Dataset · theusamaaslam
Urdu phone call audio recordings for ASR
theusamaaslam/urdu-callsqa-private
417 hours of annotated speech from Urdu phone calls for speech recognition and synthesis.
What's inside
The dataset contains 66,636 audio clips with a total duration of about 417 hours featuring phone conversation recordings of people in Urdu. Strict clips last from 1 to 10 seconds, while extended candidates last from 1 to 300 seconds; the data is accompanied by transcripts. File access is gated and requires approval.
What you can do with it
- Training and fine-tuning ASR models for Urdu telephone speech recognition
- Building text-to-speech (TTS) systems with natural conversational intonations
- Evaluating the quality of transcription models under telephone communication channel conditions
- Tasks
- Automatic speech recognition, Text → speech
- Languages
- Urdu
- Size
- 10K–100K rows
- Downloads this month
- 51.4K
- Likes
- 0
- Access
- on request on Hugging Face
How to load it
from datasets import load_dataset
ds = load_dataset("theusamaaslam/urdu-callsqa-private")