Language-Queried Target Sound Extraction Without Parallel Training Data

VLM
Re-assign community