audio-spectrogram-transformer

Patch-Mix Contrastive Learning with Audio Spectrogram Transformer on Respiratory Sound Classification (INTERSPEECH 2023)

This is the official repository of the papers "Parameter-Efficient Transfer Learning of Audio Spectrogram Transformers" [IEEE MLSP 2024] and "Efficient Fine-tuning of Audio Spectrogram Transformers via Soft Mixture of Adapters" [Interspeech 2024].

Pytorch implementation of INTEGRATED PARAMETER-EFFICIENT TUNING FOR GENERAL-PURPOSE AUDIO MODELS

Three in one. capture audio with mobile mic and then play, play songs from internal download directly, displaying spectogram (frequency) of an audio in Flutter.

Two-branch audio authenticity app that detects AI-generated speech and environmental audio using WavLM, AST, FastAPI, and React.

GenreAST is a highly accurate music genre classification model relying on the AST model

AI-powered deepfake audio detection tool built with Kotlin to analyze and verify voice authenticity in real time.

A full-stack multimodal music agent for audio analysis, playlist planning, and LLM-based music recommendation.

汎用音楽分析 CLI: BPM・拍子・コード・調性・音楽理論・ジャンル (madmom + librosa + music21 + AST/AudioSet)

Audio Spectrogram Transformer with LoRA adapter.

Code accompanying ESANN 2025 submission "Exploring Model Architectures for Real-Time Lung Sound Event Detection". Dataset used was ICBHI 2017.

High-performance multimodal pipeline synchronizing Meta's SAM 2 (Vision) and MIT's AST (Audio). Features O(1) resource management, temporal video slicing, and automated JSON metadata orchestration for edge-device AI.

Real-time acoustic safety monitor — a fine-tuned Audio Spectrogram Transformer classifies safety-critical sounds, a zone-aware severity engine ranks urgency, and an LLM writes SOP-grounded alerts, in a Streamlit control-room UI.