Graduate student in Shanghai · audio-visual foundation models · she/her
I work on models that learn from video and audio together — how a model can watch, listen, and line the two up. Most of my code is meant to actually run: small, dependency-light reference implementations first, an optional heavier backend when you need scale.
avlex — a composable framework that bridges audio-visual encoders to LLMs for video captioning and understanding.
avalign — a contrastive audio-visual alignment toolkit: log-mel + SpecAugment preprocessing, symmetric InfoNCE loss, and retrieval evaluation.
syncscope — audio-visual synchronization and active-speaker detection via cross-modal correlation.
audio-visual learning · foundation models · cross-modal alignment · video understanding · active-speaker detection · reproducible research
Open an issue on any repo — I read them. Fudan University, Shanghai.