TIGER: Time-frequency Interleaved Gain Extraction and Reconstruction for Efficient Speech Separation
Kai Li PRO
JusperLee
AI & ML interests
None yet
Recent Activity
liked a model 1 day ago
arbazsiddiqui/voxlock upvoted a paper 1 day ago
Learning Where to Split: Dynamic Band Splitting for Audio Source Separation submitted a paper 1 day ago
Learning Where to Split: Dynamic Band Splitting for Audio Source SeparationOrganizations
None yet
spaces 4
Running on Zero
Agents
2
DBS Listening Studio
🎧
Separate audio or video into speech, music, and sound‑effects stems
Running
Agents
5
BEAVER
🦫
Training-Free Hierarchical Prompt Compression
Running on Zero
Agents
3
Hive
🎵
Text-queried USS using AudioSep & FlowSep trained on HIVE.
Running
Agents
8
Dolphin: Efficient Audio-Visual Speech Separation with Discrete Lip Semantics and Multi-Scale Global-Local Attention
👀
Separate and isolate individual speakers from videos
models 19
JusperLee/DBS-TIGER-DnR
Audio-to-Audio • 1.01M • Updated • 216 • 1
JusperLee/DBS-BSRNN-DnR
Audio-to-Audio • 4.71M • Updated • 98
JusperLee/referential-dangling-detector
Text Classification • 0.1B • Updated • 164 • 4
JusperLee/Dolphin
Audio-to-Audio • 7.04M • Updated • 653 • 19
JusperLee/Real-World-AVSE-Baseline-Track2
Audio-to-Audio • 25M • Updated • 7
JusperLee/Real-World-AVSE-Baseline-Track1
Audio-to-Audio • 25M • Updated • 7
JusperLee/FlowSep-hive
Audio-to-Audio • Updated • 23
JusperLee/AudioSep-hive
Audio-to-Audio • Updated • 61 • 1
JusperLee/TIGER-speech-tiny
Audio-to-Audio • 102k • Updated • 731 • 2
JusperLee/TIGER-speech-small
Audio-to-Audio • 822k • Updated • 174 • 2
datasets 10
JusperLee/EchoSet
Viewer • Updated • 82.1k • 424 • 10
JusperLee/Hive-ALL
Viewer • Updated • 8.4M • 594 • 1
JusperLee/Hive
Viewer • Updated • 5.6M • 101 • 2
JusperLee/AudioTrust
Viewer • Updated • 8.1k • 441 • 7
JusperLee/Apollo-val-test-data
Updated • 7
JusperLee/SonicSet-v2
Updated • 168 • 6
JusperLee/LibriSpace
Updated • 13
JusperLee/LRS3-2Mix
Updated • 8
JusperLee/VoxCeleb2-2Mix
Updated • 16 • 3
JusperLee/LRS2-2Mix
Updated • 342 • 8