Qwen3-VL-4B-Instruct is a 4-billion-parameter instruction-tuned multimodal large language model launched by the Alibaba Cloud Qwen team. It is optimized for Qualcomm NPUs, integrating powerful visual-language understanding capabilities with dialogue fine-tuning functions, and is suitable for practical application scenarios such as chat reasoning, document analysis, and visual dialogue.

MultimodalGguf

YamNet is an audio event classification model trained on the AudioSet dataset. It uses the Mobilenet_v1 depthwise separable convolution architecture and is optimized for deployment on mobile devices. It can accurately identify and classify various audio events.

Audio ProcessingPytorch

An English speech recognition and translation model optimized for mobile deployment, implemented by Qualcomm.

Audio ProcessingPytorch