What changed
Hugging Face announced the release of OhanASR, a family of open-source Automatic Speech Recognition (ASR) models for Armenian. The first release includes two models: OhanASR-v0.1, optimized for accurate transcription of recorded speech, and OhanASR-Streaming-v0.1, designed for real-time speech recognition. OhanASR-v0.1 is built on a 2-billion-parameter multilingual model, fine-tuned on 820 hours of Armenian speech. OhanASR-Streaming-v0.1 builds upon this model with additional training to adapt it for streaming transcription.
What this means in practice
The models are available under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license, making them freely available for research, educational, and other non-commercial use. OhanASR-v0.1 achieves a combined Word Error Rate (WER) of 16.12% and Character Error Rate (CER) of 3.49%, ranking 2nd by WER and 1st by CER among both open and closed models on ArmBench-ASR. OhanASR-Streaming-v0.1 achieves a combined WER of 19.07% and CER of 4.47%, ranking 10th by WER and 9th by CER among open and closed models on ArmBench-ASR. Lower scores indicate better transcription accuracy.
Limitations and availability
Known limitations include occasional confusion between formal and colloquial Armenian, unreliable recognition of code-switching, and potential omission of words in unclear speech. Both models are available for non-commercial use and can be accessed through Hugging Face's platform. The models are intended for developers, researchers, and anyone building applications for the Armenian language.
