Open Source State-Of-the-Art Solution for Romanian Speech Recognition
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908630970793984 |
|---|---|
| author | Pirlogeanu, Gabriel Georgescu, Alexandru-Lucian Cucu, Horia |
| author_facet | Pirlogeanu, Gabriel Georgescu, Alexandru-Lucian Cucu, Horia |
| contents | In this work, we present a new state-of-the-art Romanian Automatic Speech Recognition (ASR) system based on NVIDIA's FastConformer architecture--explored here for the first time in the context of Romanian. We train our model on a large corpus of, mostly, weakly supervised transcriptions, totaling over 2,600 hours of speech. Leveraging a hybrid decoder with both Connectionist Temporal Classification (CTC) and Token-Duration Transducer (TDT) branches, we evaluate a range of decoding strategies including greedy, ALSD, and CTC beam search with a 6-gram token-level language model. Our system achieves state-of-the-art performance across all Romanian evaluation benchmarks, including read, spontaneous, and domain-specific speech, with up to 27% relative WER reduction compared to previous best-performing systems. In addition to improved transcription accuracy, our approach demonstrates practical decoding efficiency, making it suitable for both research and deployment in low-latency ASR applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_03361 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Open Source State-Of-the-Art Solution for Romanian Speech Recognition Pirlogeanu, Gabriel Georgescu, Alexandru-Lucian Cucu, Horia Audio and Speech Processing Artificial Intelligence In this work, we present a new state-of-the-art Romanian Automatic Speech Recognition (ASR) system based on NVIDIA's FastConformer architecture--explored here for the first time in the context of Romanian. We train our model on a large corpus of, mostly, weakly supervised transcriptions, totaling over 2,600 hours of speech. Leveraging a hybrid decoder with both Connectionist Temporal Classification (CTC) and Token-Duration Transducer (TDT) branches, we evaluate a range of decoding strategies including greedy, ALSD, and CTC beam search with a 6-gram token-level language model. Our system achieves state-of-the-art performance across all Romanian evaluation benchmarks, including read, spontaneous, and domain-specific speech, with up to 27% relative WER reduction compared to previous best-performing systems. In addition to improved transcription accuracy, our approach demonstrates practical decoding efficiency, making it suitable for both research and deployment in low-latency ASR applications. |
| title | Open Source State-Of-the-Art Solution for Romanian Speech Recognition |
| topic | Audio and Speech Processing Artificial Intelligence |
| url | https://arxiv.org/abs/2511.03361 |