TII Releases Falcon-ASR: 1.6B Arabic Speech Model Hits 20.92% WER
The Technology Innovation Institute has introduced Falcon-ASR, a 1.6B parameter speech recognition model specialized for the Arabic language with a focus on Emirati dialects. The model recorded an average word error rate of 20.92% across six test sets, though the weights and an API for public access have not yet been released.
Covered by 1 source
- MMarkTechPost↗Asif RazzaqOct 12