The code for the MAPSS measures for source separation evaluation (ICLR, 2026)
Shazam-inspired music recognition system — FAISS vector search + spectral fingerprinting, with CLAP/MuQ embeddings, FastAPI backend and React frontend.
DJTS – AI-powered music similarity engine for DJs and music collectors, combining audio embeddings, AI classifiers, and musical feature analysis.
Music mood embedding, clustering & zero-shot labeling engine (MERT + CLAP) — a pure, stateless Python MIR toolbox.
Music-emotion alignment via contrastive learning. Pairs frozen MERT audio encoder with BGE text encoder in a shared 512D space for emotion-based music retrieval.
A digital forensics pipeline designed to robustly distinguish between AI-generated music (Udio, Suno) and human-recorded audio (FMA)
Comparison of embedding-based models for electric guitar tone discrimination
Controlled benchmark comparing frozen audio foundation models (HuBERT, MERT, CLAP) for deepfake speech detection on ASVspoof5 Track 1
Exploratory music analysis using traditional audio features, MERT, and UMAP.
To associate your repository with the mert topic, visit your repo's landing page and select "manage topics."