alphacep / alphacep/vosk-api

i want hinglish voosk model for my project

Aperta
#1,860 0 commenti 2 reazioni 0 assegnatari Vedi su GitHub
Lingua principale
Jupyter Notebook
Stelle
15.1k
Fork
1.8k
Metriche di merge delle PR
Nessuna PR unita negli ultimi 30g

Descrizione

To support precise offline recognition of Hinglish—informal speech blending Hindi grammar with English vocabulary—we employ a customized Vosk speech‑recognition model. This model is built by fine‑tuning a multilingual Vosk base on curated Hinglish datasets, containing typical code‑mixed phrases like “Mujhe weather check karna hai” or “Set an alarm for 7 baje”. By training on acoustic samples from Indian speakers and adjusting the language model to accept both Hindi and English words in context, the model achieves high accuracy without needing an internet connection—ideal for privacy-sensitive or low-connectivity environments. The result is a smooth, natural user experience: users can interact naturally in Hinglish and the system will transcribe intents reliably, enabling bilingual voice apps, assistants, or transcription tools tailored for the Indian market.

Guida per i contributori

Nessuna guida per i contributori indicizzata per questo repository

Valutazione

Questa issue non è ancora stata valutata.

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.