chunked/nltk_chunking/model_doc_audio-spectrogram-transformer/chunk_6.txt

In this paper, we answer the question by introducing the Audio Spectrogram Transformer (AST), the first convolution-free, purely attention-based model for audio classification.