Multilingual Speech-to-Intent Recognition: Luganda and Runyankore

This dataset is a curated collection designed for multilingual speech intent recognition, specifically targeting low-resource languages, Luganda and Runyankore. It includes 20 distinct voice intents relevant to voice control in smart home environments, accompanied by audio samples and labelled intent classes. The dataset supports both monolingual and multilingual applications, making it ideal for developing and evaluating speech-to-intent and speech-to-text-to-intent models. Its broad applicability spans fields such as speech recognition, natural language processing (NLP), and IoT automation, offering valuable resources for researchers working in low-resource language contexts.