A Semantically Consistent Dataset for Data Efficient Query Based Universal Sound Separation Kai Li , Jintao Cheng , Chang Zeng, Zijun Yan, Helin Wang, Zixiong Su, Bo Zheng, Xiaolin Hu Tsinghua University, Shanda AI, Johns Hopkins University Equal contribution 📜 Arxiv 2026 🎶 Demo 🤗 Metadata 🤗 Hive ALL Audio 🤗 Space 💥 News [2026 05 21] Hive ALL is now also available on ModelScope for users in China who prefer faster downloads via the ModelScope mirror. 🇨🇳🚀 [2026 05 19] We release Hive ALL , the pre generated mixed audio dataset (9.17 TB, WebDataset format). No need to re mix from metadata. 🚀 [2026 03 09] We release infer audiosep.py for one command AudioSep inference in Hive. The script automatically downloads config and checkpoints from JusperLee/AudioSep hive. 🚀 [2026 03 09] We release infer flowsep.py for one command FlowSep inference in Hive, with automatic config and checkpoint download from JusperLee/FlowSep hive. 🚀 [2026 03 09] We release app.py , a unified Gradio demo that supports both AudioSep hive and FlowSep hive in a single interface. 🚀 [2026 03 09] Community Hugging Face Space is available at JusperLee/Hive for quick interactive demo. 🚀 [2026 02 09] Thanks…
Runs entirely in your browser via DuckDB-Wasm — this dataset's real data file is loaded once, then queried locally. Nothing is sent to a server.
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy