YuanGongND/ast
Code for the Interspeech 2021 paper "AST: Audio Spectrogram Transformer".
Applies Vision Transformer architecture to audio by treating mel-spectrograms as images with learnable patch embeddings and variable-length input support. Built on PyTorch with configurable model sizes (tiny to base384), it offers ImageNet and AudioSet pretraining options for transfer learning across audio classification tasks. Includes production-ready recipes for AudioSet, ESC-50, and Speech Commands datasets with attention visualization tools.
1,432 stars. No commits in the last 6 months.
Stars
1,432
Forks
244
Language
Jupyter Notebook
License
BSD-3-Clause
Category
Last pushed
May 21, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/voice-ai/YuanGongND/ast"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related tools
descriptinc/descript-audio-codec
State-of-the-art audio codec with 90x compression factor. Supports 44.1kHz, 24kHz, and 16kHz...
crlandsc/torch-log-wmse
logWMSE, an audio quality metric & loss function with support for digital silence target. Useful...
drethage/speech-denoising-wavenet
A neural network for end-to-end speech denoising
KyungsuKim42/tokensynth
The official implementation of TokenSynth (ICASSP 2025)
iver56/torch-audiomentations
Fast audio data augmentation in PyTorch. Inspired by audiomentations. Useful for deep learning.