BoquilaHUB 0.5: now it includes SOTA AI models for bioacoustics (github.com)

BoquilaHUB has launched version 0.5, introducing state-of-the-art (SOTA) AI models specifically designed for bioacoustics, enhancing functionalities like audio classification, playback, and spectrogram visualization. Users can now select an audio model and upload audio files seamlessly, marking a significant advancement for researchers and developers in the AI/ML community focused on ecological monitoring and wildlife studies. The revamped graphical user interface not only creates a more user-friendly experience when navigating audio and visual content but also includes features like an adjustable buffer for live feeds, allowing users to capture missed moments. From a technical perspective, the update includes notable architecture changes, shifting from manual image buffer manipulation to employing the egui library for improved interactivity and performance. This transition enhances the capabilities of BoquilaHUB by facilitating the integration of audio classification probabilities and embeddings through an efficient Enum structure. The latest updates, coupled with code cleanups and performance optimizations, promise to streamline user experience while expanding analytical capabilities, making this release a crucial step forward in bioacoustic data processing and analysis.