“From Masking to Merging”: rethinking SpecAugment for audio transformers
A post shares an arXiv paper by Minhee Park, Hyowon Ahn and Chanwoo Kim on rethinking SpecAugment for efficient audio spectrogram transformers.
TLDR
A post links to Minhee Park, Hyowon Ahn and Chanwoo Kim’s paper, “From Masking to Merging: Rethinking SpecAugment for Efficient Audio Spectrogram Transformer.”
Combined views
528
1 Source, first seen 16d ago
8 likes