_DataCollection_*, AB_Dummy_Calls_*), virtual-assistant SIP calls (AB_VA_SIP_*) and order calls (Orders_Batch_*).

Each <name>.wav (mono PCM16; 16 kHz, Orders_Batch 8 kHz) has a <name>.csv with one row per annotated segment:

Column Meaning
start, end segment start and end, milliseconds
text transcript; <...> sound events, [label: words] spans (repetition, error, overlap, unintelligible)
speaker speaker_1, speaker_2, ... or overlap
age, gender mostly empty
Downloads last month
20