Preprint
WER We Stand: Benchmarking Urdu ASR Models
Samee Arif et al., arXiv, 2024
Read the originalRead on October 17, 2026
- What it is
- A 2024 benchmark of Whisper, MMS and Seamless-M4T on Urdu speech, using read speech and a new conversational dataset. Its abstract compares which model does best on each kind of speech.
- How much weight it carries
- Preprint, with one dataset built by the authors. It is the only Urdu-specific benchmark we found that tests the Whisper family, so treat its ranking as one study.
Facts taken from this source
- A 2024 Urdu benchmark found that Seamless-large was best on read speech, while Whisper-large was best on conversational speech.
“seamless-large outperforms other ASR models on the read speech dataset, while whisper-large performs best on the conversational speech dataset” (Abstract)